[Ffmpeg-devel-irc] ffmpeg.log.20150505

[burek]

[09:08:14 CEST] <termos> Mavrik: okey I see, I guess I need to read the x264 source to figure out how those numbes are used there. I'm only setting them on my encoding contexts
[09:46:09 CEST] <sp00x> can someone verify if there is actually 2 or 3 prores encoders in ffmpeg? is -codec:v prores just an alias for -codec:v prores_aw, and the other has to be selected using -codec:v prores_ks ?
[10:16:29 CEST] <relaxed> sp00x: I think you're right, look at "ffmpeg -h encoder=prores_ks" and "ffmpeg -h encoder=prores_aw"
[10:24:39 CEST] <sp00x> relaxed: aha, didn't know that command.. but yes, looks that way compared to ffmpeg -h encoder=prores
[14:41:37 CEST] <waressearcher2> hi girls
[14:47:05 CEST] <waressearcher2> I got fresh source code ffmpeg-snapshot.tar.bz2 and tryed to compile it with parameters: "--enable-libvidstab --enable-gpl" but it says:  "ERROR: vidstab not found using pkg-config", is vidstab internal library for ffmpeg or should I install some 3rd party library for it to compile ?
[14:53:58 CEST] <durandal_1707> Vidstab is 3d part library
[14:54:12 CEST] <waressearcher2> durandal_1707: where to get it ?
[14:56:02 CEST] <durandal_1707> Google it
[14:56:08 CEST] <waressearcher2> https://github.com/georgmartius/vid.stab/archive/master.zip correct ?
[14:56:37 CEST] <waressearcher2> durandal_1707: y u so lazy u can't google it for me ?
[15:03:11 CEST] <Anoia> hi all
[16:10:41 CEST] <PSvils> hey all!
[16:10:51 CEST] <PSvils> has anyone had any success with compiling ffmpeg for ios with clang?
[16:11:12 CEST] <PSvils> I set up the -cc and -cxx to point to clang and clang++ respectively, yet in configure, it seems to be trying to use gcc.
[16:11:38 CEST] <PSvils> (error message: gcc is unable to create an executable file. If gcc is a cross-compiler, use the --enable-cross-compile option.)
[16:11:56 CEST] <PSvils> (I'm also passing --enable-cross-compile to configure...)
[16:25:03 CEST] <c_14> PSvils: can you pastebin your config.log ? You might also need to set --dep-cc
[16:25:42 CEST] <PSvils> I'm getting somewhere, must have been bad bash scripting, but now it's not finding the gnu gas-preprocessor.
[16:25:48 CEST] <PSvils> I'm not sure what this is.
[16:27:57 CEST] <c_14> PSvils: try setting --as to an assembler
[16:28:17 CEST] <PSvils> I tried to set it to just "as", and configure got stuck.
[16:28:23 CEST] <PSvils> (ctrl+c'd it)
[16:28:41 CEST] <PSvils> I'll try getting the gas-preprocessor manually + adding to path
[16:29:03 CEST] <PSvils> (github.com/yuvi/gas-preprocessor)
[16:29:07 CEST] <PSvils> is this correct c_14?
[16:30:15 CEST] <c_14> it looks like it might work, it's not actually GAS though
[16:32:42 CEST] <PSvils> hmm, "Error running preprocessor"
[16:33:01 CEST] <PSvils> I have it in my usr/bin so it's in path, not sure how it gets run.
[16:33:10 CEST] <PSvils> do I need some sort of interpreter for it?
[16:33:57 CEST] <c_14> For that one you'll need perl
[16:41:20 CEST] <PSvils> c_14: interesting, installed perl and confirmed working from terminal, yet still doesn't find, will hunt the issue.
[16:41:38 CEST] <PSvils> c_14: are there other as I can use on Mac for ffmpeg/
[16:41:38 CEST] <PSvils> ?
[16:44:54 CEST] <Tincho> hi! I have sent an email to the ML (https://ffmpeg.org/pipermail/libav-user/2015-May/008120.html), but want to try here too in case I have more luck.. I am being unable to reproduce RTMP streams, avformat_open_input gets stuck, and I really don't know what to try
[16:45:06 CEST] <Tincho> any tips would be very appreciated
[16:47:31 CEST] <PSvils> c_14: regarding just "as" getting stuck in configuration, is this known with "as", or should I be hunting in that area
[16:47:32 CEST] <PSvils> ?
[16:49:10 CEST] <c_14> PSvils: I don't know. You can try looking at the 2 guides for compiling FFmpeg for OSX https://trac.ffmpeg.org/wiki/CompilationGuide
[17:05:55 CEST] <PSvils> c_14: I'm not sure I understand the process - I specified --as=yasm, and it still complains about not finding gas-preprocessor in configure.
[17:06:52 CEST] <PSvils> in the first steps of configuring, it outputs: "Configured with: --prefix=/Applications/Xcode.app/Contents/Developer/usr --with-gxx-include-dir=/usr/include/c++/4.2.1"
[17:07:02 CEST] <PSvils> I never set anything like that myself - do you know where this is coming from?
[17:07:09 CEST] <PSvils> btw - super thanks for helping out and having patience :)
[17:07:19 CEST] <PSvils> I'm not exactly experienced with this kind of stuff :)
[17:37:33 CEST] <c_14> Can you pastebin your config.log? Might be something interesting in there.
[17:38:45 CEST] <PSvils> c_14: www.pastebin.com/pdW2qLKn
[17:40:45 CEST] <PSvils> I can see at the end that for some reason yasm is getting passed -arch when it should simply be -a, but not sure where that's coming from.
[17:41:16 CEST] <PSvils> or rather, I figured that the compiler etc. will need -arch, so not sure how I can change which flags get passed just to yasm........or I'm just not understanding anything of this process :)
[17:41:31 CEST] <c_14> That's from the cflags
[17:41:41 CEST] <PSvils> aha, alright
[17:43:07 CEST] <PSvils> if I change the cflags to work for asm, then clang complains "no such file or directory "armv7"
[17:44:51 CEST] <c_14> Try getting rid of everything in the --extra-cflags and --extra-cxxflags except for the -I part
[17:45:27 CEST] <PSvils> including -isysroot?
[17:46:21 CEST] <PSvils> okay
[17:46:33 CEST] <PSvils> clang now complains about no such file or directory for an include path
[17:47:10 CEST] <c_14> What's the current content of --extra-cflags
[17:47:12 CEST] <c_14> ?
[17:47:12 CEST] <PSvils> spoke too quick, my error]
[17:47:50 CEST] <PSvils> okay, now yasm complains "unrecognized object format PIC" - do I need to enable-pic? (I have it right now)
[17:48:59 CEST] <c_14> If you have PIC enabled, try disabling it and vice versa.
[17:50:06 CEST] <PSvils> disabled, now 'dynamic-no-pic' is not a valid machien for architecture x86
[17:50:56 CEST] <PSvils> I should be trying for armv7 I think.
[17:52:00 CEST] <c_14> Is this an error from yasm? clang? the linker?
[18:08:54 CEST] <PSvils> c_14: clang
[18:09:10 CEST] <PSvils> oh damn. yasm
[18:09:13 CEST] <PSvils> it's from yasm.
[18:10:20 CEST] <RaQu> Hello. I'm trying to trim a .flac file through the command prompt in Windows and I can't succeed. I'm using ffmpeg version N-71899-g6ef3426, from 20150505. The original file is 78.7 seconds long, I'm trying to cut out the first 5 seconds and the last 5 seconds. The resulting file reads that it lasts 78.7 seconds using ffmpeg -i command, instead of 68.7 seconds; but the first 5 seconds are actually gone, so when I play it it lasts 73.7 
[18:10:26 CEST] <RaQu> M:\Music\ffmpeg\bin\ffmpeg.exe -ss 00:00:05.00 -i M:\Music\AirForGString.flac -t 00:01:08.70 -acodec copy -y TrimmedFile.flac
[18:11:17 CEST] <RaQu> I was asking if I'm doing something wrong before pasting the command. Thanks a lot.
[18:14:23 CEST] <PSvils> c_14: the thing is is that I had ffmpeg building on linux for android, and from what I learned there...I don't really understand what's going on here :)
[18:16:16 CEST] <c_14> PSvils: I have no idea. You might want to ask on the user mailing list. Maybe someone else has done it before and has some pointers.
[18:16:37 CEST] <c_14> RaQu: that can happen when you copy codecs, since you're using flac you can just reencode without quality loss and it should be correct.
[18:16:52 CEST] <PSvils> Alright, gotcha. Thanks a bunch for the help eh c_14 :) much appreciated, I'll continue hacking away.
[18:18:36 CEST] <RaQu> I see. If I remove -acodec copy it actually works. Thanks a lot.
[18:24:11 CEST] <LZZ> I need help with the scale filter syntax... I want to make a thumbnail with a max size of 224x224, but keeping the aspect ratio.
[18:24:33 CEST] <LZZ> So if w > h; scale"224:-1"
[18:24:52 CEST] <LZZ> else, scale="-1:224"
[18:31:08 CEST] <LZZ> I found the vars and all, but I don't know the "functions" to make the comparison, like gt()
[18:31:16 CEST] <LZZ> If someone could point me to that documentation...
[18:32:24 CEST] <Mavrik> LZZ, how about using scale=w=224:h=224:force_original_aspect_ratio=decrease ?
[18:33:39 CEST] <LZZ> Yep, thank you Mavrik
[18:34:14 CEST] <PSvils> c_14: for future reference - to build on osx for ARM, it seems that one needs: https://github.com/ffmpeg/gas-preprocessor
[18:34:25 CEST] <PSvils> the yuvi one is outdated.
[18:36:22 CEST] <c_14> PSvils: good to know
[18:36:36 CEST] <PSvils> :)
[20:57:33 CEST] <kyleogrg> hi
[20:59:04 CEST] <kyleogrg> I've downloaded a YouTube video.  It's 29.97p, but when I step through it frame by frame in VLC, I see that roughly every other frame is blended.  Does this mean that the original video was maybe a different frame rate?
[21:13:54 CEST] <ChocolateArmpits> kyleogrg: Are you sure the original wasn't interlaced ?
[21:14:16 CEST] <ChocolateArmpits> kyleogrg: Youtube does automatic linear-blend deinterlacing on interlaced material
[21:14:39 CEST] <ChocolateArmpits> kyleogrg: so if two fields are on the same frame then you should see some frame blending
[21:41:58 CEST] <dericed> I'm using -movflags write_colr to add the color atom, but it reports as bt.709. How can I set this to bt.601 without changing the pixel values?
[21:49:58 CEST] <RaQu> Hello. I'm trying to trim a .ape file through the command prompt in Windows. I'm using ffmpeg version N-71899-g6ef3426, from 20150505. Here is my command:
[21:50:02 CEST] <RaQu> M:\Music\ffmpeg\bin\ffmpeg.exe -ss -i M:\Music\File.ape 00:00:40.00 -t 304.200 -y Test.ape
[21:50:06 CEST] <RaQu> The file Test.ape is not created at all and I get the following error message: "[NULL @ 00000000003ad080] Unable to find a suitable output format for 'Test.ape'". Does ffmpeg not encode in ape, or I'm doing something wrong? Thanks a lot.
[21:51:18 CEST] <ChocolateArmpits> RaQu: you need to put  "00:00:40:00" right after "-ss"
[21:51:55 CEST] <ChocolateArmpits> RaQu: the basic formatting is "-command [value]"
[21:52:05 CEST] <ChocolateArmpits> or rather "-parameter [value]"
[21:52:36 CEST] <RaQu> Sorry, I did put 00:00:40:00 right after -ss.
[21:52:41 CEST] <RaQu> I copied the code wrong here.
[21:53:27 CEST] <RaQu> Let me try again just in case.
[21:53:57 CEST] <ChocolateArmpits> RaQu: Who's compile are you using ? I checked Zeranoe's and ape is only for decoding but not writing
[21:54:11 CEST] <ChocolateArmpits> I mean demuxing
[21:54:38 CEST] <ChocolateArmpits> try "ffmpeg -format"
[21:54:43 CEST] <ChocolateArmpits> and then scroll up to ape
[21:54:55 CEST] <ChocolateArmpits> and see if it has letter E on the left side
[21:55:33 CEST] <RaQu> I used ffmpeg -format, it says Unrecognized option format
[21:55:35 CEST] <ChocolateArmpits> sorry, meant "ffmpeg -formats"
[21:56:04 CEST] <RaQu> I downloaded Zeranoe's, from here: http://ffmpeg.zeranoe.com/builds/
[21:56:05 CEST] <ChocolateArmpits> It then should give a list of all formats supported for demuxing and writing by the compile
[21:56:06 CEST] <RaQu> The static.
[21:56:18 CEST] <ChocolateArmpits> Check the "ape" entry
[21:56:22 CEST] <RaQu> Ape = D
[21:56:39 CEST] <RaQu> So that means only decoding?
[21:56:43 CEST] <NoTiTo_> morning all
[21:56:57 CEST] <ChocolateArmpits> Yes, if there's only letter D next to it then it's only decoding
[21:57:04 CEST] <ChocolateArmpits> with the compile that you have
[21:57:49 CEST] <RaQu> I see. Thanks a lot for the help.
[22:00:33 CEST] <NoTiTo_> I am trying to select one service / program from an MPEG TS (udp stream) and send it out to youtube but I'm battling with getting the selection right - see ffprobe output here http://pastebin.com/EPxL4k2P
[22:00:39 CEST] <NoTiTo_> any ideas?
[22:02:33 CEST] <NoTiTo_> I have tried various things as suggest on the inter webs with no luck
[22:21:50 CEST] <JackRcast> Hi all
[22:23:18 CEST] <c_14> NoTiTo_: -map p:[program id]:[stream id]
[22:30:27 CEST] <NoTiTo_> @c_14: I tried "ffmpeg -f mpegts -i udp://... -map p:1:33 -vcodec libx264 -strict experimental -map p:1:39 -acodec aac -f flv "rtmp://a.rtmp..." :(
[22:32:39 CEST] <c_14> NoTiTo_: I think it would be p:1:0 p:1:1
[22:32:42 CEST] <c_14> or just p:1
[22:33:15 CEST] <NoTiTo_> c_14: Did you see the ffprobe output on the paste bin?
[22:35:20 CEST] <c_14> yep
[22:43:35 CEST] <NoTiTo_> c_14: Codec type mismatch for mapping #0.0 -> #0.0
[22:44:14 CEST] <NoTiTo_> cool
[22:44:34 CEST] <JackRcast> Hello, is it possible to take an audio mp3 stream coming from an Icecast server (input stream) and reencode it to two new audio streams at different bitrates ?
[22:44:43 CEST] <c_14> yes
[22:45:04 CEST] <JackRcast> can you point me to a wiki page pls ?
[22:45:42 CEST] <c_14> JackRcast: just list two output files in the ffmpeg command and assign a different bitrate for each
[22:46:06 CEST] <c_14> ffmpeg -i http://foobar -c:a baz -b:a wheee out1 -c:a baz -b:a whooo out2
[22:48:27 CEST] <NoTiTo_> c_14: does the "-c:a" work or should it be "-acodec" ?
[22:48:35 CEST] <NoTiTo_> c_14: http://pastebin.com/kYUcPJRY
[22:48:43 CEST] <c_14> -c:a should work for all modern versions of ffmpeg
[22:49:26 CEST] <c_14> NoTiTo_: you shouldn't have to map p:1 twice. The single mapping should map both streams
[22:49:33 CEST] <JackRcast> thanks C_14 I'll check that
[22:49:56 CEST] <NoTiTo_> c_14: "ffmpeg version 0.8.17-4:0.8.17-0ubuntu0.12.04.1" is what I'm using
[22:50:07 CEST] <c_14> oh boy
[22:50:44 CEST] <c_14> That version is among other things ancient and broken.
[22:51:01 CEST] <NoTiTo_> LMAO
[22:51:11 CEST] <c_14> You can use a static build
[22:51:13 CEST] <c_14> http://johnvansickle.com/ffmpeg/
[22:51:24 CEST] <NoTiTo_> thats via the ubuntu package mgr
[22:53:32 CEST] <JackRcast> then can I output these two distinct streams to another remote icecast server as a source client ?
[22:54:45 CEST] <c_14> if your icecast server accepts more than one source, yes
[22:55:43 CEST] <JackRcast> so should I replace out & out 2 with the icecast server URI ?
[22:56:07 CEST] <c_14> yep
[22:56:14 CEST] <proctrap> hey there I'm trying to get the total frames of a mp4 file (DASH)
[22:56:23 CEST] <proctrap> but ffprobe -v error -count_frames -select_streams v:1 -show_entries stream=nb_read_frames -of default=nokey=1:noprint_wrappers=1 /video failed
[22:56:32 CEST] <JackRcast> C_14: can I specify a source password for the icecast server  ?
[22:57:30 CEST] <proctrap> with unrecognized option show_entries
[22:57:32 CEST] <c_14> https://ffmpeg.org/ffmpeg-protocols.html#Icecast
[22:57:48 CEST] <c_14> ^ JackRcast
[22:57:58 CEST] <grepwood> hello everyone
[22:58:31 CEST] <grepwood> does anyone know of a framework for playing things with ffmpeg codecs in sdl?
[22:58:57 CEST] <JackRcast> c_14: perfect, much thanks
[22:58:59 CEST] <grepwood> playing char * rather than FILE * would be preferable, I'm doing lots of caching :(
[22:59:35 CEST] <proctrap> because the examples I've found are using that code & select_streams should work, but doesn't ..
[23:00:08 CEST] <c_14> proctrap: what version are you running?
[23:00:21 CEST] <proctrap> 1.0.10
[23:00:42 CEST] <c_14> That's probably a wee tad bit too ancient.
[23:00:46 CEST] <c_14> Try with a recent version.
[23:01:11 CEST] <proctrap> sec I thought I've installed it from the repos
[23:01:18 CEST] <proctrap> (the site you're mentioning there )
[23:01:52 CEST] <proctrap> http://www.deb-multimedia.org/
[23:04:58 CEST] <c_14> proctrap: the static build has an ffprobe, you can use that
[23:05:00 CEST] <c_14> http://johnvansickle.com/ffmpeg/
[23:05:25 CEST] <proctrap> c_14: ah I've used that before
[23:05:36 CEST] <proctrap> but I need m4a DASH to mp3 conversion also
[23:05:48 CEST] <proctrap> and I *think* I wasn't able to do so last time I've used that..
[23:05:53 CEST] <c_14> There's an ffmpeg binary in there too
[23:06:10 CEST] <c_14> has lame, so that should work
[23:07:03 CEST] <proctrap> c_14: thanks ! I'll first try mocking around with these (if I would've used the same ffmpeg on the dev system ..)
[23:10:09 CEST] <NoTiTo_> c_14: the static libraries look promising!
[23:32:22 CEST] <proctrap> c_14: seems to work now
[23:32:31 CEST] <proctrap> except that it takes really long 1h. video :D
[23:32:48 CEST] <c_14> Well, it has to demux and decode the entire video
[23:32:52 CEST] <proctrap> (I just want the total frames, to calculate the progres.. :D )
[23:33:05 CEST] <proctrap> ofc..
[23:33:57 CEST] <proctrap> meh, what ever, I don't want to use time calculation..
[23:36:47 CEST] <proctrap> c_14: is there any problem if I'm simply extracting the XX fps & multiply by the time ?
[23:37:24 CEST] <c_14> Neither the fps nor the duration are guaranteed accurate especially if the input is vfr
[23:37:31 CEST] <c_14> It _should_ work most of the time though.
[23:37:48 CEST] <c_14> Especially if you only want to gauge progress.
[23:47:18 CEST] <kyleogrg> If anyone is still interested: https://www.youtube.com/watch?v=IO9W6FGvX5M
[23:47:57 CEST] <kyleogrg> This is the video I mentioned earlier, which has the frame blending issue
[23:48:41 CEST] <kyleogrg> It's probably a simple thing
[23:48:50 CEST] <kyleogrg> Roughly every other frame seems to be blurry
[23:51:15 CEST] <kyleogrg> Is it simply a poor frame rate conversion?
[23:52:19 CEST] <proctrap> c_14: doesn't have to be the exactest one :)
[00:00:00 CEST] --- Wed May  6 2015