@Teagan From testing I did, h264 vs hevc, I found that there was a law of diminishing returns above ~ 70mbps, little or no benefit using hevc. But there is a substantial efficiency gain in size at very low data rates using hevc.

Does this include 10 bit color depth and HDR10 metadata or is that just for 8 bit sdr?

The former will be a little higher I suppose, if that's only for 8 bit.

As with all long-GOP interframe encoders, it is 97% dependent on the motion complexity of each individual source. A slide show or talking head with straight cuts? -- Maybe 1 Mbps, if it even reaches that. A seashore with rolling waves, or the Blade Runner movie? - Easily 100 times that, or 100+ Mbps unless you tame the grain.

Once we stop thinking in terms of bitrate-centric encoding, and twist our heads around enough to grasp quality-centric encoding like CQ (HEVC) or CRF (x265), it can then become a meaningful question. With x265, for instance, CRF 23, assuming it gives a satisfactory encode for one movie, will probably give a satisfactory encode for another, but with a possibly radically different bitrate.

Not true, not even a little bit, with the bitrate metric most people embrace, since no two sources will meet the same motion criteria.

If you are doing your own tests, as @JN- has done prodigiously, your optimal CQ/CRF index occurs about when the SSIM reaches 0.995, assuming the output has not yet reached or surpassed the size of the source.

To the point that @Musicvid makes, regarding the "bitrate-centric encoding" approach, here's a link to an excellent article on rate control https://slhck.info/video/2017/03/01/rate-control.html and the authors' conclusion regarding use of vbr and why it should be avoided.

"Avoid using this mode! One of the main x264 developers himself says you should never use it. Why? As the encoder doesn’t know exactly what’s ahead in time, it will have to guess how to reach that bitrate. This means that the rate itself will vary, especially at the beginning of the clip, and at some point reach the target. Especially for HAS-type streaming, this leads to huge quality variations within short segments.

This is not a constant bitrate mode! While ABR is technically a VBR mode, it’s not much better than specifying a constant bitrate, in that it doesn’t reliably deliver good quality.

Good for: Quick (and dirty) encodes
Bad for: Almost anything"

@wwaag

The link you posted now has a big 🌟on my bookmark bar, Thanks!

@Teagan

Here is another article I like that explains the very basics of quality-based encoding.

Caution: Expect your head to be rotated exactly 180° counterclockwise.

https://handbrake.fr/docs/en/latest/technical/video-cq-vs-abr.html

Great link Wwaag, I also bookmarked it for future reference.

I read that article and had to scratch my head. If Constant Quantizer vbr should never be used, no one ever told Nvidia or Intel. Their encoders both use it. And to use "nal-hrd=cbr" the output file needs to be .ts (MPEG-2 TS)? That's just not true. I've gotten among the highest quality ffmetrics readings with that private option and never once made an mpeg-2 ts file to do it.

As with all long-GOP interframe encoders, it is 97% dependent on the motion complexity of each individual source. A slide show or talking head with straight cuts? -- Maybe 1 Mbps, if it even reaches that. A seashore with rolling waves, or the Blade Runner movie? - Easily 100 times that, or 100+ Mbps unless you tame the grain.

Absolutely correct!

And we can also refer to the bandwidth of a monitor cable, it decides the picture motion quality, sure its resolution, too.

And which codec we should take is depending on the decoding power of our hardware.

Besides, I don't think the film tachograph really follows a UHD-HDR10 video.

@Teagan if you do the straightup math, you'll find that your 4k data is exactly 5x the size of your hd data. Suggesting that if your hd range was 14-16mbps, that the matching 4k range would be 70-80mbps. Though I could see it going higher if the chroma subsampling were bumped up. But given that hevc's compression efficiency is higher than that of avc, the bottom end of the range is probably fine for most cases. One's ultimate choice is often ruled, however, by what will fit onto the disk. As a general rule, I like to fill it up if I'm under capacity by bumping up the bit rate and get quality to spare.

Regarding diminishing returns with regard to the quality/bit-rate trade off, that's totally dependent on the footage being used. For instance, my camera gives me hevc in 60, 129, and 200 mbps. If I shoot 60, there will obviously be no quality gain going above that on the bluray. I think the quality analysis is only needed if you shoot something other than hevc and need to nail down where the 2 format's quality versus bit-rate curves intersect. An alternative conservative rule-of-thumb I've heard is to discount avc by about 20%... i.e, if you shoot 100 mbps avc and transcode, go for about an 80 mbps hevc bit-rate.