What settings to choose if you want to render in 10 bit?

What settings to choose if you want to render in 10 bit?

You can render DNxHR in HQX or 444, or ProRes HQ (?) or I believe HEVC (H265) and select the 422 10-bit option. Perhaps others will chime-in with additional suggestions. There may be a Cineform option, for example. What is your source media?

You can render DNxHR in HQX or 444, or ProRes HQ (?) or I believe HEVC (H265) and select the 422 10-bit option. Perhaps others will chime-in with additional suggestions. There may be a Cineform option, for example. What is your source media?

Source media is 10-bit Intra footage.

Is the 422 10-bit option in "Encoding Profile" with "Main10"?

Assuming you are asking about H265, the encoding profile is Main 422 10. If you don't see this, perhaps it is a Studio-only option, I don't know. You should be able to use a 10-bit flavour of ProRes or DNx, though.

Assuming you are asking about H265, the encoding profile is Main 422 10. If you don't see this, perhaps it is a Studio-only option, I don't know. You should be able to use a 10-bit flavour of ProRes or DNx, though.

Yes, for H265 options, I only see: "Main", "Main 10", "Main 4:4:4", "Main 4:4:4 10".

I have the Studio version of Resolve.

So I'm assuming 4:2:2 10-bit is "Main 10"?

I don't see Main 444 10 or Main 444 - Interesting. I see both Main10 and Main 422 10 (I believe Main 10 is 10-bit, makes sense, but probably with reduced color sampling such as 4:2:0). Can you simply try it and see what you get as a result - MediaInfo should be able to tell you whether the result is 10-bit.

I use these.

HDR Deliver 2.png (67.99 KiB) Viewed 5443 times

I use these.

HDR Deliver 2.png

So you do use Main 10. Is it 4:2:2 or 4:2:0?

It depends on your hardware. Render a short test and check with MediaInfo.

Mine come out 4:2:0.

But for delivery, that's OK. I've found issues playing back 4:2:2 on anything other than a computer.

Plus, everything you watch is 4:2:0 - Blu-ray, Netflix, Amazon, YouTube, etc.

4:2:2 really only has a place during editing. And for that, I favor DNx or Cineform.

Mine come out 4:2:0.

But for delivery, that's OK. I've found issues playing back 4:2:2 on anything other than a computer.

Plus, everything you watch is 4:2:0 - Blu-ray, Netflix, Amazon, YouTube, etc.

4:2:2 really only has a place during editing. And for that, I favor DNx or Cineform.

"everything you watch is 4:2:0 - Blu-ray, Netflix, Amazon, YouTube, etc."


Ooh, I did not know that.

For 10bit use Main10 and it's 4:2:0. There is no 4:2:2 h265 encoder in Resolve (unless your GPU supports it, but I don't think there is any atm.).
Main 10 4:4:4 as name suggests is 10bit YUV 444 based and it's available on newer GPUs.

If you want to use h265 10bit 4:2:2 then this is rather unusual case (more for own usage/archival) and better off to use DNxHR/ProRes.
If you really need it then MainConcept plugin for Resolve should have it.

There is no 4:2:2 h265 encoder in Resolve (unless your GPU supports it, but I don't think there is any atm.).

MacBook M1 Max with Resolve Studio 18.1.2 - I have a 422 10-bit HEVC option. Are you referring to something else, Andrew?

If you see it, then you have it
GPU 422 support is very limited, specially for encoding. Those Mac chips may be the only ones which have it.

For 10bit use Main10 and it's 4:2:0. There is no 4:2:2 h265 encoder in Resolve (unless your GPU supports it, but I don't think there is any atm.).
Main 10 4:4:4 as name suggests is 10bit YUV 444 based and it's available on newer GPUs.

If you want to use h265 10bit 4:2:2 then this is rather unusual case (more for own usage/archival) and better off to use DNxHR/ProRes.
If you really need it then MainConcept plugin for Resolve should have it.

Good to know. Thanks.

I use these.

HDR Deliver 2.png

Why "Constant QP" instead of the default "Variable Bitrate"?

Constant QP doesn't care about bitrate, but fixed compression ratio. Easy to encode masters will end up with lower bitrate and difficult with much higher. The lower QP the higher quality and files end up with higher bitrate. Good for local playback, archive etc.
If you want to control bitrate then use VBR/CBR.

Why "Constant QP" instead of the default "Variable Bitrate"?

It offers the best quality/size balance.

If you don't care about bitrate much (eg. you want decent quality archive), but want relatively same quality for every master then you can use QP. By choosing specific QP you shift quality up and down (of course this will affect final bitrate used).
If you need to control bitrate then you use VBR/CBR.

Complex explanation in case of x264/5:

Code: Select all

Constant QP (CQP)

The Quantization Parameter controls the amount of compression for every macroblock in a frame. Large values mean that there will be higher quantization, more compression, and lower quality. Lower values mean the opposite. QP ranges from 0 to 51 in H.264, and you can easily set a fixed QP for your entire encoding process with x264 and x265. Note: libvpx does not have a fixed QP mode.

ffmpeg -i <input> -c:v libx264 -qp 23 <output>
ffmpeg -i <input> -c:v libx265 -x265-params qp=23 <output>
To know more about the idea behind QP, you can read this tutorial (if you’re not afraid of some maths).

Unless you know what you’re doing and you explicitly want this, it is suggested not use this mode! Setting a fixed QP means that the resulting bitrate will be varying strongly depending on each scene’s complexity, and it will result in rather inefficient encodes for your input video. The quality will be good if you set a low enough QP, but in contrast to CRF (see below), you may waste space, or you have no control of the actual bitrate.

Good for: Video encoding research, or if you have no CRF mode
Bad for: Almost anything else

Note that Netflix proposes using fixed-QP encoding for its per-shot encoding optimization to achieve optimal encodes for each scene. This however requires a lot of processing and careful assembly of the individual encoded shots, so it’s not a “one size fits all” method you should use unless you have the whole framework implemented.

If you don't care about bitrate much, but want relatively same quality for every master then you can use QP.

Exactly.

Here's what I mean about best quality/size balance. This is a bit oversimplified, but it gives the basic idea.

When using CQ, a frame that needs 10 bits will get 10 bits. A frame that needs 30 bits will get 30 bits.

When using CBR at 20, the frame that needs 10 will still get 20, thus bloating file size. And the frame that needs 30 also gets 20, thus reducing quality.

VBR at 20 is a little better. The frame that needs 10 will get 10, so no file bloat. But the frame that needs 30 will only get 20, thus reducing quality.

CQ offers best balance with a single pass encode.

You can think about it also as "how much I want to compress each frame". I don't think this is exactly true for codecs like h264/5, but in approximation you can say QP=3 will compress each frame 3 times, so very high quality. Each frame has different complexity in typical master, so each frame will end up with different size/bitrate needed to have compression this level. Easy master will end up with lower bitrate and difficult with much higher. It's similar to what eg. Cineform does. It tries to preserve each frame relative quality, so you end up with very VBR file (if master has frames at very different complexity). ProRes work similar, but it has restriction how far it can raise bitrate for single frame (with some average target bitrate been a reference point). DNxHD/DNxHR (in typical mode) are CBR, so they use same strict bitrate for each frame, so you perfectly know final size, but each frame relatively quality will vary.

Here's what I mean about best quality/size balance. This is a bit oversimplified, but it gives the basic idea.

When using CQ, a frame that needs 10 bits will get 10 bits. A frame that needs 30 bits will get 30 bits.

When using CBR at 20, the frame that needs 10 will still get 20, thus bloating file size. And the frame that needs 30 also gets 20, thus reducing quality.

VBR at 20 is a little better. The frame that needs 10 will get 10, so no file bloat. But the frame that needs 30 will only get 20, thus reducing quality.

CQ offers best balance with a single pass encode.

Proper VBR mode has min, avg, max, so not exactly as you said.
Like QP it will use as much as it needs, but you put restriction into it, so specially max will be fixed.

max will be fixed.

That's what I was trying to illustrate. If you set a max of 20, and the frame needs 30 to look as good as the other frames, it won't get it and the quality using VBR will be reduced.

You can overcome that by knowing what Max value will give you the quality you're looking for, but that requires more than one pass. CQ can do it in one pass.

Even better option is CRF, but of course this is not available in many encoders.
This is not trying to apply fixed compression ratio to each frame(macro-block), but tries to apply such a compression which gives same relative quality (measured by 'perceptually driven' math).

Constant Quality (CQ) / Constant Rate Factor (CRF)

I’ve talked about the Constant Rate Factor in another article in more detail. It basically gives you constant quality throughout your encoding process. It’s a “set and forget” thing—just specify the CRF and let the encoder do the rest.

ffmpeg -i <input> -c:v libx264 -crf 23 <output>
ffmpeg -i <input> -c:v libx265 -crf 28 <output>
ffmpeg -i <input> -c:v libvpx-vp9 -crf 30 -b:v 0 <output>
In H.264 and H.265, CRF ranges from 0 to 51 (like the QP). 23 is a good default for x264, and 28 is the default for x265. 18 (or 24 for x265) should be visually transparent; anything lower will probably just waste file size. Values of ±6 will result in about half or twice the original bitrate. For VP9, the CRF can be from 0 to 63. Recommended values are from 15–35.

The only downside with this mode is that you don’t know what the resulting file size or the fluctuation of the bitrate will be.

Note that a two-pass and CRF encode with the same resulting bitrates should be identical in quality. The main difference is that with two-pass, you can control the file size (if that is a requirement), whereas with CRF you just specify the quality you want.

Good for: Archival; achieving the best possible quality
Bad for: Streaming; obtaining a certain bitrate / file size

Great explanation about difference between QP and CRF:
https://slhck.info/video/2017/02/24/crf-guide.html

Mine come out 4:2:0.

But for delivery, that's OK. I've found issues playing back 4:2:2 on anything other than a computer.

Plus, everything you watch is 4:2:0 - Blu-ray, Netflix, Amazon, YouTube, etc.

4:2:2 really only has a place during editing. And for that, I favor DNx or Cineform.

I film in .mov 400Mbs 10bit 422. Using Magic Mask to roto an actor and have another actor appear behind them. Works a treat, but MM is so temperamental. You have several timelines open, try to copy the fusion clip into the main timeline. All the fusion nodes disappear. So I got that fed up of retracking the mask I ended up just rendering out a mov file, of the fusion clip, but as I want to replace the original 4K 10bit 422 footage with an exact or as close as damn it for colour work, it is so annoying you can't rely on MM to just to hold its data. But thanks for this info. A big help.

That is why I can't use a 420 file to replace the fusion clip.

For intermediate clips, I'd always use 4:2:2 in 10 bit. Cineform is very efficient under Windows.