Combining videos corrupts appended videos

[Avidummy]

"The advantage of the CRF mode is that it suits the human perception much better than the QP mode. For example it will raise the quantizers in "fast" scenes where the loss won't be visible anyway and lower the quantizers in "slow" scenes."

That's a slightly confusing statement. It seems to imply quality is not required in the fast scenes, but then goes on to correctly state bitrate is increased here (which maintains quality).

I don't know.  I think higher quantizer is a lower bitrate, so fast scenes are less quality.  It makes sense for movies, but not for videos to be analyzed frame by frame.

As I say, this is a confusing statement. If you are correct in interpreting it as higher quantizer, lower bitrate, then it is 100% incorrect to imply this is how CRF works. CRF works completely opposite to this. Data/bitrate is increased in the fast, more complex scenes, not lowered. The whole purpose of CRF is to maintain consistent quality in scenes of varying complexity. To suggest it lowers bitrate in complex scenes (because it is less likely to be noticed) is simply wrong.

What is the substantiation for your statement?  You're saying the user manual for the codec is wrong.  How can I tell who is right?

It makes sense in 1-pass encoding to conserve bitrate during fast scenes that can't be perceived anyway, except in niche cases like mine where I may want to watch frame by frame.

The only 2 pieces of evidence I have suggests the manual is correct.  1) It's the manual.  2) It makes sense.

If you carry out the test I suggested on a short clip, and also compared with a CRF zero value lossless encode, you may be surprised by how closely the quality matches in all three encodes.

I still have not done the tests, mainly because doing so risks crashing my pc, and I'm not ready to have everything shut down on me.

I think CRF could produce a good quality video, but the sacrifice would be size.  I think to gain quality without increasing size requires 2-pass.  That seems the whole purpose of 2-pass.

The reason I bought the top of the line pc with 32 cores at 5.7ghz and 128gb matched ddr5 is to do 2-pass video conversion.  Now I'm being advised to do CRF?

[Avidummy]

You can read page 4 here..

I have another very clear article "Understanding Rate Control Modes" which states:

Note that a two-pass and CRF encode with the same resulting bitrates should be identical in quality.

You just need to find (once and for all) for your type of  footage the Quality number (e.g. 18) that will result in ~10,000 kbit/s.

I suppose I'd have to test that by encoding a short clip of a fast scene using 2-pass and CRF to see if any differences are evident at the same bitrate.

A couple days ago I had a youtube video that I wanted to crop the music out at the beginning and end but the video had some frame instructions at the beginning that meant I couldn't edit the video without causing timing issues (I don't know the right words but that's the idea).  So I had to reencode, but didn't want to use 2-pass because it's a 4 hour video.  I used CRF as recommended here, but my first encode resulted in a gigantic file.  I tried again at the default 26 and got 750 bitrate when the original was 500.  Close enough, but it uses a lot of time to take guesses like that.  I don't know if 26 would always result in 750 bitrate.

[eumagga0x2a]

If you don't want quantizer being increased (equal to quality being reduced) where the CRF algorithm thinks this is permissible without the viewer noticing, just use the constant quantizer (CQ) mode.