[Ffmpeg-devel-irc] ffmpeg-devel.log.20120716

burek burek021 at gmail.com
Tue Jul 17 02:05:03 CEST 2012


[01:09] <philipl> Hi all. I'm looking at how to handle timed text styling, and I think it's going to be a real pain without using something like avio, but avio is in avformat, and moving it to avutil isn't trivail - given the dependencies on the URL code, etc.
[01:09] <philipl> Is there any ready-to-go alternative? or any other ideas?
[01:13] <durandal_1707> why you need avio at all?
[01:23] <Compn> philipl : erm, what timed text ? mp4 subs ?
[01:44] <durandal_1707> heh imagine fate with bigTIFF sample
[01:48] <philipl> Compn: yeah.
[01:48] <philipl> The styles are basically structs appened to the end of the text for each subtitle packet.
[01:49] <philipl> So I have to do low-level stream IO to work with them (or try and use actual structs but that runs into packing and endian problems.
[01:49] <Compn> are you the timed text encoder author ?
[01:49] Action: Compn isnt paying attention
[01:50] <philipl> Yeah. I just pushed the decoder, and I've been working on the encoder for a while.
[01:50] <philipl> I'm not trying to do styles as part of the initial implementation but I have to write out a basic stream header and so I still have to deal with the basic problem.
[01:51] <Compn> did you look at how other subs do styles ?
[01:51] <Compn> i would think it would be the same way...
[01:51] <Compn> since its all text
[01:54] <CIA-41> ffmpeg: 03Luca Barbato 07master * r45ff9e5b57 10ffmpeg/doc/APIchanges: doc: update api changes with the right commit hashes
[01:54] <CIA-41> ffmpeg: 03Daniel Kang 07master * r337ade52de 10ffmpeg/libavcodec/vp8.c: 
[01:54] <CIA-41> ffmpeg: vp8: refactor decoding a single mb_row
[01:54] <CIA-41> ffmpeg: This is in preperation for sliced threading.
[01:54] <CIA-41> ffmpeg: Signed-off-by: Luca Barbato <lu_zero at gentoo.org>
[01:54] <CIA-41> ffmpeg: 03Daniel Kang 07master * r17343e3952 10ffmpeg/libavcodec/ (vp8.c vp8.h): 
[01:54] <CIA-41> ffmpeg: vp8: move data from VP8Context->VP8Macroblock
[01:54] <CIA-41> ffmpeg: In preparation for sliced threading.
[01:54] <CIA-41> ffmpeg: Signed-off-by: Luca Barbato <lu_zero at gentoo.org>
[01:54] <CIA-41> ffmpeg: 03Daniel Kang 07master * r951455c1c1 10ffmpeg/libavcodec/ (vp8.c vp8.h): (log message trimmed)
[01:54] <CIA-41> ffmpeg: vp8: implement sliced threading
[01:54] <CIA-41> ffmpeg: Testing gives 25-30% gain on HD clips with two threads and
[01:54] <CIA-41> ffmpeg: up to 50% gain with eight threads.
[01:54] <CIA-41> ffmpeg: Sliced threading uses more memory than single or frame threading.
[01:54] <CIA-41> ffmpeg: Frame threading and single threading keep the previous memory
[01:54] <CIA-41> ffmpeg: layout.
[01:54] <CIA-41> ffmpeg: 03Ronald S. Bultje 07master * rc44091a9f7 10ffmpeg/libavcodec/snowenc.c: 
[01:54] <CIA-41> ffmpeg: snow: refactor code to work around a compiler bug in MSVC.
[01:54] <CIA-41> ffmpeg: This fixes the compiler error "cannot convert from 'BlockNode' to
[01:54] <philipl> Compn: Well, that's the thing, it's not text. It's basically serialized structs.
[01:55] <CIA-41> ffmpeg: vp8: Enclose pthread function calls in ifdefs
[01:55] <CIA-41> ffmpeg: This fixes building with threads disabled.
[01:55] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:55] <CIA-41> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r082829520e 10ffmpeg/libavcodec/vc1dec.c: 
[01:55] <CIA-41> ffmpeg: vc1dec: Do not use random pred_flag if motion vector data is skipped
[01:55] <CIA-41> ffmpeg: This fixes SA10143.vc1 from test-suite. Also partially fixes MC-VC1.ts
[01:55] <CIA-41> ffmpeg: from videolan streams archive.
[01:55] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:55] <CIA-41> ffmpeg: 03Ronald S. Bultje 07master * rb4675d0fbf 10ffmpeg/configure: 
[01:55] <CIA-41> ffmpeg: configure: Check for CommandLineToArgvW
[01:55] <CIA-41> (16 lines omitted)
[01:55] <philipl> The issues not how to interpret them or map them, but how to do effective IO to read and write them.
[01:58] <durandal_1707> michaelni: i think geotiff support should be mentioned in Changelog
[01:59] <michaelni> yes, please add it
[02:00] <durandal_1707> busy with b2 tiff thing
[02:08] <michaelni> ok, moment, ill add it
[02:09] <Compn> philipl : serialized struct ? like xml ? or .. i'm confused
[02:10] <philipl> No, just binary.
[02:10] <Compn> oh
[02:11] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r8b855b69a1 10ffmpeg/Changelog: 
[02:11] <CIA-41> ffmpeg: Changelog: add geotiff
[02:11] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[02:51] <durandal_1707> michaelni: it metadata stuff properly escaped when displayed?
[03:02] <michaelni> not sure what you mean by properly escaped but special chars that could cause problems should be replaced
[03:04] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r7543fd80e5 10ffmpeg/libavcodec/exr.c: 
[03:04] <CIA-41> ffmpeg: exr: check size of uncompressed buffer returned by uncompress()
[03:04] <CIA-41> ffmpeg: The actual size of uncompressed buffer returned by uncompress() may be
[03:04] <CIA-41> ffmpeg: smaller than expected, so abort decoding in such cases.
[03:04] <CIA-41> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[03:05] <durandal_1707> michaelni: tiff b2 patch on ml
[03:34] <Hawk-TAO> hi
[03:34] <Hawk-TAO> hi,gentlemen
[03:34] <Hawk-TAO> i have a question
[03:42] <Hawk-TAO> hi
[03:42] <Hawk-TAO> i have a question
[03:43] <Compn> so ask it
[03:43] <Hawk-TAO> anyone use crypto in ffmpeg?
[03:43] <Compn> dont think anyone does
[03:44] <Hawk-TAO> u said none dose?
[03:46] <Compn> right
[03:46] <Compn> try asking in #libav-devel 
[03:46] <Hawk-TAO> Compn,you said none dose?
[03:46] <Compn> yes
[03:46] <Hawk-TAO> there is none to know in that chat group
[03:50] <Hawk-TAO> sir,did you know Michael Niedermayer?
[03:50] <Hawk-TAO> is he manager of ffmpeg commuty?
[03:51] <Hawk-TAO> sir,i am a fresh in ffmpeg from China. i need your help.
[03:52] <Compn> Hawk-TAO : yes, michael niedermayer
[03:53] <Compn> is here as michaelni
[03:53] <Compn> he can help you
[03:53] <Compn> but i dont know if he is awake right now
[03:53] <Compn> sorry, i dont know the code . i am not a programmer :(
[03:54] <Compn> Hawk-TAO : greetings from USA :)
[04:00] <Hawk-TAO> ok
[04:00] <Hawk-TAO> thx
[04:05] <Hawk-TAO> Michael Niedermayer,are you online?
[04:05] <Hawk-TAO> Michael Niedermayer,sir,i need your help.
[04:06] <Hawk-TAO> compn,did you know Michael Niedermayer's e-mail?
[04:07] <Compn> yes
[04:07] <michaelni> Hawk-TAO, how can i help ?
[04:07] <Hawk-TAO> sir
[04:07] <Hawk-TAO> i wanna use crypto in ffmpeg
[04:07] <Compn> michaelni : i think Hawk-TAO already posted his question on ffmpeg-devel, asking about encrypting video using ffmpeg
[04:07] <Hawk-TAO> yes
[04:09] <durandal_1707> and michaelni replied
[04:09] <Compn> Hawk-TAO : why do you want to encrypt video ? are you broadcasting video or just want to keep it private ?
[04:10] <durandal_1707> there is gnutls and openssl which provide some kind of "crypto"
[04:10] <Hawk-TAO> yeah
[04:10] <Hawk-TAO> i wanna use ffserver
[04:13] <Hawk-TAO> and i didnt know the appropriate place to add the crypto interface. 
[04:14] <Hawk-TAO> i wanna use OpenCL to parallelize it.
[04:17] <Hawk-TAO> sir
[04:17] <Hawk-TAO> are you online?
[04:18] <durandal_1707> why crypto interface?
[04:19] <Hawk-TAO> there is crypto.c in ffmpeg source
[04:20] <durandal_1707> so you wanna add encryption?
[04:21] <Hawk-TAO> yeah
[04:22] <Hawk-TAO> i wanna add encryption in ffserver and decryption in client.
[04:23] <durandal_1707> what client, there is already decryption support in crypto
[04:23] <durandal_1707> perhaps you want some other kind of chiper?
[04:23] <durandal_1707> afaik crypto do only aes
[04:24] <Hawk-TAO> the chiper from crypto++
[04:25] <durandal_1707> actually aes encryption is already available in code, just protocol stuff is not implemented
[04:25] <durandal_1707> wtf is crypto++?
[04:26] <Hawk-TAO> yeah,but i dont know how to use it
[04:27] <durandal_1707> well using crypto++ library could be possible but you are on your own to implement it....
[04:28] <durandal_1707> because it is C++ it is even more going to be ugly
[04:28] <durandal_1707> so better pick what chipers are you want to use and implement them or use openssl/gnutls stuff which is already used by ffmpeg/libav
[04:29] <Hawk-TAO> oh,and i wanna to parallelize it
[04:30] <Compn> rot13
[04:30] <Compn> unbreakable!
[04:31] <Hawk-TAO> what is rot13?
[04:31] <Compn> just kidding, its a very simple encryption
[04:32] <Hawk-TAO> hhh
[04:33] <Hawk-TAO> and what the crypto.c is put into the ffmpeg source for?
[04:33] <Hawk-TAO> none use crypto.c in ffmpeg?
[04:33] <durandal_1707> it used
[04:34] <durandal_1707> read git log libavformat/crypto.c for more
[04:34] <Hawk-TAO> i cannt find any interface in crypto.c is called.
[04:38] <Hawk-TAO> git log?
[04:38] <durandal_1707> well, it is going to be used .... like tls.c and others
[04:38] <durandal_1707> Hawk-TAO: learn to use git
[04:41] <Hawk-TAO> is this "git-howto.txt"
[04:43] <durandal_1707> that is just short introduction
[04:44] <durandal_1707> michaelni: next nice task would be adding support for exif tags to tiff decoder
[04:44] <Hawk-TAO> sir,but what is the git log?
[04:51] <durandal_1707> http://lmgtfy.com/?q=git+log
[05:11] <Hawk-TAO> good night to all
[05:11] <Hawk-TAO> see you next day.
[13:44] <varaderoguy> Hello all
[13:44] <varaderoguy> Need to ask a compiling question
[13:45] <varaderoguy> When I configure - what is the option to allow the 'man' page to be included
[13:45] <varaderoguy> at the moment, I have ffmpeg --help - just no man ffmeg
[13:55] <Compn> make doc or make docs ?
[13:55] <Compn> it should be made already
[13:55] <Compn> assuming you have proper libs installed
[13:55] <Compn> make install will put the manual in your system
[14:09] <varaderoguy> I didn't realise you have to make the docs seperately
[14:09] <varaderoguy> Make install doesn't do it by default
[14:20] <varaderoguy> okay - still more puzzlement
[14:20] <varaderoguy> something broken here
[14:21] <varaderoguy> I've got N-41922-g9b72041
[14:21] <varaderoguy> make docs + make doc don't work
[14:36] <michaelni> varaderoguy, do you have all tools that are needed to build docs ?
[14:37] <varaderoguy> michaelni: what tools do I need then?
[14:38] <michaelni> for building whole docs: texi2html makeinfo pod2man
[14:41] <varaderoguy> Chaps - do you need a hand with writing some docs, because this kind of stuff is not really documented....
[14:42] <Compn> yes
[14:42] <Compn> help is always welcome :)
[14:42] <Compn> if configure could spit out a warning when you dont have all of the right tools, that would be good i think
[14:43] <varaderoguy> okay - well, I am a broadcast engineer, so I hope I can help you out a little
[14:43] <varaderoguy> I'm not great at coding, but I am okay at bash and php
[14:44] <varaderoguy> so - let me look at the .configure file - and see what I can do
[14:44] <varaderoguy> as soon as I've figured out how to make the doc files myself!
[14:47] <Compn> or you can write it in the docs
[14:47] <Compn> required tools
[14:47] <Compn> if its not already there
[14:47] <Compn> doesnt require any skills, except copy and paste 
[14:48] <varaderoguy> lovely
[14:48] <varaderoguy> Let me get the latest nightly build
[14:48] <varaderoguy> do you guys use SVN?
[14:49] <varaderoguy> ignore me....you use git
[14:53] <kierank> varaderoguy: it was you asking about the red thing, no?
[14:53] <varaderoguy> kierank: yes - that's me
[15:54] <CIA-41> ffmpeg: 03Tim Nicholson 07master * r45641014d3 10ffmpeg/doc/filters.texi: 
[15:54] <CIA-41> ffmpeg: doc/filters: Fix two typos
[15:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r371d37fcf5 10ffmpeg/libavcodec/ffv1.c: 
[15:54] <CIA-41> ffmpeg: ffv1: keep last_frame (to be used for error concealment)
[15:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r094845aad8 10ffmpeg/libavcodec/ffv1.c: 
[15:54] <CIA-41> ffmpeg: ffv1dec: add simple error concealment in case of CRC errors on slices.
[15:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rf5af3568f6 10ffmpeg/libavcodec/ffv1.c: 
[15:54] <CIA-41> ffmpeg: ffv1dec: keep track of errors in slice headers for EC
[15:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rb0d674ec10 10ffmpeg/libavcodec/ffv1.c: 
[15:54] <CIA-41> ffmpeg: ffv1dec: set the first slices bytestream end correctly
[15:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rac9389a663 10ffmpeg/libavcodec/ffv1.c: 
[15:54] <CIA-41> ffmpeg: ffv1dec: detect errors in bytestream end mismatches for EC
[15:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r8486c8cc9a 10ffmpeg/configure: 
[16:20] <CIA-41> ffmpeg: 03Nick Brereton 07master * r73154feec1 10ffmpeg/libavcodec/dca.c: 
[16:20] <CIA-41> ffmpeg: Fix ordering of XXCH downmix coefficients.
[16:20] <CIA-41> ffmpeg: Reviewed-by: Benjamin Larsson <benjamin at southpole.se>
[16:20] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:20] <CIA-41> ffmpeg: 03Nick Brereton 07master * re03077c432 10ffmpeg/libavcodec/dca.c: 
[16:20] <CIA-41> ffmpeg: Fix DCA-XXCH extension scaling for embedded downmixes.
[16:20] <CIA-41> ffmpeg: Reviewed-by: Benjamin Larsson <benjamin at southpole.se>
[16:20] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:52] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rb45a313e97 10ffmpeg/libavcodec/ffv1.c: 
[16:52] <CIA-41> ffmpeg: ffv1dec: fix error concealment for gop>1
[16:52] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:17] <burek> michaelni and varaderoguy, I think this is related to that docs thing: http://ffmpeg.org/trac/ffmpeg/ticket/1449 and http://ffmpeg.org/trac/ffmpeg/ticket/1515
[17:22] <burek> btw, does scale filter support constrained scaling if a user specifies only the width or height?
[17:24] <burek> oh, just ignore me, I thought there was no example in docs for that, but there is (the last line) :) -1 used :)
[17:55] <varaderoguy> Chaps - I'm off for tonight - thanks for all your support
[20:16] <Daemon404> ubitux, ping
[20:26] <ubitux> Daemon404: pong
[20:27] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r5dba4cbc2f 10ffmpeg/ffmpeg.c: 
[20:27] <CIA-41> ffmpeg: ffmpeg: fix streamcopy of some mp3 in avi
[20:27] <CIA-41> ffmpeg: Fixes Ticket1432
[20:27] <CIA-41> ffmpeg: Thanks-to: Mike Scheutzow <mike.scheutzow at alcatel-lucent.com> for some of the bug analysis
[20:27] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:28] <Daemon404> ubitux, so in 3dd2c8bbbc725bb596eca879908881d252114243
[20:29] <Daemon404> is there a reason you didnt add a check for (!st)
[20:29] <Daemon404> seems like Bad Things could still happen
[20:29] <Daemon404> (im fixing warnings in Libav, then moving to ffmpeg)
[20:30] <Daemon404> (ffmpeg warnings seem to be a semi-superset)
[20:32] <ubitux> mmh not sure
[20:32] <ubitux> i wonder if the if (rf64) was there at that time
[20:33] <Daemon404> im adding a check to libav, so i can add it to ffmpeg too
[20:33] <ubitux> and so sample_count wouldn't have been set if no fmt
[20:33] <Daemon404> it seems needed now
[20:33] <Daemon404> if im nto mistaken?
[20:33] <ubitux> yeah it looks like so
[20:34] <Daemon404> ok
[20:34] <Daemon404> ill add it
[20:51] <CIA-41> ffmpeg: 03Martin Storsjö 07master * r2ca6e9bb47 10ffmpeg/libavformat/ (network.h os_support.c): 
[20:51] <CIA-41> ffmpeg: network: Extend the fallback gai_strerror implementation to handle more error codes
[20:51] <CIA-41> ffmpeg: This is useful if a proper getaddrinfo is loaded dynamically on
[20:51] <CIA-41> ffmpeg: windows, while using the fallback implementation of gai_strerror.
[20:51] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:51] <CIA-41> ffmpeg: 03Martin Storsjö 07master * rc98e2053f5 10ffmpeg/libavformat/network.h: (log message trimmed)
[20:51] <CIA-41> ffmpeg: network: Undefine existing gai_strerror definitions
[20:51] <CIA-41> ffmpeg: This avoids warnings if there already exists a definition.
[20:51] <CIA-41> ffmpeg: This is the case on windows, where the getaddrinfo isn't available
[20:51] <CIA-41> ffmpeg: and linked to (and we use our fallbacks instead, which actually
[20:51] <CIA-41> ffmpeg: try to use the proper getaddrinfo version if found at runtime),
[20:51] <CIA-41> ffmpeg: but gai_strerror still exists as a define.
[20:51] <CIA-41> ffmpeg: 03Martin Storsjö 07master * r9e4b04f8b9 10ffmpeg/libavformat/ (network.h os_support.c): 
[20:51] <CIA-41> ffmpeg: network: Always use our version of gai_strerror on windows
[20:51] <CIA-41> ffmpeg: Even if linking directly to getaddrinfo, use our version of
[20:51] <CIA-41> ffmpeg: gai_strerror instead of the system's version. Microsoft explicitly
[20:51] <CIA-41> ffmpeg: documents that their version of gai_strerror is thread-unsafe.
[20:51] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:51] <CIA-41> ffmpeg: 03Mohammad Alsaleh 07master * rc37c383e59 10ffmpeg/libavcodec/libfdk-aacenc.c: 
[20:51] <CIA-41> ffmpeg: libfdk-aac: Check if cutoff value is valid
[20:51] <CIA-41> ffmpeg: Passing a cutoff value < sample_rate/256 will cause a crash.
[20:51] <CIA-41> ffmpeg: Also, values >20000 will have no effect and 20000 will be used anyway.
[20:51] <CIA-41> ffmpeg: Signed-off-by: Mohammad Alsaleh <msal at tormail.org>
[20:51] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:51] <CIA-41> ffmpeg: 03Diego Biurrun 07master * r20d569dd61 10ffmpeg/libavformat/mpeg.c: mpeg: remove disabled code
[20:51] <CIA-41> ffmpeg: 03Martin Storsjö 07master * r03d1b18e96 10ffmpeg/configure: 
[20:51] <CIA-41> ffmpeg: configure: Remove an unused 'have' item
[20:51] <CIA-41> ffmpeg: This is only used for checking for a certain library, but the code
[20:51] <CIA-41> ffmpeg: doesn't need to know whether the function was found.
[20:51] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:51] <CIA-41> ffmpeg: 03Jordi Ortiz 07master * r08e087ccf7 10ffmpeg/libavformat/rtmpproto.c: 
[20:51] <CIA-41> (13 lines omitted)
[21:08] <durandal_1707> can i get review for tiff b2?
[21:19] <michaelni> durandal_1707, moment, ill review it
[21:57] <Daemon404> michaelni, what are the ramifications of that SAR patchset?
[21:57] <Daemon404> will it change behavior for calling apps?
[21:58] <Daemon404> particularily on h264 with midstream sar changes
[21:58] <nevcairiel> the only change is that midstream SAR changes actually work with -mt
[21:59] <nevcairiel> sar changes could otherwise already happen (if you dont multithread)
[22:02] <Daemon404> nevcairiel, i see
[22:02] <Daemon404> neat.
[22:44] <durandal_1707> i waited enough for vivo demuxer, so i gonna write it myself
[22:45] <durandal_1707> huh ptx got broken on be
[22:47] <durandal_1707> heh this one is obvious
[22:47] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r6dac8c8327 10ffmpeg/libavutil/attributes.h: 
[22:47] <CIA-41> ffmpeg: lavu: disable av_uninit for clang
[22:47] <CIA-41> ffmpeg: This silence bunch of useless warnings like:
[22:47] <CIA-41> ffmpeg: libavformat/mpeg.c:393:37: warning: variable 'dvdaudio_substream_type' is uninitialized when used within its own initialization [-Wuninitialized]
[22:47] <CIA-41> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[22:55] <durandal_1707> hmm this looks like swscale issue, michaelni?
[23:29] <michaelni> durandal_1707, are you sure ptx is rgb555 and not bgr555 ?
[23:31] <durandal_1707> michaelni: dunno, i have no explanation for recent ptx fate failures
[23:37] <michaelni> well, it looks like bgr and -sws_flags +bitexact is missing
[23:38] <michaelni> but its very possible thats not the only issue
[23:38] <michaelni> and theres something in sws
[23:38] <durandal_1707> why bgr?
[23:38] <michaelni> because the sky is blue and not red normally
[23:38] <durandal_1707> on le?
[23:38] <michaelni> it looked wrong with ffplay
[23:39] <durandal_1707> but then other blue stuff will be red ...
[23:43] <durandal_1707> Compn: do you have vivo 1 samples?
[23:49] <durandal_1707> michaelni: colors indeed are wrong
[23:51] <durandal_1707> and you are right it is bgr555 and not rgb555
[23:56] <Compn> durandal_1707 : vivo1 in .viv container...
[23:56] <Compn> is in the /viv/ dir
[23:56] <Compn> er /vivo/
[23:56] <durandal_1707> Compn: any file i try is 2.0 version and we do not have video decoder for it
[23:56] <Compn> go for the old files 
[23:57] <Compn> from 2001
[23:57] <Compn> one sec
[23:57] <Compn> http://samples.mplayerhq.hu/vivo/off-gnaway-h.viv
[00:00] --- Tue Jul 17 2012


More information about the Ffmpeg-devel-irc mailing list