[Ffmpeg-devel-irc] ffmpeg.log.20190403

burek burek021 at gmail.com
Thu Apr 4 03:05:02 EEST 2019


[00:00:31 CEST] <bugre> faLUCE: i would like to stop ffprobe after a certain amount of time (or frames if that helps). I'm trying to automate some checkings on GOP, bitrate on our livestreams.
[00:08:47 CEST] <fella> -read_intervals
[00:10:09 CEST] <pridkett> let's say if i have IP   24.111.111.100   and let's say  my neighbor who use same ISP as me has IP 24.111.111.105 :   if somebody DDOS my ip 24.111.111.100 ,  obviously i would go offline, but would my neighbor go offline as well?
[00:29:38 CEST] <friendofafriend> pridkett: I think you're in the wrong channel, maybe ##networking instead?
[00:29:48 CEST] <pridkett> okay
[01:19:11 CEST] <bugre> fella: Thank you, you're the man!! '-read_intervals' worked.
[01:19:28 CEST] <bugre> I would never have found this one alone.
[04:56:33 CEST] <nitrxgen> just a message of appreciation for ffmpeg, idk how i would ever manipulate videos as easily as this
[06:37:42 CEST] <while> hi
[06:42:08 CEST] <while> is rfc3533 the rfc implemented by ffmpeg when encoding OGG files?
[08:26:32 CEST] <pridkett> what is the command to put just the audio test.ac3 to  .mp4 container file?
[08:27:13 CEST] <pridkett> what is the command to put just the audio "test.ac3" to  "test.mp4" container file?
[08:43:43 CEST] <JEEB> pridkett: ffmpeg -i test.ac3 -c copy test.mp4
[08:43:57 CEST] <pridkett> thanks
[08:45:40 CEST] <pridkett> wow that caused ffmpeg to crash
[08:45:57 CEST] <pridkett> never seen that befnore
[09:23:17 CEST] <pridkett> i get this error  " add '-strict -2' if you want to use it"    but if i add it  i get this error  "Unable to find a suitable output format for 'strict -2'
[09:23:17 CEST] <pridkett> strict -2: Invalid argument"
[11:07:07 CEST] <azy> how can i overlay an mp3 to an mp4 at a specific time?
[11:07:48 CEST] <tanja77> hi. I got many surveillance videos (about animals im my garden). does anyone know of a software that could "inspect" all these MANY videos and look on differences so that I could "catch" the animal automatically instead of watching all videos?
[11:08:05 CEST] <tanja77> (I am actually watching them via avidemux with more speed, but its a pain anyway)
[11:08:43 CEST] <pridkett> azy what does the mp4 file have
[11:09:10 CEST] <azy> video+audio
[11:09:17 CEST] <pridkett> azy which codec
[11:09:35 CEST] <pridkett> tanja77  what is your obsession with animals on your surveillnace video
[11:10:13 CEST] <azy> ive no idea
[11:10:19 CEST] <pridkett> ffprobe it then
[11:10:44 CEST] <azy> http://paste.debian.net/hidden/6e705203/
[11:11:44 CEST] <pridkett> does mp3 and mp4 have matching timeline?
[11:11:49 CEST] <pridkett> does mp3 and aac have matching timeline?
[11:13:19 CEST] <tanja77> pridkett, no obsession. its because the animals leaves its "rests" on our balkony and we want to know what kind of animals it is
[11:13:38 CEST] <pridkett> rests?
[11:13:45 CEST] <tanja77> for example we would have to treat a dormouse different like a mouse or so, right?
[11:14:04 CEST] <tanja77> had to google this translation:
[11:14:05 CEST] <tanja77> feces
[11:14:08 CEST] <tanja77> :-)
[11:14:11 CEST] <pridkett> i see
[11:14:32 CEST] <pridkett> azy:  Dude\ Asks\ Girl\ To\ The\ Prom\ But\ Showed\ Up\ To\ The\ Wrong\ House\!-video.mp4     that's a weird file name
[11:14:49 CEST] <azy> no @ matching timeline
[11:14:54 CEST] <azy> http://worldstarhiphop.com/videos/video.php?v=wshh8k7r15F4hBKA88lv
[11:14:57 CEST] <azy> its hilarious
[11:14:59 CEST] <pridkett> why do you have such a weird file name
[11:15:01 CEST] <tanja77> so in the first place its just about knowing who the offender is ;-)
[11:15:45 CEST] <pridkett> azy  it's not even funny
[11:15:47 CEST] <pridkett> i just watched
[11:16:06 CEST] <pridkett> azy  so what mp3 file you want to overlay?
[11:20:35 CEST] <azy> http://paste.debian.net/hidden/76eda90b/
[11:20:50 CEST] <pridkett> tanja77 sorry not sure how ffmpeg can detect animals
[11:21:05 CEST] <azy> needs to detect motion
[11:21:19 CEST] <azy> https://motion-project.github.io/
[11:21:41 CEST] <pridkett> there are face recognition
[11:21:43 CEST] <azy> any surveillance software should do it
[11:21:58 CEST] <pridkett> azy and what is the link you got the  Curb\ Your\ Enthusiasm\ Theme-Ag1o3koTLWM.webm
[11:22:30 CEST] <azy> https://www.youtube.com/watch?v=Ag1o3koTLWM
[11:22:45 CEST] <azy> i could get a different format with youtube-dl if needed
[11:22:58 CEST] <pridkett> azy webm is fine
[11:23:02 CEST] <pridkett> ffmpeg handles it
[11:23:21 CEST] <pridkett> azy are you trying to edit the video and  upload to youtube?
[11:23:29 CEST] <azy> ye
[11:23:39 CEST] <pridkett> azy what is your youtube name
[11:24:47 CEST] <azy> https://www.youtube.com/user/0xstev3/videos
[11:25:59 CEST] <pridkett> 308 subscribers  NOT bad for having only crappy videos
[11:27:13 CEST] <pridkett>  azy at what time/point does  Curb\ Your\ Enthusiasm\ Theme-Ag1o3koTLWM.webm   wants to join in
[11:29:35 CEST] <azy> 0:07
[11:31:04 CEST] <pridkett> extract the audio file from mp4
[11:31:41 CEST] <pridkett> ffmpeg -i input1.webm -i input2.aac -filter_complex amerge -ac 2 -c:a libmp3lame -q:a 4 output.mp3
[11:33:31 CEST] <pridkett> azy why are you hiding your version
[11:34:40 CEST] <azy> so i dont get haked
[11:34:49 CEST] <azy> whats the second -i meant to represent?
[11:37:05 CEST] <tanja77> pridkett, yes, I did not think of ffmpeg could do this LOL.. but I thought in an ffmpeg channel there could be perhaps persons that know other software or so
[11:37:19 CEST] <pridkett> tanja77 fair enough
[11:38:21 CEST] <pridkett> azy 99% people in freenode reveals version info
[11:38:36 CEST] <pridkett> that doesn't mean 99% people are getting hacked
[11:39:04 CEST] <azy> i was expecting something like: ffmpeg -i video.mp4 -i audio.webm -uwotm8 3:40
[11:39:51 CEST] <relaxed> tanja77: I think zoneminder would work (uses ffmpeg on the backend)
[11:40:56 CEST] <tanja77> just downloading zoneminder...
[11:41:08 CEST] <relaxed> for monitoring, not going over your past video
[11:41:16 CEST] <tanja77> ah, I see
[11:44:45 CEST] <relaxed> tanja77: maybe this?  https://www.home-assistant.io/components/binary_sensor.ffmpeg_motion/
[11:45:41 CEST] <tanja77> this looks cook, I am gonna check this
[11:45:57 CEST] <tanja77> ps: you guys see: I AM in the right channel ;-)  thanks & kisses
[11:47:43 CEST] <pridkett> hi i am doing a survey: what gender/age do you think the person on this youtube video is based on their voice:  https://www.youtube.com/watch?v=hMynODL4GtQ
[11:51:42 CEST] <azy> 14
[11:51:59 CEST] <pridkett> azy thanks and gender?
[11:52:23 CEST] <azy> male
[11:52:34 CEST] <azy> plus "ron fetter"
[11:53:36 CEST] <pridkett> is Ron, a guy only name?
[12:34:49 CEST] <JEEB> pk08: please stop privately messaging me
[12:35:25 CEST] <pk08> ok, i dont see anything related to ttl in doc/example
[12:35:27 CEST] <JEEB> there's an AVOption for ttl in the UDP protocol, set that through av_dict_set to your avformatcontext (or IO one, don't remember)
[12:35:30 CEST] <JEEB> of course not
[12:35:44 CEST] <JEEB> they are just examples on how to *set* *options*
[12:36:11 CEST] <JEEB> when you set the options you will get back a list of options that were not applied
[12:36:18 CEST] <JEEB> that way you can figure out if your options got applied or not
[12:53:57 CEST] <pk08> thanks JEEB
[14:19:27 CEST] <miyalys> Hi! I'm trying to record audio out and video simultaneously, via ffmpeg and an alsa loopback device. this command runs, but strangely the audio is only heard for maybe half a second (sounding like it should for that duration) and then it disappears?
[14:19:49 CEST] <miyalys> The command: ffmpeg -f x11grab -framerate 25 -video_size 1920x1080 -i :0.0  -f alsa -i "plug:\"dsnoop:Loopback,1\"" screencast.mkv
[14:21:22 CEST] <miyalys> Not sure if the solution is alsa or ffmpeg related, but recording just the audio separately, also with ffmpeg, it plays back just fine.
[15:06:27 CEST] <blueman24> Hello, i trying to compile ffmpeg like this "PKG_CONFIG_PATH=/usr/local/src/ffmpeg/lib/pkgconfig ./configure --prefix=/opt/ffmpeg --extra-cflags=-I/opt/ffmpeg/include --extra-ldflags=-L/opt/ffmpeg/lib --pkg-config-flags=--static --bindir=/opt/ffmpeg/bin --enable-openssl --enable-gpl --enable-nonfree --enable-libfreetype --enable-libtheora --enable-libvpx --enable-libx264 --enable-libx265"
[15:06:35 CEST] <blueman24>  i got error "ERROR: freetype2 not found"
[15:06:49 CEST] <blueman24> any thoughts?
[15:09:03 CEST] <pink_mist> try installing freetype2
[15:09:09 CEST] <pink_mist> and its development headers
[15:09:43 CEST] <blueman24> yum search freetype2 does not exist
[15:10:03 CEST] <blueman24> I installed freetype-devel
[15:10:57 CEST] <pink_mist> well, as long as the version of that is >=2, it should be fine
[15:13:15 CEST] <blueman24> my version is: 2.8
[15:19:46 CEST] <miyalys> update on my issue: Switching to a native codec solved my problem of desyncing: ffmpeg -f alsa -i "plug:\"dsnoop:Loopback,1\"" -f x11grab -r 24 -s 1920x1080  -i :0.0 -acodec aac -vcodec mpeg4 -preset medium -qscale:v 5 rec.mkv
[18:42:05 CEST] <retal> faLUCE, hi. Couple days ago you sent me message, sorry I didn't answer . Was very busy
[19:54:36 CEST] <King_DuckZ> hi, in my program I want to generate output that is readable by ffmpeg's pipe input, from a full black frame for example I generate this https://alarmpi.no-ip.org/kamokan/fk?colourless (cut to 20 lines)
[19:55:29 CEST] <King_DuckZ> then I run ffmpeg -i frame.y4m -r 1/1 deleme.bmp, but the resulting bmp image is not black not a uniform colour
[19:57:19 CEST] <King_DuckZ> deleme.bmp --> https://alarmpi.no-ip.org/kamokan/fl?colourless (cut to 30 lines)
[20:02:33 CEST] <King_DuckZ> ah fck it's interleaved, damn how did I not realize sooner
[20:02:51 CEST] <King_DuckZ> sorry, I see how to fix it now
[21:18:11 CEST] <faLUCE> hey retal ;-)
[22:49:22 CEST] <mekeor> hello. how to convert a video to MP4 (H.264) with 8 bit depth? i tried `ffmpeg -i input.mxf -c:v libx264 -preset superfast -x264opts bit_depth=8 -c:a copy -map 0 output.mp4` because i didn't see a direct way to set the bit depth but this x264-option doesn't exist
[22:50:04 CEST] <JEEB> for the basic way there's just -pix_fmt yuv420p
[22:50:23 CEST] <JEEB> but since you're dealing with MXF you probably want a high quality conversion
[22:50:33 CEST] <JEEB> I recommend building FFmpeg with zimg to get the zscale filter
[22:50:51 CEST] <JEEB> then you can do -vf "zscale=m=709,format=pix_fmts=yuv420p"
[22:51:07 CEST] <JEEB> the parameter to zscale might be unnecessary in your case
[22:51:12 CEST] <mekeor> how could i find out that yuv420p has 8 bit depth if you wouldn't tell me?
[22:51:50 CEST] <JEEB> I think in the API it has that property, and then you get a long list through -pix_fmts in ffmpeg.c
[22:52:19 CEST] <JEEB> http://up-cat.net/p/db50f53b
[22:52:21 CEST] <JEEB> a bit of it
[22:52:41 CEST] <BtbN> Well, it's an 8 bit pixel format by itself
[22:52:49 CEST] <BtbN> 10 or 12 bit ones are named accordingly
[22:53:09 CEST] <JEEB> and yes, usually if not mentioned otherwise, it's 8bit per channel by default
[22:53:41 CEST] <mekeor> cool, thank you!
[22:53:44 CEST] <mekeor> so, specifying yuv420p should be enough but the zscale option would help keep high quality, right?
[22:54:02 CEST] <mekeor> thank you ver much for your quick help :)
[22:54:27 CEST] <JEEB> not sure how much difference in pure bit depth reduction there is
[22:54:48 CEST] <BtbN> I'd say hardly any.
[22:54:56 CEST] <JEEB> mostly dithering I guess?
[22:55:07 CEST] <JEEB> which is minor
[22:55:22 CEST] <BtbN> Don't you dither when going from 8 to 10 or 12, not the other way?
[22:55:35 CEST] <BtbN> hm, no, nvm
[22:55:41 CEST] <JEEB> no, dithering is usually done to get a better result from less bits
[22:56:47 CEST] <mekeor> hmm. it seems like i'm missing some codecs... http://sprunge.us/ClPCJr
[22:58:08 CEST] <mekeor> oh, seems like that error is due to the audio stuff
[22:58:11 CEST] <JEEB> yes
[22:58:18 CEST] <mekeor> hmm. i was trying to keep all audio-tracks
[22:58:29 CEST] <JEEB> raw audio is possible in mp4 but not sure how much was implemented
[22:58:35 CEST] <JEEB> that's a really fringe thing
[22:59:52 CEST] <mekeor> raw audio?
[23:00:02 CEST] <JEEB> yes, the audio is raw pcm
[23:00:08 CEST] <JEEB> as opposed to being compressed
[23:00:40 CEST] <mekeor> and you can only have multiple tracks of audio if it's raw pcm?
[23:00:46 CEST] <JEEB> no?
[23:01:01 CEST] <mekeor> sorry, i didn't understand you then :D
[23:01:05 CEST] <JEEB> I'm saying the version of FFmpeg you have at least can't figure out which identifier to put into mp4 to put raw audio into it :P
[23:01:13 CEST] <JEEB> thus the "codec not currently supported in container"
[23:02:46 CEST] <mekeor> hmm, so might there be a solution without compiling ffmpeg on my own?
[23:03:14 CEST] <JEEB> you can use an audio format that FFmpeg can write into mp4
[23:03:29 CEST] <JEEB> or switch from mp4 to another container
[23:04:35 CEST] <mekeor> ah, okay, specifying aac works
[23:05:16 CEST] <mekeor> but in the input-file there is a "Data: none" stream and thus ffmpeg complains that the codec "none" is not supported
[23:05:47 CEST] <JEEB> yes, it's vanc data.
[23:06:02 CEST] <JEEB> which is exported with a data_type but there's no codec id for it :P
[23:06:16 CEST] <mekeor> so i need to map it, right?
[23:06:21 CEST] <JEEB> no
[23:06:29 CEST] <JEEB> most likely you don't get anything useful out of it with ffmpeg.c
[23:06:37 CEST] <JEEB> whatever SMPTE 436M VANC is :P
[23:07:23 CEST] <mekeor> so i need to tell ffmpeg to ignore that stream?
[23:07:28 CEST] <JEEB> yes
[23:07:41 CEST] <JEEB> seems like support was added to be able to copy those streams from MXF to MXF
[23:07:47 CEST] <JEEB> but nothing at all otherwise :P
[23:08:14 CEST] <mekeor> ah okay :D
[23:08:25 CEST] <mekeor> i hope i won't miss those data
[23:08:30 CEST] <JEEB> http://mogurenko.com/2015/10/30/closed-captions-and-mxf/
[23:08:39 CEST] <JEEB> seemingly closed captions
[23:10:48 CEST] <mekeor> okay, cool, i specified `-map 0:v -map 0:a`& let's see =)
[23:12:11 CEST] <JEEB> I guess if you actually read the caption data from the MXF packets you could export it as a captions track since FFmpeg has a decoder for it
[23:12:23 CEST] <JEEB> but then you'd have to add re-packaging code on the other side (writing MXF)
[23:12:37 CEST] <JEEB> which currently just happens to work because nothing is done to the data xD
[23:19:56 CEST] <mekeor> actually i could also use `-map 0` for MXF-to-MTS conversion
[23:20:13 CEST] <JEEB> not sure what it did with the data track for the record :P
[23:20:26 CEST] <JEEB> since most likely it did *nothing* to the packets, which are valid in MXF
[23:20:31 CEST] <JEEB> but not necessarily in MPEG-TS
[23:20:50 CEST] <JEEB> and most likely MPEG-TS lets you send out arbitrary data as Data streams
[23:20:51 CEST] <JEEB> :P
[23:21:03 CEST] <JEEB> as in, the FFmpeg writer
[23:40:13 CEST] <mekeor> thank you very much, JEEB and BtbN :)
[23:40:27 CEST] <mekeor> i got done what i was aiming to do :)
[00:00:00 CEST] --- Thu Apr  4 2019


More information about the Ffmpeg-devel-irc mailing list