[Ffmpeg-devel-irc] ffmpeg-devel.log.20170110

burek burek021 at gmail.com
Wed Jan 11 03:05:02 EET 2017


[10:27:18 CET] <ubitux> f450cc7bc5 h264: eliminate decode_postinit()
[10:27:20 CET] <ubitux> actually
[10:27:28 CET] <ubitux> that's where we're blocked
[10:27:34 CET] <ubitux> (forgot --reverse yesterday)
[10:41:19 CET] <ubitux> i'm working on the next merge commit
[10:42:14 CET] <wm4> nice
[10:57:04 CET] <ubitux> i'll need some testing obviously, h264 related but also wrt hwaccel
[12:32:34 CET] <ubitux> https://github.com/ubitux/FFmpeg/compare/merge-libav
[12:32:41 CET] <ubitux> can i have testing on this branch?
[12:32:49 CET] <ubitux> it passes FATE so far
[12:32:53 CET] <ubitux> michaelni ^
[12:33:07 CET] <ubitux> i moved some field reconstruction code, so you might want to test this
[12:35:44 CET] <ubitux> i might not have put it exactly in the correct position in h264_select_output_frame()
[12:35:48 CET] <ubitux> i need to think again about it
[13:17:37 CET] <michaelni> ubitux, changes ./ffmpeg  -flags2 showall -i ~/tickets/4274/sample.ts  showall.avi
[13:17:46 CET] <michaelni> also many new warnings are shown
[13:21:05 CET] <michaelni> input file should be here (link from ticket) https://docs.google.com/file/d/0BxP5-S1t9VEEY3RqUmg3NHk3aEE/edit
[13:38:43 CET] <ubitux> michaelni: thx, i'll have a look soon
[13:50:16 CET] <mateo`> michaelni: hello ! do you have any comments on the audio pool implementation for lavfi (as well as its use) ?
[13:52:30 CET] <michaelni> mateo`, i think a audio pool is a good idea
[13:58:28 CET] <michaelni> mateo`, looking at the patch but iam unsure about nb_samples 
[13:59:46 CET] <michaelni> IIUC if theres a frame with huge nb_samples then subsequent frames will also get huge nb_samples
[14:00:35 CET] <michaelni> but maybe this is a non issue, i dont know
[14:04:47 CET] <wm4> mateo`: what happens if filters often output differently sized frames?
[14:08:28 CET] <mateo`> the pool will re-allocate to the maximum size
[14:09:09 CET] <wm4> pl
[14:09:11 CET] <wm4> *ok
[14:09:41 CET] <mateo`> btw, a pool is allocated for each filter link
[14:18:38 CET] <mateo`> michaelni: it will use a bit more memory but will avoid reallocations if nb_samples is lower than the current value. I think this is a non issue (but i could be wrong)
[14:28:39 CET] <michaelni> mateo`, ok
[14:33:04 CET] <cone-620> ffmpeg 03Michael Niedermayer 07master:f48b6b8b91d6: avcodec/tiff: Perform multiply in tiff_unpack_lzma() as 64bit
[17:04:06 CET] <ubitux> michaelni: can you retry?
[17:05:56 CET] <ubitux> i don't understand why it works though
[17:06:18 CET] <ubitux>                 !(h->cur_pic_ptr->field_poc[0] == INT_MAX &&
[17:06:20 CET] <ubitux>                   h->cur_pic_ptr->field_poc[1] == INT_MAX) &&
[17:06:22 CET] <ubitux>                 !(h->missing_fields <= 1)) {
[17:06:33 CET] <ubitux> these are the conditions i needed to add before calling ff_thread_finish_setup()
[17:06:54 CET] <ubitux> which were conditions dropped in the original commit
[17:07:12 CET] <ubitux> (they were in decode_postinit, to go out quickly)
[17:07:24 CET] <ubitux> (which was preventing the call to ff_thread_finish_setup() at the end)
[17:07:49 CET] <ubitux> so while i can understand we need that at least one field_poc needs to be valid
[17:08:03 CET] <ubitux> i don't understand why we need at least one missing field to be fine
[17:08:08 CET] <ubitux> !(h->missing_fields <= 1))
[17:08:27 CET] <ubitux> but anyway, it fixes the warning flood
[17:08:35 CET] <ubitux> due to too much ff_thread_finish_setup() calls
[17:08:48 CET] <ubitux> also, the showall flags is properly honored now
[17:28:44 CET] <ubitux> michaelni: apparently i can use !h->setup_finished && h->cur_pic_ptr instead of this
[17:28:54 CET] <ubitux> which is added in later commits
[17:28:57 CET] <ubitux> so i'll do that instead
[17:30:15 CET] <ubitux> so branch updated
[18:04:10 CET] <michaelni> ubitux, it seems this changes: ./ffmpeg  -debug 16384 -vismv 7 -i ~/videos/panda.mp4 -f framecrc -an new.crc
[18:07:19 CET] <michaelni> http://samples.ffmpeg.org/mov/mp4/panda.mp4
[18:12:19 CET] <durandal_1707> Compn: ping
[18:17:17 CET] <Compn> yes
[18:17:24 CET] <Compn> durandal_1707 : i got new mplayer compiled to test
[18:20:13 CET] <durandal_1707> Compn: do you have clrvidcd.dll?
[18:20:29 CET] <durandal_1707> can be in uppercase
[18:23:41 CET] <Compn> durandal_1707 : http://samples.ffmpeg.org/drivers32/newcodecpack/clrviddd.dll and http://samples.ffmpeg.org/drivers32/newcodecpack/CLRVIDDC.DLL
[18:23:48 CET] <Compn> are those two ok ?
[18:24:37 CET] <durandal_1707> nope, it must be CD.dll
[18:24:50 CET] <Compn> oh
[18:24:58 CET] <Compn> let me see
[18:25:06 CET] <Compn> i dont have it but i can do internet search for it
[18:27:21 CET] <Compn> possible that i renamed the dll ?
[18:27:22 CET] <Compn> ehe
[18:29:15 CET] <durandal_1707> you had it at one, perhaps in old mplayer days
[18:29:28 CET] <Compn> durandal_1707 : did you check the installer zips here http://samples.mplayerhq.hu/V-codecs/UCOD/official_UCOD/
[18:29:55 CET] <durandal_1707> there is record of it in multimedia.cx wiki
[18:30:12 CET] <durandal_1707> Compn: have that already
[18:30:17 CET] <Compn> does the dll you are decompiling reference it ?
[18:30:33 CET] <Compn> i know about my fourcc list on multimedia wiki
[18:30:58 CET] <durandal_1707> Compn: this dll does encoding
[18:31:55 CET] <Compn> oh
[18:32:25 CET] <Compn> i will have to search the web archive of http://www.iterated.com/ then
[18:32:33 CET] <Compn> i do not have the clearvideo encoder dll 
[18:34:07 CET] <Compn> looks like it came out in 1996 for $1000 lol
[18:35:07 CET] <Compn> and then realplayer bought it
[18:39:48 CET] <Compn> when realplayer bought it, they took down all the files
[18:40:15 CET] <Compn> so its incredibly difficult to find
[18:40:36 CET] <Compn> http://web.archive.org/web/19971210163337/http://www.iterated.com/nettech/clearvideo/decoder.htm
[18:45:36 CET] <Compn> unfortunately it looks like the encoder demo you had to buy it first, http://web.archive.org/web/19970702025554/http://www.iterated.co.uk/ClearVideoDecoder/
[18:56:32 CET] <Compn> durandal_1707 : ylc decoder was added a while back
[18:56:40 CET] <Compn> to mplayer
[18:57:33 CET] <Compn> durandal_1707 : pixlet decoder committed
[19:00:03 CET] <Compn> argh, ftp.iterated.com is where they hosted files
[19:01:05 CET] <durandal_1707> Compn: you or someone else put wrong pix fmt to ylc
[19:01:11 CET] <Compn> probably me 
[19:01:20 CET] <durandal_1707> its yuyv422
[19:01:20 CET] <Compn> ftp.iterated.com/products/FVP11W32.EXE <~ now to search other ftp for mirror...
[19:01:33 CET] <Compn> thats for fractal viewer, not clearvid encoder btw
[19:02:04 CET] <durandal_1707> i already have decoder working fine
[19:02:18 CET] <Compn> we have ylc sample ?
[19:03:27 CET] <Compn> hmm whats yuyv422 in mplayer ? hmmmm
[19:04:18 CET] <Compn> yuyv422 is "YUY2" in mplayer 
[19:04:43 CET] <Compn> i would like to try it before committing though
[19:05:47 CET] <Compn> http://samples.ffmpeg.org/V-codecs/ylc0.avi
[19:05:49 CET] <Compn> there it is
[19:06:51 CET] <Compn> durandal_1707 : ok, fixed ylc0 decoder
[19:19:48 CET] <Compn> durandal_1707 : its also called clrvidcc.dll
[19:20:50 CET] <durandal_1707> Compn: nope. i have that dll already
[19:22:52 CET] <Compn> orly
[19:22:55 CET] <Compn> what size is it ?
[19:23:55 CET] <Compn> durandal_1707 : is it 229kb? https://www.reasoncoresecurity.com/clrvidcc.dll-e67d07b68896a3fc850f4200ae63b68b7d3bbb4f.aspx
[19:38:12 CET] <Compn> durandal_1707 : where did you get clrvidcc.dll ?
[19:38:46 CET] <durandal_1707> Compn: its on samples.ffmpeg.org
[19:39:19 CET] <durandal_1707> i just need wrapper that calls functions from that dll
[19:39:48 CET] <durandal_1707> and wrapper dll is clrvidcd.dll
[19:39:56 CET] <Compn> ah
[19:40:41 CET] <durandal_1707> it was part of realpublisher 5.0
[19:49:28 CET] <Compn> the encoder was ?
[19:50:53 CET] <Compn> durandal_1707 : can probably find this url on p2p
[19:50:55 CET] <Compn> hehe
[19:51:00 CET] <Compn> er s/url/dll
[19:51:42 CET] <durandal_1707> clearvideo
[19:52:01 CET] <durandal_1707> emojix.com
[19:52:50 CET] <Compn> ?
[20:02:11 CET] <durandal_1707> Compn: they made 1.5 version of encoder
[20:02:57 CET] <durandal_1707> its emajix
[20:03:03 CET] <Compn> oh
[20:03:40 CET] <durandal_1707> its in stored in file extradata
[20:04:54 CET] <Compn> ohhhh
[20:05:03 CET] <Compn> i found that exsnet bought clearvideo rights in 2000
[20:05:10 CET] <Compn> so emajix was exsnet 
[20:05:12 CET] <Compn> making more sense now
[20:09:02 CET] <Compn> durandal_1707 : http://web.archive.org/web/20020207093605/http://www.enxnet.com/decoder/CVdecoder.exe
[20:09:07 CET] <Compn> hue hue hue
[20:09:31 CET] <Compn> decoder only though :\
[20:19:00 CET] <Compn> durandal_1707 : its installed with soundforge 4.5
[20:19:05 CET] <Compn> durandal_1707 : http://siva.bgk.uni-obuda.hu/setup/_Audio/SoundForge%20v4,5%20and%20PlugIns/
[20:19:39 CET] <Compn> F:\setupfrg>strings clrvidcc.dll | grep -i vid
[20:19:39 CET] <Compn> ClearVideo compressor core DLL
[20:21:15 CET] <durandal_1707> lol
[20:21:27 CET] <Compn> is that it ?
[20:22:06 CET] <durandal_1707> only if there is cd.dll
[20:30:26 CET] <durandal_1707> Compn: how you found such dll in audio app, i cant, where it is?
[20:34:05 CET] <Compn> in the installer file
[20:34:46 CET] <Compn> one sec
[20:38:08 CET] <Compn> http://ffmpeg.org/~compn/clrvidcc.dll
[20:38:29 CET] <Compn> in the meta it says 'encoder'
[20:38:34 CET] <Compn> could be wrong tho
[20:42:03 CET] <durandal_1707> Compn: does it says also Core?
[20:42:31 CET] <Compn> "ClearVideo Encoder DLL"
[20:42:49 CET] <Compn> strings says "ClearVideo compressor core DLL"
[20:43:15 CET] <Compn> i cant get it to load as vfw though
[20:45:26 CET] <Compn> F:\setupfrg1>strings clrvidcc.dll | grep -i clear
[20:45:26 CET] <Compn> ClearVideo compressor core DLL
[20:45:26 CET] <Compn> F:\setupfrg1>strings clrviddc.dll | grep -i clear
[20:45:26 CET] <Compn> ClearVideo decompressor core DLL
[20:45:32 CET] <Compn> :P
[20:46:12 CET] <Compn> i think you are right , we need clrvidcd.dll
[20:59:00 CET] <Compn> durandal_1707 : to answer your original question ,the fourcc list is just what i collected from the internet, i never had clrvidcd.dll , i copied that from the avi overview page :\
[21:00:29 CET] <durandal_1707> ah
[21:01:36 CET] <Compn> ah im an idiot, i already put clrvidcc.dll here https://samples.ffmpeg.org/drivers32/UCOD-ClearVideo/ucod2/
[21:02:38 CET] <Compn> https://samples.ffmpeg.org/drivers32/UCOD-ClearVideo/ClearVideo/ mentions the 1.5 version
[21:11:27 CET] <cone-020> ffmpeg 03Steven Liu 07master:d9c2cfd31675: avcodec/bsf: fix resource leak in av_bsf_list_parse_str
[21:36:33 CET] <cone-020> ffmpeg 03Steven Liu 07master:2deafd989888: avformat/hlsenc: fix hls start and tail segment duration  problem
[21:39:22 CET] <cone-020> ffmpeg 03Steven Liu 07master:daff04bd1837: avformat/hlsenc: fix the bug when the largest segment duration pointer right value is 0
[22:51:56 CET] <Compn> durandal_1707 : probably easier to track down and contact the people who made the clear video codec, Michael Barnsley and Alan Sloan.
[22:52:59 CET] <durandal_1707> are they still alive?
[00:00:00 CET] --- Wed Jan 11 2017


More information about the Ffmpeg-devel-irc mailing list