burek021 at gmail.com
Wed May 3 03:05:01 EEST 2017
[00:02:09 CEST] <b0bby__> Hey
[00:02:19 CEST] <durandal_1707> hey
[00:02:25 CEST] <james999> yeah most info on the inet is wrong
[00:02:37 CEST] <james999> that's why i wanted to add my commandline syntax for wifi streaming with ffmpeg to the documentation
[00:02:47 CEST] <james999> but i'm not sure if that's a patch or what the procedure is
[00:25:59 CEST] <b0bby__> does anyone know whats wrong https://pastebin.com/g3duvTF1
[00:26:04 CEST] <b0bby__> ?\
[00:34:18 CEST] <durandal_1707> b0bby__: see working examples in repo
[00:34:57 CEST] <b0bby__> link?
[00:38:26 CEST] <james999> b0bby__: The error means there is a missing ; before the line it references
[00:38:40 CEST] <james999> So you'll have to check each source file line by line to discover the problem
[00:38:58 CEST] <BtbN> the source file is in an official ffmpeg header
[00:39:03 CEST] <b0bby__> Shouldn't the source file have working code
[00:39:21 CEST] <BtbN> unless you messed with them, they are fine
[00:39:30 CEST] <b0bby__> I mean the headers should be correct
[00:39:52 CEST] <james999> idk where did you get the code from?
[00:40:28 CEST] <james999> hmm well only other possibility than what BtbN is that gcc is messed up or something
[00:40:42 CEST] <BtbN> Or you just can't build something external in-tree
[00:40:57 CEST] <b0bby__> I made it. It just loads the avs
[00:41:02 CEST] <b0bby__> its a basic test
[00:44:46 CEST] <james999> well I could copy the code you have from the same source and try it on my ubuntu vm
[00:45:15 CEST] <james999> but I'm afraid halfway through furq or someone will just say it's the wrong code and the solution is trivial
[00:49:16 CEST] <furq> that should work finr
[00:49:16 CEST] <furq> e
[00:58:49 CEST] <cryptodechange> My encode output is getting q=21.0 q=-1.0
[00:59:06 CEST] <cryptodechange> How do I interpret those numbers?
[01:00:57 CEST] <kepstin> cryptodechange: using x264? ignore them, they're practically meaningless
[01:02:14 CEST] <kepstin> I think the first number will be "the qp of the current frame", which isn't really a useful measurement of anything since the x264 encoder uses different qp values for different parts of the frame.
[01:02:45 CEST] <kepstin> and the second number... ? who knows.
[01:03:23 CEST] <kepstin> (I think those numbers might mean something when using some ffmpeg internal encoders, like the mpeg1/2 video encoder)
[01:04:15 CEST] <furq> yeah it's meaningless unless you're using cqp
[01:04:24 CEST] <furq> and if you are using cqp then it's not a number you need to pay attention to
[09:45:15 CEST] <JC_Yang> 3.3 released failed to build with MSVC? so many WSA* family functions are said to redefined? anything wrong?
[10:47:01 CEST] <JC_Yang> well, --extra-cflags=-DWIN32_LEAN_AND_MEAN fix it.
[10:50:13 CEST] <JEEB> JC_Yang: I think we have MSVC FATE instances
[10:50:23 CEST] <JEEB> so that kind of failure should have been found out
[10:50:40 CEST] <JEEB> unless it's specific to some configuration
[10:50:57 CEST] <JEEB> yea, we have active 2013, 2015 and multiple 2017 configurations
[10:51:01 CEST] <JEEB> and they keep building
[10:51:07 CEST] <JEEB> http://fate.ffmpeg.org/
[10:51:11 CEST] <JEEB> ctrl+F "microsoft"
[10:51:35 CEST] <JEEB> I think we had some release branch FATE machines as well, but this interface doesn't show them (and the last I remember the fatebeta one was down)
[10:51:37 CEST] <JC_Yang> I'm using msys2, and haven't run vcvars.bat, I manually set the PATH
[10:52:14 CEST] <JEEB> msys2 is pretty much common, but generally people add one of the vcvars things into their msys2 execution environment
[10:52:23 CEST] <JEEB> you can check how the FATE instances are configured
[10:52:28 CEST] <JEEB> as in, their configure lines
[10:53:10 CEST] <crow> c_14 I compiled now ffmpeg 3.2.4 and build VDR plugin softhddevice against this and it does not crash, but with ffmpeg 3.3 it does. vdr-softhdevice plugin compiled with ffmpeg 3.2.4 can play normal Live TV and would not crash on this specified channel, but compiled against ffmpeg 3.3 it crashes on this specified channel
[10:53:14 CEST] <JC_Yang> the default msvc configure still use .o and .a *nix style output extensions? it is so here in my local build
[10:54:26 CEST] <thebombzen> JC_Yang: afaik MSVC uses .lib instead of .a and .obj instead of .o
[10:55:08 CEST] <thebombzen> or rather .lib for static libraries and .o for PE object files, but I don't think static libs are AR archives of objects
[10:55:15 CEST] <JC_Yang> so something wrong with my ./configure.
[10:55:28 CEST] <JEEB> I wouldn't be surprised if our makefiles use common output file names :P
[10:55:39 CEST] <thebombzen> msys/gcc still uses .o and .a though
[10:56:39 CEST] <JC_Yang> isn't configure for msvc default the extensions to the windows conventions?
[10:57:25 CEST] <JEEB> does it really matter what the extension is? as long as you know you've configured it to be built with cl and linked with link that's all that matters
[10:57:35 CEST] <JEEB> and you get your pdb files if you haven't disabled debug symbol creation
[11:03:32 CEST] <JC_Yang> just curious, since I doubt whether my configure has something wrong. and one more question, no ffplay build in the default configure for msvc?
[11:04:05 CEST] <thebombzen> ffplay requires sdl. do you have that?
[11:05:43 CEST] <JEEB> JC_Yang: let me see my magic 8-ball - if your configure log says you are using MSVC then it sounds like your configure line in that sense is OK
[11:06:03 CEST] <JEEB> of course my magic 8-ball cannot tell me if your configuration was sane or not otherwise
[11:08:17 CEST] <JC_Yang> no, no sdl here. thanks. JEEB, the configure says C compiler cl, and the build complete smoothly, I guess its properly done already
[11:20:54 CEST] <crow> I have here an problem application crash/coredump with vdr-softhddevice output plugin while watching Live TV on VDR. vdr-softhddevice plugin compiled agains ffmpeg 3.3 crush log url: https://defuse.ca/b/ivryaMqp pass: media ; vdr-softhddevice plugin compiled agains ffmpeg 3.2.4 no crash just log url: https://defuse.ca/b/Ape5u2QY pass: media
[11:51:00 CEST] <JC_Yang> it seems that the default configure really lack -DWIN32_LEAN_AND_MEAN, without it, my local build failed.
[11:52:16 CEST] <JC_Yang> 2017 community
[11:57:22 CEST] <BtbN> Never had to hack that in there anywhere to build on windows.
[11:57:44 CEST] <JEEB> note: he said he isn't running vcars or similar
[11:57:49 CEST] <JEEB> so he might be missing default env vars
[11:57:50 CEST] <JEEB> etc
[11:59:53 CEST] <JC_Yang> have already invoke msys2_shell within VS command prompt, and the msys2 shell do know cl, invoke cl in the shell is ok
[12:00:39 CEST] <BtbN> Is this some horribly old MSVC version?
[12:00:56 CEST] <JC_Yang> visual studio 2017 is the latest
[12:01:14 CEST] <BtbN> Works without adding any kind of define for me.
[12:01:36 CEST] <JC_Yang> maybe some requirement not matched on my side?
[12:02:16 CEST] <BtbN> From how I understand WIN32_LEAN_AND_MEAN, defining it only _disables_ stuff
[12:22:34 CEST] <c_14> crow: ok, with this backtrace it looks like ffmpeg is triggering an assertion (which it shouldn't) and the assertion is causing a malloc corruption. The malloc corruption probably isn't a bug in ffmpeg (but in your libc), but the assertion triggering seems like a probable bug in ffmpeg. open a bug report on trac.ffmpeg.org and paste the stuff from the ivryaMqp paste there
[12:23:36 CEST] <c_14> also, can you reproduce the bug on another machine or without using the vdr plugin? (using vdpau acceleration on a static video file or something)
[12:32:44 CEST] <crow> c_14 i could try to record this channel with vdr-softhddevice compiled against 3.2.4 and then play to play with ffmpeg directly. thank you about suggestion ill open bug report. and provide as much information as posible. thank you
[12:36:24 CEST] <crow> i am using archlinux arm and here is the glibc 2.25 in use.
[12:38:35 CEST] <crow> hmm i am not able to register in trac, after spam check i get this error message: No handler matched request to /register_ffmpeg_four
[12:41:10 CEST] <crow> it works now duck did not read second antispam question properly :D
[12:53:56 CEST] <crow> c_14 https://trac.ffmpeg.org/ticket/6364
[12:57:37 CEST] <crow> to reproduce the bug on another machine i need to be at home, so tonight.
[14:06:33 CEST] <termos> my AVCodecContext.time_base is 0/2 on my input decoding context, after creating it with avcodec_alloc_context3 (not using the one from the deprecated stream->codec)
[14:06:54 CEST] <termos> is there something I'm not setting properly? just following the doc/examples/transcoding.c
[14:11:05 CEST] <termos> hm, I might have a race condition
[14:17:00 CEST] <cryptodechange> Fort older movies (<2000s) that have been released as bluray, is it preferred to use grain or film tuning?
[14:33:18 CEST] <BtbN> There is never a definite answer to that
[14:33:29 CEST] <BtbN> Use what yields the best result for that specific material
[14:59:43 CEST] <aruns> Hi, have ffmpeg installed on Mac OS X 10.11.6, wondering what the -c flag is for (have checked man page and doesn't actually say) when you pass it as for e.g. -c:v
[15:00:13 CEST] <aruns> Running version 3.2.4 of ffmpeg
[15:00:58 CEST] <aruns> I am guessing the -c flag is for codec?
[15:01:15 CEST] <aruns> Oh I got it now.
[15:01:20 CEST] <aruns> Nvm.
[16:16:56 CEST] <ZippyTheChicken> hi.. I am using MCEBuddy and FFMPEG is eating up all my memory is there a reason I can findout why and fix it? I am not sure which distribution he is using but it should be only 1 or 2 years old at the most
[16:18:40 CEST] <JEEB> we don't offer support for third party software utilizing FFmpeg. if you can specify that it is an issue in FFmpeg itself rather how that application is using the libraries/cli then it goes to our area
[16:19:26 CEST] <JEEB> (as in, we don't offer support for users of third party software, if you're trying to develop a piece of software on the libraries etc it is possible to get help here if someone is around and wants to help)
[16:21:37 CEST] <ZippyTheChicken> I see... I was just wondering if I could replace his version of ffmpeg I don't know how i can find that out from his ffmpeg.exe file
[16:22:13 CEST] <ZippyTheChicken> do you distribute the source code of FFMPeg for windows? and then Authors modify it? or do you distribute the exe and dlls?
[16:22:27 CEST] <JEEB> we distribute the source code and people build it
[16:22:33 CEST] <ZippyTheChicken> ah
[16:22:39 CEST] <JEEB> the source code is common between architectures and OSs
[16:22:52 CEST] <ZippyTheChicken> c++ ?
[16:23:02 CEST] <JEEB> C, hand-written assembler and some C++
[16:23:15 CEST] <JEEB> C is generally frowned upon due to the exception system leaking etc
[16:23:19 CEST] <JEEB> *C++
[16:23:37 CEST] <JEEB> but some dependencies are C++ libraries which is why the wrappers around them are C++
[16:24:30 CEST] <ZippyTheChicken> I guess it would be impossible for me to know what he is doing then... maybe i will see if I can find some revision notes on his code and see if I can find a way to replace his version
[16:24:58 CEST] <ZippyTheChicken> thank you very much .. I know where to start now
[16:25:07 CEST] <JEEB> replacing the version can be hard if you're doing an upgrade as well (unless it's the command line version he's using an not the libraries)
[16:25:34 CEST] <JEEB> because the API could have changed between versions
[16:26:21 CEST] <ZippyTheChicken> this is what he is including in his FFMPEG Folder .... ffmpeg.dvrms.exe ffmpeg.exe ffprobe.exe libx264-56.dll pthreadGC2.dll
[16:26:40 CEST] <ZippyTheChicken> yes .. also he could be insane and editing your source... :o)
[16:28:59 CEST] <JEEB> wow
[16:29:02 CEST] <JEEB> that sounds ancient
[16:29:26 CEST] <JEEB> at least by the libx264 version
[16:29:28 CEST] <JEEB> (56)
[16:29:34 CEST] <JEEB> I think we're at over 140 now
[16:29:40 CEST] <JEEB> (that's the API version)
[16:29:55 CEST] <JEEB> ZippyTheChicken: anyways, that sounds like it's using the command line app :P
[16:30:12 CEST] <JEEB> which is not perfect for an application but at least the interfaces are generally stable as of recent
[16:30:19 CEST] <JEEB> but yea, nfi
[16:30:31 CEST] <JEEB> (this is why you can't really give support to users of 3rd party apps)
[17:05:07 CEST] <JC_Yang> invoked msys2 within visual studio dev command prompt, then run ./configure --toolchain=msvc, configure success, and then make still result in the build failed, so many winsock2 redefinition. any steps wrong? ffmpeg 3.3 here.
[17:06:10 CEST] <klaxa> did you run make clean between builds? although ./configure should kinda take care of that but not sure?
[17:06:37 CEST] <klaxa> i think i had build fails with old binaries still in place
[17:06:42 CEST] <JC_Yang> pure fresh build dir here, empty one
[17:08:07 CEST] <klaxa> hmm
[17:08:34 CEST] <klaxa> i'm not a windows guy, but having the complete output on pastebin or something would definitely help understand what's going on
[17:08:35 CEST] <JC_Yang> I invoked msys2_shell -msys2 -user-full-path, this is the way I get the vcvars envs set in msys2
[17:10:34 CEST] <JC_Yang> I've fixed this problem hours ago with -DWIN32_LEAN_AND_MEAN but I'd like to know what's wrong and why require such a fix
[17:27:09 CEST] <SouLShocK> is it possible to resize (upscale) and calculate psnr in one execution? I have a bunch of files which are downsized from source material and I'd like to calculate the psnr value for each file, compared to the source
[18:02:12 CEST] <kepstin> SouLShocK: sure, use an ffmpeg command with two inputs, and use -filter_complex to either upscale the small video or downscale the big stream before doing the psnr filter
[18:02:44 CEST] <kepstin> (I don't know whether upscaling/downscaling is better - the psnr would probably be highly dependent on the algorithm used if upscaling)
[18:06:36 CEST] <cryptodechange> Anyone familiar with mkvmerge? Trying to grab the command paramters from the GUI
[19:08:40 CEST] <SouLShocK> kepstin: https://pastebin.com/w3SRUAgs
[19:14:49 CEST] <SouLShocK> kepstin: sorry here is the full paste https://pastebin.com/8E9QPJe5
[20:23:56 CEST] <arvut> how long do you reckon it will take to process a dvd (4.0G) of VIDEO_TS through x264 to file on a old core2 (1.8Ghz)?
[20:24:24 CEST] <kerio> TIAS
[20:24:33 CEST] <kerio> also, process how
[20:25:01 CEST] <arvut> kerio: was that for me?
[20:25:14 CEST] <kerio> yes
[20:25:23 CEST] <arvut> what does TIAS mean?
[20:25:28 CEST] <kerio> try it and see
[20:25:42 CEST] <arvut> oh right, ffmpeg has commandline in it
[20:25:55 CEST] <arvut> how do you go back from the hex to regular interface?
[20:27:39 CEST] <arvut> should I move the .iso's I created from the dvd's back home to my i7 4790K and do it on that instead? (in case this takes too many hours)
[20:27:54 CEST] <arvut> I can scp the files
[20:28:06 CEST] <kerio> what is "the hex"
[20:28:15 CEST] <kerio> what are you attempting to do
[20:28:49 CEST] <arvut> at the moment lots of hex scrolling down, I can switch between raw hexdump and a more colourful output
[20:29:07 CEST] <arvut> and the cpu is on heavy load, so its processing the first dvd
[20:29:46 CEST] <kepstin> arvut: the i7 is probably gonna be 8-12x faster than the core2
[20:30:00 CEST] <furq> that seems optimistic
[20:30:22 CEST] <arvut> kepstin: nice, I'll do the next one on that, this one has ran for two hours now, load is 3.3 on two cores
[20:30:24 CEST] <kepstin> assuming a dual-core core2 quip
[20:30:29 CEST] <kepstin> less if you have a quad i guess
[20:32:14 CEST] <kerio> you still haven't told us what you plan on doing with the dvd :<
[21:42:45 CEST] <olinux> having trouble converting mp3 + still image to mp4 .. can watch locally just fine .. can upload to wistia player and it wistia player plays in chrome fine, but doesn't play well in IE/Firefox/Opera browsers
[21:43:59 CEST] <olinux> ffmpeg -framerate 1 -loop 1 -i myfile.png -i myfile.mp3 -c:a copy -c:v libx264 -shortest out.mp4
[21:55:54 CEST] <kepstin> firefox might have issues with mp3 audio? you could try aac instead
[21:56:36 CEST] <JEEB> I would actually check what colorspace got used for video
[21:56:51 CEST] <JEEB> olinux: post your full command line and the terminal output in a single pastebin or similar. then link here.
[21:56:55 CEST] <JEEB> that should tell us :P
[21:58:16 CEST] <olinux> kepstin, JEEB, thanks .. trying the aac encode now
[22:05:23 CEST] <olinux> kepstin, works! thank you
[22:45:30 CEST] <arvut> ok, I have encountered more problems, I transfered the 2nd .iso of a dvd movie (converted from vhs to dvd and then ripped to hdd)
[22:45:49 CEST] <arvut> transfered to my own machine from the laptop with the core2, but now i can't read it when its mounted
[22:46:19 CEST] <arvut> nobody is owner, nobody is group and date is 2006 (presumably when vhs was created from super8)
[22:46:31 CEST] <arvut> permission denied, is the iso damaged?
[22:46:57 CEST] <arvut> s/iso/image/
[22:47:41 CEST] <arvut> I had to mount the first image, then copy contents as root to hdd and change ownership of files, before I could access it with ffmpeg
[22:48:40 CEST] <kerio> can't you just ffmpeg -i foo.iso
[22:56:25 CEST] <arvut> I'll give it a try
[22:57:08 CEST] <arvut> invalid data found when processing input
[22:57:16 CEST] <arvut> it seems to be messed up, dd exited with errors on this one
[22:57:24 CEST] <arvut> I think we have to make a new dvd to rip from
[22:58:57 CEST] <arvut> kerio: have a look at this http://imgur.com/a/A0yY2
[22:59:24 CEST] <kerio> i guess it doesn't work :<
[22:59:57 CEST] <arvut> I'm gonna compare the sum to the original rip, but it might not actually mount on hostmachine either so I think the dvd I ripped it from is scratched or something
[23:00:25 CEST] <arvut> that ??????????? ? ? ? ? stuff is messed up when I try to ls -l /mnt/cdrom
[23:00:39 CEST] <arvut> permissions, and blinking dir
[23:20:42 CEST] <marcurling> Hello, do I have to compile my own ffmpeg for 10-bit encoding?
[23:21:49 CEST] <JEEB> depends on the encoder and what sort of binaries you're using
[23:22:34 CEST] <marcurling> I'm on W7 so I'm using Zeranoe's with default (h264)
[23:23:27 CEST] <JEEB> marcurling: if it's linked to an 8bit libx264 then yes
[23:24:19 CEST] <marcurling> I'm using the shared version so all I got to do is finding the 10-bit.dll ?
[23:26:02 CEST] <JEEB> marcurling: uhh, not sure it's that simple.
[23:26:49 CEST] <marcurling> ok, will google more. Thanks JEBB
[23:39:33 CEST] <Sashmo> ifconfig
[23:39:37 CEST] <Sashmo> oops
[23:46:06 CEST] <arvut> marcurling: try ffmpeg on gentoo, very fast
[00:00:00 CEST] --- Wed May 3 2017
More information about the Ffmpeg-devel-irc