Slashdot Mirror
Technical Objections To the Ogg Container Format
E1ven writes "The Ogg container format is being promoted by the Xiph Foundation for use with its Vorbis and Theora codecs. Unfortunately, a number of technical shortcomings in the format render it ill-suited to most, if not all, use cases. This article examines the most severe of these flaws."
Google cache link: http://209.85.229.132/search?q=cache:http://hardwarebug.org/2010/03/03/ogg-objections/&hl=en&strip=1
Because, in general, you don't only want a video stream. You want to be able to bundle multiple streams together (usually of different types, eg an audio stream and a video stream). Vorbis, for example, is audio. Theora would be video. Ogg (the container) exists to bundle them together into a single file, so you can have a movie and sound with it.
I'm not an expert in video or audio production, I just dabble in it as a hobby. but one thing I often wonder is, what is the point of these container formats?
The point of containers is to make the contained media modular. This way, you can assemble a multimedia ("many media"... whodathunkit) file from several individual pieces of media, each with codecs that may or may not be on speaking terms, into a cohesive file that i played as a unit.
If you define a format that has a video track and two channels of audio, you might think, hey, that's great, and play stuff on your computer. But what if you want a 5.1 audio track? Make another format, one for stereo and one for 5.1. Second audio program, like an alternate language? More formats: one for stereo + stereo SAP, one for 5.1 main + stereo SAP, one for 5.1 both; up to 5 formats now. Subtitles or closed captioning? More formats: up to 20 now. A different audio codec? Even assuming the SAP tracks and main use the same codec (they might not; depends on where you got the SAP track from), add 20 formats per audio codec. Multiple video codecs? The number of formats can grow exponentially. And we haven't even gotten to things like multiple camera angles and sideband info like text commentary, HTML links to things discussed in the show, or TV listings.
Or you can define a container and then, in each of those cases, only need to define a new component to be put into the container, and you're done. Containers make things much simpler and easier to implement.
Haven't read the article, because it's slashdoted, but I assume it's about the fact that the Ogg container was initially designed as a transport stream format for audio.
The article goes considerably beyond that, arguing that the container is flawed even as an audio format. Here's the money quote (emphasis mine):
Sean Daugherty "I have walked in Eternity -- and Eternity weeps."
I use Handbrake http://handbrake.fr/ for encoding. Handbrake will let you chose from MP4 or MKV for container, H.264 or Theora for video and MP3, AAC or Vorbis for audio.
However, it's not unique to OGG. AFAIK, MKV is also patent-free, and it's the standard container for torrents^Wprivate-encoded HD video. And it's a much better container anyway.
Exactly this. Matroska in general is great and a lot better than Ogg or others, but it doesn't work on any device besides PC - not on 360, not on PS3, not in mobile phones..
However, matroska support is pretty much standard in any but the most proprietary set-top boxes. For example - WDTV, TiVX, Popcorn Hour - basically anything that uses any recent Sigma Designs chipset. Similarly iRiver supports matroska on their newest portable media players and Archos's latest android based pmp also supports matroska.
JVC and Phillips have currently shipping blu-ray players that play matroska. Panasonic has announced their next generation of TVs and blu-ray players will do matroska, and the specs for NEC's next gen of video decoder chipsets (which compete with Sigma Designs) say they will include matroska support.
When information is power, privacy is freedom.
> I don't see any reason, why the version would have to change in the middle of a file in any case.
It is probably not due to the fact, that the version might change in the middle of the file, but in case, you only have a part of the file.
This makes it more robust, and better suitable for streaming: You can simply start sending from an arbitrary position, and the parser should
be able to recover at some point.
"Between strong and weak, between rich and poor [...], it is freedom which oppresses and the law which sets free"
No, the flaw is yours. The 1 bit merely says "this is not the original version" and anyone that only knows the original version just stops there. Anyone that knows the 2nd version has enough smarts to look at the 2nd version bit (or field).
In which case once there is a second version you have the exact same packet format as the current ogg, except for an extra mask, test, and one fewer flag bit. So the only gain at all is if you assume there will never be another version, and if there is even one more version then you've caused a pipeline stall for no reason. Which is stupid.
This goes along with the criticism of the checksum field as 'wasted space', but it is probably put there so you can reliably find the page header if doing a random seek. Which if you can do, then you don't need a time index because you can do a binary search to find any time index with only a tiny bit of extra seeks.
I haven't looked at these formats in depth, but it sure sounds like this guy is clueless.