continued theora improvements

May 7, 2009 in OGG, Open Web, Video | 17 comments

Monty posted another update on the work that’s been going on to improve the Theora encoder. It’s worth re-posting here because I think that it includes some compelling images and graphs that show you improvements. So I would suggest that people wander over and have a look at his update.

The headlines include:

1.  They have made substantial improvements to Theora’s encoder.  The images which I include below really show off the improvements in sharpness at the same bitrate.

2. That the encoder is now creating higher-quality streams than H.264 at many bitrates. The data includes some comparison with x264 without ffmpeg bugs which show on this test that x264 does do better than Theora in this particular test. However, there’s an important side note worth reading on this topic.

3. That the original tests that gave Theora such a bad name were done with incredibly bad tools.  See the squiggly line on the graph in monty’s post for evidence of that.

Anyway, a picture speaks a thousand words so I’ll include them here.  Open them up in two tabs and switch between them for the full effect.

Theora 1.0:



Theora.next:



Monty points out that this was largely other people’s work and they should get most of the credit. So Greg, Tim and many others – thanks. Keep up the great work.

  1. Laurent Espitallier’s avatar

    Laurent Espitallier on May 7, 2009 at 4:00 pm

    Impressive.

    Will Theora.next be compatible with Firefox 3.5 ?

  2. John (J5) Palmieri’s avatar

    John (J5) Palmieri on May 7, 2009 at 4:03 pm

    This just goes to show that people should have been listening to Monty all along. I was amazed how much push back Theora got with some quoting it as being “last generation technology”. As is often the case with many open source project the potential is there and what those who aren’t developing the technology miss is that the payoffs aren’t instant. They take time and energy just like any other project, just that Open Source projects get released early and often. I for one am glad resources are being put into a completely free codec instead of forcing us to use patented technology when creating and deploying content. Now with Mozilla supporting Theora as a baseline video format across all platforms it has become a competitive video codec.

  3. Christopher Blizzard’s avatar

    Christopher Blizzard on May 7, 2009 at 4:05 pm

    The decoder in 3.5 will be able to play video that’s encoded with the new encoder so it should be fine. Most of this work is being done on the production side to produce videos that are higher-quality. The format itself doesn’t have to change to get these improvements into the hands of users. So no software changes in players required – yay!

  4. Christopher Blizzard’s avatar

    Christopher Blizzard on May 7, 2009 at 4:07 pm

    @J5, Yeah I’m pretty happy that Red Hat and Monty were both able to invest early and over a period of years. That recent grant that Mozilla did recently has helped as well. But the web as a delivery mechanism makes sense and is likely to be a deciding factor in if the format finds a home.

  5. Speedboy’s avatar

    Speedboy on May 7, 2009 at 5:29 pm

    Wow, big change!

  6. Mibmib’s avatar

    Mibmib on May 8, 2009 at 4:39 am

    Let’s hope this Is the way out of the flash hell on youtube :)

  7. dave’s avatar

    dave on May 8, 2009 at 4:48 am

    Regarding “Theora.next” (generally called Thusnelda, or Theora 1.1) being compatible with Firefox:

    The short answer is yes it but there is a longer, perhaps more geekily interesting answer. In the not so distant past some Theora decoders were only able to handle the commonly produced Theora files and weren’t actually fully spec compliant. In preparation for using more of the spec in the Thusnelda release to maximise quality, the Theora devs made their own decoder fully spec compliant and also fixed up the Cortado Java Applet. With that work out of the way the continual improvements and tweaks can be made on the encoding side without having to update the decoder side in Firefox, portable media players etc.

    Regarding Headine 2 in your post, I think the Theora devs are trying their best to raise the level of discussion about video codec quality. Unqualified statements like that aren’t really helpful and only lead to those who favor H.264 getting defensive. It’s more correct to say that this particular clip, with those particular settings, got a higher PSNR rating than x264 (with its own particular settings) at higher bitrates. PSNR is an objective test of “quality” and is useful for tuning codecs, but it’s flawed exactly because it’s objective. Some codec tweaks can improve subjective quality as reported by viewers while reducing PSNR. Think Megahertz Myth and advertising-led increases in clock speed that led to AMD just assigning a number that looked like a clock speed which they felt reflected their true speed rating.

  8. Robert Wohlrab’s avatar

    Robert Wohlrab on May 8, 2009 at 5:59 am

    Hi,
    nice picture. As you stated before it is better than h264 (i think you mean x264 implementation). Can you please generate this picture with xvid, x264 and dirac (not schroedinger) and post it here? I would be interested.

    But big thanks to everyone who done something for Ogg Theora :)

  10. David Tenser’s avatar

    David Tenser on May 8, 2009 at 10:08 am

    Very cool. I’m suspecting that Theora is first and foremost a video codec meant to be used for real video footage/films/cameras. The still image analogy would be jpeg.

    Is there anything similar for screen projected video, e.g. screencasts? The still image analogy would be png or gif. In other words, video that has sharp pixel-accurate edges, and straight, non-blurry lines, etc.

  11. Christopher Blizzard’s avatar

    Christopher Blizzard on May 8, 2009 at 11:24 am

    I think that it’s a pretty good codec at the medium range. I think that later benchmarks will show it a little behind H.264 across a range of benchmarks, but not by much.

    I don’t know about screencasts – I suspect that it is fine there based on the ones that I’ve generated. But I also suspect that there are more appropriate codecs for that that optimize for that case.

  15. basic’s avatar

    basic on July 3, 2009 at 12:01 am

    Has anyone thought of distributing the theora quicktime components with the Mac version of Firefox?

  16. Christopher Blizzard’s avatar

    Christopher Blizzard on July 3, 2009 at 2:55 pm

    We’ve thought about it, yeah, but I think it would be a surprise for users to discover that Firefox was installing random software on their compute.r

  17. hippy3000’s avatar

    hippy3000 on July 4, 2009 at 3:28 pm

    i donit no jack-shit about wat u guyz r all talking about…peace out!