C200: C200 8 bit vs 10 bit

I believe the C700, C200, and C300 Mark II RAW are all 12 bit, and 5D photo RAW is 14 bit. ArriRAW and ProRes are 12 bit. I believe Red is 12 bit. Sony F55 is 16 bit RAW, so it wins the bit game, but doesn't win the best image quality game.
 
And, or course, the most ridiculous part is that they are "showing" the difference between 8 bit and 16 bit using a youtube video.
 
While the video wasn't professional quality, what they say is completely accurate. If one shoots with an 8-bit camera and also with a 16-bit camera, then exports everything for online viewing (JPG or MP4, both 8-bit), there will be zero differences related to the source bit depth. They make the valid point that the extra bit depth helps when doing post editing before crunching down to 8-bit for delivery (which is everywhere except perhaps digital cinema, which also supports 10- and 12-bit video). Something they didn't mention is that it's important for 8-bit video to not be too clean, via noise reduction or otherwise. Noise provides dithering and is necessary to prevent banding, especially in skies.
 
8-bit final delivery is the VHS of our era. We used to shoot on HDCAM, D-Beta and Beta SP and sweat all of the image quality concerns but often our work ended up in front of our audiences on VHS for non-broadcast. It was brutal.
 
Eric Coughlin said:
I believe the C700, C200, and C300 Mark II RAW are all 12 bit, and 5D photo RAW is 14 bit. ArriRAW and ProRes are 12 bit. I believe Red is 12 bit. Sony F55 is 16 bit RAW, so it wins the bit game, but doesn't win the best image quality game.
Click to expand...

You're mixing and comparing all over the place here. RAW bit depth doesn't compare to processed video bit depth. Linear bit depth doesn't compare to log bit depth, at least not 1 to 1.
 
YouTube supports 10-bit (and 5K and 8K even). But it says it can detect when one is connected to an HDR set and adjust accordingly if one isn't. Now, how does it do that? And do we want it detecting it?
 
DLD said:
YouTube supports 10-bit (and 5K and 8K even). But it says it can detect when one is connected to an HDR set and adjust accordingly if one isn't. Now, how does it do that? And do we want it detecting it?
Click to expand...

Thanks for the reminder on current 10-bit support. Detection is likely through the native YouTube app installed on HDR TVs. If using Chrome, it would be possible for the browser to make the detection (not many people have 10-bit displays and 10-bit video cards though).

Ultra HD Blu-ray supports H.265 with 10-bit and Rec.2020: https://en.wikipedia.org/wiki/Ultra_HD_Blu-ray. 4K/HDR on PCs has limited hardware support (looks like mostly due to copy protection): https://www.extremetech.com/computi...-ray-burners-pack-4k-playback-16x-burn-speeds
 
I posted alink on the News thread - Google Play is now streaming 4K HDR flicks. AT&T/DirecTV are planning to do so in 2018.
 
DLD said:
YouTube supports 10-bit (and 5K and 8K even). But it says it can detect when one is connected to an HDR set and adjust accordingly if one isn't. Now, how does it do that? And do we want it detecting it?
Click to expand...

It is called EDID, extended display information data. Digital video devices have a data table in firmware that lists all of the display formats, resolutions refresh, rates, bit depth, color space standards, etc, that they can support. When you connect a source and a display they "negotiate" to pick the highest quality format that both support. That information can be picked up by a streaming service so they don't waste bandwidth streaming something your system doesn't support or network bandwidth isn't available for.
 
8 bits is fine for delivery. I remember when 8-bit was touted as 16.8 million colors. And it is. There are many beautiful JPEGs of landscapes out there, of scenes that had several f-stops in real life, with no banding.

If there was a camera that had a great film LUT built in, so that I didn't have to apply a strong S-curve in post, then I might be fine even with an 8-bit camera codec.

And if all you're doing is changing colors in post, 8-bit is fine too. Banding happens when you raise the contrast. If you're pushing and pulling the image, 10-bit log is a nice safety net against banding.

Don't get me wrong. My ideal is 10-bit log raw. Just bypass the manufacturer's presets. The tastes of an international corporation seldom align perfectly with mine.
 
You must log in or register to reply here.
Back
Top