converting 4k 8-bit 420 to HD 10 bit 444

[Paul Mailath]

I know this is an old subject, I remember following it at the time and I don't recall a clear resolution (maybe I got tired of reading)

On an official Panasonic Business broadcast comparison sheet just picked up yesterday. I find the subject covered by Panasonic themselves - an official endorsement that the process works?

[Christopher Young]

Beware the marketing hype.

Basically the way I understand it is the answer is both yes and no. In Panasonic's case they are putting the emphasis on the yes part.

As it applies in the terms of chroma spatial resolution the answer is yes as you are putting four pixels of 4k/UHD chroma into one HD pixel but in terms of color bit depth no because the original chroma sampling was done with only 256 levels of encoding. Therefore still remains at 256 levels when down sampled to HD. You can down convert and re-encode 256 8-bit color to 1024 10-bit color but its color depth is still 8-bit 256. You just now have it in a bigger container. Quote from the site listed below:

"Finally, the chroma samples after conversion are essentially what you'd get from a camera with 8-bit 4:4:4 chroma, so while you do wind up with more spatial chroma resolution, you're not really getting any more in the way of subtle gradations in color. Each pixel is still going to be limited to 256 possible values each for "redness" and "blueness," even if you pad them out to 10 bits by appending zeros."

Reasonably examined here with various links to other points of view.

Can converting 8-bit 4:2:0 UHD to HD really give you 10-bit 4:4:4 video? | D Gary Grady

Chris Young
CYV Productions
Sydney

[Paul Mailath]

that was a good read - I've always doubted the premis and was surprised that Panasonic would promote it as a given

[Cary Knoop]

Originally Posted by Christopher Young

"Finally, the chroma samples after conversion are essentially what you'd get from a camera with 8-bit 4:4:4 chroma, so while you do wind up with more spatial chroma resolution, you're not really getting any more in the way of subtle gradations in color. Each pixel is still going to be limited to 256 possible values each for "redness" and "blueness," even if you pad them out to 10 bits by appending zeros."

True for the chroma but not true for the luma, the luma resolution is reduced to1/4th and thus sampled 4 times which makes it go from 8 to 10 bits.

So 4K 4:2:0 8 bit becomes HD 4:4:4 with luma being 10 bit and chroma remaining 8 bit.

[Gary Huff]

Originally Posted by Cary Knoop

True for the chroma but not true for the luma, the luma resolution is reduced to1/4th and thus sampled 4 times which makes it go from 8 to 10 bits. So 4K 4:2:0 8 bit becomes HD 4:4:4 with luma being 10 bit and chroma remaining 8 bit.

Except you won't really see any real-world benefit. Having tried this, the ProRes HQ 8-bit (in 10-bit ProRes container) 4:2:2 off the camera was far more robust for color grading than downsampling 4K 4:2:0 100Mbps material. That's why I switched to the GH5 from the A7R Mark II, to get that sweet 10-bit 4:2:2 internal recording.

[Christopher Young]

Originally Posted by Cary Knoop

True for the chroma but not true for the luma, the luma resolution is reduced to1/4th and thus sampled 4 times which makes it go from 8 to 10 bits.

So 4K 4:2:0 8 bit becomes HD 4:4:4 with luma being 10 bit and chroma remaining 8 bit.

True agreed.

Chris Young
CYV Productions
Sydney

[Christopher Young]

Originally Posted by Gary Huff

Except you won't really see any real-world benefit. Having tried this, the ProRes HQ 8-bit (in 10-bit ProRes container) 4:2:2 off the camera was far more robust for color grading than downsampling 4K 4:2:0 100Mbps material. That's why I switched to the GH5 from the A7R Mark II, to get that sweet 10-bit 4:2:2 internal recording.

Agreed again.

Chris Young
CYV Productions
Sydney