About the ME algorithm used in x264

[dy_dragon]

The UMH algorithm has been proposed many years. Why x264 always uses this method? Because these successors performance worse?

[Dark Shikari]

x264 uses HEX by default, not UMH. UMH is used as a slower search for the slower encoding presets. It uses this method because someone wrote code for it. It doesn't use something else because nobody has written a patch to implement something else.

[Adub]

ME algorithms are not exactly cake walks, both from a implementation and understanding standpoint. It's one reason why GPU accelerated ME still has yet to produce any real results. It's not easy to just whip up an ME algorithm, or C implementation of that algorithm.

As Dark Shikari said, if you want a new algorithm introduced into x264, write a patch and submit it to the devs.

[Keiyakusha]

I think you got him wrong. I think he asks why most of the users use UMH and not TESA. If this is the case, for me anything above UMH is way too slow. And while it is slow, it doesn't produce big enough improvements. If I'd want to have better quality over the settings I use most of the time - I'd probably rise some other options.