F3: 8bit XDCAM (mpeg4) vs 10bit RGB (ProRes444)

Peter G. Johnson

Peter G. Johnson

Active member
I hope these pictures demonstrate the difference in quality between internal and external recording at the absolute highest level using the F3. The difference is when you magnify the picture. This image was taken using an SCL-Z18X140 with stability control on, at most tightest zoom, recorded at (25p) 1920x1080. No bad considering it was handheld and I was panning slowly in both directions at the same time. A tripod shot would've been even better, I suppose.


8.jpg
8bit XDCAM
10.jpg
10bit RGB to Pix 240 (3G ProRes4444 12bit)
8bit xdcam.jpg
8bit XDCAM (Scaled by 400% FCPX)
10bit RGB.jpg
10bit RGB (Scaled by 400% FCPX)
 
Last edited:
Hmm. The difference is in how it handles grading...the raw images SHOULD look pretty close if not exactly the same. Add a ton of contrast to both images and re-compare. That's when you'll see the differences, mainly in the smoother gradients.
 
Great tests and thanks for posting the results.

It's hard to know if the reduction of artifacts are simply due to the increased bit-depth or the fact once out of 8-bit mode you are no longer using mpeg.

I don't have RGB mode but in my simple tests comparing 8-bit SXS to 10-bit 422 ProRes and 10-bit 422 uncompressed I found the same type of reduction in artifacting and in some cases less "dancing" artifacts on steady objects that were complex in pattern ie. fuzzy multi-coloured wool sweater (give that a go, guaranteed to "break" any codec - lol)

I guess I was somehow expecting the RGB mode to have some more wow factor to the footage, but as Richard suggested perhaps the biggest differences only come out in extensive grading. Still for me, shooting primarily 8-bit and grading it I have no complaints. It keeps the file sizes down and no one is inspecting my footage at 200-400% on pause. I might use 10-bit on SLOG or higher end projects or where multi camera setups and later matching in post will be required, but it all depends on the client and project.
 
Last edited:
starcentral said:
Great tests and thanks for posting the results.

It's hard to know if the reduction of artifacts are simply due to the increased bit-depth or the fact once out of 8-bit mode you are no longer using mpeg.

I don't have RGB mode but in my simple tests comparing 8-bit SXS to 10-bit 422 ProRes and 10-bit 422 uncompressed I found the same type of reduction in artifacting and in some cases less "dancing" artifacts on steady objects that were complex in pattern ie. fuzzy multi-coloured wool sweater (give that a go, guaranteed to "break" any codec - lol)

I guess I was somehow expecting the RGB mode to have some more wow factor to the footage, but as Richard suggested perhaps the biggest differences only come out in extensive grading. Still for me, shooting primarily 8-bit and grading it I have no complaints. It keeps the file sizes down and no one is inspecting my footage at 200-400% on pause. I might use 10-bit on SLOG or higher end projects or where multi camera setups and later matching in post will be required, but it all depends on the client and project.
Click to expand...

It's worth noting that there is NO difference in compression between 4:2:2 ProRes HQ and 4:4:4 ProRes 4444 - all of the extra bytes in the ProRes4444 files are just for the added colour resolution.

So motion artifacts etc. will be no different between ProRes HQ and ProRes 4444.
 
To me the biggest advantage recording external from SDI A and s-log is the fact that I get the best possible from the F3. No settings that can screw the image.

I would like to see the real difference between 422/444 and if it's worth getting a recorder doing 444.
 
You must log in or register to reply here.
Back
Top