Canon 5D Mark IV

[OlRivrRat]

                 @Walter

"Not sure which post you are replying to. I think it may be"

   So, Just which post might this have been a reply to ~

[dfort]

Just saw this on Youtube:
https://www.youtube.com/channel/UCFQI3-xCk_m8hWMYQwT05GQ
Log Neutral on the 5D Mark IV. Looks promising.
The video links to this thread:
http://www.magiclantern.fm/forum/index.php?topic=16299.0

Cool. Looks like the Reverse Engineering Picture Styles topic is getting some press.

[Maximo]

The mjpeg file format for video on the 5D mark IV makes HUGE files. Really bad choice. Correct me if I'm wrong but if magic lantern manages to hack the 5D4, it would potentially enable raw recording but won't be able to enable recording in mp4 H265 for example? This could only be implementend by Canon, is that correct?

[bouncyball]

MJPEG on Canon _is_not_a_bad_choice_ at all. It's just 8 bit and 4.2.0 and that is a bit dealbreaker not the size (bitrate 500mbit if I remember correctly). Raw data would be even bigger .

And I'm sure recording 4K raw will never be possible longer than available RAM allows (several secs I guess) on 5dm4 with ML.

Regards
bb

P.S. Yes you are right about H265.

[ph2007]

And I'm sure recording 4K raw will never be possible longer than available RAM allows (several secs I guess) on 5dm4 with ML.

1080 60/120fps raw I am darn happy with it, otherwise i would stay on a7s or fs5 haha

[myown]

4K mjpeg is in 4:2:2 on 5d mark iv

I did read the information about fps override. But i'm not shure about that: could there be a posibility to record full hd (1080p) with more than 60fps in h.264 with ml?

Ps: 1080p@60fps in raw....not a big hope for that. Depending on buffer size some seconds are possible. 1080p@120fps in raw-> no way :-)

[noipego]

It's just 8 bit and 4.2.0

its 422 in 4K

[Maximo]

I don't mind having huge files if they provide me with flexibility in post-production but Canon's 4K MJPEG don't.
As anyone tried the Log Neutral (link above) on the 5D4? If so, would you be OK to share some of that footage?
Having no peaking for me is just impossible, I hope ML will bring it to the 5D4.

[myown]

Having no peaking for me is just impossible, I hope ML will bring it to the 5D4.

What do you mean?

I made some tests with picture styles (not Neutral, i used my own) you'll get more details in dark or bright parts.
But it's not the same like raw Video :-P

[Maximo]

I'm talking about the lack of focus peaking (https://en.wikipedia.org/wiki/Focus_peaking).
As far as I know the 5D4 doesn't have any.

[noipego]

but Canon's 4K MJPEG don't.

im also wondering what you mean...have you tried to grade mjpeg and you find results same as h264? i dont have the camera yet but some footage i see with eoshd log looks 'different' from usual canon's compressed files
https://www.youtube.com/watch?v=XJDU4LAu194

[bouncyball]

@ noipego, myown

4.2.2. Yeah it's even better now thanx for correction.

bb

[Maximo]

@noipego
I personnaly don't mind having huge files if they provide me with flexibility in post-production: RAW or LOG.
Canon's 4K MJPEG is heavy and a lot more difficult to grade than raw or Log.
MP4 H264 or H265 wouldn't be that much different to MJPEG to grade but at least much lighter.

[djkraq]

Interesting.  I am using a Komputerbay 1066x 256GB card and can record 1080p120 for over 6 minutes to CF Card on my 1DX Mark II (although Canon says only the CFast Card can do that).  Maybe the 5D Mark IV is just limited by the buffer?  IDK. 

[AOK]

There's another thread for the processor being hacked. It's about the same ML.


Sent from my iPhone using Tapatalk

[myown]

Interesting.  I am using a Komputerbay 1066x 256GB card and can record 1080p120 for over 6 minutes to CF Card on my 1DX Mark II (although Canon says only the CFast Card can do that).  Maybe the 5D Mark IV is just limited by the buffer?  IDK.

Interessting  How we can readout the buffer size of the mark iv without ML? And whats the buffer size of your 1dx?
I'm also not sure if the mark iv is able to process 120 full hd frames per second because there is only a single processor.
And not sure either if the fps override is doing the 120 frames in 1080p.

[Prokopios]

 Could it not be possible to write RAW to an external writer, eg. Pix-E5, for the 5D IV?

[tron]

Wishing for a 5DIV port in a couple of years (not trolling, If possible, if there is time, programmers, etc).

I am thinking of Dual ISO feature since there are still shadows issues believe it or not (I am on my 2nd 5D4 and the replacement body has equal color banding with the 1st when lifting shadows, see canonrumors forum for more info). My 5D3 with ML is better than 5D4 on this!

[a1ex]

Have some relevant samples to show the repeatability of the pattern, and whether it can be corrected with changing ISO?

A few ISO 100/1600 brackets should do the trick.

BTW, the latest Eeko discovery is very useful for digic 6 ports: it's a simplified DryOS core that runs Thumb-2 instructions (probably similar architecture), all its memories are writable, it's available on all Digic 5 models, so it can serve as a playground for testing Thumb-2 stuff and understanding other low-level behaviors.

[tron]

Alex if you were referring to my post I can try. I just know that iso 50 improves slightly the situation.

Actually I have made some experiments with higher iso (400) but I tried to not burn the highlights so DR was less (of course). The banding was evident. But I believe to try to simulate a higher iso in the ML way I just have to increase ISO (say to 1600) and keep shutter and aperture constant. Highlights will be be burnt but we will get some info for the shadows and higher iso.

Please correct me if I am wrong about that.

[tron]

Color banding disappears when increasing iso and keeping other parameters the same since there is less underexposure of the shadows. But since this causes a more correct exposure for the shadows (overexposing the rest of the picture) I guess it would happen like that.

[garry23]

@tron

If you haven't read this, I thought you might find it of value : http://www.clarkvision.com/articles/iso/

Cheers

Garry

[JADURCA]

@tron

Read this too. https://en.wikipedia.org/wiki/Colour_banding

Try to add a little bit of noise to the footage, render it then verify. Look for a rasonable low noise that is almost not detectable and mask the banding. Also, as @a1ex suggest, increase the ISO when filming solid natural colors like a blue sky or sunset, , so a little bit of noise is naturally added by the internal digital circuits and natural behavior of high ISOs.

[tron]

I may be wrong of course but it does not look like posterization and certainly does not look like just chroma noise.

Also it is ALWAYS in parallel to the longest side no matter what I shoot. Since I do not shoot video I translate filming to shooting.

I would rather not increase ISO since it lowers Dynamic range. Experiments made with high iso and trying to recover shadows while keeping highlights from burning gave poor results. I have to keep exposure lower to save highlights (less DR) so shadows are even worse (and there was banding).

ISO 50 behaved better (although it is a digital iso, the way camera handles it saves us from quite a processing). EDIt: It betters 5D3 in shadow lifting too that's why I tried it with 5D4.

To tell the truth some shadows have to be kept as shadows to keep a photo natural so this can be controllable in some cases. But in really HDR demanding scenes (like an open window for example) the solution remains 5D3 + ML (or the classical bracketting and combining  )

[noipego]

hey ML ppl...im debating with myself if i should get the 5d4. despite the lack of serious video features, the camera now has good DR and the possibility of raw recording in the future. also wifi, dpaf and 60fps at 1080p are nice to have too.
Since i haven't tested myself the 1080p image quality of the camera , i'd like to ask here what ppl -who got 5d4- think of that...both mk3&4 have AA filter but some say 5d4 gives more detail, not the soft h264 file of the 5d3.
cheers