Blogging Theora

[xiphmont]

It's time for the latest Thusnelda encoder project update summary!

Although I haven't gotten much time to dive back into Thusnelda coding myself, Tim Terriberry, Greg Maxwell and others have continued the work along at a merry pace. In the past few weeks, they've finally replaced the leaky fDCT from the original VP3 and begun work on adjusting the main quantization matrices and, hopefully soon, adaptive quantization. These improvements all improve fine detail rendering and, somewhat unexpectedly, improve gradient rendering as well:

The screen caps above were produced by Theora 1.0 on the left and an experimental version of Thusnelda with early quant matrix optimization work in addition to the new fDCT on the right. Both clips were encoded in constant-quantizer mode and equal bitrates.

Other improvements, more details and the full update report here.

Comments

Vorbis ?

[(Anonymous)]

What about the vorbis improvements made by aoTuV http://www.geocities.jp/aoyoume/aotuv/ ?

Will they be included in Xiph.org vorbis? It's sad that most wikipedia audio/video use the standard libvorbis wich has lower quality than aoTuV, which it's used only by few people (most linux distributions ship official Xiph.org libvorbis).

Re: Vorbis ?

[xiphmont.livejournal.com]

Eventually. We're finishing a bugfix release of libvorbis now, and improvements will be the release after.

Re: Vorbis ?

[nullc.livejournal.com]

FWIW— Virtually none of the audio/video on Wikipedia has a low enough audio bitrate for the AoTUV improvements to shine.

(Not that merging it isn't good…)

Re: Vorbis ?

[(Anonymous)]

According to many listening tests (see this page for a summary: http://wiki.xiph.org/index.php/VorbisEncoders ) the aoTuV libvorbis has a better quality than official libvorbis with most audio rates.