This cropped up when looking at the various workflows that can be used for GH4 4K footage. I read this about the AG-DVX200:

"Increased Color Resolution And Bit Depth: One excellent benefit of downconverting UHD/4K footage to 1080 HD in post is that you can realize an increase in proportional color resolution and a notable increase in bit depth. The AG-DVX200 records 4K or UHD footage at 8 bits per pixel and utilizes 4:2:0 color sampling. After downconversion, the resulting footage has 10 bits per pixel and 4:4:4 color sampling! Yes, you can convert 3840x2160 8-bit 4:2:0 recorded footage into 1920x1080 10-bit 4:4:4 footage in post... When downconverting UHD/4K footage to 1080p HD, you also get the benefit of converting 8-bit pixel depth into 10-bit pixel depth! Since each 2x2 block of UHD/4K pixels will be summed together to create a single 1x1 pixel in 1080p HD, the individual pixel values and gradations from the source footage can be retained in the downconverted footage." ['The Benefits of Shooting in 4K' by Barry Green]

Is this dependent upon the scaler you use? What does FrameFlex do with 4K footage that needs to be scaled to HD?

Yes you can convert 8bit to 10bit but the new 10bit data has no more detail than the 8bit. its just padded.

The color space can in theory be improved as the 3840 has 1920 color samples across and 1080 samples vertically.

But I'm not sure Avid (or anything else) is smart enough to work that out. Most would resample back to 4:4:4 then down sample to the new raster size.

Pat Horridge:

Yes you can convert 8bit to 10bit but the new 10bit data has no more detail than the 8bit. its just padded.

The color space can in theory be improved as the 3840 has 1920 color samples across and 1080 samples vertically.

Hi Pat,

Your first sentence is only true when you transcode HD to HD. But if you transcode UHD to HD the situation is different. In this case 4 pixels (2 in line, and 2 in culomn) are mixed together, i.e. their individual color values are just added. With life source material (not test pattern) this leads to new color values when you convert to a 10 bit-format.

Example:  Assume you have the following 4 color values in 8-bit format in a quadruple of pixels:

160 162

161 163

The conversion process will just add the four neighbouring values to a sum of 646 for the new pixel in 10-bit-format.

This conversion is also improving noise by about 3 dB, as noise is statistic and thereby not linear added.

Joachim