xiphmont

Overview:

FFmpeg's built-in Vorbis encoder produces low enough quality output to be considered broken. This encoder is used by default in the majority of FFmpeg builds, and will produce .ogv and WebM videos with low to unusably poor audio quality.

This alert is intended for all users of FFmpeg (via the command line or GUI wrappers) and all application developers that make use of the FFmpeg command line tool. Application developers that use the FFmpeg libraries should also take care that the libavcodec built-in Vorbis encoder library is not used by accident.

Scope:

All past and present builds of FFmpeg and libavcodec up to but not including the upcoming 0.6 release. Default builds of the upcoming FFmpeg 0.6 release will not use the built-in encoder by default, but it will still be possible to accidentally use or restore the built-in encoder to default status during the FFmpeg build. It should be assumed that any build of FFmpeg and any application using FFmpeg could be producing videos with substandard Vorbis audio unless the FFmpeg build and usage is verified to be using system Vorbis libraries, such as those provided by Xiph.Org or aoTuV.

Workaround / Fix:

FFmpeg can be forced to use the external/system libVorbis library by passing:

-acodec libvorbis

as part of the FFmpeg command line.

Note that passing '-acodec vorbis' is incorrect and requests the low-quality built-in FFmpeg-internal Vorbis encoder. Also, FFmpeg may be built without libvorbis support, meaning that many FFmpeg builds only have the internal encoder available. In this case, requesting '-acodec libvorbis' will fail with the error 'Unknown encoder 'libvorbis''.

FFmpeg can be built with working libvorbis support and the internal Vorbis encoder disabled as follows:

./configure --disable-encoder=vorbis --enable-libvorbis; make; make install

Such a build completely removes the internal Vorbis encoder from libavcodec, eliminating the possibility of accidental use on the command line or in libavcodec-based applications.

Verification:

Use of a good Vorbis encoder in .ogg, .oga, .ogv and WebM files may be verified as follows. This test will work on any Ogg or WebM file to verify the encoder that produced the audio. Note that 'Vorbis' is case-sensitive:

strings file_to_be_checked | grep Vorbis

A file that was encoded using a good encoder will output a line containing 'Xiph.Org libVorbis' or 'AoTuV', such as:

boatanchor$ strings test2.ogg |grep Vorbis
Xiph.Org libVorbis I 20100325 (Everywhere)

or

boatanchor$ strings test3.ogg |grep Vorbis
AO; aoTuV b5d [20090301] (based on Xiph.Org's libVorbis)

A file encoded by FFmpeg's internal Vorbis encoder will produce no grep output as it does not set a vendor string.

Mans Rullgard has written two long rants about the Ogg container in the past few years. One made it to Slashdot apparently based on the drama potential alone. If you don't know what I'm talking about below, don't worry about it, tl;dr.

I'd not originally intended to respond to open trolling. The continued urging of many individuals has convinced me it's important to rebut in some public form. Earnest falsehoods left unchallenged risk being accepted as fact.

The rest of the gory details below the pseudo-cut

[The pseudo-cut is off-livejournal due to post-size restrictions]

Released last night: libogg 1.2.0, libvorbis 1.3.1, libao 1.0.0, and vorbis-tools 1.4.0.

*whew*

Aside from bugfixes, the major story here is the first wide release of all the surround hacking I've been doing the past two months. Previous demo pages (one and two) had documented my progress till now; there's actually one more coming that I've just started on.

In addition to that, we also tagged on a release of libogg with a new default page spill rule that reduces container overhead for higher bitrate streams like video. The change is simple, but it's a reduction in bitrate for free so why not. Applications don't have to do anything, just drop in the new lib. In addition to demo pages and blogging, I've also begun a comprehensive expansion of the Ogg documentation (the very beginnings of which are in the release-- not much yet).

The releases as usual can be had from the Xiph.Org downloads page.

It's amazing how software sidetracks can plow through years of time. I'm finally getting back to Vorbis (and other audio development) after a minor two year video distraction that saw the completion of a new experimantal Theora encoder (Thusnelda) last year, and the beginning of the next experimental Theora encoder which Tim has named Ptalarbvorm. Ptalarbvorm is already showing further large improvements over Thusnelda (honestly, I think he's already doubled again on Thusnelda).

But this post is about Vorbis.

One thing on the original Vorbis bullet list was surround encoding. Vorbis was always surround capable and the software is happy to encode as many channels as you like, but once you're past stereo, everything is encoded as entirely discrete channels. This hasn't been so bad really; despite discrete channels, Vorbis's coding efficiency compares favorably to other surround-capable formats. That said, Vorbis can do better.

When beginning surround optimization work on Vorbis, I found the encoder handled remapping the channel order of input formats, but it didn't always do it properly (FLAC and some WAV inputs were frotzed). I also found that ogg123 needed updating to handle output order. These improvements will be released in vorbis-tools 1.4.0 sometime soon.

The hard part of the tool updates turned out to be libao, which both had no concept of channel ordering, and had also bitrotted substantially while maintainerless. The idea had been that Lennart Poettering's SydneyAudio was going to replace libao so there was no sense continuing to develop libao. We'd just use libao while we waited and not put any more resources into it. Unfortunately, Lennart has been entirely occupied fighting his PulseAudio battles, and this is frankly a more important use of his time. SydneyAudio is no closer to completion than it was a few years ago. We can't continue to wait for it.

So, I'm announcing the resumption of active development and maintenance of libao until such time as there's something better to replace it.

I've made up a surround demo page (like the ones I made during Thusnelda development) with more details on vorbis-tools, libao, and Vorbis's in-progress surround hacking. It goes into alot more detail about Vorbis's surround and channel coupling system with diagrams and samples. I hope folks like it!