PerroneFord
Deals in Lead
Thanks for the great answers/information.
It sounds like the real benefit is for color correction but I still don't get how converting a file to 10 bit 4:2:2 when the camera writes the file as 8 bit 4:2:0 does anything more than just create a larger file. The only new data being added is 1's and 0's, no new information that pertains to color space.
Think of it this way.
You shoot a video of the sky on a beautiful day. The camera can record 254 distinct shades of blue (working hypothetically here people). You get the file into the editor and you decide you want to make the sky a bit more blue.
If you are working in 8-bit space, you make your changes, but since you've modified the colors, they may not be smooth and contiguous colors any more as the program is guessing which shades you really meant. So you render out and the sky has banding of colors and isn't smooth.
So you start over, but work in 10bit. Now you can define 1023 shades of blue. You make your changes and now when you render out the sky is more blue, and everything is perfectly smooth.
The goes for ANY color correction, white balance, or anything else including just changing contrast and brightness. In 8 bit you have 254 shades from white to black. In 10bit you have 1023. You can be a lot more subtle in the shadows, you can be a lot more subtle with the skin tones, with the highlights, etc.
So when you plan on doing this kind of work, you set up your camera to make sure none of the shadow information is lost, and none of the highlight information is lost. When you get to post, you can stretch that image back out and get smooth information back out if you work in 10bit.
Not sure if this makes sense to you or not.
For a super simplified example, buy a box of 8 crayons and draw a picture. Then go get a box of 100 crayons and do the same. The subtlety you can get with more colors and shades at your disposal is what makes the difference.