View Single Post
Old 06-30-2019, 01:05 PM   #8
Liverpool TV
Senior Member
Join Date: Nov 2013
Location: Liverpool, England.
Posts: 2,901

Originally Posted by Jerry View Post
The last chip I tested that had HEVC capabilities was the 7700k. This had the Intel 630 graphics package. This is the same graphics package that is on the 9900k. I could only get HEVC 10bit 4:2:0 from the hardware. I've never been able to get 4:2:2 hardware output using QS and Edius.

The graphics package is not changing until Icelake and then again with Tigerlake Intel X2 graphics.

HEVC first showed up with the Intel 530 graphics package. HEVC was only 8bit on export.

If you are able to get 4:2:2 non-HEVC and HEVC using Intel Graphics 630 hardware from within a different NLE, then it would point to an implementation limitation within Edius.

Then I have to ask, which NLE can do that? Was Premiere/Media Encoder capable of doing that with hardware?
Hi Jerry.

I've a feeling that QS H.265 4:2:2 isn't possible at all from Edius, which is why I was querying about what CPU's iGPU/QS did do it within Edius. I've tried a 630 and another one but no luck.

The odd thing is that QS H.264 from within Edius can do 4:2:2 and at either 8 or 10 Bit.

What I find more strange is why 4:2:2 wouldn't be the de facto chroma option for anything 10 Bit. I can't work out why you'd want 10 Bit 4:2:0 for anything that's 10 Bit. Is this maybe something specific for a specific target/platform?

Maybe there's a specific HDR target that's 10 Bit 4:2:0? Although I wouldn't have thought it would be too practical to give something the dynamic range capabilities of 10 Bit but without the colour resolution/accuracy of 4:2:2.

Ah, just thinking off the top of my head as I writing this post and just mentioning HDR. Maybe QS 4:2:2 only becomes active within the exporter when the project itself is set 2020 or is set to a specific HDR setting. I'll have to check that tomorrow/Tuesday when back on my main system.

I'll also let you know what did QS 4:2:2 as I can't remember right now. It definitely wasn't an NLE though. I've been doing a whole bunch of software vs hardware tests using various encoding applications and using QS, NVENC and the AMD one. It was one of the encoding apps using QS that even Edius seen as 4:2:2.

There also seems to be differences with one QS encode from one app compared to another app doing QS when selecting the same parameters. I'd always thought that QS was something that was standardised as far as options were concerned and that any encoder using QS would only have options that were available via Intel's implementation. I've had the same source file producing different outputs from different apps using the same settings.

If you're into such things let me know and I'll PM you some links to some of the preliminary test files from my tests so far.

I will be doing a couple of YouTube videos with my final results that I'll post here on the forum for anyone interested in such things.

"There's only one thing more powerful than knowledge. The free sharing of it"

If you don't know the difference between Azimuth and Asimov, then either your tapes sound bad and your Robot is very dangerous. Kill all humans...... Or your tape deck won't harm a human, and your Robot's tracking and stereo imagining is spot on.

Is your Robot three laws safe?
Liverpool TV is offline   Reply With Quote