Author Topic: Danne's crop_rec_4k experiments for EOS M  (Read 369448 times)

amitkattal

  • New to the forum
  • *
  • Posts: 34
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2950 on: January 01, 2021, 06:18:25 PM »
Only in the step when transcoding to DNG, so in MLVApp or MLVFS. Resolve can't remove that.
So if i output as cinema dng from mlvapp, there is no way i can remove the focus pixel maps later? because any edit made in mlvapp doesnt apply to cinemadng

masc

  • Contributor
  • Hero Member
  • *****
  • Posts: 2104
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2951 on: January 01, 2021, 06:29:12 PM »
So if i output as cinema dng from mlvapp, there is no way i can remove the focus pixel maps later?
Right.

Because any edit made in mlvapp doesnt apply to cinemadng
Then you did something wrong. All RAW corrections will be applied. You probably exported via "fast pass" ( = bypass without any changes ).
5D3.113 | EOSM.202

ytolun

  • New to the forum
  • *
  • Posts: 3
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2952 on: January 07, 2021, 11:17:12 AM »
Hello everybody, I have two problems with ML NOV20 build.

1) I can't use my lens in AF mode while using 5K frtp. It just freezes the LiveView and doesn't focus on anything.
2) I can't use HDMI output to my phone with a capture card for 1080p mode or for 5K frtp mode. On 1080P mode, I get a black screen with no overlays. ON 5K frtp, the overlays work but I see pink distorted lines instead of the video feed.

Do you know any solutions for these problems?

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2953 on: January 08, 2021, 08:22:56 PM »
1) Skip af when working movie modes. Treat eos m as a manual focus cinema camera for best results.
2) Skip HDMI out altogether when filming with eos m for best results.
If any of above works with af on or hdmi out consider yourself lucky.

tupp

  • Member
  • ***
  • Posts: 130
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2954 on: January 09, 2021, 09:01:35 AM »
2) I can't use HDMI output to my phone with a capture card for 1080p mode or for 5K frtp mode. On 1080P mode, I get a black screen with no overlays. ON 5K frtp, the overlays work but I see pink distorted lines instead of the video feed.

@ZEEK posted a video just a few hours ago that includes how to get an HDMI monitor working with the EOSM.

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2955 on: January 09, 2021, 10:38:05 AM »
Yeah, that guy is really testing those presets out 8).
https://www.youtube.com/c/ZeekEOSZ/videos


tupp

  • Member
  • ***
  • Posts: 130
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2956 on: January 09, 2021, 09:47:05 PM »
Yeah, that guy is really testing those presets out 8).
Well, instead of just discouraging someone by telling them to skip altogether trying to use an HDMI monitor with ML, @ZEEK helpfully:
  • advises which ML presets are ideal for HDMI;
  • shows how to adjust the ML settings within those presets to eliminate HDMI problems;
  • and instructs on how to get the best framing from a monitor.
     ;)

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2957 on: January 09, 2021, 09:57:47 PM »
Well, that´s what I like with magic lantern. When users finds ways and do the unexpected. Personally I still think af and monitors with eos m is way too much hassle than recording without.
Then again he asks for usage with his phone. Don´t know if that´s possible.

ZEEK

  • Senior
  • ****
  • Posts: 329
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2958 on: January 10, 2021, 12:08:14 PM »
I have to give much appreciation to the developers who have done a fantastic job coding these things for considerably some time. It can be a lot of work for sure! And a lot of things with ML can be unpredictable.

We can't expect things to work straight off the bat, which is why these builds are experimental. Feedback from users with uploaded samples is highly recommended to fix some of the issues.

As for HDMI stuff with Monitors, it can be like playing with fire. The 2.5K HDMI Preset is the best for high resolutions as I have yet to see a dropped frame after a few 1-minute or so recordings. 5K FRTP is also quite good. 1080 mode has it's ups and downs.

Using your smartphone as a monitor with the EOS M is good for 2.5K/2.8K RAW, zero dropped frames and clear image. Other modes go blank. The best would be to test with different capture cards.

But as Danne said, HDMI Monitors will present more issues, I have just tried to exemplify what I have done as I like to use monitors to record myself when sampling lenses for skin tones on different RAW Modes.  8)


EOS M

ytolun

  • New to the forum
  • *
  • Posts: 3
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2959 on: January 10, 2021, 12:29:18 PM »
Thanks, everyone for all your replies. My post wasn't approved for a while and in the meantime, Zeek has replied to my comment on YouTube.

As for the result, I guess I knew what I was going into when I decided to shoot with an EOS M as opposed to my smartphone  :) Even though it's cumbersome at times, I learned a lot. I didn't even know what an aperture was before all this :D

Walter Schulz

  • Contributor
  • Hero Member
  • *****
  • Posts: 8659
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2960 on: January 10, 2021, 02:10:22 PM »
@Zeek: I'm interested how you manage audio for your youtube content. Too much content (definitely not yours) suffers in this regard. If you can setup a how-to (what works and what doesn't) about good sound on a budget I think the audience will listen!
Of course if you have a bunch of Neumann mics ... forget it! ;-)

ytolun

  • New to the forum
  • *
  • Posts: 3
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2961 on: January 10, 2021, 02:55:57 PM »
@Zeek: I'm interested how you manage audio for your youtube content. Too much content (definitely not yours) suffers in this regard. If you can setup a how-to (what works and what doesn't) about good sound on a budget I think the audience will listen!
Of course if you have a bunch of Neumann mics ... forget it! ;-)

For YouTube, I tape my very cheap lavelier mic under my t-shirt. Cheap mics don't pick up much reverb and taping it under my shirt keeps it reasonably close to my mouth.

Give it a listen (it's not in English but still) and if you like what you hear, give it a try:

NateM

  • Just arrived
  • *
  • Posts: 1
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2962 on: January 19, 2021, 10:25:52 PM »
Hello! After purchasing an EOS M in late 2020 (under the influence of DSLR Video Shooter and Zeek's YouTube videos) I began learning ML. I successfully shot some raw clips and was impressed by the dynamic range when color grading. But extremely frustrated by the camera frequently freezing or the video preview being nonexistent.

I honestly had given up on ML until I came across this video (https://www.youtube.com/watch?v=-r-tI93GZBk&t=34s&ab_channel=PilYang). I was amazed that the filmmaker had been able to make the shots look sharp and there was almost zero noise in the night scenes. In the description he mentioned it had been shot in h.264 not raw.

So I dove back in ML on my EOS M and realized maybe he had just cranked the CBR to 3x and maxed out the data he could record in h.264. I tried it and loved the fact the camera was more reliable and the clips played back just fine. But then I wanted to confirm that the data rate was actually 3x. I struggled to find the manufacturers bit rate for 1080p at 24fps for the EOS M.

But I found that the T4i shared the same exact sensor and that camera records 1080p at 24fps at a vbr of 48 mbps (See the second photo). So I analyzed the clips I recorded on my camera (1080p at 24fps) with a CBR set at 3x and the bit rate was only 55 mbps (see image). Then I recorded another clip with the CBR turned back to 1x with the same settings. And the bit rate was 47 mbps.

So my question is 55 mbps the max bit rate on the EOS M in H.264? And do we think his video had good image quality simply because he used fast glass and a CBR instead of a VBR?

On a side note I have been renting the new Canon R6 for professional photo and video work and it has been amazing.

Thanks


Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2963 on: January 20, 2021, 12:51:47 PM »
Interested to why you are getting freezes? What build are you using and what settings? When recording h.264 it should be as stable as without magic lantern.
I am yet to find proof that image will get better cranking this setting. I do not think it will make a difference using this setting. Not when I tested it. Better yet to film raw if you want some real magic to happen.

eosmrocks

  • Just arrived
  • *
  • Posts: 1
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2964 on: January 20, 2021, 06:57:08 PM »
Hello everyone,

I have a very simple question regarding the zoom assist via half-shutter. You see, I am using manual lenses and from time to time I need to zoom on the live view to check my focus, and there are many instances when the focus peaking feature does not do the job quite enough.
In Canon's GUI for the live view, there is the magnifier button and once I click that I can choose a point in the screen and zoom there, check the focus, zoom back, check my composition and shoot. This process is long. The best solution is to set the half-shutter button to zoom but unfortunately with that I can only check the focus on center. Is there any way to make it such that I would be able to touch on the screen thus moving the focus box to a certain point and then once I press the half-shutter it would zoom at that specific point rather than the center?

Thank you for any help.

Janke

  • Freshman
  • **
  • Posts: 97
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2965 on: January 24, 2021, 01:22:38 PM »
One quick question about the intervalometer and/or low fps settings:

If the battery runs out during shooting, will the file always be saved correctly? (I tried a search, didn't find anything...)

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2966 on: January 24, 2021, 04:42:14 PM »
Please do testing and report back.

2blackbar

  • Hero Member
  • *****
  • Posts: 514
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2967 on: February 01, 2021, 09:25:28 PM »
Wheres now working link to recent build ? Due to recent split of the thread , links go nowhere in camera specific thread.
OK, its first page now, i guess thats better.
Also 2.8k lasts longer than previous builds, what kind of magic happened ?

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2968 on: February 01, 2021, 09:56:15 PM »
Yes, better. Amazing we are at 119 pages here.
Recently I fixed a bug when using halfshutter while using anamorphic frtp preset.
Also added an experimental "shutter average" function. It will record while changing shutter every 7th frame. An average tool.

2blackbar

  • Hero Member
  • *****
  • Posts: 514
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2969 on: February 01, 2021, 10:22:07 PM »
Yeah, You did amazing work with it.
M is a special cam, really great colours, i cant get them from 5D2 and i keep coming back to M over and over again.
As for shutter, in 1080mcm i usequite often shutter of 1/41 but only with fps override exact fps, it gives me nice non stuttery motion blur while not looking like muddy camcorder blur yet.
 It wont work with other fps override methods like lowlight or fps override off. Sometimes i also use 1/45 but only with exact fps.
I noticed that i cant use exact fps in 2.8k and other crop modes , i have to use 1/50, i have lights at 50hz and with fps override exct fps and shutter 1/45 i get no banding , on other modes or fps override off i have to set it to 1/50 to get rid of banding , so exact fps is a special mode where i can get less stuttering on panning and moving objects, motion blur just looks different on it, not sure why and how but it does, without fps override exact fps 1/45 looks too blurry.

I made a color test with M and 5D2, both same light and same exposure, in mlvapp colours look different, especially on 5D2 its kinda orangey where it should be reddish.
Heres 5D2 frame:
https://drive.google.com/file/d/1RKpYa74qClOBkwEOspcC9jj3HnrG997T/view?usp=sharing
M frame:
https://drive.google.com/file/d/1OiF8FNmcuHAJs1kqnnuhw2NRrnaECH_5/view?usp=sharing

Colours from both cams should look like on M.Maybe ill ask ilia and masc whats happening there.

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2970 on: February 01, 2021, 10:43:59 PM »
I never use fps override with my crop_rec presets. Could you show me comprable examples of when you are getting banding vs not getting banding modifying shutter, fps override as you describe?

Your comparison between 5D2 and eos m. Is it white balance related? I matched wb and they look pretty similar to me. Adobe camera raw.






fine tuning shutter works with 2.8k


2blackbar

  • Hero Member
  • *****
  • Posts: 514
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2971 on: February 01, 2021, 11:01:47 PM »
Yes in adobe camera raw it looks similar on both but in mlvapp i get different colours on 5D2 , im sure theres something going on, both colors above yellow should be red and redder, not orangey like on 5D2, i cant get nice red colour from 5D2 no matter how i white balance it in mlvapp.
5D2:

M:


As for fps override, im talking about rolling shutter timing, it has faster readout in exact fps, 18.7ms rolling shutter value . i can get it to 17.8ms , id like it to stay like that by default( on 17.8 ), i might have to edit sourcecode again.

I like the look of motion blur with fps override.Banding is not an issue for me, its just i want fastest readout from sensor and different looking motion blur.

Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2972 on: February 01, 2021, 11:22:40 PM »
I guess I never learn the difference with this rolling shutter ;)
You can also fiddle the timer A and B from the crop mode sub menu.
reg_6014 = timer B
reg_6008 = timer A
Maybe you can optimize from there?

2blackbar

  • Hero Member
  • *****
  • Posts: 514
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2973 on: February 01, 2021, 11:51:03 PM »
I dont see these regs affecting rolling shutter value, but i see that rolling shutter lower than 23ms in crop modes will get me no preview/glitchy image.I have it in fps override menu at A ft-106 and B FT+5) but anything lower than 23 and theres no image, i dont know if it was like that long time ago but i remember using fps override in crop modes with success but on other hand iwanst able to record for so long so maybe its a trade off.


Danne

  • Developer
  • Hero Member
  • *****
  • Posts: 7701
Re: Danne's crop_rec_4k experiments for EOS M
« Reply #2974 on: February 02, 2021, 05:51:53 AM »
Timers affects performance and evidently output. Question is by how much and when. Would be nice to see examplas from comparison scenes.