Author Topic: Danne's crop_rec_4k experiments for EOS M  (Read 239646 times)

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1750 on: October 01, 2019, 06:06:19 AM »

mesebar2

  • New to the forum
  • *
  • Posts: 31
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1751 on: October 01, 2019, 07:15:00 AM »
You just download the focus pixel maps from the link below, unzip them and drag them onto the black screen on MLV App. If it's still there, 3x3 chroma smooth shouldn't hurt ;)
MASC Focus Pixel Tutorial Video found on this page towards the bottom:
https://www.magiclantern.fm/forum/index.php?topic=16054.675
Focus Pixel Maps Download:
https://www.youtube.com/redirect?q=https%3A%2F%2Fbitbucket.org%2Fdaniel_fort%2Fml-focus-pixels%2Fdownloads%2F&redir_token=XAGCeaNtGohw2MmJkXwYfOzsvPN8MTU2OTk3MDE3NEAxNTY5ODgzNzc0&v=zbOEAD9RfJ0&event=video_description

Thank you so much, ZEEK!
I'll try and let you know if it worked as soon as I'll be back home.
Thank you again!

EDIT: Thanks ZEEK, it worked perfectly!!!

ngemu

  • Freshman
  • **
  • Posts: 83
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1752 on: October 01, 2019, 08:38:43 AM »
Absolutely loving the new aperture up and down custom buttons. Sometimes, clicking the buttons too fast or soon causes a glitch where the aperture shoots all the way up or down.

When it gets dark, i switch back to the iso custom buttons.

Makes shooting so much more enjoyable :)

Johnny Yul

  • New to the forum
  • *
  • Posts: 3
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1753 on: October 01, 2019, 03:41:36 PM »
10bit raw is a great solution, but there is a problem, 10bit is linear right? so it does not contain enough data for mid tone and shadow area, worse than 10bit log (not raw file), i think it should be recorded in log profile, not in post-production in MLVApp (data has already lost at that time), it should be done within the progress something like cmos data to 10 bit.
i think 10bit log profiled raw is the perfect solution, it can replace 14bit linear raw in most cases, 10bit linear raw has a obvious gap to 14bit linear, especially when grading mid tone an dark area, mid tone is so important!

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1754 on: October 01, 2019, 03:56:02 PM »
10bit raw is a great solution, but there is a problem, 10bit is linear right? so it does not contain enough data for mid tone and shadow area, worse than 10bit log (not raw file), i think it should be recorded in log profile, not in post-production in MLVApp (data has already lost at that time), it should be done within the progress something like cmos data to 10 bit.
i think 10bit log profiled raw is the perfect solution, it can replace 14bit linear raw in most cases, 10bit linear raw has a obvious gap to 14bit linear, especially when grading mid tone an dark area, mid tone is so important!
What is you point? 10bit is a compromise to get continuous recordings. If 12bit or 14bit works use that but we are talking eos m here. Not the most obvious cpu powerhorse  :-*

Johnny Yul

  • New to the forum
  • *
  • Posts: 3
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1755 on: October 01, 2019, 04:07:00 PM »
What is you point? 10bit is a compromise to get continuous recordings. If 12bit or 14bit works use that but we are talking eos m here. Not the most obvious cpu powerhorse  :-*
thank you for your great job, first! i am using your build.
i mean current 10bit raw is linear, right? if it is log profied, it will be almost perfect, just within the progress down sampling to 10bit, i don't know whether the cpu can afford it (i use eosm too) , just a suggestion :)

Luther

  • Senior
  • ****
  • Posts: 323
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1756 on: October 01, 2019, 04:13:01 PM »
10bit raw is a great solution, but there is a problem, 10bit is linear right? so it does not contain enough data for mid tone and shadow area, worse than 10bit log (not raw file), i think it should be recorded in log profile, not in post-production in MLVApp (data has already lost at that time), it should be done within the progress something like cmos data to 10 bit.
i think 10bit log profiled raw is the perfect solution, it can replace 14bit linear raw in most cases, 10bit linear raw has a obvious gap to 14bit linear, especially when grading mid tone an dark area, mid tone is so important!
Raw data is raw data. There's no "log" in raw. I think you have a reading to do. As Danne said, you can use 12-14bit if you need more dynamic range.

Johnny Yul

  • New to the forum
  • *
  • Posts: 3
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1757 on: October 01, 2019, 05:38:22 PM »
Raw data is raw data. There's no "log" in raw. I think you have a reading to do. As Danne said, you can use 12-14bit if you need more dynamic range.
raw data is also numbers array in memory (a kind of voltage value? ), number can of course be remapped by a non-linear function curve (such as log), just value modification, nothing else.
A log profiled data does not need special data or treatment, only the remapped value taking place of the original one, same size and structure as the original, the only thing need to do is to let the cpu remap the value when down sampling to 10 bit, it is a kind of tiny calculation, i wish cpu can afford it, although, it can already do the down sampling calculation, right?
Anyway, thanks to all the contributors :)

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1758 on: October 01, 2019, 05:44:02 PM »
My bitdepth implementation is a crude, simple one. Lowering analog gain registers from around 400(14bit) and then to 12/10bit values. If there are other values to tweak I don´t know how to do that. This modification was realized from an idea from Levas. Code in crop_rec.c:
Code: [Select]
            if (bitdepth == 0x3)
            {
                /* 10bit roundtrip only not applied here with following set ups */
                adtg_new[13] = (struct adtg_new) {6, 0x8882, 60 + reg_gain};
                adtg_new[14] = (struct adtg_new) {6, 0x8884, 60 + reg_gain};
                adtg_new[15] = (struct adtg_new) {6, 0x8886, 60 + reg_gain};
                adtg_new[16] = (struct adtg_new) {6, 0x8888, 60 + reg_gain};
            }
           
            if (bitdepth == 0x4)
            {
                /* 12bit roundtrip only not applied here with following set ups */
                adtg_new[13] = (struct adtg_new) {6, 0x8882, 250 + reg_gain};
                adtg_new[14] = (struct adtg_new) {6, 0x8884, 250 + reg_gain};
                adtg_new[15] = (struct adtg_new) {6, 0x8886, 250 + reg_gain};
                adtg_new[16] = (struct adtg_new) {6, 0x8888, 250 + reg_gain};
            }

Luther

  • Senior
  • ****
  • Posts: 323
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1759 on: October 01, 2019, 05:52:03 PM »
Not sure if I'm following you guys. Isn't ML just reducing the amount of possible tones per channel from 16.384 (14-bit) to 1024 (10-bit)? How could you apply "log" into this if it's an absolute value?

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1760 on: October 01, 2019, 05:59:22 PM »
Log or not this is how it's done on my end  :-*.
Log is done in post not in cam so I only wanted to share the analog gain regs. Maybe it gives some ideas

AF-OFF

  • Member
  • ***
  • Posts: 109
  • 50Dc EOSM
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1761 on: October 01, 2019, 06:51:02 PM »
just did a little testing with the actual build and noticed that photos (accidentaly) taken in videomode aren't black anymore but regular .CR2 files Which is nice
The file size of 20.4 MB is identical to a photo taken in either mode, but the resolution is different
With the 2.5k 1:! preset enabled, it's height is 2916px instead of the 3456px of the photo mode cr2, the width is the same 5184px

Thank's to Danne and others

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1762 on: October 01, 2019, 06:58:30 PM »
Yup. Fixed that. You can get full height by changing metadata in post.

AF-OFF

  • Member
  • ***
  • Posts: 109
  • 50Dc EOSM
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1763 on: October 01, 2019, 08:43:56 PM »
Hey Danne! that is great
Didn't complain about it in anyway more it's the opposite
But hey, You know what is weird? I complain to myself since years, not to having an icon of .mlv files on a computer
And this could be the solution. Just autotake a photo before recording a video
Then we would have an image in "finder" to discern which clip is what..
So anyway, thanks a lot for Your work

2blackbar

  • Hero Member
  • *****
  • Posts: 507
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1764 on: October 02, 2019, 12:41:39 AM »
Danne i dont know what magic You did there but its working A LOT better now, previously i could not record for 20 seconds without stopping even if icon was green, now i recorded over 1 minute without stopping , twice in a row and no issues, ill test more but i already see a difference.
 Thank You !

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1765 on: October 02, 2019, 10:25:43 AM »
New build:
https://www.magiclantern.fm/forum/index.php?topic=9741.msg208959#msg208959

- Fast access to 5k anamorphic full sensor readout

kuchka94

  • New to the forum
  • *
  • Posts: 17
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1766 on: October 02, 2019, 02:09:09 PM »
Hi! Thanks for your work! Latest builds works better and better=)
But I can't find in this builds video record restart, I need to take some footage more than 30 min.
Movie restart under raw tab doesn't work, I use H264 record.

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1767 on: October 02, 2019, 02:17:09 PM »
Movie restart only for raw in this build if I recall correctly. Might add back for h264

kuchka94

  • New to the forum
  • *
  • Posts: 17
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1768 on: October 02, 2019, 03:48:29 PM »
Movie restart only for raw in this build if I recall correctly. Might add back for h264
yes, please, add for h264, it will be great! thank you!

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1769 on: October 02, 2019, 04:45:00 PM »
New version:
https://bitbucket.org/Dannephoto/magic-lantern/downloads/crop_rec_4k_mlv_snd_raw_only_2019Oct02.EOSM202.zip

- Movie restart always enabled when recording h264. Don´t see any reason for having it disabled?

Please test @kuchka94


Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7392
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1771 on: October 02, 2019, 04:53:56 PM »
Up now, sorry...

turnlemons2lemonade

  • New to the forum
  • *
  • Posts: 17
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1772 on: October 03, 2019, 09:57:19 AM »
Thank you Danne! Been testing out and putting content with your latest builds on my channel.

AWESOME WORK! Danne, give me the link to donations, I wanna return the favor for your work.

Here are a few things I noticed:
1. When recording MCM Rewire 10bit after I start recording I get some green lines on the screen.
2. Sometimes I get dropped frames (pink overlays) when recording with a external monitor - I have a FEELWORLD MA5 (which is 1080p and can't be switched to 480) - is there a way around it to reduce the amount of CPU usage so I can still use the external monitor?
3. It seems that I get dropped frames with both 14 / 10 bit MCM Rewire. That's a pity as I can't use this monitor when recording, just check framing/focus and turn it off.
4. I usually use Histogram with Exposure to the right and Kill GlobalDraw OFF (as you told Zeek) and Turn GlobalDraw OFF after 30 secs

Been pushing this little thing to create YouTube content and I have to say, the quality is better then my A6400.

Another thing that I would improve to be easier for new users (even though the presets are just amazing and makes life so much easier) would be to have a better workflow for converting RAW to ProRess directly.

Keep it up and looking forward to it!

masc

  • Contributor
  • Hero Member
  • *****
  • Posts: 2011
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1773 on: October 03, 2019, 10:11:25 AM »
3. It seems that with 10bit I don't get that much dropped frames is this easier on the CPU?
Not to the CPU, but to the card interface.

Another thing that I would improve to be easier for new users (even though the presets are just amazing and makes life so much easier) would be to have a better workflow for converting RAW to ProRess directly.
There will never be a ProRes encoder in this cam. You'll always have to convert it on another machine. Use Switch, MLVProducer or MLVApp. It can't be easier.
5D3.113 | EOSM.202

kuchka94

  • New to the forum
  • *
  • Posts: 17
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #1774 on: October 03, 2019, 03:42:38 PM »
New version:
https://bitbucket.org/Dannephoto/magic-lantern/downloads/crop_rec_4k_mlv_snd_raw_only_2019Oct02.EOSM202.zip

- Movie restart always enabled when recording h264. Don´t see any reason for having it disabled?

Please test @kuchka94

Works perfect! Thank you!