gmane.comp.multimedia.ogg.theora.devel http://blog.gmane.org/gmane.comp.multimedia.ogg.theora.devel hourly 1 1901-01-01T00:00+00:00 Gmane http://gmane.org/img/gmane-25t.png http://gmane.org Ogg Theora files player for Windows XP http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3224 <pre>Hello, Can anyone help me write Ogg Theora files player in Visual C++ 2008? I built and added to the CLR Empty Project: - libogg-1.3.0, libtheora-1.1.1, libvorbis-1.3.2 (http://www.xiph.org/downloads/) - PortAudio_v19_20111121 (http://portaudio.com/archives/pa_stable_v19_20111121.tgz) - SDL-1.2.15 (http://www.libsdl.org/release/SDL-devel-1.2.15-VC.zip) What next? First I would like to play the *.ogv file. I know that in revision 11431 of Theora (http://svn.xiph.org/trunk/theora) there is splayer.c (win32\experimental\splayer\), but when I try to compile it Visual C++ shows many errors. Besides this in the revision 11431 there aren't all required libraries (for example ogg, vorbis, PortAudio). Best regards, Maciej_______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> Maciej Mączyński 2012-05-11T13:07:24 Ogg Theora files player http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3223 <pre>Hello, I'd like to compile splayer.c from revision 11431 in Visual Studio 2008 on Windows XP. Module splayer.c requires PortAudio. I'm trying to build Portaudio for Windows using Microsoft Visual Studio: http://portaudio.com/docs/v19-doxydocs/compile_windows.html I did step 1 and part of step 2. Step 2 says: "Copy the entire ASIOSDK2 folder into src\hostapi\asio\." Where is the src\hostapi\asio\ folder? Should I create it? If so, where exactly? Point 4 says: "If you have Visual Studio 2005, Visual C++ 2008 Express Edition or Visual Studio 2010, double click the portaudio.sln file located in build\msvc\." Where are the portaudio.sln file and the build\msvc\ folder? There aren't these elements in the \theora\win32\experimental\splayer\portaudio\ folder and anywhere else on my hard disk. Best regards, Maciej_______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> Maciej Mączyński 2012-04-25T08:37:38 Ogg Theora files player http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3220 <pre>Hello, I try to compile player_example.c from http://downloads.xiph.org/releases/theora/libtheora-1.1.1.zip in Visual Studio 2008 on Windows XP. I built and added to the project libogg-1.3.0, libtheora-1.1.1, libvorbis-1.3.2 (http://www.xiph.org/downloads/) and SDL-1.2.15 (http://www.libsdl.org/release/SDL-devel-1.2.15-VC.zip). When I compile player_example.c Visual Studio shows: fatal error C1083: Cannot open include file: 'unistd.h': No such file or directory fatal error C1083: Cannot open include file: 'sys/soundcard.h': No such file or directory fatal error C1083: Cannot open include file: 'sys/ioctl.h': No such file or directory I found out that these files are part of Linux, not Windows. Are there equivalents of these files for Windows? How can I change player_example.c, so that it compile on Windows? Best regards, Maciej Mączyński_______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> Maciej Mączyński 2012-04-11T07:04:36 theora streamer for Android anyone? http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3214 <pre>ola, i am one of the developers of http://giss.tv, some people here might know about it. we use since 5 years some tools ( gstreamer, ffmpeg2theora, pure data, ... ) to make Ogg/Theora/Vorbis streams related to media-hacktivism and social movements. lately, we realized how important it becomes to have lighter tools in some contexts where you cannot set a medialab and a decent network, like in pacific occupation of public spaces ( plaza catalunya, puerta del sol, ... ). that's why we are trying now to develop some streaming solutions for mobile phones and, logically, we try right now to make this work on Android. i compiled libogg, libvorbis and libtheora for ARM and linked it to the Java interface using JNI... but alas, no luck, it crashes in various places of the native code, so i guess i need a proper port of the libraries to Android ( that should work on ARM processors with fixed point processing no ? ). i'm writing now this message to know if anybody have worked on this already and if they suceeded in making an Ogg/Theora encoder that actually works on Android. thanks in advance for any track that could be useful in that research. saludos, sevy&lt; at &gt;giss.tv </pre> ydegoyon< at >gmail.com 2011-10-15T06:47:47 ffmpeg2theora 0.28 released http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3205 <pre>ffmpeg2theora 0.28 - http://v2v.cc/~j/ffmpeg2theora/ Some changes that went into this release: - select resize method via commandline - fix 5.1 channel ordering - update to FFmpeg 0.7 branch - additional bug fixes Source: http://v2v.cc/~j/ffmpeg2theora/downloads/ffmpeg2theora-0.28.tar.bz2 svn co https://svn.xiph.org/tags/ffmpeg2theora/ffmpeg2theora-0.28  Binaries: Linux 32bit http://v2v.cc/~j/ffmpeg2theora/ffmpeg2theora-0.28.linux32.bin Linux 64bit http://v2v.cc/~j/ffmpeg2theora/ffmpeg2theora-0.28.linux64.bin Mac OS X http://v2v.cc/~j/ffmpeg2theora/ffmpeg2theora-0.28.pkg Windows http://v2v.cc/~j/ffmpeg2theora/ffmpeg2theora-0.28.exe _______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> j< at >v2v.cc 2011-07-31T16:28:50 media.xiph.org y4m rearrangement http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3204 <pre>All, Just to let you know, we've flattened the file hierarchy in Tim's (derf's) collection of y4m-format test clips. Previously, the HD content was sorted into separate directories. Now it's all in the same directory as the SD and CIF-resolution files. If you have a mirror, you can save yourself some bandwidth by manually moving everything up a level: 'cd y4m; mv *p/* .' before rsyncing again. If you don't have a mirror and would like one, the archive is available over anonymous rsync: rsync rsync://media.xiph.org/y4m/ # for a file listing rsync -a --progress rsync://media.xiph.org/y4m/ y4m/ # for a complete copy If you're using test scripts which assume the old layout, they will need updating. -r </pre> Ralph Giles 2011-05-26T23:01:14 Can't compile libtheora vs2010 http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3186 <pre>I'm getting errors like so on initial build of libtheora - 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\x86_vc\x86stat.c': No such file or directory 1&gt; mmxstate.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\x86_vc\mmxstate.c': No such file or directory 1&gt; mmxloopfilter.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\x86_vc\mmxloopfilter.c': No such file or directory 1&gt; mmxidct.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\x86_vc\mmxidct.c': No such file or directory 1&gt; mmxfrag.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\x86_vc\mmxfrag.c': No such file or directory 1&gt; state.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\state.c': No such file or directory 1&gt; quant.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\quant.c': No such file or directory 1&gt; internal.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\internal.c': No such file or directory 1&gt; info.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\info.c': No such file or directory 1&gt; idct.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\idct.c': No such file or directory 1&gt; huffdec.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\huffdec.c': No such file or directory 1&gt; fragment.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\fragment.c': No such file or directory 1&gt; dequant.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\dequant.c': No such file or directory 1&gt; decode.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\decode.c': No such file or directory 1&gt; decinfo.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\decinfo.c': No such file or directory 1&gt; decapiwrapper.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\decapiwrapper.c': No such file or directory 1&gt; bitpack.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\bitpack.c': No such file or directory 1&gt; apiwrapper.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\dec\apiwrapper.c': No such file or directory 1&gt; recon_mmx.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\x86_32_vs\recon_mmx.c': No such file or directory 1&gt; fdct_mmx.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\x86_32_vs\fdct_mmx.c': No such file or directory 1&gt; Generating Code... (TaskId:16) 1&gt; Compiling... (TaskId:16) 1&gt; dsp_mmx.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\x86_32_vs\dsp_mmx.c': No such file or directory 1&gt; reconstruct.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\reconstruct.c': No such file or directory 1&gt; mode.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\mode.c': No such file or directory 1&gt; mcenc.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\mcenc.c': No such file or directory 1&gt; mathops.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\mathops.c': No such file or directory 1&gt; frinit.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\frinit.c': No such file or directory 1&gt; frarray.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\frarray.c': No such file or directory 1&gt; encoder_quant.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\encoder_quant.c': No such file or directory 1&gt; encoder_toplevel.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\encoder_toplevel.c': No such file or directory 1&gt; encoder_idct.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\encoder_idct.c': No such file or directory 1&gt; encoder_huffman.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\encoder_huffman.c': No such file or directory 1&gt; encode.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\encode.c': No such file or directory 1&gt; encapiwrapper.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\encapiwrapper.c': No such file or directory 1&gt; dct_encode.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\dct_encode.c': No such file or directory 1&gt; dct_decode.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\dct_decode.c': No such file or directory 1&gt; dct.c (TaskId:16) 1&gt;c1 : fatal error C1083: Cannot open source file: '..\lib\enc\dct.c': No such file or directory 1&gt; Generating Code... (TaskId:16) 1&gt; The command exited with code 2. (TaskId:16) 1&gt;Done executing task "CL" -- FAILED. (TaskId:16) 1&gt;Done building target "ClCompile" in project "theora_static.vcxproj" -- FAILED.: (TargetId:37) _______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> andrew aspden 2011-04-22T19:04:01 using Kate for WebVTT encapsulation http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3175 <pre>(sending this publicly so the discussion about this is in the open) Hi OggK, all, OggK, you've probably seen that the bug in Firefox for adding Kate support has been closed as WONTFIX: https://bugzilla.mozilla.org/show_bug.cgi?id=481529 I've just posted a reply there mentioning how I envisage in-band WebVTT support to be provided through Kate. I want to ask your opinion. In case you haven't heard about WebVTT yet, it's the time-aligned text format that has been developed by the WHATWG as the baseline text format for HTML5 media elements, see http://www.whatwg.org/specs/web-apps/current-work/webvtt.html . Also see http://www.whatwg.org/specs/web-apps/current-work/multipage/video.html#the-track-element for how to include it with a audio or video element. My idea is that - because WebVTT is so similar to SRT - it would be simple to support WebVTT in Ogg by encapsulating it in a Kate track and making sure that it can be extracted again without loss of information through the kate libraries. There would only be a few changes necessary: * WebVTT has a header which needs to be parsed and re-created. * Also, there is a suggestion for inclusion of name-value Metadata at the top right after the header, which we'd want to retain (maybe in a header or a first packet). * Then there are cue settings, which are position modifiers on each cue (segment of timed text). They need to be recreated, too. The marked-up text inside cues needs to be retrieved unchanged. For an old, but still mostly correct explanation with examples of some of the things see: http://annodex.net/~silvia/tmp/WebSRT/ (WebSRT got renamed to WebVTT) and for something a bit newer see: http://www.html5videoguide.net/presentations/WebVTT/ I'm just wondering what you think about this kind of approach. Cheers, Silvia. </pre> Silvia Pfeiffer 2011-04-12T00:44:42 Inverse DCT in coding stage http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3173 <pre>Why in the procedure oc_enc_block_transform_quantize() is used function oc_idct8x8()? After all, Inverse DCT only need to decoding stage, rather than coding? </pre> Alexander Lubyagin 2011-04-08T12:54:36 binary stream after tokenizer and huffman http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3171 <pre>Which procedure/function or structure returns finally formed binary stream? I see lib/tokenize.c and lib/encode.c. But it is too hard to understand novice. Theora algorithm differ in coding with JPEG: http://www.pcs-ip.eu/index.php/main/edu/8 Which function/structure returns stream like this: 100000100011001101101101111101011111111010110011010 </pre> Alexander Lubyagin 2011-04-07T20:33:24 quantize after fdct, _dequant table, and idct http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3170 <pre>1) What are you doing "mathematically" in a procedure x86/x86enquant: oc_enc_quantize_sse2()? This - the assembler code, and I do not understand mathematically - that's going on there. --- A: 120 121 28 73 -20 -99 -98 -100 123 122 112 108 73 -32 -102 -98 123 123 117 121 100 74 -15 -97 124 125 125 124 123 116 74 -33 125 125 125 125 124 124 100 74 124 124 124 125 125 124 121 114 125 125 124 125 125 125 120 117 125 125 125 125 125 125 123 113 --- _dct: 2767 1165 -439 154 15 24 -87 -53 -1254 1080 -245 0 125 -15 -101 -67 -512 260 212 -207 200 -49 -80 -55 -111 -111 300 -166 125 -17 -70 -45 -54 -69 127 18 -31 60 -82 -30 -50 -30 40 83 -74 72 -68 -35 -31 -33 37 48 -40 40 -28 -44 -21 -13 29 12 1 -3 11 -44 --- _dequant: 24 32 32 36 32 32 40 36 36 36 40 40 40 44 48 56 48 48 48 48 48 56 56 56 52 56 60 56 60 60 60 56 60 56 60 64 64 64 60 60 64 60 56 60 64 64 64 64 64 64 64 64 64 60 64 64 64 64 64 64 64 64 64 64 --- _qdct: 115 36 -39 -14 34 -14 4 -7 7 -3 -1 -3 5 0 0 0 3 -4 6 -1 -1 -1 -1 2 -3 4 0 -2 -1 -2 -1 2 0 1 -1 0 0 1 1 -1 0 -1 -1 -1 -1 1 -1 1 0 0 -1 1 -1 -1 0 -1 1 0 0 0 -1 -1 0 -1 So, _qdct[i][j] = _dct[i][j] / _dequant[i][j] ? 2) Second question - you are using: 24 20 20 20 20 16 24 20 20 20 28 24 24 28 32 40 32 32 28 28 32 44 40 40 36 40 44 44 44 44 44 44 44 44 44 48 48 48 44 44 48 44 44 44 48 48 48 48 48 48 48 48 48 44 48 48 48 48 48 48 48 48 48 48 , or 24 32 32 36 32 32 40 36 36 36 40 40 40 44 48 56 48 48 48 48 48 56 56 56 52 56 60 56 60 60 60 56 60 56 60 64 64 64 60 60 64 60 56 60 64 64 64 64 64 64 64 64 64 60 64 64 64 64 64 64 64 64 64 64 for luma plane, and 28 28 28 36 32 36 52 40 40 52 60 56 52 56 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 60 for chroma plane? 3) Third question. How do you restore the original matrix after fdct and quantize ... Used dequantize + idct? </pre> Alexander Lubyagin 2011-04-05T11:25:43 RLE for DC/AC-coefficients separately for the blocksor at whole? http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3167 <pre>It is proper way - to code DC/AC-coefficients for all macro-blocks at whole? I think that coding for DC/AC-coefficients for blocks separately gives lesser RLE-compression ratio. It is right? For example, at first we code DC-coefficients for all blocks. Compress they by RLE. Then, AC(1) coefficients. Compress it by RLE. And ... AC(63) coefficients for all blocks. Compress by RLE. It is better for RLE-coding? </pre> Alexander Lubyagin 2011-04-04T09:39:53 DCT-coefficients in http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3166 <pre> Why for white-pixel image, oc_enc_fdct8x8_c() returns the matrix with non-zero AC-coefficients? _x[64]: * 107 107 107 107 107 107 107 107 * 107 107 107 107 107 107 107 107 * 107 107 107 107 107 107 107 107 * 107 107 107 107 107 107 107 107 * 107 107 107 107 107 107 107 107 * 107 107 107 107 107 107 107 107 * 107 107 107 107 107 107 107 107 * 107 107 107 107 107 107 107 107 _y[64]: * 3426 1 1 1 0 0 0 0 * 1 2 1 1 0 0 0 0 * 1 1 1 1 0 0 0 0 * 1 2 1 1 0 0 0 0 * 1 1 1 1 0 0 0 0 * 0 1 1 1 0 0 0 0 * 1 1 1 1 1 0 1 0 * 0 0 0 0 0 0 0 0 This is - error of butterfly-method? P.S. My program dct.py returns: 1016 0 -0 0 0 0 -0 -0 0 -0 0 0 0 0 0 -0 0 -0 0 0 0 -0 0 0 0 0 -0 0 0 0 -0 0 0 -0 0 -0 0 0 -0 -0 0 0 0 0 0 0 0 0 -0 0 -0 -0 0 0 -0 -0 -0 -0 0 0 -0 -0 0 0 </pre> Alexander Lubyagin 2011-03-31T17:22:26 DCT in Theora and DCT "by hands" http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3165 <pre>Hello. I try to perform DCT using oc_enc_fdct8x8_x86_64sse2(), and by Python-script dct.py. Results differ. Why? Original pixel data block: Morig = [ [28,28,28,28,28,28,28,28], [28,28,28,28,28,28,28,28], [28,28,28,29,29,29,28,28], [28,29,28,29,29,29,28,28], [28,29,28,28,28,28,28,28], [29,28,29,28,28,28,28,28], [29,28,29,29,29,29,28,28], [29,29,29,29,29,29,29,29], ] Result by libtheora: * -3187 3 -3 2 0 -1 1 -1 * -5 -1 0 1 0 -1 -1 -1 * 4 0 3 0 1 2 2 2 * -4 4 5 -1 -1 1 1 2 * 2 0 2 0 -1 -2 -3 -2 * 2 1 -4 3 2 0 2 -1 * 2 1 2 0 1 1 1 1 * -1 0 0 1 0 0 0 -1 Result by dct.py: -1479 55 50 47 39 30 21 10 52 -2 -1 -1 -1 -1 -1 0 52 -2 -1 -1 -1 0 0 0 44 -1 0 -1 -1 0 0 0 40 -1 -1 -1 -1 -1 -1 0 32 -1 -2 0 0 0 0 0 21 0 0 0 0 0 0 0 10 0 0 0 0 0 0 0 ---- Source code of dct.py: #!/usr/bin/python # pixels Morig = [ [28,28,28,28,28,28,28,28], [28,28,28,28,28,28,28,28], [28,28,28,29,29,29,28,28], [28,29,28,29,29,29,28,28], [28,29,28,28,28,28,28,28], [29,28,29,28,28,28,28,28], [29,28,29,29,29,29,28,28], [29,29,29,29,29,29,29,29], ] from math import sqrt,cos,pi from sys import stdout M = 8 N = 8 def dct(a): print "---" F = [] for u in xrange(N): F = F + [[]] for v in xrange(M): s = 0.0 for i in xrange(N): for j in xrange(M): #print a[i][j] li = 1 if (i == 0): li = 1/sqrt(2) lj = 1 if (j == 0): lj = 1/sqrt(2) d = li*lj*(a[i][j]-128) #d = (a[i][j]-128) d *= cos(((pi*u)*(2*i+1))/(2*N)) d *= cos(((pi*v)*(2*j+1))/(2*M)) s += d s *= sqrt(2.0/N) s *= sqrt(2.0/M) F[u] = F[u] + [s,] stdout.write(" %6d" % s) stdout.write("\n") return F dct(Morig) # END Source code based on http://www.cs.cf.ac.uk/Dave/Multimedia/node231.html formulas for 2D-DCT. </pre> Alexander Lubyagin 2011-03-31T14:53:42 Examples of matrices before FDCT http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3164 <pre>Why do we need a function x86/mmxencfrag:oc_enc_frag_sub_128_mmx()? ^ 28 28 28 28 28 28 28 28 ^ 28 28 28 28 28 28 28 28 ^ 28 28 28 29 29 29 28 28 ^ 28 29 28 29 29 29 28 28 ^ 28 29 28 28 28 28 28 28 ^ 29 28 29 28 28 28 28 28 ^ 29 28 29 29 29 29 28 28 ^ 29 29 29 29 29 29 29 29 After _sub_128_mmx: ^ -100 -100 -100 -100 -100 -100 -100 -100 ^ -100 -100 -100 -100 -100 -100 -100 -100 ^ -100 -100 -100 -99 -99 -99 -100 -100 ^ -100 -99 -100 -99 -99 -99 -100 -100 ^ -100 -99 -100 -100 -100 -100 -100 -100 ^ -99 -100 -99 -100 -100 -100 -100 -100 ^ -99 -100 -99 -99 -99 -99 -100 -100 ^ -99 -99 -99 -99 -99 -99 -99 -99 After FDCT, x86/sse2fdct:oc_enc_fdct8x8_x86_64sse2(), it is: * -3187 3 -3 2 0 -1 1 -1 * -5 -1 0 1 0 -1 -1 -1 * 4 0 3 0 1 2 2 2 * -4 4 5 -1 -1 1 1 2 * 2 0 2 0 -1 -2 -3 -2 * 2 1 -4 3 2 0 2 -1 * 2 1 2 0 1 1 1 1 * -1 0 0 1 0 0 0 -1 It is right? So, we need 13 bits for storing DCT coefficients (for 8-bit pixels)? See formulas - http://www.cs.cf.ac.uk/Dave/Multimedia/node231.html F(i,j) = 64*255*2/8 = 4080 = 12 bits; + Sign = 13 bits (-4096...+4095)? </pre> Alexander Lubyagin 2011-03-31T09:29:30 DCT in Theora http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3161 <pre> Yes, I forgot about the files x86/sse2*.c Function oc_enc_fdct8x8_x86_64sse2() now visible in my debug messages. Thanks, Timothy. </pre> Alexander Lubyagin 2011-03-28T17:11:32 DC prediction http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3150 <pre>Hi, The specification points out a problem in the dc prediction : "This procedure describes the complete process of undoing the DC prediction to recover the original DC values. Because it is possible to add a value as large as 580 to the predicted DC coefficient value at every block, which will then be used to increase the predictor for the next block, the reconstructed DC value could overflow a 16-bit integer." And explains the solution : "This is handled by truncating the result to a 16-bit signed representation, simply throwing away any higher bits in the two’s complement representation of the number." which is done in the algorithm: * Truncate DC to a 16-bit representation by dropping any higher-order bits. But why dropping higher-order bits ? With a simple exemple with an unsigned integer 100001 If I truncate this number to a 4-bit representation by dropping any higher-order bits I get 0001 which is far from the initial value. I think the value could be rounded to 1111 which is better ? To round DC, which is signed with a 17-bit representation, to a 16 bit signed value, this gives : if (DC&lt;− (2^15)) DC = -2^15 else if (DC&gt;2^(15)-1) DC = -2^(15)-1 Am I wrong ? Regards J.F. _______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> Jonathan Fabrizio 2011-03-25T11:38:25 inter-frame vs intra-frame http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3149 <pre>Hi, The specification does not give any indication on when recording intra-frame. Obviously the first frame is an intra frame but after what is the strategy used in Theora ? I think an intra frame is recorded whenever consecutive frames become too different ? Is there any criterion ? What's happen in case of periodic changes (*blinking *lights for exemple) ? Thanks J.F. _______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> Jonathan Fabrizio 2011-03-25T11:34:03 Various typo in spec http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3144 <pre>Hi, I have few questions on the specification. * Just to understand : - What is the interest of super blocks ? Is it to save place when recording coded block flags (7.3) ? - What is the advantage of using the coded order ? it is more often easier with raster order (especially to find neighbor in 7.8.1 for exemple). Is it to simplify the correspondence between block, macro block and super block ? * A problem with motion vectors : - How motion vector are decoded ? In the specification, it is said : "Each component can take on integer values from -31 . . . 31, inclusive, at half-pixel resolution, i.e. -15.5 . . . 15.5 pixels in the luma plane. For each sub- sampled axis in the chroma planes, the corresponding motion vector component is interpreted as being at quarter-pixel resolution, i.e. -7.75 . . . 7.75 pixels. The precise details of how these vectors are used to compute predictors for each block are described in Section 7.9.1. " But nowhere, decoded values are divided by two. Algorithm in 7.5.1 gives MVX and MVY in -31..31. Algorithm in section 7.5.2 keep these values. In section 7.9.1.3, the use of MVX2 and MVY2 seem to be the solution but in algorithm 7.9.4, these two variables are initialized with the same value : B. Assign MVX the value |MVECTS[bi ]x | ? sign(MVECTS[bi ]x ). C. Assign MVY the value |MVECTS[bi ]y | ? sign(MVECTS[bi ]y ). D. Assign MVX2 the value |MVECTS[bi ]x | ? sign(MVECTS[bi ]x ). E. Assign MVY2 the value |MVECTS[bi ]y | ? sign(MVECTS[bi ]y ). As MVECTS are integer, floor and ciel functions give the same result. I do not understand. Maybe MVECTS[bi ] should be replaced by MVECTS[bi ] /2 in all these expressions ? * Few typos : in section 7.8.1 Variables used : PDIV (...) The valud -&gt; the valu*e* mbi-&gt;mb*j* (...) The index ... block bj 7. If block bi is not along the *the* bottom in table 7.47 : p[0] p[1] p[2] p[3] w[3] w[1] w[2] w[3] -&gt; p[0] p[1] p[2] p[3] w[*0*] w[1] w[2] w[3] Regards J.F. _______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> Jonathan Fabrizio 2011-03-24T14:24:32 FPGA encode stages flow diagram http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3138 <pre>Good day! I create diagram of encoder process. Using it i create implementation of encoder in FPGA (Xilinx/Altera). Please critique it. Is there missing stages? Here is blog http://developer-fpga.blogspot.com/ Here is picture of encoding stage 1 https://lh4.googleusercontent.com/-NV8o9DG3jvE/TYjYXr-dYGI/AAAAAAAAAos/U06O-YvhSI0/s1600/stage1.jpg Here is picture of encoding stage 2 https://lh5.googleusercontent.com/--1U5TaiVAEU/TYjYhW4n2OI/AAAAAAAAAow/vRFbzObFhww/s1600/stage2.jpg _______________________________________________ theora-dev mailing list theora-dev&lt; at &gt;xiph.org http://lists.xiph.org/mailman/listinfo/theora-dev </pre> digital design 2011-03-22T17:42:59 Motion vector decode http://comments.gmane.org/gmane.comp.multimedia.ogg.theora.devel/3136 <pre>Hi, In the specification page 142 1.(d). it is said : If FTYPE is non-zero (inter frame), using PF, NMBS, MBMODES, NBS, and BCODED, decode the motion vectors into MVECTS using the procedure given in Section 7.5.1. I think, there is a little mistake as the correct procedure is not in section 7.5.1 but in section 7.5.2 (?) Regards J.F. </pre> Jonathan Fabrizio 2011-03-17T11:46:44 Search Engine Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.multimedia.ogg.theora.devel