--- "Tabuleiro" wrote: Hi, guys. I was working with the VP3 code, planning to integrate it with Ogg for a self-contained solution for video playback, platform independant, no installation required, no QT, no AVI runtimes. My main interest is CD-based work using local files and portability/performance, not really into streaming or anything fancy (network error correction, prediction, etc.) Not sure if Theora will cater for my needs or if I should continue with my current development plans... --- end of quote --- This sounds really cool! We get our video from CDs and webservers--nothing fancy--and would definitely be interested in a simple, portable framework. The existing systems (QuickTime, VFW, Real) are very complex, frequently buggy, and messy to install. I think there's a real market for simple, self-contained VP3+Vorbis playback library. I'd love to see Theora evolve in this direction. How much work do you think would be required to get from here to there? Cheers, Eric --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'theora-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
I don't know where Theora is going, but I believe it is going to be more powerful than what I need at this time. I agree with you that for simple, basic video playback you don't need a lot of stuff that is now part of Real, Windows Media and Quicktime. I would love to get a 100K library that only gets a stream of data (one audio and one video track), plays the audio (Vorbis compressed) and attempts to retrieve and decompress the current frame in synch with it as best as it can. Kind of AVI but with a self contained runtime. I believe Smacker and Bink catered for this market for a while, and were (are still being) used by a large number of game developers. That's what I was working on, and I was leaning towards the following simplistic plan: 1) Interleave audio and video using an AVI like RIFF structure. 2) "Emulate" the VFW behavior, which would give me a hook into the existing VP3 codec code. The basic VFW API is very simple, and only does a couple of calls to the codec, basically passing a bag of bytes and requesting a decompressed frame in a given format (negotiated in advance.) I was planning to use only one type of surface (probably RGB because it is easier to port, but I understand the VP3 likes YUV better, and needs additional steps to decompress to RGB. Is this right?) This is of course is the simplest possible approach, and some will say it is too simple and flawed, and the AVI file structure is terrible. On the other hand I really liked the idea of the Ogg framework and how streams are organized, though, but I must confess I have not understand it completely :) , so it seems a little complicated... (able to resume in cases of network congestion, etc.) Maybe it will be possible to wait and contribute in this area, so we would use a common and flexible file format. Regards, Mauricio Piacentini Tabuleiro>>>>The existing systems (QuickTime, VFW, Real) are very complex, frequently buggy, and messy to install. I think there's a real market for simple, self-contained VP3+Vorbis playback library. I'd love to see Theora evolve in this direction. How much work do you think would be required to get from here to there? <p>--- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'theora-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
On Wednesday, July 17, 2002, at 06:51 PM, Eric M. Kidd wrote:> The existing systems (QuickTime, VFW, Real) are very complex, > frequently buggy, and messy to install. I think there's a real market > for simple, self-contained VP3+Vorbis playback library. I'd love to > see Theora evolve in this direction.Well, the general plan is to extract a portable version of the vp3 codec, develop and ogg encoding for it, and wrap it with an api probably very like the vorbis one. That should be a simplification over the other wrappers you mention. I don't see why we wouldn't be interested in hosting continued development of the quicktime and vfw plugins; however, I've not looked closely enough at the current packet structure to be able to guess how much (if at all) we'd want to change it for ogg. If there are significant bitstream changes, that might complicate maintenance. Slightly later, Tabuleiro wrote:> Count me in if you need help with tools for this project, I have some > experience in software development for Windows/Mac/MacOSX (no Linux, > though... have to learn it)We certainly do need help with tools. I believe Monty has volunteered for the codec cleanup and ogg embedding, but we'll also need a encoding and playback examples and apps for a variety of platforms and formats, likely including maintainers for the quicktime and vfw plugins. But this is an open source project. All reasonably congruent help is appreciated. :-) -r <p>--- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'theora-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
--- Ralph Giles wrote: I don't see why we wouldn't be interested in hosting continued development of the quicktime and vfw plugins; however, I've not looked closely enough at the current packet structure to be able to guess how much (if at all) we'd want to change it for ogg. If there are significant bitstream changes, that might complicate maintenance. --- end of quote --- The various existing VP3 codecs all seem to run in a "packet-per-frame" mode. QT or VFW justs tosses them a blob of data (and requests an output pixel format), and a thin glue layer munges stuff around a bit and calls into VP3 itself. I can think of two advantages of preserving this architecture: 1) It's probably less work. 2) Nearly all professional video editing systems support QuickTime/MPEG-4 container files. If Ogg Theora uses the existing VP3 packet format, it would probably be possible to convert QuickTime container files to Ogg container files, and leave the data unchanged. This would be a quick&dirty way to get some decent tool support for Theora. VP3 internally supports quite a few pixel formats, with a couple of caveats: A) The supported pixel formats are slightly different on Mac and Windows. B) The assembly code for many of the pixel formats is heavily optimized and CPU-dependent. Under Windows, you need to install a variety of assembler upgrades and MSVC service packs just to compile the code. So porting the assembly code to GCC/gas is likely to be a pain, if that's the plan. But anyway, that's enough theory--on to development stuff. :-) If you don't mind hosting the VP3 QuickTime codecs (at least for now), to whom would you like me to send patches? Once again, thank you for all your cool multimedia work! Cheers, Eric --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'theora-dev-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.