5d3 full res (6k) 24 p canon all-i codec

Started by mlrocks, June 03, 2018, 08:29:04 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

mlrocks

Here is the idea:

5D3's stock flow is
1. full res (6k) raw video 30 p stream
2. pixel binning 3x3 to 1
3. compressing to all-i h264 codec

If we can bypass step 2 of pixel binning, we can get full res all-i h264 video clips, which should be within the CF card writing limit of 90 MBS. The computer power of the camera should be able to handle the new workflow, due to the skipping of the pixel binning step.

Cheers, ML really rocks.

Levas

Step 2 makes it possible to read the sensor out fast enough for 30fps.
As far as I know, once you skip the binning part the read out speed drops down to 7.4 fps.
6k would be nice, but the sensor read out speed is the bottleneck I guess :-\

mlrocks

Even if it is JUST UHD ALL I compression, it is a bless. Imagine 60 fps of UHD ALL-I on 5d3. Then I don't need to upgrade for a very very long time, maybe until 5d5 with 500 MB/s CF3 writing speed. LOL.

Kharak

Guys. The h264 encoder is a dedicated chip in the camera. It is hardwired for 1920x1080  and 720p 59.976.

Unless someone with intimate knowledge of the chip and how to access it cones forth, dont expect anything higher than 1920x1080 60p. For what I understand, the only thing that can be done is adjusting what is fed to the chip e.g. 3x crop 1920x1080. And even if one managed to access the chip, chances are it has no overhead to do anything much higher than

Same goes for dreaming of Prores or Prores raw, there simply is no overhead or hardware in the camera capable of doing such calculations.

Even though Raw recording has way higher bandwith than h264, it is made possible because it basically is a diversion from the normal canon image processing pipeline. Basically the h264 file you get goes through a lot of steps before hitting your card. The raw stream a1ex discovered and managed to diverge, skips almost all the image processing pipeline and goes straight to card and leaves the image processing of the RAW data to 3rd party programs. Because it skips most of the Pipeline it is only limited by SD/CF card interface speed (and digic processor in higher regards).

Basically...
once you go raw you never go back