[Ffmpeg-devel-irc] ffmpeg.log.20160929

[burek]

[00:05:53 CEST] <ac_slater> TwinTailed: are you doing -acodec copy?
[00:06:15 CEST] <ac_slater> oh wait
[00:06:20 CEST] <ac_slater> I see, you're using the libs
[00:06:38 CEST] <ac_slater> first create an ffmpeg command line to do what you want (to make sure it's possible)
[00:12:57 CEST] <TwinTailed> ac_slater: It is possible
[00:12:59 CEST] <TwinTailed> tried it
[00:13:27 CEST] <ac_slater> good
[00:13:54 CEST] <ac_slater> used either `mediainfo` or `ffprobe -show_streams` on the output files from the command line tool and your app to see if there are differences
[00:14:14 CEST] <ac_slater> remember that windows doesn't have all of the codecs that VLC and quicktime implement at an application level
[00:14:56 CEST] <ac_slater> H.264 is a weird one since it's licensed
[00:15:07 CEST] <TwinTailed> ac_slater: Why is my output file almost 3 times larger than ffmpeg's output?
[00:15:17 CEST] <ac_slater> could be literally 1000 reasons ;)
[00:15:26 CEST] <TwinTailed> i mean, the original file is 8MB, ffmpeg's is 9MB, mine is 21MB
[00:15:43 CEST] <TwinTailed> and if I use octx->bitrate = ictx->bitrate
[00:15:48 CEST] <ac_slater> you're transcode isn't organizing the data in the container properly
[00:15:49 CEST] <TwinTailed> my file becomes 74MB
[00:16:30 CEST] <ac_slater> so, my advice: first get ffmpeg's command line tool to spit out EXACTLY what you want (make sure it plays in your desired players). Then try to replicate the transcode in the APIs
[00:17:07 CEST] <ac_slater> remember though, ffmpeg's command line tool supports the PIPE input/output. So you could simply run the command line tool as a daemon/service and have your application feed it
[00:17:32 CEST] <TwinTailed> ac_slater: i'd rather use libav for other reasons :)
[00:17:40 CEST] <TwinTailed> i know it works as a daemon
[00:17:41 CEST] <ac_slater> if you;re using h.264, remember that it has a BUNCH of options
[00:17:55 CEST] <ac_slater> the simple examples on the wiki might not be enough
[00:18:24 CEST] <ac_slater> I'd run ffmpeg/ffplay/ffprobe with the MAX loglevel to see how ffmpeg is reading the input file
[00:18:27 CEST] <TwinTailed> ac_slater: They are not enough, I totally agree
[00:18:39 CEST] <ac_slater> could have some non-reproducable encoding options (hopefully not
[00:18:41 CEST] <ac_slater> )
[00:19:36 CEST] <TwinTailed> so should i now use ffmpeg to remux & transcode the video file and also use my program to do the same and compare them with ffprobe?
[00:19:51 CEST] <ac_slater> I totally would
[00:20:04 CEST] <ac_slater> send the output to paste.debian.net or something similar
[00:20:29 CEST] <ac_slater> I'd actually use `mediainfo` as well. It lists nice things about mp4/h264
[00:20:29 CEST] <TwinTailed> can u tell me what options in ffmpeg cli should i use? Im not so experienced in ffmpeg's cli
[00:20:35 CEST] <ac_slater> there has to be a better analyze application...
[00:20:45 CEST] <ac_slater> well. What's your input file?
[00:20:49 CEST] <ac_slater> encoding, etc
[00:21:15 CEST] <TwinTailed> my input file is vid.mp4
[00:21:15 CEST] <ac_slater> (also, dont forget to look at this https://trac.ffmpeg.org/wiki/Encode/H.264)
[00:21:25 CEST] <TwinTailed> did look at it
[00:21:28 CEST] <ac_slater> ah good
[00:21:49 CEST] <ac_slater> if nothing else, it's a nice example as to how to organize your parameters to ffmpeg
[00:23:09 CEST] <TwinTailed> ac_slater: might it be that my video file is not pix_fmt yuv420p?
[00:23:17 CEST] <ac_slater> could very well be
[00:23:21 CEST] <ac_slater> you can change that though
[00:23:34 CEST] <TwinTailed> im using  ctx->pix_fmt = encoder->pix_fmts[0];
[00:23:37 CEST] <ac_slater> actually, I'm sure the h264 encoder (assuming libx264) will change that for you
[00:23:50 CEST] <ac_slater> what does ffprobe say the color format is?
[00:23:58 CEST] <ac_slater> (on the output file you created with your app)
[00:24:22 CEST] <TwinTailed> Video: h264 (High) (avc1 / 0x31637661), yuv420p
[00:24:41 CEST] <ac_slater> that's right, it should convert it for you
[00:25:08 CEST] <TwinTailed> so pixel format is not a problem
[00:25:49 CEST] <ac_slater> doesn't seem like it
[00:26:05 CEST] <ac_slater> you'd see some red flags if the encoder couldnt handle the color format
[00:26:35 CEST] <ac_slater> I'm going to assume your issue is either h264 profile (baseline vs Main for example), bitrate, and possibly options like CABAC
[00:26:41 CEST] <ac_slater> but
[00:27:00 CEST] <ac_slater> you haven't said EXACTLY which type of transcoding you're doing. I assume it's to shrink the file
[00:27:26 CEST] <TwinTailed> ac_slater: just basic transcoding, not real goal
[00:28:04 CEST] <TwinTailed> ac_slater: OK, I have used ffmpeg -i input -c:v libx264 -c:a copy output.mp4
[00:28:21 CEST] <ac_slater> you should spend some time with just h264 then, no container and no audio. See if you can get something that way - then add audio and containerization later
[00:28:40 CEST] <ac_slater> I mean, that MIGHT do something. Provided libx264 has some defaults
[00:28:43 CEST] <ac_slater> not sure if it does
[00:28:53 CEST] <ac_slater> it might just play stupid and do whatever the fuck it wants
[00:30:25 CEST] <TwinTailed> ac_slater: MediaInfo for file generated using FFMPEG http://pastebin.com/hWxWbFHu
[00:31:13 CEST] <TwinTailed> ac_slater: my program's generated video http://pastebin.com/xAazQLdx
[00:32:03 CEST] <ac_slater> see the differences?
[00:32:06 CEST] <ac_slater> I do
[00:32:32 CEST] <TwinTailed> I do, but I don't know which corrupt the media :P
[00:33:21 CEST] <SchrodingersScat> kepstin: had one that was very off, it was around 7954 pts/dts, but was about 120 seconds in, is there a reason that happens?
[00:33:26 CEST] <ac_slater> does the ffmpeg command line tool generated file play where you want it to play?
[00:33:38 CEST] <ac_slater> TwinTailed: ^
[00:33:50 CEST] <ac_slater> ie - not corrupt
[00:33:53 CEST] <TwinTailed> ac_slater: yes
[00:34:06 CEST] <TwinTailed> ac_slater: it does work properly and its size is much smaller
[00:34:44 CEST] <kepstin> SchrodingersScat: make sure you don't cross the streams - e.g. a file with both video and audio streams will probably have different timebases for each
[00:34:48 CEST] <ac_slater> look at a few things in your output... "Overall bit rate" and "Encoding library/settings"
[00:35:03 CEST] <ac_slater> A) you're file's bitrate is very high
[00:35:14 CEST] <furq> Duration_LastFrame                       : -20 ms
[00:35:17 CEST] <furq> well that doesn't look good
[00:35:26 CEST] <ac_slater> B) It doesn't seem like you're invoking libx264 specifically, maybe ffmpeg's internal h264 encoder, TwinTailed
[00:35:39 CEST] <ac_slater> and what furq said
[00:35:44 CEST] <furq> ffmpeg doesn't have an internal h264 encoder
[00:35:47 CEST] <kepstin> ffmpeg doesn't have an internal h264 encoder, libx264 is the only supported encoder
[00:35:48 CEST] <ac_slater> good
[00:35:50 CEST] <ac_slater> I didnt think it did
[00:35:52 CEST] <kepstin> er, actually not true
[00:36:01 CEST] <kepstin> i think you can use openh264 in some builds
[00:36:10 CEST] <TwinTailed> furq: what does duration_lastframe mean?
[00:36:12 CEST] <ac_slater> I just didnt see mediainfo's specific handler for libx264... meaning the file is really messed up
[00:36:19 CEST] <kepstin> but you probably shouldn't unless you need to for license reasons :)
[00:36:21 CEST] <TwinTailed> ac_slater: How can i decrease the bitrate?
[00:37:30 CEST] <ac_slater> TwinTailed: the example programs have some usage
[00:37:32 CEST] <furq> if you want to match what ffmpeg does then set preset to medium and crf to 23
[00:37:42 CEST] <furq> that shouldn't make any difference though
[00:37:45 CEST] <ac_slater> but, it's a field on you AV*Context
[00:38:07 CEST] <furq> it won't make any difference to whether the file is playable, that is
[00:38:09 CEST] <ac_slater> first step should be able to reproduce some thing CLOSE to what ffmpeg spits out.
[00:38:17 CEST] <TwinTailed> if I set bitrate to the same as input's bitrate, file size is multiplied by 4
[00:38:25 CEST] <ac_slater> why would you do that
[00:38:27 CEST] <furq> well yeah that's not a thing you should do
[00:38:30 CEST] <TwinTailed> i get a 74 MB file out of 8MB
[00:38:43 CEST] <ac_slater> think about it for a sec
[00:38:49 CEST] <SchrodingersScat> kepstin: ooh, crossing the streams...
[00:39:07 CEST] <ac_slater> TwinTailed: try to set it to half or something
[00:39:15 CEST] <furq> don't set the bitrate at all
[00:39:17 CEST] <furq> set crf to 23
[00:39:23 CEST] <SchrodingersScat> kepstin: https://www.youtube.com/watch?v=jyaLZHiJJnE
[00:39:41 CEST] <TwinTailed> furq: av_dict_set(&param, "crf", "23", 0);
[00:39:43 CEST] <relaxed> TwinTailed: you're using cavlc instead of cabac, which will greatly affect the size
[00:39:45 CEST] <TwinTailed> furq: like that 6?
[00:39:51 CEST] <ac_slater> TwinTailed: hint, crf is in the H264 wiki I sent, it should explain it
[00:39:56 CEST] <TwinTailed> relaxed: how to use cabac :P
[00:40:30 CEST] <ac_slater> TwinTailed: don;t think about advanced shit like that right now... just focus on getting it to play
[00:40:38 CEST] <ac_slater> g2g.. good luck
[00:40:53 CEST] <TwinTailed> ac_slater: thanks for everything :)
[00:41:01 CEST] <TwinTailed> furq: DONT U DARE LEAVE ME ALONE!
[00:43:25 CEST] <TwinTailed> furq: how to set it to cabac?
[00:43:40 CEST] <ac_slater> (hint: cabac isnt your issue ;))
[00:44:07 CEST] <TwinTailed> i know but i want to decrease the size
[00:45:44 CEST] <relaxed> libx264 has presets that control compression
[00:46:44 CEST] <TwinTailed> relaxed: i know
[00:47:05 CEST] <TwinTailed> why does it say
[00:47:09 CEST] <TwinTailed> Duration_LastFrame: -20 ms
[00:47:11 CEST] <TwinTailed> ?
[00:47:18 CEST] <furq> afaik if you just don't set any codec settings then it'll use the libx264 defaults
[00:48:13 CEST] <TwinTailed> furq: alright
[00:48:14 CEST] <furq> https://ffmpeg.org/doxygen/trunk/group__lavc__core.html#gae80afec6f26df6607eaacf39b561c315
[00:48:21 CEST] <TwinTailed> furq: what to do about the high bitrate?
[00:48:31 CEST] <furq> you do the thing i just said
[00:49:00 CEST] <TwinTailed> furq: i already allocate context for encoder
[00:49:37 CEST] <furq> then don't set any codec options after that
[00:50:12 CEST] <TwinTailed> furq: i didn't until u told me to set crf and preset
[00:50:25 CEST] <TwinTailed> furq: but I now I removed them, still problem is there
[00:50:58 CEST] <TwinTailed> furq: by codec options height, width, samplerate, timebase are included?
[00:51:17 CEST] <furq> i doubt it
[00:51:55 CEST] <TwinTailed> ok so whats the problem with file, you all confused me
[00:53:21 CEST] <TwinTailed> furq: and what made that -20ms?
[00:53:49 CEST] <furq> i have no idea what's going on with that -20ms but it doesn't look right
[00:54:01 CEST] <furq> i'd guess a timestamp is set wrongly
[00:55:35 CEST] <TwinTailed> furq: the libav log doesn't mention the last frame
[00:55:56 CEST] <TwinTailed> furq: they are 3021 video frames, it mentions till 3020 (I flush)
[00:56:08 CEST] <furq> does it start at 0
[00:56:45 CEST] <TwinTailed> furq: ah, thats why, its an index
[00:58:43 CEST] <TwinTailed> furq: i set these options http://pastebin.com/1p7CjKxW
[01:33:39 CEST] <TwinTailed> furq: whats the Format Profile option name in av_dict_set?
[01:34:05 CEST] <kingsley_> How do you tell ffmpeg to maintain video quality, even if it means not compressing?
[01:39:10 CEST] <DHE> -c[:type] copy will keep the same codec and copy the encoded data. otherwise use a lossless codec. ffv1 for video, flac for audio are simple but functional choices
[01:47:09 CEST] <kingsley_> DHE: If I ask you nicely, could you please ELI5 (explain it like I'm 5) with a simple command line example?
[01:47:42 CEST] <DHE> oh my god, where are your parents!?
[01:49:20 CEST] <DHE> to answer your question, I feel like I need to know what it is you're trying to accomplished. converting a video to lossless or uncompressed isn't something useful by itself
[01:50:16 CEST] <kingsley_> DHE: OK, that all makes sense.
[01:50:58 CEST] <kingsley_> I'd like to preserve the quality of video as I processes it several times and ways.
[01:51:26 CEST] <DHE> so you're going to be doing several editing passes with filters and stuff
[01:52:04 CEST] <kingsley_> For what it's worth, the following expanded a very short .mp4 from 38129 bytes to 353641.
[01:52:57 CEST] <kingsley_> $ ffmpeg -i in.mp4 -qp 0 out.mkv
[01:53:20 CEST] <DHE> sounds about right
[01:53:36 CEST] <kingsley_> Yes, you're perceptive and correct: I'm doing several editing passes with filters and stuff.
[01:54:16 CEST] <kingsley_> That's interesting.
[01:55:45 CEST] <DHE> well, a few things I'd consider. first you can pipeline filters so it might be possible to do your edits in a single shot of ffmpeg.
[01:57:19 CEST] <DHE> otherwise, I guess you'd want "-c:a copy -c:v ffv1" before your output file for intermediate files if you're doing video editing and leaving the audio intact. then give a proper codec (eg: libx264) for the final encode
[01:57:26 CEST] <TwinTailed> DHE: in LibAV, using av_opt_set(&param, "profile", "high", 0); is not changing the profile to high, it stays baseline
[01:58:20 CEST] <DHE> TwinTailed: if the encoding parameters you specify are effectively baseline, you get baseline. the profile is more a constraint macro applied on your settings
[01:58:48 CEST] <DHE> did you ensure the profile key was consumed when initializing the codec?
[01:59:32 CEST] <TwinTailed> DHE: no
[01:59:49 CEST] <TwinTailed> DHE: I just set width, height and some other basic stuff
[02:00:33 CEST] <DHE> as a simple sanity check, you should check your dictionary comes back empty when you're done with it. any keys left over were not recognized by the libav stuff that it was given to
[02:00:36 CEST] <TwinTailed> DHE: http://pastebin.com/xAazQLdx
[02:00:46 CEST] <DHE> I do not have time to do code reading tonight
[02:00:52 CEST] <TwinTailed> not a code
[02:00:56 CEST] <TwinTailed> its a mediainfo output
[02:02:22 CEST] <kingsley_> DHE: Thank you for your thoughts.
[11:15:33 CEST] <golserma> Hi I'm wondering what the best format for whatsapp is? x264 baseline AAC and mp4?
[12:26:27 CEST] <spacemadman> kepstin: hey.. sorry got disconnected yesterday and didnt have connectivity to log back in.
[13:52:48 CEST] <MichaelVB> Is there a way to make ffmpeg stretch my motion jpeg data to fit a duration?
[13:53:07 CEST] <MichaelVB> Currently I've got data with a few missing frames every now and then and plays at higher speeds
[13:58:11 CEST] <MichaelVB> And considering the motion jpegs don't contain any time metadata its a bit of a PITA
[14:35:53 CEST] <flux> I don't know, but I solved similar issue a long time ago by going through the time-stamped frames and inserting duplicate frames where needed, and then running video encoder
[14:37:14 CEST] <BtbN> you can use setpts to change the framerate to whatever you like
[14:39:38 CEST] <gahan> how can I test streams instead of specifying input file from command line on linux?
[14:40:01 CEST] <BtbN> specify the stream url instead?
[14:42:46 CEST] <gahan> BtbN: won't it download it to fs? I'm about testing input streams
[14:43:20 CEST] <BtbN> why would it do that?
[14:43:28 CEST] <gahan> I'm just guessing, thanks :)
[15:14:56 CEST] <spacemadman> Hi, does ffmpeg do m2v's?
[15:24:44 CEST] <kriNon> Hey, I am getting this error: http://pastebin.com/ZPuTBxFf in a video file, what does this error mean?
[15:26:41 CEST] <BtbN> it has broken timestamps. Looke like each timestamp is repeating once.
[15:33:46 CEST] <nonex86> the error means exactly what it says, you provided not increased dts :/
[15:35:58 CEST] <nonex86> if you mux from your code you can try set dts to AV_NOPTS_VALUE and ffmpeg will make them for you, but this functionality was marked deprecated (at least in version 3.0 it is)
[15:45:02 CEST] <spacemadman> hey guys
[15:45:12 CEST] <spacemadman> i have an mov in 25fps with audio
[15:45:16 CEST] <spacemadman> how do i convert it to 24?
[15:51:13 CEST] <furq> do you want the duration to remain the same
[15:53:37 CEST] <furq> if you're doing a pal to film conversion then you probably want -vf setpts=25025/24000*PTS -af atempo=24000/25025
[15:53:44 CEST] <furq> you'll need to reencode the video and audio for that though
[15:55:38 CEST] <Anova> hi guys. I'm learning about HLS. My goal is to be able to download a HLS stream. I've managed this with ffmpeg, but its quite slow, ie about 1.75x (way lower than my internet connection speed). I'm assuming the reason for slow speed is the single thread downloading each ts file one at a time and combining it into mp4 at the end. Is this correct? If so, is it possible to download all the ts files with another tool that does mult-threaded 
[15:55:45 CEST] <Anova> and then simply combine them with ffmpeg?
[15:56:20 CEST] <Anova> any tips/advice will be much appreciated :)
[15:59:31 CEST] <BtbN> Anova, sounds like someone forgot -c copy
[16:09:35 CEST] <spacemadman> furq: can you help with the full command?
[16:09:58 CEST] <spacemadman> furq: also, my mov is uncompressed qt .. how do i maintain that as the format in the output?
[16:11:53 CEST] <pgorley> hi, can i set an AVCodecContext's callbacks during decoding? or do i need to reinitialize it for this to take effect?
[16:12:01 CEST] <superdump> given an AVFormatContext that is going to be passed into avformat_write_header() how do i set the equivalent of the "-fflags +genpts" command line option?
[16:13:50 CEST] <superdump> i'm guessing i create an AVDictionary and set some key value pair on it
[16:14:10 CEST] <superdump> but it's not clear what it's meant to be... unless the key is fflags and the value is genpts
[16:27:54 CEST] <superdump> can one do av_opt_set(avfctx, "fflags", "genpts", NULL); for example?
[16:30:11 CEST] <superdump> or does one need to create an AVDictionary and av_dict_set(mydict, "flags", "genpts", NULL); or something?
[16:34:20 CEST] <nonex86> AVFMT_FLAG_GENPTS?
[16:35:28 CEST] <superdump> yes, i want to set that
[16:36:24 CEST] <nonex86> i mean maybe you should set in some flags field
[16:36:46 CEST] <nonex86> this works for demux for sure, but dont know
[16:36:59 CEST] <superdump> well why both with the AVDictionary and av_opt_set() and stuff if you can just access the flags member of AVFormatContext directly?
[16:36:59 CEST] <nonex86> will it work in case of muxing
[16:37:34 CEST] <superdump> hopefully
[16:37:36 CEST] <superdump> i'll try it
[16:37:41 CEST] <nonex86> not sure, i think some of options can be codec specific
[16:37:51 CEST] <nonex86> and not have equivalent in ffmpeg structs
[16:38:05 CEST] <nonex86> so they are passed through dictionary to codec
[16:38:10 CEST] <nonex86> but this is just my guess
[16:39:28 CEST] <nonex86> at least if you check options_table.h, genpts option is relasted to AVFMT_FLAG_GENPTS
[16:39:45 CEST] <nonex86> *related
[16:39:50 CEST] <superdump> yup
[16:41:37 CEST] <nonex86> in avformat.h: Flags modifying the (de)muxer behaviour. A combination of AVFMT_FLAG_*.
[16:41:59 CEST] <nonex86> so looks like this should work even for muxer
[16:42:40 CEST] <nonex86> but my guess you should put AV_NOPTS_VALUE in pts
[16:43:02 CEST] <nonex86> before writing packet by av_write/interlived
[16:43:15 CEST] <nonex86> *interleaved
[16:55:47 CEST] <superdump> nonex86: i'm basically trying to replicate whatever ffmpeg -i infile.mp4 -vcodec copy -fflags +genpts outfile.mp4 does
[16:57:40 CEST] <nonex86> superdump, yeah, i understand, i used pts generation only in demux codepath, one more question, why dont you provide pts yourself?
[16:58:22 CEST] <nonex86> superdump, pts generation is an easy task
[16:58:56 CEST] <superdump> i'm guessing there's a bug in this code i'm looking at with regard to timestamps
[17:00:26 CEST] <nonex86> you just take pts from source file, convert it using source timebase
[17:00:49 CEST] <nonex86> then convert this real pts using destination time base
[17:01:06 CEST] <nonex86> and you good to go with the packet
[17:03:19 CEST] <nonex86> also maybe you should take an account the stream start time
[17:04:30 CEST] <nonex86> *in
[17:17:50 CEST] <Anova> BtbN i have included the copy command
[17:37:07 CEST] <TwinTailed> in libav, av_opt_set(ctx>priv_data, "profile", "high", 0); is not changing my h264 mp4 video's profile to high, its always baseline
[17:40:39 CEST] <kepstin> TwinTailed: that option just sets a max limit. Other options will cause x264 to not use all the features of high, in which case it'll make baseline
[17:41:06 CEST] <kepstin> (for example, the 'ultrafast' preset usually won't make anything higher than baseline)
[17:41:18 CEST] <nonex86> what about AVCodecContext->profile ?
[17:41:23 CEST] <kepstin> unless you need to limit profile for some reason - e.g. device support - you should basically never set it.
[17:42:21 CEST] <nonex86>      * profile
[17:42:22 CEST] <nonex86>      * - encoding: Set by user.
[17:49:14 CEST] <TwinTailed> kepstin: isn't using "high" will make it use cabac?
[17:49:24 CEST] <TwinTailed> and also much less size
[17:49:33 CEST] <kepstin> TwinTailed: no. If you use cabac, then it will say the output is high
[17:49:43 CEST] <kepstin> you have the cause and effect backwards
[17:49:50 CEST] <TwinTailed> ah
[17:49:56 CEST] <TwinTailed> then how do i use cabac?
[17:50:03 CEST] <kepstin> you have some options which say "don't use cabac", you need to remove them
[17:50:15 CEST] <TwinTailed> kepstin: the bitrate is being high and thus size is also big
[17:50:25 CEST] <kepstin> (possibly indirectly, for example because you're using a fast preset)
[17:50:36 CEST] <kepstin> iirc only ultrafast disables cabac tho
[17:51:04 CEST] <TwinTailed> kepstin: any of these options force cabav? http://pastebin.com/hKqkJwFx
[17:51:08 CEST] <TwinTailed> thats not a code ^
[17:51:15 CEST] <TwinTailed> just a snippet
[17:51:49 CEST] <TwinTailed> kepstin: and here are the audio's options http://pastebin.com/At1aS9zu
[17:51:59 CEST] <kepstin> TwinTailed: not clear from that code - are you setting any options elsewhere on the video
[17:52:02 CEST] <TwinTailed> kepstin: other than that, I haven't set anything
[17:52:29 CEST] <TwinTailed> kepstin: i have no av_opt_set in my code
[17:52:38 CEST] <kepstin> TwinTailed: you might want to try explicitly setting preset to medium and crf to e.g. 23 to replicate the defaults, but those are the defaults so it should be using them unless you've overridden something iirc :/
[17:53:49 CEST] <TwinTailed> kepstin: they are being overridden, but i have no idea why
[17:54:18 CEST] <kepstin> TwinTailed: also, make sure libav* debug logging is enabled and look at the log output to see if there's anything obviously wrong there
[17:54:40 CEST] <TwinTailed> http://pastebin.com/RTYiT2cb <-mediainfo output
[17:55:35 CEST] <TwinTailed> kepstin: here is the debug log output http://pastebin.com/Q5qbYqyT
[17:56:42 CEST] <kepstin> TwinTailed: looks fine there, x264 is even saying that it's using high profile.
[17:57:24 CEST] <TwinTailed> kepstin: then why the output is baseline O.o
[17:57:53 CEST] <kepstin> TwinTailed: are you looking at the correct file? have you looked at the code of this 'mediainfo' tool to see how it actually determines that value?
[17:58:16 CEST] <TwinTailed> kepstin: the guys here told me to use mediainfo
[17:58:24 CEST] <TwinTailed> kepstin: yes, im sure its the correct file
[18:02:46 CEST] <TwinTailed> kepstin: what does this error mean: [mov,mp4,m4a,3gp,3g2,mj2 @ 0059bc20] shifted frame pts, curr_cts: 25807030 @ 539
[18:02:47 CEST] <TwinTailed> 7, ctts: 26277, ctts_count: 28196
[18:02:47 CEST] <TwinTailed> ?
[18:03:56 CEST] <TwinTailed> kepstin: no one here told me how to set the pts, they all tell me its automatic, but if i don't set it using av_rescale_q I get non-monotonic dts
[18:04:17 CEST] <kepstin> TwinTailed: no idea, it's something mp4 specific and i'm not familiar with the format. It might be an issue with the pts values you get from your input. Note that the 'ffmpeg' command-line tool does a lot of cleanup on input PTS values by default.
[18:05:51 CEST] <kepstin> But yeah, you really shouldn't have to touch pts values when doing a transcode, unless needed due to different requirements between source and destination containers.
[18:58:26 CEST] <spacemadman> kepstin: hey man, how are u?
[18:58:35 CEST] <spacemadman> sorry my internet died yesterday and I couldnt log back in.
[19:00:01 CEST] <JEEB> define "m2v"
[19:00:11 CEST] <JEEB> uhh, sorry for that
[19:00:17 CEST] <JEEB> was scolled way up in the buffer
[19:21:27 CEST] <spacemadman> hey JEEB
[19:21:29 CEST] <spacemadman> was that for me?
[19:47:39 CEST] <newuser87123> second video iis not playing at all in combined video of 2 parts. audio is fine. vlc codec info tells that both parts have same codec, bitrate, resolution. framerate,etc
[19:50:01 CEST] <newuser87123> kepstin: ^^^
[19:50:46 CEST] <kepstin> newuser87123: um, who are you? I have no context
[19:51:29 CEST] <newuser87123> I came yesterday with an rmtpe stream and you directed me to #youtube-dl
[19:53:24 CEST] <newuser87123> kepstin: I downloaded that stream with youtube-dl but the problem is that its a 2 part stream and after joining them with ffmpeg, 2nd part is not playing while 1st is playing fine. audio is also playing fine. codec info tool of vlc shows that both parts have same codec, bitrate, framerate, resolution, etc
[19:53:47 CEST] <kepstin> how did you "join them" with ffmpeg?
[19:54:14 CEST] <kepstin> (please use a pastebin-like service to post the full command and output if possible)
[19:54:46 CEST] <newuser87123> ffmpeg -f concat -i lst.txt -c copy output.mp4 (lst - file '1st file' file '2nd file')
[19:56:36 CEST] <kepstin> hmm. well, when that works it works; when it doesn't you might just have to do a re-encode (and use the concat filter rather than concat demuxer)
[19:58:43 CEST] <newuser87123> how to use filter than demuxer ?
[19:58:59 CEST] <spacemadman> kepstin: hey man, anyway to encode an mov to m2v keeping in mind that the source is an uncompressed 8bit RGB mov and retaining as much detail as possible?
[19:59:34 CEST] <spacemadman> also, if i were to squeeze in a 25 to 24 and a pitch shift, what would be the right flags to use?
[19:59:53 CEST] <kepstin> newuser87123: https://www.ffmpeg.org/faq.html#Concatenating-using-the-concat-filter
[20:00:35 CEST] <kepstin> spacemadman: you had the pitch shift correct in the command I saw, you just had '-r 24' in the wrong place (had to move it to an input option instead of output option)
[20:01:14 CEST] <kepstin> newuser87123: I guess also https://www.ffmpeg.org/ffmpeg-filters.html#concat which has examples
[20:01:42 CEST] <kepstin> (although they're complicated examples; the simple "two files, one a+v each" is easier)
[20:01:50 CEST] <newuser87123> maybe using a different version of ffmpeg would be of any help ? kepstin
[20:02:22 CEST] <kepstin> newuser87123: I dunno about a "different version", but you should be always keeping ffmpeg up to date for general improvements and fixes.
[20:02:34 CEST] <kepstin> it's unlikely to help this specific issue tho
[20:02:51 CEST] <kepstin> (which is probably that the two video streams are incompatible in some subtle way)
[20:04:44 CEST] <spacemadman> kepstin: umm.. input option?
[20:05:10 CEST] <kepstin> spacemadman: input options are options which go before an input file (before the -i filename), and apply to a specific input
[20:05:25 CEST] <newuser87123> kepstin: http://www.tiikoni.com/tis/view/?id=8c9382f
[20:05:45 CEST] <kepstin> spacemadman: when -r is used as an input option, it causes the fps in the file to be ignored, and it'll be rewritten at the requested framerate
[20:06:02 CEST] <ChocolateArmpits> Were there any suggestions to have "-ac" as a filter too for straightforward downmixing? Now the only way  seems to be using channelsplit and specifying the input layout option as the output downmix
[20:06:15 CEST] <spacemadman> kepstin: ffmpeg -r 24 -i input.mov -filter:v "setpts=1.0417*PTS" -y output.m2v
[20:06:44 CEST] <spacemadman> this good?
[20:06:45 CEST] <kepstin> spacemadman: i'm not sure what you're doing with that... you'd want to either use the -r input option or a setpts filter, not both...
[20:07:06 CEST] <kepstin> spacemadman: that reads the input file as if it was 24fps, then stretches it to 25ish fps
[20:07:19 CEST] <spacemadman> ok i am totally lost.
[20:07:32 CEST] <kepstin> spacemadman: what is the actual fps of your source file, and what is the output fps you want?
[20:07:42 CEST] <spacemadman> 25 and its an uncompressed mov.
[20:07:47 CEST] <furq> -vf setpts=25025/24000*PTS -af atempo=24000/25025
[20:07:57 CEST] <spacemadman> i need it converted to 24 but in a m2v
[20:08:20 CEST] <spacemadman> retaining as much bitrate and resolution as possible
[20:08:23 CEST] <ChocolateArmpits> use -vf fps=24
[20:08:29 CEST] <kepstin> spacemadman: alternatively to furq's command, "-vf settb=24,setpts=N"
[20:08:54 CEST] <kepstin> spacemadman: both do the same thing. ChocolateArmpits's command doesn't do what you desire.
[20:08:55 CEST] <spacemadman> where do these flags go?
[20:08:57 CEST] <furq> fps=24 will drop frames
[20:09:24 CEST] <spacemadman> i dont want it to drop or interpolate frames.. i want it to stretch the duration
[20:09:24 CEST] <kepstin> spacemadman: "-vf ..." is an output option, it goes in the space after the input files and before an output file.
[20:09:45 CEST] <furq> the options i posted will stretch the duration and the audio to match
[20:10:11 CEST] <furq> it'll keep the audio pitch the same
[20:10:17 CEST] <furq> there's another filter which will pitch shift but i forget the name now
[20:10:22 CEST] <kepstin> furq: why the '25025/24000' rather than '25/24'? I'm just curious.
[20:10:30 CEST] <furq> for 24000/1001fps
[20:10:43 CEST] <kepstin> hmm, but the input might *actually* be 24fps if it's film :/
[20:10:59 CEST] <spacemadman> nope
[20:11:03 CEST] <furq> when someone says 24fps i assume they mean 23.976
[20:11:05 CEST] <spacemadman> its an ad film and its 25
[20:11:14 CEST] <spacemadman> from a PAL country
[20:11:27 CEST] <kepstin> spacemadman: your source is 25, and you want NTSC 24 (aka 24/1.001)?
[20:11:35 CEST] <spacemadman> i want film 24
[20:11:46 CEST] <spacemadman> to go to a theater
[20:11:53 CEST] <ChocolateArmpits> ok so in PAL the telecine is applied by slowing the footage playback rate
[20:12:00 CEST] <furq> yes it is
[20:12:20 CEST] <ChocolateArmpits> So he only needs -r 24 as the input option
[20:12:26 CEST] <furq> no
[20:12:29 CEST] <ChocolateArmpits> Yes
[20:12:34 CEST] <furq> oh as an input option
[20:13:10 CEST] <spacemadman> guys, so my source is definitely 25 with sync audio
[20:13:31 CEST] <spacemadman> i need an m2v cause OpenDCP accepts only m2v to make MPEG2 MXF Interop DCP's
[20:13:50 CEST] <furq> does -r as an input option work reliably
[20:13:53 CEST] <furq> i was under the impression it didn't
[20:13:54 CEST] <kepstin> spacemadman: ok, so either use the '-r 24' input option OR '-vf setpts=24/25*PTS' OR '-vf settb=24,setpts=N'
[20:14:32 CEST] <spacemadman> all three will work?
[20:14:36 CEST] <spacemadman> what about the audio?
[20:14:44 CEST] <furq> -af atempo=25/24
[20:14:46 CEST] <spacemadman> will it do the necessary pitch shift as well?
[20:14:59 CEST] <furq> actually those are both the wrong way around
[20:15:04 CEST] <ChocolateArmpits> Audio will have to be processed separately if you use the input option
[20:15:11 CEST] <furq> it should be -vf setpts=25/24 -af atempo 24/25
[20:15:17 CEST] <furq> -vf setpts=25/24 -af atempo=24/25
[20:15:21 CEST] <kepstin> hmm, right, you're slowing down
[20:15:33 CEST] <furq> or -vf setpts 25025/24000 -af atempo=24000/25025 if you want 23.976fps
[20:15:39 CEST] <newuser87123> kepstin: its weird, old version of ffmpeg inbuilt in 'thekmplayer' is joing both the parts and playing the video perfectly, but newer version is not, its strange, why so ?
[20:16:12 CEST] <kepstin> spacemadman: if your ffmpeg has it, you can use the 'rubberband' filter to adjust audio speed without changing pitch.
[20:16:31 CEST] <furq> atempo doesn't change pitch
[20:16:37 CEST] <ChocolateArmpits> furq: 24000/1001 is only relevant for ntsc, as back in the days it was used as a telecine midpoint between 24p and 30000/1001 dropframe
[20:16:40 CEST] <spacemadman> furq: so ffmpeg -i input.mov -vf setpts=25/24 -af atempo=24/25 -y output.m2v?
[20:16:47 CEST] <furq> sure
[20:16:56 CEST] <furq> that'll use whatever the default encoding settings for m2v are
[20:17:00 CEST] <furq> which you probably don't want
[20:17:07 CEST] <kepstin> hmm, atempo doesn't? ok. It might be worthwhile comparing the two filters for quality.
[20:17:09 CEST] <spacemadman> so how do i specify?
[20:17:18 CEST] <furq> shrug
[20:17:22 CEST] <spacemadman> sorry man
[20:17:23 CEST] <spacemadman> new to this
[20:17:24 CEST] <ChocolateArmpits> What are your encoding requirements?
[20:17:26 CEST] <furq> i've never encoded mpeg-2 with ffmpeg
[20:17:41 CEST] <kepstin> spacemadman: find out the codec specs for the format - you should normally have a bitrate and gop size given to you, maybe some vbv limits.
[20:17:49 CEST] <spacemadman> ChocolateArmpits: well i need it to make a cinema MPEG2 MXF Interop DCP
[20:17:58 CEST] <spacemadman> and found a tool called OpenDCP that allows me to make it
[20:18:16 CEST] <ChocolateArmpits> probably uses ffmpeg one way or another behind back
[20:18:25 CEST] <newuser87123> kepstin: atempo what ? are you talking to me ?
[20:18:30 CEST] <spacemadman> Nothing on the OpenDCP manual. Looked everywhere on their site/forums since you asked me that yesterday
[20:18:44 CEST] <kepstin> newuser87123: no, that's not relevant to you.
[20:19:27 CEST] <spacemadman> i have one example MXF Interop DCP and i can media info it if you guys would like to see
[20:19:42 CEST] <newuser87123> kepstin: what might be the reason ? older version inbuilt 'thekmplayer' is joing both files perfectly but newer version directly from zerone is not !
[20:19:54 CEST] <furq> actually sorry
[20:19:56 CEST] <kepstin> newuser87123: I have no idea.
[20:20:03 CEST] <furq> -vf setpts=25/24*PTS -af atempo=24/25
[20:20:45 CEST] <spacemadman> furq: ffmpeg -i input.mov -vf setpts=25/24*PTS -af atempo=24/25 -y output.m2v
[20:20:52 CEST] <furq> yeah
[20:21:07 CEST] <spacemadman> ffmpeg cant do mxf's directly right?
[20:21:14 CEST] <furq> apparently it can
[20:21:20 CEST] <furq> there's an mxf muxer
[20:21:20 CEST] <spacemadman> whaat?
[20:21:28 CEST] <newuser87123> kepstin: okay, if I want then from where I can get the older version of ffmpeg which is inuilt 'thekmplayer' ? also how to verbosely paste concat cmd output so that you haave an idea
[20:21:32 CEST] <kepstin> spacemadman: we told you this when you first asked :/
[20:21:32 CEST] <spacemadman> then i have an example MXF
[20:21:36 CEST] <spacemadman> should i just mediainfo that?
[20:21:52 CEST] <furq> run ffprobe on it i guess
[20:21:58 CEST] <spacemadman> kepstin: woot woot.. thats awesome.. sorry i didnt read that bit
[20:21:58 CEST] <furq> ffprobe -show_streams -show_format foo.mxf
[20:22:01 CEST] <kepstin> all I can find about mxf interop mpeg is that the max bitrate is 250mbit :/
[20:22:01 CEST] <spacemadman> must have missed it
[20:22:20 CEST] <spacemadman> furq: hang on.. doing that
[20:22:26 CEST] <spacemadman> kepstin: i'll give you all the info i can man
[20:22:56 CEST] <kepstin> i'm curious whether they're maybe doing intra-only encoding or something silly like that (the newer standard with jpeg2000 are intra-only)
[20:23:26 CEST] <spacemadman> MPEG2 MXF Interop DCP's are an old standard. God knows why its even being used.
[20:24:09 CEST] <spacemadman> furq: Failed to set value '-show_format' for option 'show_steams': Option not found
[20:24:46 CEST] <spacemadman> i have the mediainfo though
[20:24:50 CEST] <spacemadman> should i just paste bin?
[20:25:03 CEST] <newuser87123> kepstin: I'm thinking of deleteing newer version and going for older version ! should I or not ?
[20:25:07 CEST] <kepstin> spacemadman: that was just a typo, try '-show_streams' :)
[20:26:22 CEST] <spacemadman> paste it here?
[20:26:54 CEST] <kepstin> huh, apparently MXF-Interop MPEG2 is actually limited to 80mbit/s, the higher limit is for the newer stuff
[20:27:48 CEST] <spacemadman> kepstin: yeah its shit.
[20:28:00 CEST] <spacemadman> but some theaters demand it.. thats why no option
[20:28:18 CEST] <kepstin> 80mbit mpeg2 should do a really good job for 1080p, as long as there's no stupid vbv limits or really low keyframe interval
[20:28:47 CEST] <spacemadman> it doesnt look great in the theaters in comparison with the jp2k versions
[20:28:49 CEST] <kepstin> mpeg 2 really isn't that bad of a codec, it's just the bitrate limits on dvds and broadcast really bring out its worst
[20:29:26 CEST] <furq> and the fact that dvds are 576i at best
[20:29:47 CEST] <kepstin> like, if you tell ffmpeg to make an 2-pass 80mbit encode of a 1080p movie in mpeg 2 with default options, it's probably gonna say "i couldn't figre out how to use all that bitrate, the file's gonna be smaller than you wanted"
[20:30:19 CEST] <kepstin> (I paraphrased, but there is a log message to that effect which ffmpeg will print)
[20:30:20 CEST] <furq> and also the fact that a lot of dvds were incredibly badly mastered
[20:30:53 CEST] <spacemadman> oh..
[20:30:53 CEST] <furq> it seems a bit harder to fuck up blu-ray but i'm sure you're all about to tell me otherwise
[20:31:38 CEST] <kepstin> blu-ray still has vbv limits and such, but the combination of a newer codec (h264) and higher bitrates + larger buffers means it's harder to screw up.
[20:32:00 CEST] <furq> i've seen a lot of real trash on dvd that has little to do with the codec
[20:32:02 CEST] <kepstin> you're more likely to see issues on tv shows where they try to fit lots of content on a disk, i guess.
[20:32:06 CEST] <furq> mostly bad deinterlacing and stuff
[20:32:23 CEST] <furq> and insane indecisive cropping
[20:33:02 CEST] <furq> a lot of early-2000s bbc dvds are terrible
[20:33:29 CEST] Action: kepstin has seen some weird stuff in anime upscales for bd
[20:33:47 CEST] <kepstin> if you ever want to see a really strange upscale bd, see if you can find the japanese "last exile" release
[20:33:57 CEST] <kepstin> where it's an *interlaced* upscale of telecined animation
[20:34:07 CEST] <furq> nice
[20:34:08 CEST] <spacemadman> guys, so m2v.. any chance i can retain all that niceness from the uncompressed mov?
[20:34:15 CEST] <kepstin> (still looks better than the dvd, of course)
[20:34:34 CEST] <kepstin> spacemadman: the answer is "maybe, but you have to look up the encoding specs to see what settings are allowed"
[20:34:54 CEST] <spacemadman> no info on opendcp.. so i am assuming the sky is the limit?
[20:35:05 CEST] <newuser87123> how to pause ffmpeg -help ? a lot of commands are buried up high in cmd window itself ! also how to copy cmd verbose out so as to paste in paste bin ? kepstin
[20:35:23 CEST] <spacemadman> it just says picture input - m2v
[20:35:28 CEST] <kepstin> spacemadman: all i can see is that the bitrate is max 80mbit, but i dunno if that's average/vbr or if there's restrictions :/
[20:35:46 CEST] <spacemadman> thats the interop mxf dcp ur talking about right?
[20:35:52 CEST] <kepstin> spacemadman: if you can, just do a 2-pass encode with "-b:v 80M" and see what happens.
[20:35:56 CEST] <spacemadman> i am asking about the source m2v opendcp expects to make the mxf
[20:36:07 CEST] <furq> does this tool accept the uncompressed mov you have
[20:36:09 CEST] <kepstin> spacemadman: it looks like opendcp doesn't re-encode, it just remuxes
[20:36:12 CEST] <furq> oh
[20:36:14 CEST] <furq> i guess not then
[20:36:14 CEST] <kepstin> but that's just a guess
[20:36:23 CEST] <spacemadman> nope.. not for the interop mxf
[20:36:42 CEST] <furq> is ffmpeg's mpeg-2 encoder good these days
[20:36:56 CEST] <kepstin> furq: it's not particularly bad?
[20:36:58 CEST] <furq> i remember the one time i had to author a dvd i ended up using some dedicated tool, but that was years ago
[20:37:25 CEST] <furq> i remember it being CCE but apparently that costs several thousand dollars, so it probably wasn't that
[20:37:35 CEST] <kepstin> furq: if you have to encode interlaced stuff, it's apparently tricky to work with, and I don't think you can do soft telecine (repeatfield marks) without an external tool
[20:37:51 CEST] <kepstin> so, kinda painful for dvds :/
[20:38:04 CEST] <furq> this would have been 5+ years ago
[20:38:29 CEST] <furq> i was probably just using ffmpeg wrong but i remember it gave bad results
[20:38:40 CEST] <furq> from a pretty clean 1080p source
[20:41:09 CEST] <kepstin> if you do a 2-pass encode with the vbv options and gop size set correct for dvd, nothing will look great, but I haven't heard of anything *significantly* better than ffmpeg.
[20:44:35 CEST] <furq> i would hope CCE is significantly better considering that it costs $1000
[20:48:16 CEST] <superdump> nonex86: the timebase is the same in the source and destination
[20:48:32 CEST] <superdump> start time could be an issue, will need to investigate
[20:48:38 CEST] <spacemadman> thanks guys
[20:48:45 CEST] <furq> i think it was hcenc that i used
[20:48:59 CEST] <furq> but yeah i'm not recommending it or anything
[20:49:09 CEST] <superdump> nonex86: also, is avcodec_open2() needed if the codec is on a stream created with avformat_new_stream() ?
[21:09:55 CEST] <lindylex> If I download a file in a bash script calling youtube-dl how I get the filename?  Example in bash: $( ./youtube-dl --extract-audio --audio-quality 0 --audio-format mp3 --restrict-filenames --ffmpeg-location ./ffmpeg $url )
[21:15:51 CEST] <lindylex> Sorry wrong chanel.
[21:53:37 CEST] <fenfir_> Im pretty sure when I try and send a stream with rtmps ffmpeg is actually sending rtmpts. The pcap shows my rtmp server complaining about malformed rtmpt packets and when I run it with rtmp as the protocol and point it through stunnel in client mode it works with out any issues. As far as I understand and what it sounds like in the documentation is that rtmps is just rtmp wrapped in ssl and rtmpts is rtmpt wrapped in ssl.
[22:22:49 CEST] <transhuman> hi I am getting a peculiar problem here is my ffmpeg line "$path_to_ffmpeg  -y -i \"/home/<user>/mp4input/$file\" -c:v libx264 -vf scale=$vidSizeXY  -c:a aac -strict -2 \"/home/<user>/mp4final/$file\"";   -----the error its giving me is Error, both -y and -n supplied. Exiting. but I am only specifying -y what gives?
[22:43:17 CEST] <BtbN> probably some stuff in one of your variables.
[22:43:29 CEST] <BtbN> Can't really help without the actuall command that's being invoked.
[22:43:50 CEST] <BtbN> also, -strict expermiental is no longer needed for the aac encoder.
[22:46:50 CEST] <transhuman> ok thanks
[22:47:08 CEST] <transhuman> whats the 2 mean it was in some of the examples
[22:47:15 CEST] <transhuman> i mean -2
[22:47:44 CEST] <transhuman> I think whats happening BtbN is its interpreting the long file name as switches
[22:48:12 CEST] <BtbN> experimental.
[22:48:18 CEST] <transhuman> oh ok
[22:48:26 CEST] <BtbN> Then you aren't propperly escaping spaces
[22:48:34 CEST] <BtbN> Not an issue with ffmpeg, but with your script.
[22:48:42 CEST] <transhuman> yes I realize that
[22:49:00 CEST] <BtbN> if ffmpeg complains about the aac encoder without the strict, you really should update
[22:49:14 CEST] <transhuman> ok
[22:49:16 CEST] <BtbN> it's no longer experimental for quite a while now
[23:09:16 CEST] <transhuman> is there any way to tell ffmpeg to batch convert a list of files in parallel?
[23:16:34 CEST] <BtbN> run it multiple times
[23:24:33 CEST] <tdr> transhuman, script it with a loop and background each ffmpeg command
[23:24:49 CEST] <DHE> well, you could do multiple inputs and multiple outputs that are completely disjoint, but I wouldn't bother with that method. several ffmpeg instances is better.
[23:35:15 CEST] <transhuman> ok thanks
[23:38:37 CEST] <transhuman> one last question what does Protocol name not provided, cannot determine if input is local or a network protocol, buffers and access patterns cannot be configured optimally without knowing the protocol    mean?
[23:41:11 CEST] <transhuman> tdr I actually tried to use parallel processing code, but it doesn't seem to parallelize for some reason I thought maybe you can only run one instance
[23:41:33 CEST] <transhuman> maybe its a bug in my code
[23:47:30 CEST] <tdr> transhuman, for unix shell, something like for x in `cat list`; do ffmpeg <some stuff here>  $x & done     should work
[23:48:07 CEST] <transhuman> right ok that might be simpler thanks
[23:48:14 CEST] <transhuman> what about the other question tdr?
[23:48:25 CEST] <transhuman> the protocol question
[23:48:46 CEST] <transhuman> seems like what ever is causing it is causing my program to exit
[23:49:34 CEST] <tdr> transhuman, however you're feeding it your input, it doesn't know if they are local or network or what.  what command/syntax are you using
[23:50:07 CEST] <transhuman> "$path_to_ffmpeg -y -i $inputpath1 -c:v libx264 -vf scale=$vidSizeXY  -c:a aac $outputpath1";
[23:50:28 CEST] <transhuman> this is in perl
[00:00:00 CEST] --- Fri Sep 30 2016