[Ffmpeg-devel-irc] ffmpeg.log.20150526

[burek]

[02:09:31 CEST] <jeffshanab> How can I build and install into a directory with internal name @rpath on OSX
[04:42:01 CEST] <brettk80> Hi all - wondering if someone could lend a hand with FFMpeg and clip trimming errors
[08:03:09 CEST] <chungy> Is there a method to dump/load chapters with ffmpeg? eg, to txt or xml files, or whatever else there is.
[08:03:40 CEST] <chungy> Also, perhaps metadata in general
[08:14:22 CEST] <chungy> looks like I want -f ffmetadata
[08:17:39 CEST] <chungy> Yea that'll work for scripting. Does there happen to be a way to replace metadata in-place (without needing to write out a whole new file)?
[08:26:09 CEST] <chungy> Hmm. doesn't dump the metadata of streams. bummer.
[10:17:36 CEST] <Filarius|2> Hello, is it posible to feed to ffmpeg both video and audio with same anon pipe? Some time ago I find how to convert raw video frames using anon pipe (frames created in realtime). Is here way, as example, to switch raw video frames and raw sound frames and make ffmpeg underdand it, like in data stream odd frames is video and even frames is audio ?
[10:20:10 CEST] <chungy> If your OS supports them, you can use named pipes (mkfifo)
[10:23:51 CEST] <JEEBsv> Filarius|2: you usually want to either completely separate them or use a container like NUT or so
[10:24:12 CEST] <JEEBsv> (NUT is what I generally hear as being used for raw audio/video data)
[10:26:19 CEST] <Filarius|2> OS is Win, and it support.  Thanks forreply, I know about it is possible with named pipes. But I curious if here simple way to push this data with single pipe
[10:27:17 CEST] <Filarius|2> oh, container. I wish something simple what I implement myself or there be code for language I use.
[10:27:29 CEST] <Filarius|2> *what I can implement
[10:39:28 CEST] <ocx> anyone familair with polycom video conf products? [11:37] <ocx> i am looking for an opensource product that is compatible with them [11:37] <ocx> so i have a polycom on one end and need to come up with someone opensource on the second end
[12:42:24 CEST] <xender> Can I cut AAC audio losslessly (without re-encoding)?
[12:43:01 CEST] <JEEBsv> yes
[12:43:06 CEST] <JEEBsv> -ss -t before -i
[12:43:13 CEST] <JEEBsv> after -i requires decoding
[12:44:07 CEST] <c_14> you'll also need -c:a copy (as an output option)
[12:44:36 CEST] <JEEBsv> yeah, I expected the person to know that in case of copying streams ;)
[12:45:52 CEST] <xender> I've done concatenating the streams with ffmpeg -f concat -i file-list.txt -c copy previously
[12:46:26 CEST] <c_14> Some people use -c:a aac because they want to keep it as aac. I just tend to add it. You never know.
[12:47:01 CEST] <xender> c_14: But that would mean re-encoding, wouldn't it?
[12:47:15 CEST] <c_14> Yes, it would.
[12:47:45 CEST] Action: c_14 was just referencing what some people use accidentally instead of copy when they actually want copy.
[12:47:46 CEST] <xender> And re-encoding is always lossy, unless I don't know something and it actually isn't?
[12:47:52 CEST] <xender> Oh
[12:48:38 CEST] <xender> Do decoders handle concatenated streams of different bitrate well?
[12:48:50 CEST] <xender> It's going to be played on some Windows probably
[12:49:30 CEST] <xender> Probably with codecs already included with the OS.
[12:49:40 CEST] <xender> Preparing stuff for non-tech people.
[12:50:07 CEST] <c_14> reencoding with a lossy codec is always lossy.
[12:50:34 CEST] <c_14> Most decoders should handle it well, it's basically just variable bitrate.
[12:52:04 CEST] <xender> Well, there's also matter of metadata. I suspect decoder won't rely on bitrate encoded in metadata then?
[12:53:20 CEST] <xender> Huh,so it seems that even concatenating different codecs can work? Interesting
[12:53:44 CEST] <xender> Well, I once accidentally concatenated stereo streams with mono streams, and decoder seemed to be a bit lost
[12:56:08 CEST] Action: c_14 isn't sure different codecs will work
[12:57:35 CEST] <c_14> Whether the decoder relies on metadata depends on the decoder. I'd assume most don't though. The metadata can lie.
[12:59:27 CEST] <chungy> Not sure AAC would play nice with that. Opus is one of the few audio codecs designed to change bitrates/channels mid-stream :P
[13:03:25 CEST] <xender> Heh, thank you for help, going out. \o
[13:29:17 CEST] <jeffshanab> `How can I build and install into a build directory with internal name @rpath on OSX instead of /usr/local/bin
[13:38:55 CEST] <BtbN> Just patch the rpath after building.
[13:40:58 CEST] <jeffshanab> That is what I was trying to avoid. I have done that in the past and most build systems have corrections for that now.
[13:41:34 CEST] <BtbN> rpath stuff on OSX is a horrible mess
[13:41:46 CEST] <BtbN> You can propably do something by passing custom c/ldflags
[13:42:52 CEST] <jeffshanab> I used to love apple products, but they decided they must changed things unneccesarily and now I hate including OSX in my builds
[13:50:17 CEST] <BtbN> Hm? the rpath nonsense has always existed on OSX
[13:50:28 CEST] <BtbN> beeing able to actualy use non-absolute paths for library is a quite recent addition
[13:50:36 CEST] <BtbN> +s
[17:15:59 CEST] <xender> Do I need -avoid_negative_ts when operating on audio only (AAC, m4a files)?
[18:14:32 CEST] <animax> hi all. http://www.pasteall.org/58665. why should I use this command (with frame rate factors '1' and '25')? if I do, each input frame gets duplicated 24 times. I use png sequences as input.
[18:15:59 CEST] <chungy> You should only need to specify the frame rate on one side (input or output) with -f image2.
[18:35:56 CEST] <jbmcg> This might be a dumb question, but is FFMPEG supposed to insert a frame at the beginning with a timestamp of 0 when re-encoding a video? Here's what I mean: http://pastebin.com/2MXdmzhf
[18:36:41 CEST] <jbmcg> When I compare the videos side by side in After Effects, it seems like everything is one frame off
[19:00:43 CEST] <Trenak> Hello, how convert from HD to SD?
[19:07:12 CEST] <JodaZ> chose a resolution and bitrate?
[19:11:07 CEST] <Trenak> ffmpeg -i SouthPark.avi -ac 2 -ab 256k -s 576x432 -f avi -qscale 5 SouthPark.avi right??
[19:15:45 CEST] <chungy> You should only need "ffmpeg -i southpark-hd.avi southpark-sd.avi"
[19:15:59 CEST] <chungy> er, with the "-s 576x432" in the middle too
[19:17:59 CEST] <Trenak> ok thanks
[19:22:45 CEST] <Trenak> but this lose quality as fuck
[21:33:00 CEST] <jesseg> Where is CONFIG_ALSA_INDEV supposed to be defined and is it always defined?
[21:36:21 CEST] <jesseg> oh never mind I be in wrong channel..!
[22:11:30 CEST] <realies> hi
[22:13:03 CEST] <realies> anyone with experience streaming audio from ffmpeg to icecast?
[22:13:31 CEST] <realies> I'm getting an End of file error
[22:14:55 CEST] <realies> http://pastebin.com/6HryHsNS
[23:51:21 CEST] <moontails> Hi guys!
[23:52:01 CEST] <moontails> I was wondering if anyone can give me pointers on how to extract volume levels for each frame in a video
[23:56:05 CEST] <llogan> moontails: maybe you can hack astats filter do give you want you want
[23:58:13 CEST] <realies> turned out I need -legacy_icecast :)
[00:00:00 CEST] --- Wed May 27 2015