Alister Chapman explained this pretty well some years back in a response to a poster's question as to why he was having problems with the 12-bit Linear RAW with his Odyssey Q7 so I'll guide you to his comments, facts and figures. I don't think he would mind. If interested the range of posts can be followed at the link below. As DJ said there is no comparison when comparing 12-bit Linear to Sony's 16-bit RAW on the R5 recorder. If I wanted Sony RAW then an F5/F55 with an R5 and its 16-bit RAW would be the value path to go IMHO. For ease of post-work on everyday type jobs, I find the 10-bit S-LOG3 with its much more even data distribution works well for my kind of output. Not doing Hollywood here
Chris Young
"For 14 stops: 12-bit linear raw is a compromise. The 10-bit log is a compromise. Once upon a time, we used to use the 10-bit recording for just 6 to 10 stops. But now we want to record 14 stops in a 10 bit. So along comes log which instead of recording each successively brighter stop with more data (to mimic real-world light) we record each stop with the same, very limited amount of data. The net result is that while in many cases it is perceptually lossless, the brighter stops are recorded with less and less data relative to the true light in the scene being shot and overall there is less data per stop than there would be if we were only capturing 6 to 10 stops. The compromises in log start just above middle grey and continue all the way to the very brightest stop. Then add to that compression artefacts as the 10-bit recordings are highly compressed in the case of XAVC or perhaps a little less highly compressed if you take the SDI or 10 bit HDMI (4K 30fps max) and record it externally with something like ProRes. Linear raw tries to record the scene with the correct amount of data for each stop. So that means each stop requires twice as much data as the previous one. The problem is that if we use just 2 code values for the first stop we need 2048 code values to record stop 11. 12-bit data only has a maximum of 4096 code values, so by stop 11 we have run out of data. 14 stop linear raw actually needs 16 bits. So Sony cheat, In all probability, they use something called floating-point math. In effect, this rounds the data by a factor of 4 or 8. In the highlights, the brightest stop instead of using 8192 code values only gets 2048 code values. That's actually no big deal, 2048 code values is a heck of a lot of data. But down in the shadows where a stop might only have 16 code values, divide that by 4 and it's only 4 and that is a huge compromise. Log likes to be exposed brightly, it helps reduce noise in the end result, and gives better colour. 12-bit linear raw, really, really needs to be well exposed as anything down in the lower end will lack code values so will appear coarse and grainy. But there is a huge amount of data up in the highlights of 12 bit raw. Bright scenes work really well with 12 bit linear, low key can be a disaster. But it's uncompressed so there are no compression artefacts to deal with, just massive file sizes to cope with. it's Bayer sensor data so there is more scope for white balance shifts in post-production. If you really want raw, look for a used F5/R5. 16 bit linear raw, compressed visually losslessly 3:1. Files that are no bigger than ProRes HQ, but with the correct amount of code values for 14 stops. If you are serious about HDR then think about an F55. The FS5/FS7/F5 sensors just don't have the colour range to really take full advantage of Rec-2100. Sure you can fill up Rec-2020 with a saturation boost but you won't have the depth of colour and range that an F55 can produce, and after all, HDR is all about showing a greater range, it's not just a case of turning up the volume."
https://us.community.sony.com/s/ques...language=en_US
https://www.xdcam-user.com/2017/08/t...-raw-to-s-log/
Data level distributions.jpg
Leave a comment: