ok now I see why we need another packet in the header. The first one is basically for ID purposes (theora_decode_header()). I've jammed my huffman trees in there but it sucks because I need them later on when I actually build the huffman tree structs (theora_decode_init, calls InitHuffmanSet()). At that point the original header packet is gone so I've been keeping it around in a buffer. Gross!! Someone please help. This is likely the last bitstream change before we lock it down for Beta, so I want to get it right. ___ Dan Miller (++,) Founder, CTO, On2 Technologies --- >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.
Dan Miller wrote:> ok now I see why we need another packet in the header. The first one is basically for ID purposes (theora_decode_header()). I've jammed my huffman trees in there but it sucks because I need them later on when I actually build the huffman tree structs (theora_decode_init, calls InitHuffmanSet()). At that point the original header packet is gone so I've been keeping it around in a buffer. Gross!! > Someone please help. This is likely the last bitstream change before we lock it down for Beta, so I want to get it right. > ___ Dan MillerDan, if you need your trees for decoding, make sure you can read them from an element that can be easily accessed in the container from outside. If you plan to have DirectShow support for Theora on Windows, bare in mind DShow needs to know this stuff *BEFORE* building the graph. You can not build the DShow graph and then read some chunks of data to initialize your decoder correctly, unless this stuff doesnt change anything vital in the decoding graph affecting the graph itself ( resolution, colour space output, etc. ).... this is, assuming you plan to have a Theora splitter/parser filter on DShow and a Theora decoder filter as separate elements in the playback chain .... Christian --- >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 Tue, 2003-03-18 at 21:40, Dan Miller wrote:> Someone please help. This is likely the last bitstream change before we lock it down for Beta, so I want to get it right.Attached is a partial patch. I believe it's doing the right thing on the encode side, but there's no decode support yet (what's in the patch is useless and broken). Maybe this will be a helpful starting place. I'll work on the other half when I get a chance, but not before tomorrow evening. Note we also want to add a metadata header, identical in format to the vorbis comment header. I recommend we do that as a separate step, though. -r <p> -------------- next part -------------- A non-text attachment was scrubbed... Name: header-g1.diff__charset_UTF-8 Type: text/x-patch Size: 6325 bytes Desc: header-g1.diff__charset_UTF-8 Url : http://lists.xiph.org/pipermail/theora-dev/attachments/20030319/051b421f/header-g1-0001.bin
On Tuesday, March 18, 2003, at 09:40 pm, Dan Miller wrote:> ok now I see why we need another packet in the header. The first one > is basically for ID purposes (theora_decode_header()). I've jammed my > huffman trees in there but it sucks because I need them later on when > I actually build the huffman tree structs (theora_decode_init, calls > InitHuffmanSet()). At that point the original header packet is gone > so I've been keeping it around in a buffer. Gross!!This is now done, along with the vorbis-style comment header. What else remains? I remember some discussion about there being a better way to pack the tables in the third header packet. Dan, do you still want to do that? Have you done it already? I'd also like to add a colourspace flag to the info header. I don't think that's contentious; I'm thinking of just an 8 bit field with three or so defined values. Is there anything else that affects the bitstream, particularly the headers? if not I think we can proceed with an alpha2 release. The big holdup for beta is the spec. Obviously the bitstream needs to settle before that gets going, but maybe now is time to start. How's the schedule looking on that for those working on it? Finally, I should mention for those who don't already know, we've been maintaining todo lists and idea pages for various xiph projects in a wiki. Please check out http://wiki.xiph.org/TheoraTodo and add anything you think is relevant. Cheers, -r --- >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.
hmm..> From: Ralph Giles [mailto:giles@xiph.org] > > ok now I see why we need another packet in the header. The...> > This is now done, along with the vorbis-style comment header. > What else > remains? I remember some discussion about there being a better way to > pack the tables in the third header packet. Dan, do you still want to > do that? Have you done it already?Thanks for the work. I did implement a more efficient packing scheme for the huff trees. I will need to check out your build with the multiple header packets & merge them.> > I'd also like to add a colourspace flag to the info header. I don't > think that's contentious; I'm thinking of just an 8 bit field with > three or so defined values. Is there anything else that affects the > bitstream, particularly the headers? if not I think we can > proceed with > an alpha2 release.We really should get the huffman changes in there; there were some other things I did at that time, I need to review the code.> > The big holdup for beta is the spec. Obviously the bitstream needs to > settle before that gets going, but maybe now is time to start. How's > the schedule looking on that for those working on it?I guess that's me. Unfortunately with the change in my employment status at On2, I haven't put as much time into this as I would like. I've gone from being paid to work on it to a volunteer status. That doesn't mean it won't get done, but I can't put an exact date on it. I figure about a month of reasonably consistent work will get us to a reasonable place for beta release. I'm hoping to put that month in starting in a couple weeks, so perhaps July 1 is a reasonabe date to look for. Isn't the plan to integrate Ogg2 into this beta release? How is that looking?> > Finally, I should mention for those who don't already know, > we've been > maintaining todo lists and idea pages for various xiph projects in a > wiki. Please check out http://wiki.xiph.org/TheoraTodo and > add anything > you think is relevant. > > Cheers, > -r > > --- >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. >___ Dan Miller (++,) Founder, On2 Technologies --- >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.
hi -- (apologies in advance for any formatting problems with this post, I'm stuck on a web email app) -----Original Message----- From: Ralph Giles [mailto:giles@xiph.org]>> Thanks for the work. I did implement a more efficient packing scheme > for the huff trees. I will need to check out your build with the > multiple header packets & merge them.>You're welcome. The merge shouldn't be too hard. We're still callingyour read_/write_QTables/FrequencyCounts() from theora_encode/decode_tables(). Cool -- there may need to be some code reorg though, since I really need the information at a different point in the process (for playback). Right now I stick the bits in an array so I can use them later (yuck)>> We really should get the huffman changes in there; there were some > other things I did at that time, I need to review the code.>I agree. Lets hold alpha 2 for those.> I guess that's me. Unfortunately with the change in my employment > status at On2, I haven't put as much time into this as I would like. > I've gone from being paid to work on it to a volunteer status. That > doesn't mean it won't get done, but I can't put an exact date on it. > I figure about a month of reasonably consistent work will get us to a > reasonable place for beta release. I'm hoping to put that month in > starting in a couple weeks, so perhaps July 1 is a reasonabe date to > look for.>A couple of others had mentioned interest as well. Would it help ifothers contributed or would you prefer to work on your own for the initial draft? That timescale sounds fine I think. There's lots of code polish that can be done in the meantime which will make for a shorter beta. I have no objection to getting some help on the spec. However I've taken a rather unorthodox approach to it, so it might be a bit of a frankenstein unless I can convince others of the wisdom of my ways. If anyone wants to contribute to the spec, please mail me directly at dan@on2.com (it's still my email for now)>> Isn't the plan to integrate Ogg2 into this beta release? How is that > looking?>...In the sense that it's a new api for multiplexing, the line is lessclear. I don't know what monty had in mind here; both libogg2 and OggFile have been mentioned in this sense. I don't feel I have sufficient understanding of the issues at this point to judge a good API design for that. Hopefully experimenting with tools will help us in this direction. we need to get Monty's input on this issue. I think there are some things that he really needs to address as author of the Ogg format. IMHO, -r --- >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. -------------- next part -------------- A non-text attachment was scrubbed... Name: winmail.dat Type: application/ms-tnef Size: 5894 bytes Desc: winmail.dat Url : http://lists.xiph.org/pipermail/theora-dev/attachments/20030513/4cbf3d11/winmail-0001.bin
sounds good to me. Please suggest a time, we can meet in #theora ___ Dan Miller (++,) Founder, On2 Technologies <p>> -----Original Message-----> From: Mauricio Piacentini [mailto:mauricio@tabuleiro.com] > Sent: Tuesday, May 13, 2003 2:54 PM > To: theora-dev@xiph.org > Subject: Re: [theora-dev] bitstream changes > > > Can we schedule maybe a quick IRC meeting later today or > tomorrow, and trace > a plan that will lead to an alpha2 or beta1 release? Monty > would be able to > tell use exactly how to help with this and the missing parts > in the Ogg > integration, and if we can help at all. Maybe it is something > only he can > do, and in this case we can tackle other stuff while he is busy. > > Regards, > Mauricio > > > In the sense that it's a new api for multiplexing, the line is less > > clear. I don't know what monty had in mind here; both libogg2 and > > OggFile have been mentioned in this sense. I don't feel I have > > sufficient understanding of the issues at this point to judge a good > > API design for that. Hopefully experimenting with tools > will help us in > > this direction. > > --- >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. >--- >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.
Maybe Matching Threads
- bitstream versioning
- Bitstream encoded huffman tables always the same
- autoconf problem
- Theora integration question
- theora encoder reordering, order of puting data from DCT 8x8 blocks to huffman compressor, and puting result of huffman compressor to buffer bitstream memory