Magic Lantern Forum

Experimental builds (WIP) => crop_rec and derived builds => Topic started by: theBilalFakhouri on June 06, 2022, 12:50:10 AM

Title: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 06, 2022, 12:50:10 AM
Welcome to:
(https://i.ibb.co/47fVSm4/Bilal-Build-for-100-D.png)


-What is this?
It's a port from my custom builds for 650D/700D (https://www.magiclantern.fm/forum/index.php?topic=25784.msg231049#msg231049) to 100D, please read 650D/700D thread for more details then come back here and read the differences.

-Port status:
This port is a work in progress, expect some missing things compared to 650D/700D builds, please report bugs if you found some.


-What's missing:
-crop_new module (Crop mode V2), not ported yet (didn't make tries), will work on it soon (should work after making some tweaks).

-crop_rec (Crop mode):
25 FPS isn't working in UHD 1x3 preset (hardware limit), only 23.976 FPS works.

"One more hack" doesn't work on 100D, it freezes LiveView same as EOS M.

-What's working:
All other things not mentioned above are working fine :) (well, HDMI output not tested yet).

-Quirks (not related to my build, but I have noticed it):

When selecting 3K and 1440p presets you need to turn off then on RAW video multiple times to let RAW video update RAW video resolution parameters.

Recording stops very early on 100D e.g. when recording high resolutions like 2560x1440p @ 23.976 FPS at 10 bit lossless with 192 MHz overclock compared to 700D (on 100D it's 1-2 seconds while on 700D it's up to 11 seconds, same settings, same scene, same everything).

Edit: Fixed the second issue, also provided more details for the first issue, read this (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239224#msg239224).

-Download:
crop_rec_4k.2022Jun19.100D101.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jun19.100D101.zip)

-Focus Pixel Maps:
I have added them too (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239140#msg239140).

-Other notes:
-Will update this OP and add more details, also make it nicer with more photos and showcasing short videos as porting progressing.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 06, 2022, 10:21:56 AM
Updated the build, same link:

Quote from: theBilalFakhouri on June 06, 2022, 12:50:10 AM
-Download:
crop_rec_4k.2022Jun06.100D101.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jun06.100D101.zip)

-Fixed white bar in the right in 1920x1280 1:1 preset.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on June 06, 2022, 11:52:11 AM
Probably possibble to reuse existing pixel maps. Trial and error in mlv app should be helpful.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 06, 2022, 12:58:29 PM
Quote from: Danne on June 06, 2022, 11:52:11 AM
Probably possibble to reuse existing pixel maps. Trial and error in mlv app should be helpful.

I have tried, changed focus maps name to match my 3K and 1440p presets in terms of width and height (with black areas, I got width and height from mlv_dump), but MLVApp crashes for some reason
Edit: it seems I have downloaded wrong files, existing pixel maps appear to work! I have just added them in MLVApp repo.

On 700D I have captured maximum resolution in each Binning mode, e.g. 5208x3478 1:1, 1736x3478 in 1x3, then I sent them to dfort and he created the focus pixel maps for them, this way the same pixel map will work in any resolution, you just need to rename the same map to match mlv_dump height and width.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 06, 2022, 02:30:06 PM
Added focus pixel maps for 100D in MLVApp repository, should be downloaded automatically when you open a MLV clip.
You can also Download (https://drive.google.com/file/d/1LJaOYqqZT3kREext4EQPVcER-kinkXaD/view?usp=sharing) them manually (add them in MLVApp folder root).

No need to use Chroma Smoothing anymore.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 06, 2022, 06:05:00 PM
Thanks for bring back magic lantern to canon 100D with fresh feature  :)
I am so happy that I still have this great camera  8)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: DeafEyeJedi on June 06, 2022, 07:15:06 PM
THIS is the SHIT!  8)

I AM game.

Gonna test this one out along with the LiveView Hack builds that you did.

Thank you guys!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 13, 2022, 12:52:53 AM
Thanks guys @GianlucaM83 @DeafEyeJedi!
Looking forward for your reports/results :D

Thanks to donors  (https://www.magiclantern.fm/forum/index.php?topic=26250.msg239078#msg239078) who gave me the opportunity to work on 100D.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 13, 2022, 01:10:42 AM
I figured out what's causing these:

Quote from: theBilalFakhouri on June 06, 2022, 12:50:10 AM
-Quirks (not related to my build, but I have noticed it):

When selecting 3K and 1440p presets you need to turn off then on RAW video multiple times to let RAW video update RAW video resolution parameters.

Recording stops very early on 100D e.g. when recording high resolutions like 2560x1440p @ 23.976 FPS at 10 bit lossless with 192 MHz overclock compared to 700D (on 100D it's 1-2 seconds while on 700D it's up to 11 seconds, same settings, same scene, same everything). WTF? aren't both cams have same buffer size? (will look if there is a solution in other branches).


It's from DEFAULT_RAW_BUFFER_SIZE (in raw.c) which is too small, on 100D it's 5.2 MB vs 650D/700D/EOSM it's 24 MB!
I tried to increase it (I changed 100D buffer size value to 700D buffer size value) and above problems are no longer there, but of course this won't work, the camera will freeze (I made this test just to make sure if above issues are coming from DEFAULT_RAW_BUFFER_SIZE).

I doubled checked if 5.2 MB default buffer size on 100D might be a typo or something, and it seems correct :(
Here was the commit  (https://foss.heptapod.net/magic-lantern/magic-lantern/-/commit/9936e350480223bf1da5a9e198698d97a3049421)for it, details how we got it (https://www.magiclantern.fm/forum/index.php?topic=5601.msg196632#msg196632), I run a1ex test and it was reporting 5.2 MB in all video modes.

So for now:
-Can we change DEFAULT_RAW_BUFFER to another one (which has bigger size)? or it must be Canon RAW buffer?
-Can we increase DEFAULT_RAW_BUFFER_SIZE a little more (until above issues are no longer there), would this cause problem? how can we increase it in a safe way? from where 5.2 MB limit is coming from?

I need some help here from other Devs.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 13, 2022, 02:40:26 AM
Quote from: a1ex on March 07, 2018, 11:36:03 PM
Unless somebody starts porting the 4K and full-res LiveView on these cameras, it won't make any difference. ...

100D/EOSM2 x5: 2592 * 1108 * 14/8 = 5025888 (smaller than 5.2MB).

...

So now we need at least 3144 * 1337 * 14/8 = 7,356,174 = ~ 7.4 MB buffer size.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 15, 2022, 12:03:36 PM
-Some info:
Total shoot memory size is ~91 MB
Total  SRM  memory size is ~31 MB
Total:                             ~122 MB 

We are using both of these for RAW video recording (saving frames to memory before writing it to SD card), beside there is 24 MB from LV RAW buffer on 700D (5.2 MB on 100D) which is being used for RAW buffer.

I remembered that we actually allocate buffer size as needed in very high resolution (when RAW frame size is higher than LV RAW buffer size a.k.a DEFAULT_RAW_BUFFER)
like Full-Res LiveView on 700D or 1440p and 3K on 100D (we are not restricted to 24 MB on 700D or 5.2 MB on 100D, see (https://foss.heptapod.net/magic-lantern/magic-lantern/-/commit/7ad3754768b26a106d057550ca359089ba837fcf)):

On 700D when I use Full-Res LiveView preset (5208x3478), new RAW buffer gets allocated from shoot memory, and allocation size "RAW_LV_BUFFER_ALLOC_SIZE" is defined as "SRM_BUFFER_SIZE - 0x1000"
You can see how much is SRM_BUFFER_SIZE in consts.h for each model, on 700D/100D it's ~31 MB, so in this case LV RAW buffer will take ~31 MB from shoot memory and the remaining ~59 MB will get allocated by mlv_lite (for RAW video recording), and SRM memory won't be used for both LV RAW buffer and RAW video recording which is weird! why SRM memory won't be allocated for RAW video recoding?

which mean it's waste of ~31 MB of memory, probably the code in raw.c need some modifications (for SRM memory issue)?

Okay, back to 100D, when selecting 1440p (needs ~6.7 MB RAW buffer size) or 3K (needs ~7.4 MB RAW buffer size) presets, the new RAW buffer will get alloctaed from shoot memory, and you might already know that allocation size is defined as "SRM_BUFFER_SIZE - 0x1000" which is ~31MB which is higher than what we actually need!

So we are wasting ~24 MB of shoot memory, remaining would be 59 MB, and for some reason SRM memory won't be allocated by mlv_lite (already said that), so we are wasting total of ~55 MB of memory for RAW video recording, and that's why RAW video recording stops very early on 100D when recording at 3K or 1440p.

So after figuring out above things, I changed "RAW_LV_BUFFER_ALLOC_SIZE" to match RAW frame size (a tiny bit higher), from "SRM_BUFFER_SIZE - 0x1000" to "raw_info.width*raw_info.height*14/8 + 0x1000", by doing that we have left of ~83 MB from shoot memory, and also SRM memory will get allocated by mlv_lite (RAW video) :)

Which mean --> longer recording times at higher resolutions on 100D :D

-And regarding why you will need to turn off and on RAW video multiple times to update RAW video parameters:
That's because RAW video will allocate all shoot memory when it's on, turning off RAW video will free RAW video buffer and set shoot memory free then this thing will let crop_rec (and raw.c) allocate new RAW buffer from shoot memory (for high resolutions), then by turning on RAW video again will let it allocate remaining shoot memory.

More like a race condition, for sure this is fixable, need some code modifications for raw.c and mlv_lite.c and probably crop_rec.c.

-By figuring above things will open the door for some enhancements for Full-Res LiveView recording.

-Test build?
Sure: crop_rec_4k.2022Jun19.100D101.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jun19.100D101.zip)

Note: You will need to turn off and on RAW video multiple times after selecting 1440p or 3K presets, to let RAW video update its parameters!!
Please report for any issues
.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 15, 2022, 01:17:34 PM
-Uploaded the source code for Devs:
magic-lantern-bilal-100d (15-6-2022).7z (https://drive.google.com/file/d/1evYxe7KXE5tEKR1UBLGHzlfRXvfXy3HT/view?usp=sharing)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 15, 2022, 06:42:53 PM
Quote from: theBilalFakhouri on June 15, 2022, 12:03:36 PM
-Some info:
Total shoot memory size is ~91 MB
Total  SRM  memory size is ~31 MB
Total:                             ~122 MB 

Above results were when "Image quality" was set to JPG from Canon settings, read Image quality and its effect on memory (more RAM 4 free) (https://www.magiclantern.fm/forum/index.php?topic=26521.msg239231#msg239231) for more details.

So by setting "Image quality" to RAW then restarting the camera, you get:

Total shoot memory size is     ~95 MB
Total  SRM  memory size is     ~62 MB
Total:                                 ~157 MB 

(do your calculations for the rest of reply #10 (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239224#msg239224) :P)

--> Which has positive impact on RAW recording times.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 16, 2022, 03:35:48 PM
-Results on 100D (because of extra RAM (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239232#msg239232)):

-2560x1440p @ 23.976 FPS 10-bit lossless, 192 MHz SD overclock, sound recording on, lvface + aewb hacks are on, GD off (~70 MB/s required write speed reported by RAW video):

Before: ~17 seconds.
After:  ~50 seconds.

-Other scene (same settings):

Before: First try: ~30 seconds,  Second try: ~33 seconds.
After:   First try: 1 Minuets and 57 seconds,  Second try: 1 Minuets and 27 seconds.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Skinny on June 16, 2022, 04:20:54 PM
@theBilalFakhouri congrats! very nice results! :) almost two minutes is impressive for 2560x1440 resolution.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 16, 2022, 10:44:41 PM
Thanks @Skinny, I am happy that I could enhance recording times on 100D with high resolutions (it was only 1-2 seconds maximum at 2560x1440p with same above settings on 100D :P).

Could you do this test (https://www.magiclantern.fm/forum/index.php?topic=26521.msg239231#msg239231) on 5D2? also 550D because . . why not :).
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Skinny on June 18, 2022, 06:38:55 PM
@theBilalFakhouri sure, done: https://www.magiclantern.fm/forum/index.php?topic=26521.msg239289#msg239289
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 19, 2022, 03:43:04 AM
UHD 1280x2160 1x3 16:9 preset is alive on 100D :) , also 25 FPS now works in 4.3K 1x3 preset:

New build: crop_rec_4k.2022Jun19.100D101.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jun19.100D101.zip)

-UHD 1x3 probably won't have 25 FPS, this time it's a hardware limit (slow sensor speed).
-Also a new focus pixel map should be created for UHD 1x3 preset, some focus pixels in the bottom of frame aren't covered with exiting focus pixel map (editing current pixel map might be easy? sorry never played with them).
But there is a tutorial somewhere in the forum for creating focus pixel map, currently I will upload the exiting map to MLVApp repo, also it will be included here (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239140#msg239140).




It was CMOS 7 issues, not FPS Timer B:
-UHD 1x3: changed it from 0xB86 to 0xB66.
-4.3K 1x3: changed it from 0xB87 to 0xB67. (0xB87 works fine at 23.976 FPS, but 0xB67 was required to remove artifacts when using 25 FPS, I don't have such problem on 700D?)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: ArcziPL on June 19, 2022, 11:35:04 AM
Well done! You're doing a fantastic work!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: SebastianC on June 24, 2022, 03:06:10 AM
Thank you, Bilal! Great job!

Eos 100d 4k crop video(4.3k mode)
Up to 5-9 seconds every record.

https://youtu.be/qFi05EyY87o

Camera: Canon 100d
Lens: EF 35mm f2 is
crop_rec_4k.2022Jun19.100D101
SD card: Sandisk Ex pro 128GB 170mb.
4K Crop mode: 4080 × 1836

Music: The Easy Winners - E's Jammy Jams

I hope long to record because no mode has long record times, all modes have up to 15 seconds per time.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 24, 2022, 07:36:09 PM
@SebastianC

No problem, have fun :)

Quote from: SebastianC on June 24, 2022, 03:06:10 AM
I hope long to record because no mode has long record times, all modes have up to 15 seconds per time.

Actually all 1x3 presets are continuous but with these settings:

-Enable 192 MHz SD overclock
-Turn on the new hacks which are "lvface + aewb" from RAW video submenu

You might get continuous recording with these at 14-bit lossless (mostly depends on scene), but if you want 100% continuous recording set "bit-depth" to 12-bit from "Crop mode" submenu (you must keep 14-bit lossless in RAW video submenu).

3K and 1440p presets won't be continuous, you can lower its resolution or aspect ratio from RAW video submenu, also use lower bit depths in these presets.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: SebastianC on June 24, 2022, 08:07:40 PM
Quote from: theBilalFakhouri on June 24, 2022, 07:36:09 PM
@SebastianC

No problem, have fun :)

Actually all 1x3 presets are continuous but with these settings:

-Enable 192 MHz SD overclock
-Turn on the new hacks which are "lvface + aewb" from RAW video submenu

You might get continuous recording with these at 14-bit lossless (mostly depends on scene), but if you want 100% continuous recording set "bit-depth" to 12-bit from "Crop mode" submenu (you must keep 14-bit lossless in RAW video submenu).

3K and 1440p presets won't be continuous, you can lower its resolution or aspect ratio from RAW video submenu, also use lower bit depths in these presets.

I will try later today,Great! excited!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: SebastianC on June 25, 2022, 03:49:50 AM
Quote from: SebastianC on June 24, 2022, 08:07:40 PM
I will try later today,Great! excited!
I tested today:

4.3k :1440x1836,12bits,23.973fps,record up to 4m 45s.

1440p :2560x1440,10bits,record up to 1m 24s.

1736-976: record up to 2m 23s

Thank you for your help, Bilal! You are a genius!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 26, 2022, 10:56:24 AM
hi.
i performed some tests with the latest build "crop_rec_4k.2022Jun19.100D101".

This is the SD card i use:
SanDisk Extreme Pro, 32Gb, 95 MB/S UHS-1, Class 10, U3, V30

These are the settings i used:
SD overclock 192 MHz
Small hacks on + more hack "lvface + aewb".
Preview: realtime
Global draw: On, all modes + focus Peak on + histogram: RAW RGB, Log

these are the results:

-UHD 1X3 1280x2160 - No audio
12 bit Lossless, 23.973 fps
Frames 64, Duration 2 Sec

-4K 1x3: 1x3 Binning 1360x1976 - No audio
12 bit Lossless, 23.973 fps
Frames 65, Duration 2 Sec

-4K 1x3: 1x3 Binning 1360x1976 - No audio
10 bit Lossless, 25 fps
Frames 71, Duration 2 Sec

I suspect that my sd card is not compatible with the sd overclock module at all  :(

if you need I can upload the example mlv files and provide more details


sorry for my bad English
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on June 26, 2022, 11:09:42 AM
Load bench.mo in modules tab. Change overclocking frequency to 160 MHz. Restart camera.
Access Debug tab -> Benchmarks -> Card benchmarks -> Quick R/W benchmark (1 min) and start it.
If card defaults to 19-21 MByte/s -> Lower frequency, restart and retest.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 26, 2022, 12:35:41 PM
Quote from: Walter Schulz on June 26, 2022, 11:09:42 AM
Load bench.mo in modules tab. Change overclocking frequency to 160 MHz. Restart camera.
Access Debug tab -> Benchmarks -> Card benchmarks -> Quick R/W benchmark (1 min) and start it.
If card defaults to 19-21 MByte/s -> Lower frequency, restart and retest.
Thanks for the help
In the tests I get a write and read speed of about 21 mb/s in all three modes "160 MHz - 192 MHz - 240 MHz"
If I do the test in video mode I get these errors: Alloc error: buffer=16777216

(https://i.postimg.cc/x8d4nQQR/MVI-7866-MOV-snapshot-02-04-877.jpg)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 26, 2022, 01:04:38 PM
@GianlucaM83

160MHz overclocking is working fine here using my Sandisk Extreme PRO 32GB UHS-I U3 95MB/s.
You must restart your camera after changing SD overclocking frequency, if you didn't do that the setting won't be applied.

Let's start fresh:
-Unload all modules, load only bench.mo, restart the camera.
-Switch to Photo mode (no LiveView)
-Run cards benchmarks, share the result.
-Load sd_uhs.mo, restart the camera.
-Select 160 MHz, restart the camera.
-Run cards benchmarks, share the result.

Quote from: GianlucaM83 on June 26, 2022, 12:35:41 PM
If I do the test in video mode I get these errors: Alloc error: buffer=16777216

Turn off RAW video before running benchmarks in movie mode.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 26, 2022, 01:54:09 PM
Quote from: theBilalFakhouri on June 26, 2022, 01:04:38 PM
@GianlucaM83

160MHz overclocking is working fine here using my Sandisk Extreme PRO 32GB UHS-I U3 95MB/s.
You must restart your camera after changing SD overclocking frequency, if you didn't do that the setting won't be applied.

Let's start fresh:
-Unload all modules, load only bench.mo, restart the camera.
-Switch to Photo mode (no LiveView)
-Run cards benchmarks, share the result.
-Load sd_uhs.mo, restart the camera.
-Select 160 MHz, restart the camera.
-Run cards benchmarks, share the result.

Turn off RAW video before running benchmarks in movie mode.

ok. I followed your instructions. 
these are the results:

only bench.mo loaded:
Write speed 41.8 mb/s
Read speed 43.6 mb/s

(https://i.postimg.cc/NybGTcCD/IMG-Fresh-sd-off.jpg) (https://postimg.cc/NybGTcCD)

sd_uhs.mo Loaded (restarted camera), 160 MHz selected (restarted camera)
Write speed 21.1 mb/s
Read speed 21.8 mb/s

(https://i.postimg.cc/v1GHcBhD/IMG-SD160mhz-on.jpg) (https://postimg.cc/v1GHcBhD)

I cannot understand the reason for this result

further info I can provide
the sd is formatted in FAT32 and I did a deep format in camera before installing magiclantern

Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 26, 2022, 02:14:35 PM
Okay, thanks for re-testing.

In my custom 100D build I am using my 160 MHz preset which worked fine using my SD cards, let's test a1ex's 160 MHz preset with this build:
crop_rec_4k.2022Jun19.100D101_160MHz.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jun19.100D101_160MHz.zip)

Replace all ML files with the provided build (or make a clean install, format SD card in PC then install the provided build), test 160 MHz preset and see if it will drop to 21 MB/s or not, report back please.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 26, 2022, 02:52:50 PM
Thanks for your patience  :)
I did a clean install
unfortunately i get the same speed even with this version.
i think the problem is just my SD, maybe it is damaged.
i will buy a new one and come back with new information.
(https://i.postimg.cc/LqmCwK9r/IMG-7886.jpg) (https://postimg.cc/LqmCwK9r)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on June 26, 2022, 02:57:07 PM
Maybe giving this version a try first:
https://builds.magiclantern.fm/jenkins/job/crop_rec_4k/80/artifact/platform/100D.101/magiclantern-crop_rec_4k.2018Jul22.100D101.zip
There are several sd_uhs.mo variants to try.
Just copy it to ML\modules directory.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 26, 2022, 03:20:53 PM
@GianlucaM83

One last test from my side, test 160 MHz again with this build:
crop_rec_4k.2022Jun19.100D101_160MHz_SD_Configuration.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jun19.100D101_160MHz_SD_Configuration.zip)

Quote from: GianlucaM83 on June 26, 2022, 02:52:50 PM
i think the problem is just my SD, maybe it is damaged.

Maybe your 95 MB/s variant is different from mine . . or requires different overclock settings.

Quote from: GianlucaM83 on June 26, 2022, 02:52:50 PM
i will buy a new one and come back with new information.

Make sure it's Sandisk Extreme PRO 170 MB/s UHS-I U3 version which works with 192 MHz overclock.



Quote from: Walter Schulz on June 26, 2022, 02:57:07 PM
Maybe giving this version a try first:
https://builds.magiclantern.fm/jenkins/job/crop_rec_4k/80/artifact/platform/100D.101/magiclantern-crop_rec_4k.2018Jul22.100D101.zip

Let's test that build with official sd_uhs from a1ex:
https://drive.google.com/file/d/1xWYubqqr0DceaqMhXqiIovyF_7ep5vfc/view?usp=sharing

After running a1ex's sd_uhs version, a log would be saved in ML/LOGS folder in SD card (SD_UHS.LOG), please upload it somewhere.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 26, 2022, 03:35:53 PM
Ok done. this is the .LOG file
https://drive.google.com/file/d/1sP4CJKMZlBXwItMUhhkYXB1fLYEjs23E/view?usp=sharing (https://drive.google.com/file/d/1sP4CJKMZlBXwItMUhhkYXB1fLYEjs23E/view?usp=sharing)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 26, 2022, 03:46:20 PM
I used the sd_uhs from a1ex.  because i hadn't seen the first link with your build, sorry.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 26, 2022, 03:56:58 PM
Cool, so 160 MHz does actually work with your SD card. You need to check if it stable (doesn't drop to 21 MB/s), you can do this by recording RAW video or run Benchmarks multiple times in movie mode and photo mode.



Last thing I need you to test is:

Quote from: theBilalFakhouri on June 26, 2022, 03:20:53 PM
One last test from my side, test 160 MHz again with this build:
crop_rec_4k.2022Jun19.100D101_160MHz_SD_Configuration.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jun19.100D101_160MHz_SD_Configuration.zip)

If that worked, then I would know what causing the issue (the drop to 21 MB/s when using 160 MHz in my custom build).
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 26, 2022, 04:15:29 PM
I installed the latest test build you provided. Something has changed. it looks better now

Write speed 67.3 MB/s
Read speed 71.9 MB/s

(https://i.postimg.cc/LJDG3BLk/IMG-7889.jpg) (https://postimg.cc/LJDG3BLk)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 26, 2022, 04:18:40 PM
Nice, please make some RAW recording tests and see if it stable (doesn't drop to 21 MB/s).
Will post more details later.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 26, 2022, 04:56:09 PM
ok I'm back  :P
I recorded two raw videos. duration one minute per video, because the battery was getting low.
It's really stable, I think it could keep shooting continuously.

These are the settings

-UHD 1X3 1280x2160
12 bit Lossless, 23.973 fps
Frames 1504, Duration 1 Min, But I think it's continuous

-4K 1x3: 1x3 Binning 1360x1976
12 bit Lossless, 23.973 fps
Frames 1493, Duration 1 Min, But I think it's continuous

Bilal you are great  8)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 29, 2022, 09:01:44 AM
I had the opportunity to try the build more thoroughly.
The speed of the SD seems constant. I can record for several minutes without problems.
One problem I noticed is that, in every RAW video I have recorded, on the top of the frame only, there is some kind of intermittent noise and strange vertical stripes.
This is most visible in outdoor footage, instead in the recordings made at home the problem is very little noticeable.
This happens in both mlvapp and davinci resolve using the uncompressed dng files extracted with mlvapp.
The problem is also visible by opening the mlvs files with the FastcinemaDNG program.
I upload some MLVS test files.
Note that. The mlvs files I have uploaded contain only a few frames extracted from the original mlvs files.
I exported them from mlvapp using the MLV flatpass method. this is because the original files are too big to upload directly to google drive.
if you need other files or other tests I can do it
Thanks for the help

https://drive.google.com/drive/folders/1-COgSnAwgprsIif1iA1tzZ2iQcoonMWL?usp=sharing (https://drive.google.com/drive/folders/1-COgSnAwgprsIif1iA1tzZ2iQcoonMWL?usp=sharing)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 29, 2022, 10:21:02 AM
@GianlucaM83

Can you cut MLV files even further? file sizes are still a little huge to me.
2 Seconds per clip should be enough.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 29, 2022, 10:42:43 AM
Yes of course, sorry
Link to the same folder but with smaller files:
https://drive.google.com/drive/folders/1-COgSnAwgprsIif1iA1tzZ2iQcoonMWL?usp=sharing (https://drive.google.com/drive/folders/1-COgSnAwgprsIif1iA1tzZ2iQcoonMWL?usp=sharing)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 29, 2022, 11:07:58 AM
Thanks for reporting, it's CMOS 7 (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239303#msg239303) issue, a little tweak should fix the problem, I will release a new build soon.

Do you know if this problem happens in other presets? or when using 25 FPS with some presets?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 29, 2022, 11:21:01 AM
No problem.
unfortunately i only got to try these two presets at 23.973 fps. Maybe later or tomorrow i will be able to try the other presets at 25 fps too and I will report the results  :)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 29, 2022, 11:33:06 AM
Thanks, it would be cool if you test every preset so we can fix all issues if there are ones.
I did only test presets indoor, didn't notice problems. as you said, it seems to happen (the artifacts) at certain lighting conditions.

Here is a new build, hopefully will fix artifacts for 4K/UHD 1x3 presets:
crop_rec_4k.2022Jun29.100D101.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jun29.100D101.zip)

Please re-test mentioned presets, no need to hurry, take your time.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on June 30, 2022, 10:58:13 AM
I did some more tests with new build. I shot some raw video under the same lights conditions as last time.
Used only UHD 1X3 and 4k 1x3 presets in both 25fps and 23,973 fps.
Unfortunately, the results are the same as last time. On the top of the frame I have that strange effect. videos shot indoors have no problem.
I noticed a few videos with corrupted frames at the beginning.
I share the link to the folder where I uploaded all the mlvs files. Inside it you will find subfolders with the name of the used presets.
Take a look when you can
If you need more tests just ask  :)
https://drive.google.com/drive/folders/145oB6eJZbft6W7xLb768njG7Il9LAoUR?usp=sharing (https://drive.google.com/drive/folders/145oB6eJZbft6W7xLb768njG7Il9LAoUR?usp=sharing)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 01, 2022, 03:27:43 AM
Weird, I am unable to reproduce these artifacts on my 100D, tried with a lot of lighting conditions, there are no artifacts.

Could you share your *camera and ML settings? does these artifacts appear in LiveView too?
Are you using FPS override? or any other settings?

Is focus box centered on screen? e.g. turn off crop mode, press zoom button, make sure x1 printed on screen (by Canon overlays), press SET button to center the focus box, select UHD or 4K preset, artifacts are still there?

Quote from: GianlucaM83 on June 30, 2022, 10:58:13 AM
...videos shot indoors have no problem.

It actually does have the same problem, look closely at all indoor clips (above silver Coca Cola), in some areas the artifacts are there but they are not very visible.

Quote from: GianlucaM83 on June 30, 2022, 10:58:13 AM
I noticed a few videos with corrupted frames at the beginning.

How to reproduce the issue? :-\ .. I never had corrupted frames . .
*Your camera/ML settings please?

*For ML settings, please zip SETTINGS folder inside ML folder in SD card and upload it somewhere, but first make sure your settings do produce the artifacts.

I think I have another solution for the artifacts, used by Danne on EOS M for the same presets which is changing vertical image offset a little to the top, this way both UHD/4K 1x3 presets won't have accurate centering on sensor (I am trying to avoid this option) but at least it will clear the artifacts . . let's try this option as last resource, will share images show the current offset and the new offset if we decided to do it this way.




@100D users:
Do you have these artifacts when using 4K/UHD 1x3 presets?

It looks like this (vertical lines in the top of frame):

(https://i.ibb.co/hYJZ8BL/Artifacts.png)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 01, 2022, 09:20:07 AM
I am not using any particular setting.
Almost all settings are the default ones except these:
I only enabled the "focus peack" and " raw histogram" in the overlay section. In the RAW video section I activated the two options "small hack" and "more hacks lvface+aewb". Fps ovveride is not active.
In the canon menu I set the video format to PAL and the recording format to 1920x1080 25 fps.
I have tried disabling crop mode, putting the focus rectangle in the center in 1x mode and re-enabled crop mode but the problem remains.
I noticed that these artifacts are also visible in the preview on the live view, before I start recording and while I am recording.
These are the modules I have enabled:
crop_rec - dual_iso - mlv_lite - mlv_play - mlv_snd - sd_uhs.

This is the configuration file:
https://drive.google.com/file/d/1H-n_Q9_rn-ioQV_99a5FQO4MHXrV3TyP/view?usp=sharing (https://drive.google.com/file/d/1H-n_Q9_rn-ioQV_99a5FQO4MHXrV3TyP/view?usp=sharing)

Thanks
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 01, 2022, 12:21:27 PM
Disable histogram.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 01, 2022, 07:06:32 PM
@Danne
I did it but nothing changes. I also disabled everything in the overlay section. I also tried a fresh installation but the artifacts remain.
I leave a short clip with the operation I usually do to record raw video. to see if I missed something.



One more thing, I tried the 1920x1280p 1:1 preset it seems there are no artifacts with it.
https://drive.google.com/file/d/1YUVXB5CODWrKc8VtWEDrEE326ojRB6hn/view?usp=sharing (https://drive.google.com/file/d/1YUVXB5CODWrKc8VtWEDrEE326ojRB6hn/view?usp=sharing)


I personally think it would be helpful if some other users would also provide their feedback on this.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 01, 2022, 11:17:31 PM
Thanks for the video, I will provide some more tests to run later, but for now please try:

-Set White Balance to specific value instead of Auto.
-Switch to Photo mode, go to ML settings, Expo tab and turn on ExpSim.
-Restart camera.

Now try UHD or 4K preset, artifacts are still there?
If yes, I am going to provide another set of tests.

-Does 4.5K and 4.3K 1x3 presets work without these artifacts? try 23.976 and 25 FPS.
-If you want to upload MLV clips, please cut them to 5 frames instead of two seconds, that would be enough too . .

I started to suspect that it could be your camera sensor is slightly different from mine (slight hardware differences?), I could be wrong of course . . BTW, where was your 100D manufactured? (you can read this on sticker in the bottom of camera)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 02, 2022, 08:17:57 AM
Okay I did this:
- I switched to photo mode.
- I turned on expsim mode in the expo tab.
-I set the white balance to a personal value (5200k).
-restarted the camera
- I recorded some raw video in all 1x3 modes, both 24.973 fps and 25 fps
Unfortunately, artifacts are visible in all modes  :(
I put the link to mlvs file, this time 5 frames per clip.
https://drive.google.com/drive/folders/1IQ1bpWKOI9T0aRATwlj_1tfRJxkmpViS?usp=sharing (https://drive.google.com/drive/folders/1IQ1bpWKOI9T0aRATwlj_1tfRJxkmpViS?usp=sharing)

Regarding the possibility that our cameras have a slightly different sensor, perhaps that could be true, but I don't know much about it.
On the bottom of my camera is a little label that says: Made in TAIWAN, and the serial number.
I don't know if this might be relevant, maybe it doesn't make any difference but, I think it is a camera for the European market. i think this because some time ago, I sent the camera to a service center for auto focus calibration of the kit lens.
The repair was under warranty and the camera was sent by Canon to a lab in Germany.

Edit: I did some more tests at home with low artificial light, the artifacts are almost not visible  :o
I think this is going to drive me crazy :P
https://drive.google.com/drive/folders/1_acY0asl8sB-pPAq7ju74X2mASjPJP71?usp=sharing (https://drive.google.com/drive/folders/1_acY0asl8sB-pPAq7ju74X2mASjPJP71?usp=sharing)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 02, 2022, 12:56:32 PM
Could you test this version?
https://bitbucket.org/Dannephoto/magic-lantern_dannephoto_git/downloads/crop_rec_4k_mlv_snd_isogain_1x3_presets_100D_2022Jul02.100D101.zip

Doesn´t include bilals later stuff but you could try if sensor behaves good in general.

Source:
https://bitbucket.org/Dannephoto/magic-lantern_dannephoto_git/src/master/

Go to this branch:
crop_rec_4k_mlv_snd_isogain_1x3_presets_100D
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 02, 2022, 04:19:11 PM
Hi Danne
I installed your build.
I did some testing, at home for now.
But I need help to understand which presets I should use and how I should set them. Because the camera crashed a few times during recording and in the clips I made there are many corrupted frames.
I think it's my fault, I got used to bilial's build which had slightly different settings.
link with the various crash logs:
https://drive.google.com/drive/folders/1flNTGxPTwwTybtRWXZYDTQ5kQqRcHn1U?usp=sharing (https://drive.google.com/drive/folders/1flNTGxPTwwTybtRWXZYDTQ5kQqRcHn1U?usp=sharing)

However, I was able to record some clips.
With your build, in these lighting conditions there seems to be no artifacts on the top of the frame, but I think it is better to do some more tests in other lighting conditions as well, and set more appropriate settings to avoid the corrupted frames and crashes
Some MLVs test files:
https://drive.google.com/drive/folders/12Y-iMQ1-m7xHrJPqAFiZaQN4dLbHsJG9?usp=sharing (https://drive.google.com/drive/folders/12Y-iMQ1-m7xHrJPqAFiZaQN4dLbHsJG9?usp=sharing)

Thanks guys
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 02, 2022, 04:37:46 PM
Did you test regular 14bit vs 12 or 10bit in Bilals build? Looks a little like gain registry issue but mostly guessing here.
Would be nice to have access to bilals code online for faster checking.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 02, 2022, 04:59:54 PM
@GianlucaM83

Thanks for testing, I will provide more things to test (later today or tomorrow).
Hardware differences was a guess, could be wrong, my 100D made in Taiwan too. I don't think it's something related to auto focus calibration.

@Danne
My 100D doesn't have any artifacts and my tests was done with the same build and settings used by GianlucaM83.
I don't think it's Analog gain problem, well, @GianlucaM83 could redo your tests with in 14-bit lossless only:

-Set 14-bit lossless from RAW video submenu
-Turn off bit-depth option from "Crop mode"
-Restart camera
-Test UHD and 4K 1x3 presets

-Artifacts are still there?

Also, could you test 3K and 1440p 1:1 presets?
So far the issue seems affect only 1x3 Binning modes and only GianlucaM83's 100D.

More things are coming from my side soon, currently a little busy.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 02, 2022, 05:07:07 PM
I'll grab my 100 soon and join the testing squad  8)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 03, 2022, 10:30:14 AM
-I tested UHD 1X3 and 4k 1x3 presets in 14 bit lossless. artifacts still appear  :(
-I tested the 3k and 1440p 1:1 presets. There are no visible artifacts. At least I can't see them. maybe the 1:1 mode don't give problems on my camera.  :)

Obviously to get a good recording time I had to reduce the bits to 10, use 23.976 fps and lower the resolution.
As usual I put some sample mlv files made with both presets at different resolutions and bit depht. take a look when you can.
Thanks

1440p 1:1 preset:
https://drive.google.com/drive/folders/1kgkTOUrFuPfsMWtJ3_cgCHkFS2CAspjh?usp=sharing (https://drive.google.com/drive/folders/1kgkTOUrFuPfsMWtJ3_cgCHkFS2CAspjh?usp=sharing)

3k 1:1 preset:
https://drive.google.com/drive/folders/1_pfTlm_dCFb3NkO-dw6yzbkbsJ3USpaE?usp=sharing (https://drive.google.com/drive/folders/1_pfTlm_dCFb3NkO-dw6yzbkbsJ3USpaE?usp=sharing)

Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 03, 2022, 10:51:39 AM
Tested all 1x3 presets. Hats off to you bilal and all 100D owners. THose presets you created are perfectly working both 24 and 25fps. Seems very stable too. Not like my build ;).
I keep nagging about code but you should keep a fresh codebase along with your builds. Equally imprtant as posting builds.
Keep up the good work!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 03, 2022, 11:16:58 AM
By the way. I think I know how to fix GianlucaM83 issue. Had a similar issue with 5d3. I need the crop_rec.c code to check.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 03, 2022, 11:41:58 AM
Guys your work is amazing  8)
I am very happy that the build works well on your cameras and for all the users who have this great camera  :)

Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 03, 2022, 03:56:52 PM
Thanks @Danne :)

Quote from: Danne on July 03, 2022, 10:51:39 AM
I keep nagging about code but you should keep a fresh codebase along with your builds. Equally imprtant as posting builds.

Yeah, you are right, currently I am preparing for a Github solution (will re-write 650D/700D code then apply the new changes maybe in separate branch for 100D), that would fix this problem.

Quote from: Danne on July 03, 2022, 11:16:58 AM
By the way. I think I know how to fix GianlucaM83 issue. Had a similar issue with 5d3. I need the crop_rec.c code to check.

Here is the new fresh code:
magic-lantern-bilal-100d (29-6-2022).7z (https://drive.google.com/file/d/1dLCSrN5YXb0jBukYW3tE9jwkOxuMhSY1/view?usp=sharing)

I am aware of the fix could be changing the offset on sensor a little to the top, but I am trying to avoid this (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239495#msg239495) currently.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 03, 2022, 04:13:24 PM
Not offset, other regs.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 03, 2022, 04:17:43 PM
Okay, waiting your Magic :)

@GianlucaM83

I am looking into 1440p/3K right now, thanks for the tests again. Also I am preparing new tests as backup plan.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 03, 2022, 05:41:06 PM
Please try this build @GianlucaM83
https://bitbucket.org/Dannephoto/magic-lantern_dannephoto_git/downloads/crop_rec_4k_2022Jul03.100D101.zip

I changed registry to this in crop_rec code:
        /* PowerSaveTiming registers */
        /* after readout is finished, we can turn off the sensor until the next frame */
        /* we could also set these to 0; it will work, but the sensor will run a bit hotter */
        /* to be tested to find out exactly how much */
        adtg_new[4]  = (struct adtg_new) {6, 0x8172, nrzi_encode(readout_end + 1) }; /* PowerSaveTiming ON (6D/700D) */
        adtg_new[5]  = (struct adtg_new) {6, 0x8178, nrzi_encode(readout_end + 1) }; /* PowerSaveTiming ON (5D3/6D/700D) */
        adtg_new[6]  = (struct adtg_new) {6, 0x8196, nrzi_encode(readout_end + 1) }; /* PowerSaveTiming ON (5D3) */
        adtg_new[7]  = (struct adtg_new) {6, 0x8173, nrzi_encode(fps_timer_b - 5) }; /* PowerSaveTiming OFF (6D/700D) */
        adtg_new[8]  = (struct adtg_new) {6, 0x8179, nrzi_encode(fps_timer_b - 5) }; /* PowerSaveTiming OFF (5D3/6D/700D) */
        adtg_new[9]  = (struct adtg_new) {6, 0x8197, nrzi_encode(fps_timer_b - 5) }; /* PowerSaveTiming OFF (5D3) */
        adtg_new[10] = (struct adtg_new) {6, 0x82B6, nrzi_encode(readout_end - 1) }; /* PowerSaveTiming ON? (700D); 2 units below the "ON" timing from above */
       
        /* ReadOutTiming registers */
        /* these shouldn't be 0, as they affect the image */
        adtg_new[11] = (struct adtg_new) {6, 0x82F8, nrzi_encode(readout_end + 1) }; /* ReadOutTiming */
        adtg_new[12] = (struct adtg_new) {6, 0x82F9, nrzi_encode(fps_timer_b - 1) }; /* ReadOutTiming end? */



Sidenotes. I see you work from  crop_rec_4k branch. I would prefer working from crop_rec_4k_mlv_snd as it has g3gg0 mlv_snd fixes. Anyway. Please test and report back. Very short of time over here unfortunately.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 03, 2022, 05:49:41 PM
Quote from: Danne on July 03, 2022, 05:41:06 PM
Sidenotes. I see you work from  crop_rec_4k branch. I would prefer working from crop_rec_4k_mlv_snd as it has g3gg0 mlv_snd fixes.

I have added them to crop_rec_4k branch a long time ago (look at previous commits).



I will check your fix and play with it on my 100D.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 03, 2022, 08:28:48 PM
Thanks Danne
I downloaded the build and repeated the tests.
I noticed that there are some new modules in this build.
I only loaded the ones I had already loaded in the previous builds.
Unfortunately the artifacts are still there in the 1x3 modes.
One thing I noticed but maybe it is not important. In all the builds I have tried.
When I activate the crop mode in the video mode. On the liveview of the camera I see the same artifacts that are in the recorded video.
While recording a raw video I see at the top of the live view the same colored, flashing bands that I see later in the video I recorded.
Currently the preview is set to realtime. if I try to change it to another option it resets itself when I exit the magic lantern menu.
i don't know if i explained myself correctly. my English sucks.

Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on July 03, 2022, 08:55:05 PM
Hm, I think it's related to powertime registry but hard to know for sure. Is it always happening or only sometimes?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: SebastianC on July 03, 2022, 08:57:49 PM
Thank you Danne! I will try new !
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 03, 2022, 09:04:58 PM
Danne.
It happens every time and only when i enter the 1x3 crop mode in video mode.
if i activate the 1440p 1:1 preset there are no visible artifacts on the liveview or even in the video i recorded. maybe it is just a coincidence
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 03, 2022, 09:29:51 PM
Quote from: GianlucaM83 on July 03, 2022, 08:28:48 PM
I noticed that there are some new modules in this build.
I only loaded the ones I had already loaded in the previous builds.

The new modules are there in repo, it either not supported yet or it's for development stuff, don't use them.
I always delete them before uploading a new build for 100D, it's okay for now because we are providing testing build, and these builds won't be used after we solve the problem.

Quote from: GianlucaM83 on July 03, 2022, 08:28:48 PM
One thing I noticed but maybe it is not important. In all the builds I have tried.
When I activate the crop mode in the video mode. On the liveview of the camera I see the same artifacts that are in the recorded video.
While recording a raw video I see at the top of the live view the same colored, flashing bands that I see later in the video I recorded.

Seems normal, these type of artifacts should be shown in both LiveView and recorded RAW data. (why you have these artifacts on your 100D isn't normal)

Quote from: GianlucaM83 on July 03, 2022, 08:28:48 PM
Currently the preview is set to realtime. if I try to change it to another option it resets itself when I exit the magic lantern menu.

Yeah, for presets which have correct real-time preview there is no need for Framing preview, I made it like that.
Probably I should make an option for the user too, to let him select which preview he want to use.



Quote from: SebastianC on July 03, 2022, 08:57:49 PM
Thank you Danne! I will try new !

Do you have the same artifacts  (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239495#msg239495)which GianlucaM83 has?
If no, there is nothing new to test here.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 03, 2022, 09:43:05 PM
just another little test I don't know if it can make sense.
maybe it's the summer heat playing tricks on me  :o

-UHD 1X3 preset, 12 bit, 23,976 fps.

400 ISO, artifacts are visible
https://drive.google.com/file/d/1P8o22RTubL9hY1vmsD2z7yQ7BGjsephS/view?usp=sharing (https://drive.google.com/file/d/1P8o22RTubL9hY1vmsD2z7yQ7BGjsephS/view?usp=sharing)

800 iso, artifacts are not visible
https://drive.google.com/file/d/16Q3pZ2neZvQb5dPC3Xa9_S-RVlQL9dCu/view?usp=sharing (https://drive.google.com/file/d/16Q3pZ2neZvQb5dPC3Xa9_S-RVlQL9dCu/view?usp=sharing)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 03, 2022, 09:58:43 PM
@GianlucaM83 let's try these:

First, backup your camera ROM files, you will find them in ML/LOGS folder, copy them to a safe place (also let me have them, zip them then upload it somewhere and send me the link via personal message), please Don't share the link publicly, it's copyrighted material.

Tests:

1- Let me take a look into your 100D registers:

-Download this adtg_gui.mo (https://drive.google.com/file/d/1QfBQkORaKQ-cT9-Y4QOm-mpsHhecFOmm/view?usp=sharing), put it in ML/modules folder in SD card
-Unload all modules, and load only adtg_gui.mo, restart camera
-Set video mode to 1080p24 to gain extra speed

-Follow the following video, you can see in the video I am entering Canon menu and toggling between x1/x5/x10 modes, you should do that too to let "ADTG registers" bring the registers into it:
https://www.youtube.com/watch?v=Q-Tdu9YZ4fc

-After you finish, you will find "adtg.log" file in ML/LOGS folder in SD card, upload it somewhere and share it.

2- Try this build:
crop_rec_4k.2022Jul03.100D101_delay.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jul03.100D101_delay.zip)

-Just install the build and try UHD or 4K 1x3 presets and feedback please.
-In this build I made the delay longer before refreshing LiveView from 100 MS to 500 MS (MS = Milliseconds), I need to make sure if it's not a race condition.

3- Update your camera firmware to this one:
https://pel.hu/down/v101-sl1-100d-x7-win.zip

-Format your SD card in PC (remove ML)
-Put CCF16101.FIR in SD card
-Make sure battery is fully charged
-Run firmware update process

Edit:
Note: I don't know if this would affect or reset autofocus calibration data (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239517#msg239517) for your kit lens!!

-Then install ML build and see if UHD or 4K 1x3 presets still have the artifacts or not.


4- I will prepare a build for manual CMOS 7 fine-tuning, will post it with more details later.

Please make these tests in order.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 03, 2022, 11:58:37 PM
Artifacts were reproduced on my 100D :) :

Thanks to ADTG registers LOG provided by @GianlucaM83, I noticed 5 ADTG registers which have different values on GianlucaM83's 100D, which are:

ADTG2[c00d]:    0x5249
ADTG2[c00e]:    0x1
ADTG2[c00f]:     0x1
ADTG2[c010]:    0x2
ADTG2[c011]:    0x2

-On my 100D it's:

ADTG2[c00d]:    0x5000
ADTG2[c00e]:    0x53
ADTG2[c00f]:     0x52
ADTG2[c010]:     0x52
ADTG2[c011]:     0x52

I copied GianlucaM83's 100D values to my 100D, then the artifacts appeared in 1x3 presets.
These registers are gain registers.

-Let's try my 100D values on GianlucaM83's 100D using this build:
crop_rec_4k.2022Jul04.100D101_ADTG.zip (https://bitbucket.org/bilal_fakhouri/magic-lantern/downloads/crop_rec_4k.2022Jul04.100D101_ADTG.zip)

@GianlucaM83
Please try provided build and 1x3 presets then feedback please.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 04, 2022, 12:30:26 AM
With this latest build the artifacts seem to be gone.
I will try tomorrow in other lighting conditions. But I think this is the right way.
Thanks Bilal and also to the others who helped me  :)
You guys are great  8)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on July 04, 2022, 08:32:38 AM
Today I tested the new build:"crop_rec_4k.2022Jul04.100D101_ADTG".
In the same lighting conditions where I had artifacts before.

-Tested all 1x3 presets. 12 bit, 23.976 fps, 25 fps.
The artifacts are no longer visible  :)
I put the link with the test MLV files:
https://drive.google.com/drive/folders/19E3IRpgkBfjoIsFjB7MfDkf6fDS4xKcO?usp=sharing (https://drive.google.com/drive/folders/19E3IRpgkBfjoIsFjB7MfDkf6fDS4xKcO?usp=sharing)

Bilal you did it  8)
Thanks!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: DeafEyeJedi on July 08, 2022, 12:44:56 AM
Wonderful progress guys!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on September 10, 2022, 03:53:18 AM
Quote from: theBilalFakhouri on June 06, 2022, 12:50:10 AM
-What's not working:

SD Overclocking:
During my tests 192 MHz SD overclock will only work with Sandisk Extreme PRO UHS-I U3 170 MB/s version, and it seems stable.

240 MHz doesn't work even when using Sandisk Extreme PRO UHS-I U3 170 MB/s.
Both 192 MHz and 240 MHz don't work when using Sandisk Extreme PRO UHS-I U3 95MB/s.

New sd_uhs update:
https://www.magiclantern.fm/forum/index.php?topic=26634.msg240128#msg240128

State:
240 MHz does work to some point (https://www.magiclantern.fm/forum/index.php?topic=26634.msg240127#msg240127) using Sandisk Extreme PRO UHS-I U3 170 MB/s.
192 MHz now works on my Sandisk Extreme PRO UHS-I U3 95MB/s.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on September 10, 2022, 09:44:08 AM
First of all, thank you Bilal and everyone who contributed to make this new module free for everyone  :)
I did a bench with my new sd card: Sandisk Extreme Pro 170 MB/s U3.
Seems that it can handle the 192 MHz preset correctly, instead with the 240 MHz preset it drops to 20MB/s.
Before this update, the same SD only worked with the 168 MHz preset. So this seems like a huge improvement to me.
Later I will also try it with my old  Sandisk Extreme PRO UHS-I U3 95MB/s
Thanks again!

Sandisk Extreme Pro 170 MB/s U3: New sd_hs module.192MHz preset:
(https://i.postimg.cc/7L0gJfB7/bench192mhz.jpg)

Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on September 10, 2022, 09:57:42 AM
Quote from: GianlucaM83 on September 10, 2022, 09:44:08 AM
I did a bench with my new sd card: Sandisk Extreme Pro 170 MB/s U3.
Seems that it can handle the 192 MHz preset correctly, instead with the 240 MHz preset it drops to 20MB/s.
Before this update, the same SD only worked with the 168 MHz preset. So this seems like a huge improvement to me.

Cool, what is your card capacity for Sandisk Extreme Pro 170 MB/s U3?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on September 10, 2022, 10:12:46 AM
I bought the 64 GB version
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: SebastianC on September 10, 2022, 04:21:28 PM
My 100D 3K can record more than 20-30seconds now with new 240 MHz.

Before I just only can record less than 5-10 seconds.

Good job Bilal!
Thank you for your work!
:)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on September 15, 2022, 10:51:41 AM
Another better 240 MHz for 100D:
https://www.magiclantern.fm/forum/index.php?topic=26634.msg240328#msg240328

Quote from: theBilalFakhouri on September 15, 2022, 08:28:05 AM
-It works better at high temps using my Sandisk 170 MB/s and 100D, it didn't drop to 21 MB/s after 5 minuets of recording :D
-SDR104 patch which required to enable high OC for some cards now works with the newest 240 MHz preset, and it's stable
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on September 16, 2022, 05:27:23 AM
I did a bench with the newest_sd_uhs_SDR104_patchmodule.
The 240 MHz preset now works on my SD as well  8)
Thanks!

This is the result:
240 MHz Preset
Sandisk Extreme Pro 170 MB/s U3 64GB

Write speed: 83.7 MB/s
Read speed:  85.7 MB/s
Write speed: 83.5 MB/s
Read speed:  85.7 MB/s

(https://i.postimg.cc/d1sJX8sJ/bench-240-mhz-newestmodule-16-09-2022.jpg)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on September 16, 2022, 06:27:07 AM
@Bilal: 10 MByte/s less than rest of the bunch? Did you get simiiar results?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on September 16, 2022, 08:59:59 AM
@GianlucaM83 Awesome, thanks for the test!

@Walter
Yeah it seems a little slow at 240 MHz (but still higher than 192 MHz (https://www.magiclantern.fm/forum/index.php?topic=26511.msg240231#msg240231)), for me I can get up to ~90 MB/s using my 170 MB/s SD card on 100D (same on 700D, my SD card limit is ~90 MB/s).

(https://i.ibb.co/VSgLgKT/bench0.png) (https://ibb.co/7WnSn9K)


@GianlucaM83
Your SD card should perform around ~90 MB/s (or a little higher) with 240 MHz. I suggest to run benchmarks with fast card reader on PC.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on September 16, 2022, 10:05:54 AM
I did a beanchmark of my SD card on my PC as you suggested. I used My Kingston FCR-HS4 card reader which is usb 3.0, connected directly to a usb 3.0 port on the motherboard.
I used the CrystalDiskMark program with the default settings and the aja system software. Unfortunately, I do not know any other softwares.
These are the results:
(https://i.postimg.cc/s2f9wgBb/IMG-20220916-084432.jpg) (https://i.postimg.cc/wvyNGy94/bench-PC.jpg)
what do you think about these results, do you think they are in the normal range for my SD?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on September 16, 2022, 10:28:36 AM
Yes, they are.
Your Kingston FCR-HS4 (as mine) does not know anything about overclocking and runs pure UHS-I protocols as specified by SD Card Association. Means: UHS-104 is the limit.
You can compare your numbers with https://www.cameramemoryspeed.com/reviews/sd-cards/sandisk-extreme-pro-170mbs-uhs-i-u3-v30-64gb-sdxc-memory-card/
In this test all cardreaders but one are running non-proprietary transfer modes. The only one using overclocking stands out.

Maybe Bilal can get behind it.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on September 16, 2022, 10:35:04 AM
Quote from: Walter Schulz on September 16, 2022, 10:28:36 AM
Yes, they are.
Your Kingston FCR-HS4 (as mine) does not know anything about overclocking and runs pure UHS-I protocols as specified by SD Card Association. Means: UHS-104 is the limit.
You can compare your numbers with https://www.cameramemoryspeed.com/reviews/sd-cards/sandisk-extreme-pro-170mbs-uhs-i-u3-v30-64gb-sdxc-memory-card/
In this test all cardreaders but one are running non-proprietary transfer modes. The only one using overclocking stands out.
Thanks Walter Schulz for the explanation  :)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on September 16, 2022, 12:09:30 PM
Thanks for the benchmarks.

I don't always trust CrystalDiskMark benchmarks (at least for fake SD cards), it wasn't really accurate (https://www.magiclantern.fm/forum/index.php?topic=24315.msg218540#msg218540) when I bought fake (https://www.magiclantern.fm/forum/index.php?topic=24315.msg218496#msg218496) Sandisk Extreme PRO 170 MB/s SD cards.
The other software (AJA) benchmark seems closer in term of write speed benchmarks compared to camera benchmark. Of course ignore read benchmarks (because we are using 192 MHz for reads on 100D when using 240 MHz).

Could you also test with Black Magic Disk Speed Test? Here is a download link (https://drive.google.com/file/d/1OsKTt-GJdGb0ngPBURmZdXlP2nncTtug/view?usp=sharing) from my Google Drive (I don't know where is the official download link)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: GianlucaM83 on September 16, 2022, 02:07:04 PM
I did a test with Disk Speed Test.
Write speed is 85.5 MB/s
(https://i.postimg.cc/PJMGDMPB/Disk-Speed-Test.png)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Skinny on September 16, 2022, 08:35:02 PM
@theBilalFakhouri congrats with new achievements, cool stuff guys! Do 300mhz, why stop at 280 :D
@GianlucaM83 nice virus ti :) I can see a lot of musicians here on ML forum for some reason :o
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on September 16, 2022, 08:41:34 PM
Quote from: Skinny on September 16, 2022, 08:35:02 PM
Do 300mhz, why stop at 280 :D

Because next stop is not 300 but 384 and that is a tough call.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on September 17, 2022, 01:59:04 AM
@GianlucaM83
Thanks for the benchmarks, your SD card is doing on average (88 + 82 + 85.5) / 3 = ~85 MB/s which is close to 83.7 MB/s from camera.
If we exclude CrystalDiskMark, (82 + 85.5) / 2 = 83.7 MB/s. Well, I don't think this is an accurate way to do measurements, also because SD card frequency is unknown using PC card reader.

Last suggested test:
Perform low level format in camera, do benchmarks in Play mode at 240 MHz.

But anyway, in all cases 240 MHz is working fine here, the slower write speed in 240 MHz could be from your SD card or your 100D.
The only accurate way to verify is to test another Sandisk 170 MB/s SD card on your 100D and test your current Sandisk 170 MB/s SD card on other DIGIC 5 model at 240 MHz.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on September 17, 2022, 02:05:52 AM
@Skinny

Thank you!

Yeah as Walter said, other frequencies above 240 MHz are 280 MHz and 384 MHz. Will double check later if we can select something among 280-384 MHz range.
We are so lucky that our Canon cameras (DIGIC 5 models) from ~10 years ago support unofficial UHS-I speeds! Impressive work Canon ;D

Well, I must get 280 MHz stable (if possible) then think about higher than 280 MHz frequencies possibilities.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on September 17, 2022, 02:27:31 AM
Next plans for 100D port:

At current stage I think I got most things from my 650D / 700D port to 100D, main things were:

-Crop mode (crop_rec) presets
-New hacks for RAW video
-And recently 240 MHz SD overclocking
-Memory allocation fix for high resolutions
-And other things

I need to check HDMI state too for some crop mode presets.

Remaining thing is Crop mode V2 (crop_new) module, more likely I will not port it. on 700D it's still confuse users, also the module isn't 100% stable (causes corrupted frames at some settings).
My alternative idea for this is to make more stable presets for Crop mode (crop_rec). Will discuss later what new presets to include in crop_rec, it will be something like more Time-Lapse modes and high frame rate modes.

Well, I might give crop_new a try on 100D.

100D users: feel free to give suggestions, your opinions and to discuss. please don't ask for new features, only if it related to my 650D / 700D port.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: thatguyjay on November 17, 2022, 12:08:48 AM
Hey guys, I'm a noob poking into this subject matter. I had an SL1 in my closet and just heard about all this RAW shooting capability added. Thank you for all your work on this. I can't seem to get my settings to work optimally and I thought I could use some advice.

I installed the latest build I could find: crop_rec_4k.2022Jul04.100D101_ADTG

I'm using a 128GB SD card like the one in the pic above: Sandisk Extreme Pro

I was only able to find one setting that worked to record more than 5 seconds before buffering/crashing the camera or showing pink stripes in LiveView: 1280x720 10-bit.

I also tried turning on crop mode but that seemed to be an additional X factor.

I guess there are some modes that work and some that don't and it has to do with the speed of the SD card? Any advice on recommended settings that will get me to at least UHD would be very helpful.

I'm trying to get to at least 1920x1080 to use this as a B camera with some interviews IDK if I'll have an answer by then, but I'm shooting Monday. I would be happy to support development on the SL1 if it would help get us over the finish line for this camera's capabilities.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 22, 2023, 03:50:40 PM
Dear all,

first of all thank you for creating such incredible additional value to the Canon cameras! This is fantastic.

Although I'm completely new to this, I managed to successfully install ML, update it with this crop_rec build from June 19th (is this the latest, recommended build?), tune the speed of my SD card by overlocking to 85MB/s write speed (Sandisk Extreme Pro 128GB 200MB/s @240MHz, thanks to the latest sd_uhs update).

What I still don't understand is how to achieve the maximum resolution for RAW videos. No sound needed and also low frame rates are acceptable for me (astro photography). But many crop settings result in slow camera response, corrupted live view, reduced field of view and even freeze the camera.
Sometimes a mode seems to work, I can access higher resolutions and make continuous video. The next time I try to do the same it says max resolution limited to xy, no matter what I try.

Is there a tutorial for dummies like me on how to set up the camera correctly for high res raw videos? I'm quite sure it is all about the right settings or sequence of setting options to make this work.

I really tried to read through the forum and go deeper into each thread that seems to point into the right direction, but with many threads having >50 pages of answers it is kind of hopeless.

Of cause I tried the forum search as well, but either my problems are kind of unique (which I doubt) or I searched for the wrong key words.

Can someone point me into the right direction, please?

Best regards,

Mike
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on February 22, 2023, 04:55:33 PM
Quote from: soundfan on February 22, 2023, 03:50:40 PM
Sometimes a mode seems to work, I can access higher resolutions and make continuous video. The next time I try to do the same it says max resolution limited to xy, no matter what I try.

https://wiki.magiclantern.fm/faq#raw_mlvcan_t_set_resolution_to_full-hd1920_is_not_possible_in_current_video_mode_max_1728
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 22, 2023, 06:39:23 PM
Thanks for the lightning fast reply, Walter.

The Wiki I read of cause but as I understand it this is the way to (slightly) higher resolutions in the original ML nightly build, right? This is also limited to max 2520.

Beside that it records only 2 seconds now, although my SD supports up to 85MB/s @240MHz. Is the higher write rate only supported in the experimental crop mode?

I was trying to do 4K recording with the crop_rec module. And this caused the problems mentioned before.

Also I recognized that the camera heats up very fast when testing with live view and playing with the ML menu. Is that normal?

Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 22, 2023, 08:02:59 PM
Well, I think I solved the riddle. For sure I read the introducing post of Bilal's crop_rec_4k experiments for 650D / 700D (T4i / T5i) ten times but didn't pay attention to the probably most important sentence of all:

QuoteGeneral notes:
-You can load both crop_rec and crop_new modules, but don't activate them at the same time (use "Crop mode" or "Crop mode V2").
-Use maximum aspect ratio and and resolution from mlv_lite when using crop_rec presets with real-time correct preview, otherwise mlv_lite won't crop the RAW data correctly.

With this pre-settings everything works now as expected.

Only thing I recognized is that write speed is only stable as long as camera temp is below ~45°C. Beyond that anything can happen. Now that I don't need to try 'n error for hours anymore I guess that overheating will no longer be an issue.

I was sure that the solution was lieing right in front of me and I was just too blind to see...  ::)

Thanks anyway for the great work.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on February 23, 2023, 11:59:59 AM
Hello @soundfan
Thanks for your words!

Quote from: soundfan on February 22, 2023, 03:50:40 PM
I managed to successfully install ML, update it with this crop_rec build from June 19th (is this the latest, recommended build?)..

Yes, that build should work for you. If you have artifacts in 1x3 presets (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239495#msg239495), try this build (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239557#msg239557).

Quote from: soundfan on February 22, 2023, 03:50:40 PM
But many crop settings result in slow camera response, corrupted live view, reduced field of view and even freeze the camera.
Sometimes a mode seems to work, I can access higher resolutions and make continuous video. The next time I try to do the same it says max resolution limited to xy, no matter what I try.

Could you share a video shows what are you doing then upload it on Youtube?
If you are using "Crop mode" and some high resolution presets like 1440p and 3K, you need to turn off then on RAW video multiple time to apply the new preset:

Quote from: theBilalFakhouri on June 15, 2022, 12:03:36 PM
Note: You will need to turn off and on RAW video multiple times after selecting 1440p or 3K presets, to let RAW video update its parameters!!

Will try to fix this issue later this year.


Quote from: soundfan on February 22, 2023, 06:39:23 PM
Beside that it records only 2 seconds now, although my SD supports up to 85MB/s @240MHz. Is the higher write rate only supported in the experimental crop mode?

I was trying to do 4K recording with the crop_rec module. And this caused the problems mentioned before.

Which preset are you using which give you 2 seconds?

You need to pay attention to required write speed, if it was too high (higher than the camera can actually writes) you will have limited recording times.
You can use hacks under RAW video submenu, also use lower bit-depths (from Crop mode submenu, such as 12/11/10 bit-depths) and lossless compression (always select 14-bit lossless in RAW video submenu) to reduce required write speed and increase write speed in LiveView.

Also, set Picture quality to RAW from Canon menu, then restart the camera (for once), this will free 31 MB of memroy. All of these tips are covered in 650D / 700D topic (https://www.magiclantern.fm/forum/index.php?topic=25784.msg231049#msg231049).

Again, if you could share a video shows what are you doing that would be helpful to me.


Quote from: soundfan on February 22, 2023, 06:39:23 PM
Also I recognized that the camera heats up very fast when testing with live view and playing with the ML menu. Is that normal?

Quote from: soundfan on February 22, 2023, 08:02:59 PM
Only thing I recognized is that write speed is only stable as long as camera temp is below ~45°C. Beyond that anything can happen. Now that I don't need to try 'n error for hours anymore I guess that overheating will no longer be an issue.

I don't what "very fast" means, that depends on many factors like ambient temp, camera power type (Battery, AC adapter, power bank) . . Does it heat up in same way without ML?
Usually iirc my 100D / 700D are around 45 to 55 when using Battery, ~21 ambient.

But I have noticed that with SD overclocking especially at 240 MHz makes SD card heat up faster (also the surrounded components), on 100D at some point this affect write/read (https://www.magiclantern.fm/forum/index.php?topic=26634.msg240328#msg240328) speeds (when SD card get too hot):

Quote from: theBilalFakhouri on September 15, 2022, 08:28:05 AM
-Side effect I noticed (at least using my card and 100D):

● It happens after SD controller/card become too hot (I mean it) write speed performance will drop from 90 MB/s to 72 MB/s in benchmarks (Play mode), or from ~80 MB/s to ~62 MB/s for RAW video recording.
● After a camera restart SD card will restore full write speed (90 MB/s in Play mode, ~80 MB/s in RAW video recording) and you should be able to record more minuets until it drop again to 72 MB/s in Play mode or 62 MB/s while RAW video recording.

On 700D / 5D3 no such issue there. You can make sure if this happening to you by running benchmarks in Play mode when you notice the drop, compare it to when SD card is at cool temp.
Other speed drop issue: perform low level format in camera, at least for once.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on February 23, 2023, 12:02:05 PM
Quote from: soundfan on February 22, 2023, 08:02:59 PM
QuoteGeneral notes:
-You can load both crop_rec and crop_new modules, but don't activate them at the same time (use "Crop mode" or "Crop mode V2").
-Use maximum aspect ratio and and resolution from mlv_lite when using crop_rec presets with real-time correct preview, otherwise mlv_lite won't crop the RAW data correctly.

With this pre-settings everything works now as expected.

The note in red color should only affect how RAW data is being cropped, shouldn't be related to issues which you are having.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 23, 2023, 08:41:56 PM
Hi Bilal,

thanks for your answers.

I'm not sure what I'm doing wrong. I wanted to make a video showing my problems and at that point everything worked. No matter which resolution or bit depth. Always continuous recording.

Next minute write speed drops to 20MB/s and nothing works anymore. I got the feeling that the sd_uhs module is not working stable.

Here are two videos anyway. Perhaps you see something I don't.

https://youtu.be/xScs1yYDfcw

https://youtu.be/Jqk7S25dpmw


Many thanks for your support in advance.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 23, 2023, 09:19:14 PM
one more thing I found when bench testing SD: It now runs first write test @83.4MB/s but during first run of read test speed dros at frame 59 to 20MB/s, also for the 2nd read and write cycle.

I tried to copy sd_uhs module again, but this didn't work. Somehow I think the crop high res writing is not stable on my camera.

Is there something I can do to support debugging?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on February 23, 2023, 10:30:20 PM
Thanks for sharing the videos, everything seem fine, except for SD card speed drop. Yes, it seems 240 MHz OC isn't stable for you, that's strange.

Which sd_uhs version are you using?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 24, 2023, 07:01:42 AM
Actually I used your latest one just yesterday:

https://www.magiclantern.fm/forum/index.php?topic=25841.msg234228#msg234228 (https://www.magiclantern.fm/forum/index.php?topic=25841.msg234228#msg234228)

What I didn't try yet is a low level format of SD inside the camera before testing. This I will try next this afternoon.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on February 24, 2023, 07:14:57 AM
Hmmm, okay
low level format won't help in case if speed drop to 21 MB/s (switch to 48 MHz).

Things to try:

1- Benchmark SD card in PC using fast card reader, share the result
2- Try 192 MHz in camera, run benchmarks, does it work? (don't forget to restart the camera after selecting 192 MHz)
3- Redo step 2 but with 160 MHz
4- Switch AccessMode to SDR50 in camera, select 240 MHz, restart camera, run benchamrks, does it work?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 24, 2023, 09:45:55 AM
Yes, I will try those again.

Thing is benchmark worked already with 84MB/s write speed and 86MB/s read speed couple of times. Trying again after several tests of different resolutions leads to the drop to 21MB/s. Unfortunately I'm not yet able to put a handle on it.

Anyway, I will continue trying this afternoon for sure.

And thanks for your suggestions and continuous support!  8)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on February 24, 2023, 10:58:51 AM
Did you switch access mode?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 24, 2023, 11:00:32 AM
Quote from: Danne on February 24, 2023, 10:58:51 AM
Did you switch access mode?

In fact I did. Same result.  :'(
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 24, 2023, 05:27:32 PM
Okay, I did a lot of test now but results don't make any sense (at least to me).

When I started (cold) camera and run benchmark twice at 240MHz, everything was fine, see pictures below.

https://drive.google.com/file/d/1vW79Px5pmOV-1GP6scQEY2XvRNvK2Qqe/view?usp=sharing (https://drive.google.com/file/d/1vW79Px5pmOV-1GP6scQEY2XvRNvK2Qqe/view?usp=sharing)

https://drive.google.com/file/d/1xTXHToLQFlgfWJ3HLdjtO8910EFV3jh-/view?usp=sharing (https://drive.google.com/file/d/1xTXHToLQFlgfWJ3HLdjtO8910EFV3jh-/view?usp=sharing)

After that I made a recording @ 4k/14bit. It recorded continuously. Then I changed resolution to 4.5K and it went back to 20MB/s. No matter what I tried next (other resolution, different SD speed, turning off and on camera, even remove battery in between), no change.

Next I tried 240Mhz at SDR50 . This resulted in the best benchmark ever, see picture.

https://drive.google.com/file/d/19eeOJvfZhc60fhVAj0CR_RChjPnrw_k-/view?usp=sharing (https://drive.google.com/file/d/19eeOJvfZhc60fhVAj0CR_RChjPnrw_k-/view?usp=sharing)

With this setting I managed to record continuously 4K/14bit, 4.5K/14bit, UHD/14bit without problem. Then I tried another benchmark. First write was 90MB/s again, but already the next reading cycle dropped to 20MB/s again. I switched off camera, unplugged battery, waited a while. Restarted benchmark with this result:

https://drive.google.com/file/d/1AxaCyyTnZCiCbBnMWAmoAkC6ucepC4Eb/view?usp=sharing (https://drive.google.com/file/d/1AxaCyyTnZCiCbBnMWAmoAkC6ucepC4Eb/view?usp=sharing)

At the second reading cycle it dropped to 23MB/s again.

I start to run out of ideas... :'(

Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on February 24, 2023, 05:55:10 PM
Thanks for the tests, but you missed the first three steps here, waiting the results :)

Quote from: theBilalFakhouri on February 24, 2023, 07:14:57 AM
Things to try:

1- Benchmark SD card in PC using fast card reader, share the result
2- Try 192 MHz in camera, run benchmarks, does it work? (don't forget to restart the camera after selecting 192 MHz)
3- Redo step 2 but with 160 MHz
..

Try recording with 192 MHz and 160 MHz, does it also drop to 21 MB/s after some time?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 24, 2023, 06:30:27 PM
Unfortunately I don't have a fast SD card reader for PC. The one integrated in my laptop was telling me like 26MB/s for the Sandisk Extreme Pro 128GB 200MB/s with the disk speed test from your earlier post. This is kind of ridiculous.

Card benchmark with

160MHz (SDR50): 64.3/71.9/64.8/71.9
190MHz (SDR50): 75.1/drop to 22.7/20.8/21.8

redo 160MHz (SDR50): 63.8/71.9/64.1/71.9

Another test with SDR104:
160MHz: 67.0/71.9/66.8/71.8
192MHz: 78.4/drop to 22.9/20.8/21.7

redo 160MHz: 66.9/71.9/66.7/71.9

160MHz seems to be the only stable frequency right now. No matter if SDR50 or SDR104.

Does this generate any new ideas?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 24, 2023, 06:39:59 PM
Recording in 160MHz seems the be continuous as long as predicted data rate is adequate.

Recording in 192MHz drops almost right away to ~20MB/s. Back to 160MHz it works continuous again immediately.

At least I now know a mode I can work with. Would be nice to get the last MB/s as well, but so far it works.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on February 24, 2023, 06:45:28 PM
@soundfan: AF is off, correct?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 24, 2023, 06:47:05 PM
Yep.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on February 24, 2023, 07:38:00 PM
Thanks for the tests, it seems to me it's a SD card issue, but not quite sure.
Tweaking GPIO registers in sd_uhs.c might help or might have no effect.

It's also possible that your SD card needs some kind of configuration which I am not aware of.
Do you have access to other SD card or other DIGIC 5 camera?

Could you share CID info from "SD Overlock" submenu?

For now, please try formatting SD card using SD Memory Card Formatter (https://www.sdcard.org/downloads/formatter/) on PC.
That might help fixing SD card issues if there were any.

Then redo tests at 192 MHz or 240 MHz, does it work better after formatting with mentioned tool?

I will provide sd_uhs.mo with tweakable GPIO values from submenu later, so you can try some values.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on February 24, 2023, 08:28:21 PM
In my build I think it needed some pause when applied on 100d. Might be some patch routine similar to 5diii here. I know because I tried some faster patching that works on eos m but didn't on 100d.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on February 24, 2023, 08:31:20 PM
@Danne

Have you tried my latest sd_uhs (https://www.magiclantern.fm/forum/index.php?topic=25841.msg234228#msg234228) on your 100D with your SD cards?
Latest sd_uhs works fine on my 100D, also other 100D users reported the same thing in eariler replies.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on February 24, 2023, 08:32:47 PM
Did the formatting with SD Memory Card Formatter. Same result as before. Cold camera works for a while @240MHz with 84MB/s, then drops to 20MB/s.

@160MHz works perfectly again, benchmark same as recording.

Here the CID info from that SD card:
MID: 0x03
OID: 0x5344
PNM: SR128
PRV: 8.6
PSN: 0xc0512c54
MDT: 2023/01
CRC: 0x89

whatever that means...  ::)

Tried another card (Intenso SDHC UHS-I Preimium 45MB/s 32GB)

standard speed: 22.0/43.5/22.1/43.5
SDR104:
160MHz: 24.5/71.9/26.7/71.9
192MHz: 15.4/21.7/15.4/21.7  Card seems not to support higher frequencies

SDR50:
160MHz: 26.8/71.4/26.7/71.4
192MHz: 15.3/21.7/15.3/21.7  Card seems not to support higher frequencies

I know that this card is not usable for RAW video recording, therefore I bought the fastest available (Sandisk Extreme Pro 128GB 200MB/s).
Perhaps it is useful to point into the right direction anyway.

Unfortunately I don't have another camera on hand (yet).
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on February 24, 2023, 08:46:17 PM
Quote from: theBilalFakhouri on February 24, 2023, 08:31:20 PM
@Danne

Have you tried my latest sd_uhs (https://www.magiclantern.fm/forum/index.php?topic=25841.msg234228#msg234228) on your 100D with your SD cards?
Latest sd_uhs works fine on my 100D, also other 100D users reported the same thing in eariler replies.
No time for my baby 100d :(.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 01, 2023, 05:30:45 PM
Back with some more test results:

SD card seems not to be the problem.

I just bought a new Sandisk Extreme Pro 64GB 170MB/s, the most recommended SD card (becomes pretty difficult to get these days).

First benchmarks were promising, see below:
160MHz:
https://drive.google.com/file/d/13MapLJ4ZCe9JpHnSjvgEEl2MLnbhc5qE/view?usp=share_link (https://drive.google.com/file/d/13MapLJ4ZCe9JpHnSjvgEEl2MLnbhc5qE/view?usp=share_link)

192MHz:
https://drive.google.com/file/d/160_bb9xJS6dwOqaOWUloEs5PE5ORBLkT/view?usp=share_link (https://drive.google.com/file/d/160_bb9xJS6dwOqaOWUloEs5PE5ORBLkT/view?usp=share_link)

240MHz (first attempt):
https://drive.google.com/file/d/1K5kmn8IfyB23vPMsJQNU7J33EnteDCP6/view?usp=share_link (https://drive.google.com/file/d/1K5kmn8IfyB23vPMsJQNU7J33EnteDCP6/view?usp=share_link)

Here the CID info from that new SD card:
MID: 0x03
OID: 0x5344
PNM: SR65G
PRV: 8.6
PSN: 0x3fa84F98
MDT: 2022/02
CRC: 0x76

After a few test and camera getting just a little warmer (~38°C, no idea if that is relevant), the recording stopped again and benchmark shows drop to 20MB/s again:

240MHz (second attempt):
https://drive.google.com/file/d/1eKaKcPtwb8Q41pEUbGrxWZEezmRXOomh/view?usp=share_link (https://drive.google.com/file/d/1eKaKcPtwb8Q41pEUbGrxWZEezmRXOomh/view?usp=share_link)

So actually the same behaviour like with the other Sandisk Extreme Pro 128GB 200MB/s card.

BTW: How can I post pictures directly in my posts? Most forums allow upload of pictures. I couldn't find any option here. No problem though, can upload to Google drive anyway, but only can share the link.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 01, 2023, 06:45:13 PM
Okay, it seems like your 100D could be a little different from mine, either from SD card controller in camera, or some SD configuration.
I need to look into your 100D ROM files, also I need a DebugLog from your 100D:

1- zip ROM0.BIN and ROM1.BIN in ML/LOGS folder and upload it somewhere (like Google Drive) then please send a link to me via personal message (don't post the link here).

2- DebugLog:
    Make a clean install (fromat SD card in PC first, then format it in camera) and install io_trace_full.2023Mar01.100D101.zip (https://drive.google.com/file/d/1klqcOTGWfXvBzc3McUMaND4IeRptni8f/view?usp=share_link) build on Sandisk 170 or 200 MB/s SD card, switch the camera to photo mode (no LiveView),
    click on "Don't click me!" from "Debug" tab in ML menu, wait until it finish, it will print "saved" and name of log file on screen, then turn off the camera, you will find the log file in SD card,
    upload it somewhere and share it please.

-The camera will take a few seconds with LED ON until it turn on, that's normal with this io_trace build.


Quote from: soundfan on March 01, 2023, 05:30:45 PM
BTW: How can I post pictures directly in my posts? Most forums allow upload of pictures. I couldn't find any option here. No problem though, can upload to Google drive anyway, but only can share the link.

You need to convert .ppm files first to either PNG or JPEG using some kind of image processing software, Photoshop can do it for example.
Then you will need to upload the PNG or JPEG to image host website like https://imgbb.com/ (https://imgbb.com/)

After uploading it there click on the picture from host website, so you can find these:

(https://i.ibb.co/XSyH7ft/embe.png) (https://ibb.co/nz3p7qj)

Copy either Full or Thumbnail image BBCode to your post here, then you will have a posted picture on the forum.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 01, 2023, 08:34:02 PM
Thanks again for your help.

Registered to imgbb.com

Test:
(https://i.ibb.co/R3mVwsg/bench2.jpg) (https://ibb.co/R3mVwsg)

Sent you pm.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 01, 2023, 08:56:52 PM
Thanks for the LOG and the ROMs, SD configuration (i.e registers which covered in sd_uhs.c) seems identical to my 100D.
Perhaps there is other configuration which work better for 192 MHz and 240 MHz across 100D bodies (assuming it's not a hardware difference), let's start with simple thing:

1. Downclocking read speed to 160 MHz instead of 192 MHz in 192 MHz / 240 MHz presets:
    Please try this sd_uhs.mo (https://drive.google.com/file/d/1wVMNIORu7Bv5xz8-53TDPUskUEcSzjLw/view?usp=share_link), try 192 MHz and 240 MHz, does write speed drop to 21 MB/s?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 02, 2023, 04:54:08 PM
Tried this sd_uhs.mo. Same result.

Cold camera benchmark works on all three speeds, after some playing in the menues and some test recording speed drops to 20MB/s again.

160MHz:

(https://i.ibb.co/Q8R5b6K/bench4.jpg) (https://ibb.co/Q8R5b6K)

192MHz:

(https://i.ibb.co/cvL7tgR/bench5.jpg) (https://ibb.co/cvL7tgR)

240MHz (first attempt):

(https://i.ibb.co/7QfFYhP/bench6.jpg) (https://ibb.co/7QfFYhP)

second attempt:

(https://i.ibb.co/n1vNxSq/bench7.jpg) (https://ibb.co/n1vNxSq)


Seems not the be the sd_uhs.mo file...
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 02, 2023, 06:43:29 PM
Thanks for the test.

The temperature isn't the only factor for 21 MB/s speed drops, some SD configurations perform better at high temps than others. See (https://www.magiclantern.fm/forum/index.php?topic=26634.msg240328#msg240328).
What you want is to find a better SD configuration than the one I found. I suggested to try some different GPIO values:

Quote from: theBilalFakhouri on February 24, 2023, 07:38:00 PM
Tweaking GPIO registers in sd_uhs.c might help or might have no effect.

Current sd_uhs.c is using:

    MEM(0xC022C634) = 0x599;  <-- GPIO1

    MEM(0xC022C638) = 0x555;  <-- GPIO2
    MEM(0xC022C63C) = 0x555;
    MEM(0xC022C640) = 0x555;
    MEM(0xC022C644) = 0x555;
    MEM(0xC022C648) = 0x555;


Some info I wrote about them can be found here (https://www.magiclantern.fm/forum/index.php?topic=26634.msg240128#msg240128) (under "New GPIO registers") , what you want for 100D:

Quote from: theBilalFakhouri on September 07, 2022, 05:46:28 PM
for my 100D:
0x599 for GPIO1, 0x555 GPIO2 worked, and it seems the best for 240 MHz preset.

0x588 for GPIO1, 0x555 GPIO2 works too on 100D
0x599 for GPIO1, 0x566 GPIO2 works too on 100D
0x577 for GPIO1, 0x544 GPIO2 (700D values) don't work very well.

Without tweaking these registers, 240 MHz never been stable on 100D/EOS M, so they are required and they are a part of SD puzzle.

Try the other suggested combos, or try some new combos based on these info (https://www.magiclantern.fm/forum/index.php?topic=26634.msg240128#msg240128). if none of them worked, sorry it's hard to know exactly how to make it work. I might provide some tests later.




Okay, here are some tests, try these values (you can change them from SD Overclock submenu, every time after you change them restart the camera):

1- GPIO1: 0x588,   GPIO2: 0x555
2- GPIO1: 0x599,   GPIO2: 0x566
3- GPIO1: 0x577,   GPIO2: 0x544
4- GPIO1: 0x555,   GPIO2: 0x555
5- GPIO1: 0x566,   GPIO2: 0x555

This sd_uhs.mo (https://drive.google.com/file/d/1G9qkIisI65BD9UbYtaZhZmj3NBhIvJtw/view?usp=share_link), does any of above combos work better? try at 192 MHz first, if it give the same results, more likely there is no need to do the tests at 240 MHz.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 03, 2023, 05:44:13 PM
Wow, that were a lot of proposals and a lot to test, too.

I did various combinations. All of your proposals and a few from my guts feeling.

Infact many work absolutely stable at 192MHz (I ended up using GPIO1: 0x5dd,   GPIO2: 0x599).

But unfortunately none of them keeps working at 240MHz. Even during recording with green data rate displayed camera stops out of the sudden. after that benchmark is at 20MB/s again.

Anyway, at least 192MHz is now stable with benchmark rates like these:

(https://i.ibb.co/G94zKCk/bench47.jpg) (https://ibb.co/G94zKCk)

So far I'm pretty satisfied at the moment. Recording in UHD with 14bit works continuously.

Others like 3K, 1440p or Full-Res LV stop immediately with raw detect error.

UHD, 4K, 4.3K, 4.5K and 1920px1280p work perfectly.

I was actually hoping to get the Full-Res to work with whatever framerate is possible.

Do you know if there is a way to get there?


Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 03, 2023, 06:10:52 PM
Cool, thanks for running the tests!

Quote from: soundfan on March 03, 2023, 05:44:13 PM
Infact many work absolutely stable at 192MHz (I ended up using GPIO1: 0x5dd,   GPIO2: 0x599).

So some of these combos I shared eariler worked fine at 192 MHz on your 100D?

Quote from: theBilalFakhouri on March 02, 2023, 06:43:29 PM
1- GPIO1: 0x588,   GPIO2: 0x555
2- GPIO1: 0x599,   GPIO2: 0x566
3- GPIO1: 0x577,   GPIO2: 0x544
4- GPIO1: 0x555,   GPIO2: 0x555
5- GPIO1: 0x566,   GPIO2: 0x555

I need to know which ones exactly worked fine at 192 MHz from above list on your 100D, so I can test them on my 100D at both 192 MHz and 240 MHz.
Although "GPIO1: 0x5dd,   GPIO2: 0x599" worked fine for you at 192 MHz, it doesn't really match Canon default patterns. I have made some tests like these in the past, some of them works yes.

I have one more left test to give for 240 MHz, might make it work on your 100D or might not, but firstly, please give what I asked for :)




Quote from: soundfan on March 03, 2023, 05:44:13 PM
Others like 3K, 1440p or Full-Res LV stop immediately with raw detect error.
..
I was actually hoping to get the Full-Res to work with whatever framerate is possible.
..
Do you know if there is a way to get there?

Sure, from first post:

Quote from: theBilalFakhouri on June 06, 2022, 12:50:10 AM
-Quirks (not related to my build, but I have noticed it):

When selecting 3K and 1440p presets you need to turn off then on RAW video multiple times to let RAW video update RAW video resolution parameters.

Also:

Quote from: theBilalFakhouri on June 15, 2022, 12:03:36 PM
Note: You will need to turn off and on RAW video multiple times after selecting 1440p or 3K presets, to let RAW video update its parameters!!

Same thing applies for Full-Res preset, this issue will be fixed in the next release from my side. However, commit (https://github.com/bilalfakhouri/magiclantern_hg_02/commit/b6f1d8c6a79cfd448c828064ab139682181fb782) is available online which fix the issue.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 03, 2023, 07:48:11 PM
Quote from: theBilalFakhouri on March 01, 2023, 08:56:52 PM
1. Downclocking read speed to 160 MHz instead of 192 MHz in 192 MHz / 240 MHz presets:
    Please try this sd_uhs.mo (https://drive.google.com/file/d/1wVMNIORu7Bv5xz8-53TDPUskUEcSzjLw/view?usp=share_link), try 192 MHz and 240 MHz, does write speed drop to 21 MB/s?

Well, the one from this post worked also stable on my 100D at 192MHz, but also crashed at 240MHz.

Quote from: theBilalFakhouri on March 03, 2023, 06:10:52 PM
Same thing applies for Full-Res preset, this issue will be fixed in the next release from my side. However, commit (https://github.com/bilalfakhouri/magiclantern_hg_02/commit/b6f1d8c6a79cfd448c828064ab139682181fb782) is available online which fix the issue.

That is a little too much for me yet. Well, I tried many times to switch on and off the raw video. I found that jumping to live view and back (trash button) worked also well und made the modes update immediately. But not for the mentioned modes 3K, 1440p and Full-Res. Well, actually it worked in live view but didn't update the frame dimensions in RAW video.

Anyway, I will continue tying.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 03, 2023, 08:09:28 PM
Quote from: soundfan on March 03, 2023, 07:48:11 PM
Well, the one from this post worked also stable on my 100D at 192MHz, but also crashed at 240MHz.

That's make things totally different :) , in this case we need to redo GPIO tests without downclocking read speed to 160 MHz in 192 MHz preset:

Download this sd_uhs.mo (https://drive.google.com/file/d/1sy3IlFaH38A_JPCtouv7qs8k1L5Y4cZc/view?usp=share_link): (has tweakable GPIO menu, read speed set to 192 MHz in 192 MHz and 240 MHz presets)
Just try the following values using 192 MHz overclock, does any of them perform better at 192 MHz (speed doesn't drop to 21 MB/s)?

1- GPIO1: 0x588,   GPIO2: 0x555
2- GPIO1: 0x599,   GPIO2: 0x566
3- GPIO1: 0x577,   GPIO2: 0x544
4- GPIO1: 0x555,   GPIO2: 0x555
5- GPIO1: 0x566,   GPIO2: 0x555

Perform a camera restart after you change GPIO1 and GPIO2, report back please.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 03, 2023, 09:03:48 PM
Okay, let's see:

1- GPIO1: 0x588,   GPIO2: 0x555 drops to 20MB/s @ 192MHz
2- GPIO1: 0x599,   GPIO2: 0x566 drops to 20MB/s @ 192MHz
3- GPIO1: 0x577,   GPIO2: 0x544 drops to 20MB/s @ 192MHz
4- GPIO1: 0x555,   GPIO2: 0x555 drops to 20MB/s @ 192MHz
5- GPIO1: 0x566,   GPIO2: 0x555 drops to 20MB/s @ 192MHz


So 192MHz frequency for reading seems not to work on my 100D
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 03, 2023, 09:58:09 PM
Okay, thanks, that was helpful, it seems GPIO values aren't the problem here.

Let's try three more tests for now before I give up, please make the test at 192 MHz preset:

1- I set 0xC0400624 and  0xC0400628 to 0x100 instead of 0x201 in 192 MHz, this sd_uhs.mo (https://drive.google.com/file/d/17G0sWG1TwedTcvDh4b2G7D5BPv4OSsUi/view?usp=share_link)
2- I set 0xC040061C to 0x200 instead of 0x100                           in 192 MHz, this sd_uhs.mo  (https://drive.google.com/file/d/1Zu17-16b2TvuJwWJcO7xhhGmjcJb0xrD/view?usp=share_link)
3- I set the changes from step 1 and step 2                                 in 192 MHz, this sd_uhs.mo (https://drive.google.com/file/d/1SdTPMyDCJEkKzKg-v61_PtspGZ3X8L9k/view?usp=share_link)

Does any of these sd_uhs work better at 192 MHz on your 100D?
(sorry, too many sd_uhs.mo copies)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 04, 2023, 08:47:03 AM
Quote from: theBilalFakhouri on March 03, 2023, 09:58:09 PM
Okay, thanks, that was helpful, it seems GPIO values aren't the problem here.

Let's try three more tests for now before I give up, please make the test at 192 MHz preset:

1- I set 0xC0400624 and  0xC0400628 to 0x100 instead of 0x201 in 192 MHz, this sd_uhs.mo (https://drive.google.com/file/d/17G0sWG1TwedTcvDh4b2G7D5BPv4OSsUi/view?usp=share_link)
2- I set 0xC040061C to 0x200 instead of 0x100                           in 192 MHz, this sd_uhs.mo  (https://drive.google.com/file/d/1Zu17-16b2TvuJwWJcO7xhhGmjcJb0xrD/view?usp=share_link)
3- I set the changes from step 1 and step 2                                 in 192 MHz, this sd_uhs.mo (https://drive.google.com/file/d/1SdTPMyDCJEkKzKg-v61_PtspGZ3X8L9k/view?usp=share_link)

Does any of these sd_uhs work better at 192 MHz on your 100D?
(sorry, too many sd_uhs.mo copies)

Okay, did the tests.

First only once each with cold camera:

Test 1:
(https://i.ibb.co/VxXcNtZ/bench55.jpg) (https://ibb.co/VxXcNtZ)
Test 2:
(https://i.ibb.co/fGh4Nrd/bench56.jpg) (https://ibb.co/fGh4Nrd)
Test 3:
(https://i.ibb.co/58cs654/bench57.jpg) (https://ibb.co/58cs654)

There seemed to be a slight improvement with camera getting warmer so I did another 3 benchmark tests each and indeed a slight improvement was visible. So here are the results after the third test each:
Test 1:
(https://i.ibb.co/dWBh3XL/bench60.jpg) (https://ibb.co/dWBh3XL)
Test 2:
(https://i.ibb.co/x27fKk0/bench63.jpg) (https://ibb.co/x27fKk0)
Test 3:
(https://i.ibb.co/1TtpS3m/bench66.jpg) (https://ibb.co/1TtpS3m)

In total I'd say the first versions shows the best results. Didn't do any recording though.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 04, 2023, 08:52:41 AM
Thanks for running the tests,

Quote from: soundfan on March 04, 2023, 08:47:03 AM
In total I'd say the first versions shows the best results.

maybe I wasn't clear in my previous reply, I am not looking for benchmarks numbers, I mean does 192 MHz drop to 21 MB/s after some time?
When using these sd_uhs versions:

Quote from: theBilalFakhouri on March 03, 2023, 09:58:09 PM
1- I set 0xC0400624 and  0xC0400628 to 0x100 instead of 0x201 in 192 MHz, this sd_uhs.mo (https://drive.google.com/file/d/17G0sWG1TwedTcvDh4b2G7D5BPv4OSsUi/view?usp=share_link)
2- I set 0xC040061C to 0x200 instead of 0x100                           in 192 MHz, this sd_uhs.mo  (https://drive.google.com/file/d/1Zu17-16b2TvuJwWJcO7xhhGmjcJb0xrD/view?usp=share_link)
3- I set the changes from step 1 and step 2                                 in 192 MHz, this sd_uhs.mo (https://drive.google.com/file/d/1SdTPMyDCJEkKzKg-v61_PtspGZ3X8L9k/view?usp=share_link)

Quote from: soundfan on March 04, 2023, 08:47:03 AM
Didn't do any recording though.

Looking forward for recording tests! (to check if it drop to 21 MB/s or not).
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 04, 2023, 10:59:37 AM
Alright! I did some excessive testing now. Finally I found something reproduceable  :D

Version 1 drops to 21MB/s almost immediately, version 3 after a while. Even if recording timer is green, at some point it drops out of the sudden.

Version 2 seems to run stable until camera becomes too hot (beyond 10 minutes UHD 14bit @ 24°C indoor temp). I tested this several times, so it can be reproduced.

What I found during this testing is that camera temperature isn't updated on the live screen during recording. I started at 43°C and when I stopped after recording more than 10 minutes it jumped to 60°C. Is that a known issue (don't know if I should call it that)?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on March 04, 2023, 12:02:54 PM
Not temperature, correct registry setting needed. Seems version 2 is close.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 05, 2023, 06:24:01 PM
Okay, thanks for running tests again!

Quote from: soundfan on March 04, 2023, 10:59:37 AM
Version 2 seems to run stable until camera becomes too hot (beyond 10 minutes UHD 14bit @ 24°C indoor temp). I tested this several times, so it can be reproduced.

So does it drop to 21 MB/s when it's become not stable?
What does benchmarks show after SD becomes too hot (when it become not stable)?




Quote from: soundfan on March 04, 2023, 10:59:37 AM
What I found during this testing is that camera temperature isn't updated on the live screen during recording. I started at 43°C and when I stopped after recording more than 10 minutes it jumped to 60°C. Is that a known issue (don't know if I should call it that)?

It could be a bug (or maybe RAW video disables some Canon updates while recording), didn't pay much attention to these stuff, will make tests later.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 05, 2023, 06:30:45 PM
Quote from: theBilalFakhouri on March 05, 2023, 06:24:01 PM
Okay, thanks for running tests again!

So does it drop to 21 MB/s when it's become not stable?
What does benchmarks show after SD becomes too hot (when it become not stable)?

No, it runs stable at 192MHz. Only camera shut down due to overheating once. Second time it ran completely stable until SD was full  8)
Benchmark still at 79MB/s writing...

Quote from: theBilalFakhouri on March 05, 2023, 06:24:01 PM

It could be a bug (or maybe RAW video disables some Canon updates while recording), didn't pay much attention to these stuff, will make tests later.

Not very important, but thanks anyway to have a look at it.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 11, 2023, 05:20:44 PM
Hi Bilal,

it's me again. I'm actually pretty happy with the recording capabilities of my 100D at 192MHz now. Didn't have any instability yet.

Anyway, is there a chance we may find a way to get the 240MHz running, too?

Many thanks in advance!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 13, 2023, 12:20:07 AM
Hello soundfan,
Glad you are happy, thanks for the report :)

Unfortunately, more likely I won't be able to provide more tests this month. Maybe next month.
Anyway, I will let you know when I come up with new tests for 240 MHz.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: iaburn on March 21, 2023, 11:39:18 PM
Hi! I saw this thread and I couldn't resist. I got myself a 100D now that there is no "Zeek" channel in Youtube about it and it's still very cheap :D

On mine I noticed the vertical stripes on 1x3 modes, but solved with the "ADGT" version, and for now I only tried the Sandisk Pro 95MB/s and works well at 192Mhz (around 70MB/s).

theBilalFakhouri I'm still understanding how your build works and I would have a couple of questions:
Is it normal that, when I'm on the menu, at fixed intervals the live view flashes for an instant?
Is there any reliable way to set 3K mode? I was lucky just once, Now no matter how many times I activate/deactivate RAW, or go into Canon menu, I cannot get the RAW resolution to change.

And something extra about the camera itself. Is there a proper way to go from OFF to video mode and turn the camera off from video mode? I mean, do I have to wait a few seconds on the "ON" mode before going to Video or shutting it down?  ::)
Ah, and also, is it possible to use a dummy battery like in the EOS M? It has no hole and it forces you to close the battery door Found a suitable dummy battery  :D

I'm super impressed with this build. I know there is a lot of work behind, but I have the feeling that it came from nowhere when no one was expecting it, amazing!!
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on March 23, 2023, 01:30:21 PM
Quote from: iaburn on March 21, 2023, 11:39:18 PM
Hi! I saw this thread and I couldn't resist. I got myself a 100D now that there is no "Zeek" channel in Youtube about it and it's still very cheap :D

Hello!
Cool, 100D is a good choice too.

Quote from: iaburn on March 21, 2023, 11:39:18 PM
On mine I noticed the vertical stripes on 1x3 modes, but solved with the "ADGT" version, and for now I only tried the Sandisk Pro 95MB/s and works well at 192Mhz (around 70MB/s).

Thanks for the report, I will delete other builds and keep ADTG version, seems to work fine on different 100D bodies.

Quote from: iaburn on March 21, 2023, 11:39:18 PM
Is it normal that, when I'm on the menu, at fixed intervals the live view flashes for an instant?

For now, yes, this problem does exist. I think it can be fixed, but didn't try yet. the fix could be similar to this one:

Quote from: theBilalFakhouri on April 25, 2022, 04:19:52 AM
-Menus:
-fixed interference between ML menu and Canon overlays in LiveView, I took this fix from here (https://foss.heptapod.net/magic-lantern/magic-lantern/-/commit/692eb50142bfdfa52a86253bdc46c06cf52a38f9), solves this issue (https://www.magiclantern.fm/forum/index.php?topic=7473.msg198464#msg198464).




Quote from: iaburn on March 21, 2023, 11:39:18 PM
Is there any reliable way to set 3K mode? I was lucky just once, Now no matter how many times I activate/deactivate RAW, or go into Canon menu, I cannot get the RAW resolution to change.

Here is what I am doing to make 3K preset work (I know, it's a dumb workaround :P):

https://www.youtube.com/watch?v=qPjsqeoyvFo

Quote from: iaburn on March 21, 2023, 11:39:18 PM
And something extra about the camera itself. Is there a proper way to go from OFF to video mode and turn the camera off from video mode? I mean, do I have to wait a few seconds on the "ON" mode before going to Video or shutting it down?  ::)

I didn't get what you mean. You can set the power switch directly from OFF to Video mode and vice versa, what's the problem with this? that's how Canon made the camera work.

Quote from: iaburn on March 21, 2023, 11:39:18 PM
I'm super impressed with this build. I know there is a lot of work behind, but I have the feeling that it came from nowhere when no one was expecting it, amazing!!

Thanks, glad you liked it :)

I am actually discarding this branch (beside my 650D / 700D one). Will not work on them. There will be a new successor branch which is better in many ways.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: iaburn on March 23, 2023, 01:58:06 PM
Thanks a lot for the video, following these exact steps it works every time!  ;D

The question about the off/on/video switch was very dumb. You are right that Canon did it that way, so there should be no problem moving the switch between modes fast  ::)

This branch is already very usable, can't wait to see what's coming  :o
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: gabriielangel on March 23, 2023, 04:52:46 PM
This is quite nice! How long can clips be recorded at 3K?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: soundfan on March 23, 2023, 05:07:34 PM
Quote from: gabriielangel on March 23, 2023, 04:52:46 PM
This is quite nice! How long can clips be recorded at 3K?

Until the card is full (if card is fast enough)  :D
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: ouzijoe on June 07, 2023, 09:41:42 PM
Hello Everyone,

I just hop on the magic lantern adventure buying a 100d and putting crop mood without letting him time to breath :)
I am still exploring the option even though i did my fair share of research.

1/ I can not get rid of focus pixel, i have tried downloaded and installed all pixel map from reddit, and it stills does not work
2/ i can not record 5.2k 14 bit more than 1 seconds even in interior (i go a sandisk pro 170 m/s sd); should i change my sd to something better or the results are random from one device to the other?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on June 08, 2023, 08:28:56 PM
Quote from: ouzijoe on June 07, 2023, 09:41:42 PM
1/ I can not get rid of focus pixel, i have tried downloaded and installed all pixel map from reddit, and it stills does not work
2/ i can not record 5.2k 14 bit more than 1 seconds even in interior (i go a sandisk pro 170 m/s sd); should i change my sd to something better or the results are random from one device to the other?

1: Check MLVapp -> View -> Show installed Focus Pixel maps
2: Benchmark card in cam.
Access Modules tab and activate bench.mo
Restart cam in photo mode
Access Debug tab -> Benchmarks -> Card benchmarks -> Quick R/W benchmark (1 min)

Share results.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: ouzijoe on June 13, 2023, 11:08:38 AM
Hi everyone,
sorry for late reply, i thought my post was still under review

Concerning the focus pixel, which are my main issue, per my understanding i have understood that the map for crop mood for 100d have not been implemented yet?
Here are the focus pixel maps installed :

(https://image.noelshack.com/minis/2023/24/2/1686646214-pixel-map.png) (https://www.noelshack.com/2023-24-2-1686646214-pixel-map.jpg)

As for the recording, i have managed to get pretty stabled at 4.4k 14 bits, i guess it is pretty good. I use the sandisk pro 200 mb/s (not the 170 mb/s as i thought). In recording i get around 69 mb/s. Benchmark in photo mode are 83.3 mb/S write speed (for both) and 85.6 mb/s read speed (for both). Benchmark in video mode give me 78.9 mbs and 71.7 mbs write speed (and i got this message malloc error : buffer=16777216). Does it worth investing in a sandisk 170 mb/s

One last thing, i got frequent lines on my pictures, i have understood that it may comes for underexposure? is there a way to avoid them?

(https://image.noelshack.com/minis/2023/24/2/1686647032-lines.png) (https://www.noelshack.com/2023-24-2-1686647032-lines.jpg)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 13, 2023, 07:36:38 PM
Quote from: ouzijoe on June 13, 2023, 11:08:38 AM
One last thing, i got frequent lines on my pictures, i have understood that it may comes for underexposure? is there a way to avoid them?

(https://image.noelshack.com/minis/2023/24/2/1686647032-lines.png) (https://www.noelshack.com/2023-24-2-1686647032-lines.jpg)

Hi @ouzijoe, which ML build are you using?
Also, could you share short MLV sample have the problem?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: ouzijoe on June 14, 2023, 11:26:28 AM
i am using the last version 101 from 9 may
What do you want me to share? An entire clip, or a dng extract showing the problem?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Danne on June 14, 2023, 11:32:26 AM
Mlv clip.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: ouzijoe on June 14, 2023, 07:57:53 PM
Here i uploaded a small MLV, you will see the problem every 4/5 frames : https://file.io/0ouV2z9kipwU

As for the focus pixel maps, do i miss something or there is no current solution?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 15, 2023, 04:09:22 PM
@ouzijoe

Thanks for the clip, related issue (https://www.magiclantern.fm/forum/index.php?topic=26511.msg239495#msg239495).

For now, I want to look into your 100D registers, please follow this guide:

Quote from: theBilalFakhouri on July 03, 2022, 09:58:43 PM
Tests:

1- Let me take a look into your 100D registers:

-Download this adtg_gui.mo (https://drive.google.com/file/d/1QfBQkORaKQ-cT9-Y4QOm-mpsHhecFOmm/view?usp=sharing), put it in ML/modules folder in SD card
-Unload all modules, and load only adtg_gui.mo, restart camera
-Set video mode to 1080p24 to gain extra speed

-Follow the following video, you can see in the video I am entering Canon menu and toggling between x1/x5/x10 modes, you should do that too to let "ADTG registers" bring the registers into it:
https://www.youtube.com/watch?v=Q-Tdu9YZ4fc

-After you finish, you will find "adtg.log" file in ML/LOGS folder in SD card, upload it somewhere and share it.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: ouzijoe on June 16, 2023, 12:11:26 AM
https://file.io/6TQSuJuHzLwU here is the log

Can you or someone else please answer me, is there or no pixel maps for crop mood resolution?? i have downloaded them all, from reddit and the other ones on google drive, drag them successfully on mlv but it still does not work
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 16, 2023, 09:44:39 AM
Quote from: ouzijoe on June 16, 2023, 12:11:26 AM
https://file.io/6TQSuJuHzLwU here is the log

I came too late? The file was deleted and I can't download it.


Quote from: ouzijoe on June 16, 2023, 12:11:26 AM
is there or no pixel maps for crop mood resolution??

Currently there is no focus pixels maps for crop mood presets on 100D.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: ouzijoe on June 16, 2023, 11:07:22 AM
https://file.io/VphbHSUBg56J
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 16, 2023, 11:35:51 AM
Please upload it to another host like Google Drive (or send it through Discord (https://discord.gg/uaY8akC)). The file get deleted from "file.io" after one download attempt? and you are sharing the download link in public.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: ouzijoe on June 16, 2023, 12:08:49 PM
https://drive.google.com/file/d/1mCuMHMWTvYgGScMgp8sgVbl4fWKMrzKl/view?usp=sharing
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on June 18, 2023, 10:41:18 PM
@ouzijoe

Thanks for the log, I need another one but this time:

-Keep "Show" set to "Everything"
-Keep "ENGIO Registers" turned OFF

And the follow other steps in the video and previous post.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 03, 2023, 03:26:52 PM
Man oh man. I'd totally forgotten about ML and just stumbled into this forum. Installed lastest.
Now,,, guys.. .ML virgin again really here... ML running fine, how do I kick off this high res video recording ?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 03, 2023, 03:35:19 PM
Quote from: edt on July 03, 2023, 03:26:52 PM
ML running fine, how do I kick off this high res video recording ?

Follow video (Build Walkthrough) and text: https://www.magiclantern.fm/forum/index.php?topic=26851.0
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 05, 2023, 10:24:41 AM
Not the easiest then! (for me at least)
I've tried a few setting but get poor results + very short clips.

May  I just check before I try again.. with the right combo of setting in this ML build, is my 100D & a decent fast card capable of recording video higher than std 1080p ? For 5 mins say.
Not after 4K (though would be nice)  :)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 05, 2023, 11:16:48 AM
I tested some cards with 650D (slightly faster than 100D) and got uncompressed 14 bit just a little bit below 1080p and for at least 5 minutes. Rough calculation:
1920 x 1080 x 24 fps x 14 bit / 8 = 83 MByte/s. That's more or less the maximum bandwidth the cam can do for continuous recording.
That is *without* compression. For standard scenes you may get up to factor 0.4. Which tunes it down to a mere 33 MByte/s.

Benchmark your card.
https://www.magiclantern.fm/forum/index.php?topic=26941.msg244189#msg244189
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 05, 2023, 12:03:24 PM
For me, it's a hunt for more resolution that's the lure... not fussed about extra bit depth. If possible, would stop me eyeing up Sony Alphas on thebay.
This achievable
FYI my card benchmarked at 44MBs, happy to get a faster one though, if needed.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 05, 2023, 12:22:51 PM
44 MB/s is without overclocking and for read, I assume.
What is write showing?

Please lookup link and Bilal's primer video.
A card compatible with 240 MHz mode should show write numbers above 90 MByte/s.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 05, 2023, 12:39:39 PM
Write 41.3 MB/s, read 43.6 MB/s

Thought realise now that's with the dial in Photo mode.

Run now with 100D in movie mode & oddly different results!
36.7MB/s for write
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 05, 2023, 12:50:33 PM
Lower numbers in movie mode are unavoidable.
In older times (= last year ;-) ) movie mode benchmark numbers were a good indication for cam's performance for recording.
And photo mode (esp. PLAY-UNK) just told the cam's physical limits without overhead by cam's task during recording.

So: Don't waste your time in movie mode. Switch back to photo mode and try overclocking modes.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 05, 2023, 04:12:06 PM
Before I buy a card that'll do ~ 90MB/s, does this new ML allow for higher than HD recording, but direct to MP4? Not fussed about increased bit depth, just like to go beyond 1080 dots.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 05, 2023, 04:18:58 PM
Did you run benchmark with SD_UHS loaded, Overclocking On and set to 240 MHz?

And nope. ML cannot manipulate H.264 resolution.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 05, 2023, 04:26:14 PM
Yes ..tried the 3 diff clock speeds, different modes. All run SLOWER. Useless little card!
OK so new card ordered. 140MB/s :)
And.. if I invest the time to work out settings, what resolution above HD can I reliably achieve, for a recording length of over 5 mins length?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 05, 2023, 04:29:21 PM
You haven't tried to change Access Mode to "SDR50", right? But you should. Setting in Overclock submenu.
Card speed will drop into 20 MByte area if card cannot deal with oc settings.

If you take care with exposure you can do the math on your own:

Vertical pixel x horizontal pixel x bit depth x fps x compression factor / 8 < 80 MByte/s
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 05, 2023, 04:53:05 PM
Yup dropped to 20! New card coming anyhow :)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 05, 2023, 05:00:31 PM
May I ask which one you ordered?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 05, 2023, 05:33:00 PM
SanDisk Ultra SD Memory Card 64GB  SDHC Class 10 For Cameras
It's rated 140MB/s should do the trick
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 05, 2023, 05:59:05 PM
Incompatible with sd_uhs.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: names_are_hard on July 05, 2023, 06:32:50 PM
Quote from: edt on July 05, 2023, 05:33:00 PM
It's rated 140MB/s should do the trick

Manufacturers list the read speed, because it's bigger.  You care about the write speed.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: edt on July 05, 2023, 07:34:16 PM
Quote from: Walter Schulz on July 05, 2023, 05:59:05 PM
Incompatible with sd_uhs.

How do I tell what cards ARE ok?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 05, 2023, 07:38:07 PM
https://www.magiclantern.fm/forum/index.php?topic=25841.0
In my area Samsung EVO Plus are frequently on sale giving them best price/performance ratio.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Wannabe on July 09, 2023, 08:46:47 PM
Hi guys, I just found this thread.
I got myself an EOS 100D and I have the following problem: Crop Mood does not work on it.
Crop Mood interrupts the recording after about one second and then it turns the camera off.

Here is a video that showcases the problem:

https://www.youtube.com/shorts/WOl37HwIeds
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: names_are_hard on July 09, 2023, 09:12:37 PM
First off: the video shows ML working on the cam.

It looks like you're trying to record in 5.2k.  This requires a card capable of fast writing.  What do benchmarks show you for this card?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Wannabe on July 09, 2023, 09:24:49 PM
Quote from: names_are_hard on July 09, 2023, 09:12:37 PM
First off: the video shows ML working on the cam.

It looks like you're trying to record in 5.2k.  This requires a card capable of fast writing.  What do benchmarks show you for this card?

(https://i.postimg.cc/2jWcxbTG/bench1.jpg)
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: names_are_hard on July 09, 2023, 09:30:02 PM
Confirms that the card doesn't suck, that's something.

Do you get long recordings at lower resolutions?
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Wannabe on July 09, 2023, 09:36:30 PM
Quote from: names_are_hard on July 09, 2023, 09:30:02 PM
Confirms that the card doesn't suck, that's something.

Do you get long recordings at lower resolutions?
It only worked once out of many tries. I was able to record about 20 seconds of video at 4.8K.
My card's volume is only 64 GB. I was told that these small volume cards can be problematic and that I should buy a new card that is at least 128 GB.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Walter Schulz on July 09, 2023, 10:11:38 PM
There are cards types affected by size (low storage size with reduced performance) but this would show in benchmark. Your benchmark runs are ok which means your card is not affected.

Your card is not the fastest but a mere 5 percent difference to best performers should only have marginal effects on resolution/recording duration. Just take care not to overexpose which will break compression thus causing buffer overflow in an instant.
Make a test record in an environment with controlled lighing.
You should be able to get continuous recording somewhere between 70 to 80 MByte/s.

Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Wannabe on July 09, 2023, 10:27:16 PM
Quote from: Walter Schulz on July 09, 2023, 10:11:38 PM
There are cards types affected by size (low storage size with reduced performance) but this would show in benchmark. Your benchmark runs are ok which means your card is not affected.

Your card is not the fastest but a mere 5 percent difference to best performers should only have marginal effects on resolution/recording duration. Just take care not to overexpose which will break compression thus causing buffer overflow in an instant.
Make a test record in an environment with controlled lighing.
You should be able to get continuous recording somewhere between 70 to 80 MByte/s.
Thanks. It is already night where I live and I will try to test it in a dark room where there is not much light.
Will come back with an update in a few minutes.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Wannabe on July 09, 2023, 11:16:55 PM
I tried to record a video in a dark room. No luck again.
Crop Mood/ML crashes all the time.
I tried so many times, that I became an expert at pulling out the battery in record time.

Then the aperture dial stopped working.

Then the ISO cannot be changed and it is stuck at 6400. No way to lower it.

Then the custom buttons cannot be adjusted from the ML settings.
I tried to set Up/Down arrows for ISO control and Left/Right arrows for aperture control. I then restarted the camera. It did not work.

Then I got the following error screen:

(https://i.postimg.cc/bwd4BXX9/IMG20230709235452.jpg)

Nothing works, folks.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 09, 2023, 11:47:11 PM
The crashes you are having are related to crop mood module because of changing preview configuration or in other words something related to LiveView (that what I can tell from this log (https://www.magiclantern.fm/forum/index.php?topic=26947.msg244287#msg244287)).
Please share your ML settings, which preset exactly are you using at which aspect ratio (from crop mood menu)? Do all crop mood presets cause crashes?

The SD card isn't related to camera crashes (lock up) in this case.

Customize buttons only work while crop mood is active and in normal preview.
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: theBilalFakhouri on July 09, 2023, 11:55:19 PM
Also please consider to report any problem related to crop mood build in Crop mood - WYSIWYG Edition (650D / 700D / EOS M / 100D) (https://www.magiclantern.fm/forum/index.php?topic=26851.msg242974#msg242974) thread to keep things more organised.
I might lock my previous threads for my previous 100D (https://www.magiclantern.fm/forum/index.php?topic=26511.0) / 650D / 700D (https://www.magiclantern.fm/forum/index.php?topic=25784.0) custom builds, and keep them as archive (I will no longer support them).
Title: Re: Bilal's crop_rec_4k experiments for 100D / SL1
Post by: Wannabe on July 10, 2023, 12:08:45 AM
Quote from: theBilalFakhouri on July 09, 2023, 11:55:19 PM
Also please consider to report any problem related to crop mood build in Crop mood - WYSIWYG Edition (650D / 700D / EOS M / 100D) (https://www.magiclantern.fm/forum/index.php?topic=26851.msg242974#msg242974) thread to keep things more organised.
I might lock my previous threads for my previous 100D (https://www.magiclantern.fm/forum/index.php?topic=26511.0) / 650D / 700D (https://www.magiclantern.fm/forum/index.php?topic=25784.0) custom builds, and keep them as archive (I will no longer support them).
Alright, thanks.
I will stop posting in this thread and will use the one you linked instead.