WebM support

[Dukers]

https://trac.handbrake.fr/ticket/140

Thank you thank you thank you

* It's a VP8 encoder (the only one to date is libvpx I think). WebM is the container.

[s55]

Hate to break this to you, but no-one is actually working on this, and no-one has expressed any interest in working on it.

All that ticket is, is a "If you submit a patch, we *might* accept it". In other words, is seems unlikely that VP8 will be in HandBrake anytime soon, if ever.

[Dukers]

Hate to break this to you, but no-one is actually working on this, and no-one has expressed any interest in working on it.

I know, s55. But just knowing that a remote chance of happening might exist is better than nothing. libtheora is totally worthless.

[s55]

There was always a remote chance. Nothing has changed.

And yes, I totally agree with you. VP8 would be a nice alternative and allow us to remove theora without causing too much of a stir.

[s55]

Actually theora is looking rather dead now. Don't look like much is happening with it these days. Wonder if anyone actually uses it in HB.

[Dukers]

libtheora is dead long ago. The 1.2 version is taking ages to complete and doesn't improve substantially the already very bad compression ratio. FMP4 from libavcodec is probably better than libtheora...

libvpx is not so great, but is miles ahead libtheora. Google is putting man power into it to improve the encoder since version 0.9.5.

[nexradix]

Please don't remove Theora! I use Handbrake with Theora exclusively.

While I agree it isn't as good as h.264, it is currently the ONLY viable format that isn't patent-encumbered. While VP8 has great potential, it isn't supported out-of-the-box in most GNU/Linux distros yet, plus it hasn't been around long enough to provide full confidence that submarine patents might still be out there. Plus while Theora might not have the best compression ratio, at larger file sizes it looks great, and can be decoded on almost any hardware without difficulty.

Don't get me wrong, I'd love to see a VP8 encoder in Handbrake, but wouldn't want Theora removed until VP8 is mature enough to be a suitable replacement.

[s55]

Literally had to collect myself off the floor.... someone choosing to use theora...

I'll expand on what I said earlier. I don't expect to see VP8 support soon, which means theora won't be going anywhere.

It's sole purpose for being there, is for the crazy people like yourself

plus it hasn't been around long enough to provide full confidence that submarine patents might still be out there

Pretty sure MpegLA now has a site that discusses the patents that apply to it. Also, there have been patent pools brandied around for Theora, it's just no-one actually cares about it since there isn't enough people using it to make any money with lawsuits.

[musicvid]

If one compares x264 encodes side-by-side with VP8, it is apparent why I hope Handbrake will never support it, and why Google is shooting itself in the foot by choosing to use VP8 exclusively.

Just my opinion, you understand, but VP8 just doesn't measure up.

[Dukers]

libvpx needs a lot of work, sure. But compare libvpx vs libtheora...

[musicvid]

So in essence, you are saying that VP8 is good because something worse exists??
Thanks for a good chuckle.

[Dukers]

Yes.
Something very worse. Encode anything at the same bitrate and see how incredibly bad libtehora performs compared to the "less bad" libvpx.
But you have nothing to worry about. The crappy libtheora will likely stay for a long time in HB.

[musicvid]

My point is, I don't use either. x264 is working just fine for me even at lower bitrates.

[Dukers]

Hey, s55, libvpx 0.9.6 is out. Improves encoder. When your patches stop producing garbage, hopefully the new version will give us better image.

[s55]

It's not a problem with libvpx, it's a problem with my code. I'm no expert on libhb or C for that matter. It's more a project to see how difficult it would be to integrate, not, provide a full implementation. Looking at a few different implementations at the moment.
So, at this point in time, it's undetermined whether anything will come of my work, or if someone will pickup where ever I leave it.

Time will tell.

[Dukers]

I know. I'm hoping that you and other developers find a way to implement it. Thanks for the effort.

[s55]

Just wanted to make it clear to others that nothing is set in stone.

It's a curiosity thing on my part with may or may not turn into full support.