What Camera Shoots Prores 4444

There is 'large' advantage to shrink RAW, not final debayered paradigm. Past doing this y'all are saving a lot of data/space.
By compressing RAW bayer data you shrink just 4000x2160 pixels in their RAW land.
When this gets debayered to create RGB data (missing RGB color pixels are created) and you end upwardly with 3x more data to compress (4000x2160 for each RGB component). In case of ProRes this is what gets compressed. Fifty-fifty if ProRes XQ pinch is more like iv.iii:1 you are endig up with much higher data rate as you have 3x more to compress.

RAW=4000x2160*12bits*30fps/3=998Mbits/8=124MBytes/sec
RAW data without pinch would be 3x124=372MB/sec
This 372MB/s becomes 3x372MB/sec=1116MB/sec (after debayer equally RGB data) which gets compressed by ProRes XQ about iv.3 times, so nosotros finish up with 1116/4.3=260MB/sec. Because debayer procedure interpolates and creates 'more color information' we end up worse, even with higher pinch in place.
ProRes boilerplate bitrate tin can exist around -+five%, so you may stop up with bit different concluding bitrate.

3x compression on RAW bayer information or 4.3x on final RGB should requite nigh the same final result in terms of quality (even if I have non seen any report on how compression on RAW affects debayered image).
3x compression with decent algorithm tin can be definitely treated as visually lossless (Sony F65 uses likewise virtually iii:i compression on their RAW). It'southward not like ZIP compression, but lossy, which decides what information to remove. In well-nigh cases these are high frequencies (dissonance), and so your footage is bit softer than information technology could exist (pixels are non that edge perfect), which in some case can be even desired effect

The only advantage of using ProRes in this case is simplicity of the workflow. You are getting debayered, high quality 12bit iv:4:iv slightly compressed file which will work in almost every NLE/grading software.
RAW reward is that you accept full room to play with grading, virtually likely better debayering algorithm than inside camera and of course lower information rate. You just have to be fix to handle RAW files and need more than powerful machine.

Then past uncomplicated rule ProRes XQ is nearly 9/iv.3 bigger than 3:1 RAW (3x3, as ProRes encodes uncompressed RAW and has 3x more to compress as it deals with RGB information).

The only intermediate codec which has back up for raw information is Cineform. Nil stops ProRes to encode RAW bayer data as at the cease this tin be treated equally 'whatever other video'. The outcome would exist later as this ProRes would take to be treated in special way and get through debayering process, similar CinemaDNG. There is no such a matter at the moment (except for Cineform). Would this exist much better than CinemaDNG? At just 3:1 compression I don't retrieve it'southward worth. ProRes is proprietary technology with possessor beingness a bit problematic when it comes to licensing, so why bother to 'push' such a technology Creating open source ProRes equivalent should exist trivial as ProRes is actually very manifestly codec (Cineform is way more avant-garde and interesting engineering science), it's just difficult to make it mainstream without having big $ behind yous.

Selection is yours

Source: https://forum.blackmagicdesign.com/viewtopic.php?f=2&t=37139

Posted by: flynncrue1941.blogspot.com