[00:17:44 CET] <hanna> eh I was just interested for the purpose of using this in mpv's vo_gpu [00:17:52 CET] <JEEB> yea [00:18:10 CET] <JEEB> as far as I know there's no proper scenechange detection stuff in FFmpeg [00:18:16 CET] <JEEB> well, as "proper" as such things are [00:18:33 CET] <JEEB> in the subtitling world what I linked used to be the general level [00:20:08 CET] <JEEB> hanna: basically you could have seen what sort of algo was used, and then seen during playback when a "scene change" was hit [00:23:42 CET] <paule32> hello, need help [00:23:43 CET] <paule32> https://pastebin.com/VhEFjUv8 [00:24:14 CET] <JEEB> paule32: you need to specify the output container [00:24:38 CET] <JEEB> it can't guess it by the "ffn" extension [00:28:53 CET] <paule32> JEEB: thx [00:28:55 CET] <paule32> but [00:28:57 CET] <paule32> sudo ./ffmpeg -f x11grab -r 5 -s 1024x720 -i :0.0 -preset ultrafast -f mp3 http://localhost:8181/feed1.ffn [00:29:03 CET] <paule32> give me next errors [00:29:24 CET] <JEEB> you're trying to put video into mpeg-1 layer 3? :) [00:29:45 CET] <paule32> ehm, yes [00:29:48 CET] <paule32> stupid [00:29:51 CET] <paule32> ffserver [00:29:59 CET] <paule32> directive? [00:30:26 CET] <BtbN> ffserver is dead, don't use it [00:42:44 CET] <paule32> JEEB: https://pastebin.com/c3L3krzC [00:43:07 CET] <JEEB> still [00:43:11 CET] <JEEB> you're taking in video, and outputting mp3 [00:43:20 CET] <JEEB> mp3 cannot have any video so it tells you there's no streams [00:44:32 CET] <paule32> and why do i can watch over the half internet video.mp3 ? [00:48:10 CET] <c_14> you're probably thinking of mp4 [00:57:01 CET] <paule32> https://pastebin.com/sK0CsnL9 [00:57:14 CET] <paule32> i get other error's, now [00:57:47 CET] <JEEB> because non-fragmented mp4 doesn't work [00:57:51 CET] <JEEB> for streaming [00:58:19 CET] <JEEB> you probably wanto -movflags [00:58:29 CET] <JEEB> -movflags frag_keyframe [00:58:34 CET] <JEEB> for fragmented mp4 [00:58:48 CET] <JEEB> but really, if you want to do web streaming then just use nginx-rtmp [00:59:02 CET] <JEEB> and then feed rtmp to it and you can get HLS/DASH [01:03:02 CET] <kerio> you can get hls with just ffmpeg too, tho [01:03:12 CET] <JEEB> sure [01:03:38 CET] <JEEB> I guess you can even do the shared fragments thing with HLS and DASH now? [01:03:43 CET] <JEEB> no idea how well it works, though ^^; [01:04:36 CET] <kerio> lmao dash [01:04:57 CET] <JEEB> also I should patch the HLS master playlist generator to also take maxrate into mention :P [01:05:07 CET] <JEEB> currently it requires one to set bit_rate [01:10:40 CET] <paule32> so, ther server+client run [01:10:44 CET] <paule32> but [01:10:46 CET] <paule32> av_interleaved_write_frame(): Broken pipe time=00:00:49.40 bitrate= 0.0kbits/s speed=0.992x [01:10:46 CET] <paule32> Error writing trailer of http://localhost:8181/feed1.ffm: Broken pipe 0.0kbits/s speed=0.992x [01:10:57 CET] <JEEB> your server gave up [01:10:58 CET] <paule32> the error message above [01:11:09 CET] <paule32> to low memory? [01:11:29 CET] <paule32> of MaxFileSize=5M [01:11:32 CET] <paule32> ? [01:11:36 CET] <JEEB> I have no idea [01:11:57 CET] <JEEB> if that's ffserver then you will unfortunately not get any support on it here [01:12:06 CET] <JEEB> it was a long-discussed thing that nobody was maintaining [01:12:10 CET] <JEEB> so it finally got removed some time ago [01:12:25 CET] <paule32> ok [01:12:37 CET] <kerio> put ffmpeg in cgi-bin and use mjpeg [01:12:39 CET] <kerio> ez [01:12:41 CET] <JEEB> as I said, if you just want to do streaming that you can point a browser to, nginx-rtmp is probably the way to go (or just plain nginx with file PUTs working with HLS) [01:13:30 CET] <paule32> then i have to leave apache? [01:13:42 CET] <kerio> well [01:13:49 CET] <kerio> i mean [01:13:51 CET] <JEEB> you can do the HLS post thing with apache too ,sure [01:13:53 CET] <kerio> it is 2018 [01:14:09 CET] <JEEB> but I bet yhou will get less help with configuring that :) [01:14:23 CET] <kerio> JEEB: nginx-rtmp doesn't even do hls by itself anyway [01:14:28 CET] <kerio> it just puts fragments in a directoryt [01:14:28 CET] <JEEB> it does [01:14:40 CET] <JEEB> at least the README says it does [01:14:41 CET] <kerio> and then it's your responsibility to host them [01:14:46 CET] <kerio> yea sure but [01:15:07 CET] <JEEB> basically looked at https://github.com/sergey-dryabzhinsky/nginx-rtmp-module/ [01:15:20 CET] <furq> if only nginx-rtmp had some kind of http server [01:15:22 CET] <furq> but it doesn't [01:15:31 CET] <kerio> furq: but paule32 wants to use apache [01:15:33 CET] <kerio> :\ [01:15:41 CET] <kerio> no idea why [01:15:44 CET] <furq> rip [01:15:53 CET] <kerio> ikr [01:16:15 CET] <JEEB> well effectively if it generates the files then just making a webroot for that I think is good enough? [01:16:25 CET] <JEEB> I don't really see the issue here other than it doesn't have its own buffer format [01:16:30 CET] <kerio> but anyway, there's nothing simpler in using nginx to host the hls files generated by nginx-rtmp rather than using apache [01:16:33 CET] <JEEB> from which it generates those things on the fly [01:16:40 CET] <kerio> other than the fact that nginx is better than apache, that is [01:16:54 CET] <JEEB> sure [01:17:14 CET] <kerio> nginx-rtmp is neato [01:20:08 CET] <BtbN> kerio, nginx-rtmp is still an nginx, so hosting http is kind of what it does anyway. [01:36:14 CET] <paule32> there is somewhere wrong [01:36:17 CET] <paule32> https://pastebin.com/cQnAAps8 [01:41:24 CET] <JEEB> whatever you're feeding to broke the write [01:41:42 CET] <JEEB> I have no idea what you have behind that HTTP end point [01:42:20 CET] <paule32> desktop stream [02:34:41 CET] <n0xlf> furq you still around? [02:43:57 CET] <aphirst> tomorrow i need to fix the audio on some files [02:44:07 CET] <aphirst> the L audio track is 0.02 s earlier than the R audio track [02:44:10 CET] <aphirst> and it makes it sound awful [06:00:38 CET] <stockstandard> Hey everyone, could someone please help me diagnose this issue? Unrecognized option 'max_muxing_queue_size'. Error splitting the argument list: Option not found [06:01:12 CET] <stockstandard> I am trying to get this script running and getting that error: https://github.com/Collisionc/sickbeard_mp4_automator [06:01:26 CET] <stockstandard> Not sure which parameter that is coming from though because i dont believe I am passing anything called max_muxing_queue_size [11:39:31 CET] <gusto> hi [11:40:03 CET] <gusto> someone an idea why vc2 is so slow? normally I have no problems, but I tried encoding x11grab fullhd 60fps with vc2 and it gives me like 30% of it [11:41:20 CET] <gusto> libx264 performs much better giving me almost 60 at default settings and CPU utilisation is about 30% [11:54:33 CET] <atomnuker> no simd [13:27:49 CET] <paule32> hello, need help [13:27:50 CET] <paule32> https://pastebin.com/dx4wGyWC [13:28:20 CET] <paule32> the stream seem to be streamed, but i can't see picture/video in vlc [13:28:50 CET] <paule32> when i open network stream: rtp://127.0.0.1:8182/test1.mp4 [13:35:43 CET] <eenp^> notice the SDP output from ffmpeg? [13:36:06 CET] <eenp^> i.e. for yours, this is it: https://pastebin.com/eFA7dhrU [13:36:26 CET] <eenp^> paste that into a file, named say, f.sdp [13:36:43 CET] <eenp^> now, open that .sdp file in vlc - appears to work well [13:36:49 CET] <eenp^> @paule32 [13:38:43 CET] <eenp^> also, ffplay should be able to view it too: `ffplay -protocol_whitelist file,udp,rtp -i ./f.sdp` [13:40:32 CET] <paule32> thank you [13:42:05 CET] <paule32> but [13:42:24 CET] <paule32> how is ffplay know about the stream? [13:42:41 CET] <paule32> lets assume, when i have 2 streams/feeds ? [13:42:45 CET] <eenp^> the .sdp file contains info about where its located, and some other protocol nonsense [13:42:58 CET] <eenp^> have a gander inside the sdp output [13:44:27 CET] <paule32> i can see o,c = ip4 127.0.0.1 [13:44:35 CET] <paule32> and port [13:44:44 CET] <paule32> that's all? [13:45:52 CET] <eenp^> line 7 specifies the port [13:46:08 CET] <eenp^> other than that - don't know too much about the protocol [13:50:28 CET] <paule32> cool [13:50:32 CET] <paule32> very fast [13:50:40 CET] <paule32> at 320x200 [13:52:34 CET] <eenp^> sweet [13:58:02 CET] <eenp^> oh btw - not sure if anyone was here from when I was asking about why ffmpeg couldn't transcode some files (with an error msg of `[ogg @ 0x5564abcbb7e0] Codec not found`) [13:58:40 CET] <eenp^> turns out the .ogg had a begin of stream and end of stream at the head of the file, which rogg_pagedump dutifully reported [13:58:56 CET] <eenp^> just ignoring the first 147 bytes or whatever fixed it [14:02:27 CET] <cableguy> hello team [14:02:38 CET] <durandal_1707> hello [14:02:43 CET] <cableguy> how do i loop video to the length of the audio [14:02:59 CET] <cableguy> i found same question from 5 months ago https://video.stackexchange.com/questions/22301/loop-webm-video-for-the-whol... and nobody replied [14:03:19 CET] <durandal_1707> stream-loop [14:04:57 CET] <cableguy> you mean -stream_loop [14:10:13 CET] <cableguy> im using this [14:10:14 CET] <cableguy> ffmpeg -y -loop 1 -r 6 -i 1.jpg -i 1.mp3 -shortest -c:v libx264 -preset ultrafast -tune stillimage -pix_fmt yuv420p -c:a copy video.mp4 [14:10:22 CET] <cableguy> to loop an image to audio [14:10:31 CET] <durandal_1707> yes [14:10:40 CET] <cableguy> but it does not work for video instead of image [14:10:41 CET] <durandal_1707> for images [14:12:32 CET] <cableguy> which switch do i change [14:20:50 CET] <paule32> hello [14:20:58 CET] <paule32> when i start ffmpeg: [14:21:01 CET] <paule32> sudo ./ffmpeg -f x11grab -r 5 -s 1000x600 -i :0.0 -preset ultrafast -f rtp rtp://[2a00:c1a0:8502:2800:1000:1000:1000:10]:8182/test1.mp4 [14:21:05 CET] <paule32> i get: [14:21:23 CET] <paule32> c=IN IP4 2a00:c1a0:8502:2800:1000:1000:1 [14:21:26 CET] <paule32> for sdp [14:21:36 CET] <paule32> see: IP4 [14:21:46 CET] <JEEB> sounds like a bug in the code that writes out the SDP manifest [14:21:54 CET] <JEEB> although you'd have to look at the specification for it [14:21:58 CET] <JEEB> if there even is an IP6 [14:23:03 CET] <paule32> for http host, i have to write (in browser): http://[2a00:c1a0:8502:2800:1000:1000:1]/index.html [14:35:14 CET] <BtbN> Why do you run ffmpeg as root? [14:36:43 CET] <artha> hi can anyone tell me if I can change the filter arguments of ffmpeg in runtime? [14:37:49 CET] <BtbN> some of them, usually very limited, probably not the ones you want to [14:39:29 CET] <artha> @BtbN, I am trying to to use the equalizer filter with time varying arguments while playing the audio. is it possible/ [14:40:06 CET] <BtbN> which equalizer? [14:40:27 CET] <BtbN> There seem to be 3 [14:40:39 CET] <artha> ffplay -af "equalizer=f=1000:width_type=h:width=200:g=-10" -i foo.wav [14:40:50 CET] <artha> this is the command [14:40:51 CET] <BtbN> no idea which of the 3 equalizers that maps to, hm [14:41:51 CET] <furq> use mpv? [14:41:58 CET] <furq> you can definitely change mpv's eq at runtime [14:42:10 CET] <furq> and also it's an actual media player and not a proof of concept [14:42:55 CET] <artha> ok I'll try that. thanks [14:43:19 CET] <BtbN> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavfilter/af_biquads.c;h=d5... it does support some runtime reconfig [14:45:07 CET] <artha> @BtbN, great this is what I was actually looking for. [14:45:35 CET] <BtbN> I'm not 100% sure how to send those commands when using ffmpeg.c. But I think there is some way? [14:46:10 CET] <BtbN> Also, there are aparently 4 equalizer filters in ffmpeg. [14:46:51 CET] <artha> maybe I need to create a callback function. check for changed arguments before the buffer is created [14:47:04 CET] <BtbN> if you are using the API, it's not a problem at all [14:47:19 CET] <BtbN> there is a function to send those commands [14:47:39 CET] <furq> you can drop to an interactive prompt to send commands to sendcmd, but i've never been able to figure out how to operate it [14:47:46 CET] <furq> and also it'll pause the output while you're at the prompt [14:48:06 CET] <furq> you could also potentially use the zmq filter [14:49:23 CET] <artha> Yeah. I dont want the audio to pause. And zmq filter?? I will check that out too [14:49:41 CET] <BtbN> Are you using the ffmpeg cli, or the libav* libraries directly? [14:51:09 CET] <artha> till now I have been using the cli. But I got nowhere with it. Now I want to try the libavfilter [14:51:34 CET] <BtbN> yeah, if you're using the API there will be no extra pause [14:52:00 CET] <BtbN> And if you're using the ffmpeg cli in some "embedded" way in some other application you can also send commands pretty much instantaneously. [14:52:04 CET] <artha> I am using python. so now I have to bind the lib to be able to use it' [14:52:12 CET] <BtbN> But to me it sounds like you're better off using the API [14:53:32 CET] <artha> I guess however instantaneous i send them, there will be an observable pause in the stream wont it?? [14:53:44 CET] <artha> if I use cli [14:54:15 CET] <BtbN> I doubt it. Will take a few ms at worst [14:55:56 CET] <artha> can you recommend a good cli for it? [14:56:34 CET] <BtbN> hm? [14:56:43 CET] <BtbN> ffmpeg.c is the cli application [14:57:15 CET] <artha> I meant for python [14:58:43 CET] <BtbN> I'm not sure what you mean [15:01:19 CET] <artha> no prob. I will check out the process_command() api and zmq filter furq suggested. thanks for helping out [15:58:33 CET] <IsntFunny> Is there some quick way to strip out any audio track without encoding the full video? [15:58:49 CET] <JEEB> yes, just map only video [15:58:58 CET] <JEEB> -map v [15:59:07 CET] <IsntFunny> and -c copy? [15:59:09 CET] <JEEB> this will map all the video track(s) from all inputs [15:59:15 CET] <JEEB> which usually is one, but yunno [15:59:27 CET] <BtbN> -c:v copy -an -sn [15:59:53 CET] <IsntFunny> welp, awesome. thanks :) [16:00:05 CET] <JEEB> oh, right -an -sn is the other way [16:04:00 CET] <IsntFunny> damn [16:04:05 CET] <IsntFunny> now i messed up all of my footage [16:04:32 CET] <IsntFunny> And i of course have no backup for those [16:09:31 CET] <IsntFunny> ffmpeg doesn't have some kind of temp dir where i could try to rescue my files? [16:09:57 CET] <JEEB> no [16:10:06 CET] <DHE> did you set your input and output as the same file? [16:10:08 CET] <JEEB> I mean, if you set your output as input then rip [16:10:24 CET] <IsntFunny> DHE yep [16:10:35 CET] <IsntFunny> All my drone footage gone :D [16:10:55 CET] <DHE> nothing on other media, like the drone itself maybe? [16:12:44 CET] <IsntFunny> nope, i always move instead of copy. A lot of data and just 64gb of sd card... Even for backups it'd be kind of big. Aw snap [16:13:13 CET] <IsntFunny> Well, lesson learned [17:01:46 CET] <Fyr> guys, is there a way to optimize matroska container besides -reserve_index_space? [17:02:24 CET] <Fyr> mkclean produces smaller files than FFMPEG. [17:03:21 CET] <Fyr> it means the authors know something. [17:03:27 CET] <Fyr> (= [17:03:38 CET] <durandal_1707> by how much? [17:04:09 CET] <Fyr> not so much [17:05:25 CET] <Fyr> but the thing about index space should be optimized, like in MP4 files. for now I have to mux it, then remux it, since I don't know how much space is required to write the index. [17:05:46 CET] <Fyr> then muxing into MP4, FFMPEG mux it and remuxes it automagically. [17:06:18 CET] <Fyr> ='( [17:06:54 CET] <durandal_1707> open feature request? [17:07:14 CET] <next1> close [17:08:58 CET] <classssic> hi, somebody know how to decoder h264 to MJPEG using vaapi hw accel? [17:10:41 CET] <atomnuker> jkqxz has a patch but he's waiting on durandal_1707's patches [17:11:51 CET] <durandal_1707> how so? look im volunter [17:18:51 CET] <atomnuker> the yuvj removal [17:19:03 CET] <atomnuker> cmon, you said its all fixed [17:19:17 CET] <atomnuker> just resubmit [17:33:06 CET] <stockstandard> Hey everyone, could someone please help me diagnose this issue? I'm Not sure which parameter that is coming from because i dont believe I am passing anything called max_muxing_queue_size [17:33:10 CET] <stockstandard> Unrecognized option 'max_muxing_queue_size'. Error splitting the argument list: Option not found. I am trying to get this script running and getting that error: https://github.com/Collisionc/sickbeard_mp4_automator [17:36:11 CET] <sfan5> https://github.com/Collisionc/sickbeard_mp4_automator/blob/master/mkvtomp4.p... [17:36:22 CET] <sfan5> application devs fault [17:42:08 CET] <stockstandard> SFAM5 what exactly is the issue? [17:43:13 CET] <stockstandard> sfan5* [17:43:40 CET] <sfan5> the application developer is making the assumption that the used ffmpeg version supports -max_muxing_queue_size [17:43:57 CET] <sfan5> this assumption isn't holding, which is why you're getting that error [17:44:35 CET] <stockstandard> is that a newer option or something that has been depricated? [17:45:31 CET] <sfan5> former [17:45:43 CET] <stockstandard> Right now I am just using 'apt-get ffmpeg' in my docker create, is there a version i should be passing? [17:46:00 CET] <sfan5> debian-based distributions often ship outdated ffmpeg versions [17:46:25 CET] <sfan5> if you read the README of that repo, the author suggests downloading ffmpeg from https://ffmpeg.zeranoe.com/builds/ [17:46:34 CET] <furq> that'll work well on linux [17:46:49 CET] <furq> stockstandard: https://www.johnvansickle.com/ffmpeg/ [17:47:18 CET] <furq> with that said debian stable has 3.2.10 and testing has 3.4.1 [17:47:40 CET] <furq> which should be new enough [17:48:20 CET] <stockstandard> yeah i just checked, apt-get is pulling version 2.8.11 [17:48:30 CET] <furq> is that ubuntu 16 or something [17:48:31 CET] <stockstandard> ubuntu [17:48:33 CET] <furq> right [17:48:36 CET] <furq> get a newer distro [17:48:45 CET] <furq> or just use those static builds [17:50:32 CET] <stockstandard_> Is there a way to reference the newer version in the dockerfile? [17:50:34 CET] <stockstandard_> https://pastebin.com/BC1waExt [17:52:01 CET] <sfan5> extract it to /usr/local or something [17:55:03 CET] <stockstandard_> apologies sfan5 - I am fairly new to all of this... any examples you could show on how to do that [17:56:48 CET] <sfan5> tar -xf ffmpeg-release-64bit-static.tar.xz -C /usr/local/bin --strip-components=1 ffmpeg-3.4.1-64bit-static/ffmpeg [17:58:31 CET] <stockstandard_> like this? https://pastebin.com/MZwgGShR [17:59:59 CET] <sfan5> you'd need to download the file first, but yes [18:02:34 CET] <stockstandard_> Is there a way to do that in the dockerfile? I have a separate cron job that re-flashes the docker image every week so ideally there would be no manual downloading involved [18:03:22 CET] <sfan5> ...just invoke wget [19:12:28 CET] <whytho> Hi! I'm trying to use the ffmpeg libs in a project but I am running into problems. Basically what I'm trying to do at this point is to be able to play both video and audio from the same file. So far I can play video, that is not a problem. Playing the audio of the file at the sime time though, I just can't get it to work. [19:12:58 CET] <whytho> I've looked at drangers tutorial but it's outdated, I've looked at lei xiaohua's tutorials but all of them seem to be audio or video, not audio and video. I also found a github repo that claims to be working for ffmpeg >=3.2.4 but same there, no sound is playing [19:13:49 CET] <paule32> hello [19:14:12 CET] <paule32> my samsung smartphone can't open stream.sdp in chrome browser [19:14:17 CET] <whytho> What I'm asking here is, can anyone point me in the direction of a guide that shows how to play video and audio from the same video file for ffmpeg 3.4? Preferably using SDL2, but hopefully I can manage that last bit by myself worst comes to worst [19:14:20 CET] <paule32> what can i do? [19:14:24 CET] <whytho> I'm using windows 10 BTW [19:17:09 CET] <JEEB> whytho: the proof-of-concept ffplay thing uses SDL2 [19:17:18 CET] <JEEB> so you could say it's an "example" [19:23:04 CET] <whytho> I will check that out, thank you! [19:25:42 CET] <ans_ashkan> Hi all, I'm trying to overlay a sequence of images on top of a video and loop images. this is what I use (ffmpeg -i hq26368874313361335.mp4 -i "Seq\seq_%5d.png" -filter_complex "[1:v]loop=-1:size=496[top];[0:v][top]overlay=shortest=1" temp.ts) [19:26:33 CET] <ans_ashkan> the problem is loop starts at about frame 400 at second loop [19:27:27 CET] <ans_ashkan> any idea what is wrong with this? [19:36:53 CET] <kepstin> ans_ashkan: probably a framerate mismatch, you should specify the "-framerate" input option on the png input to the same as the framerate of the video [19:37:02 CET] <kepstin> if you don't specify, it's 25fps, iirc. [19:39:15 CET] <paule32> hi JEEB [19:39:21 CET] <paule32> have you a idea? [19:42:58 CET] <ans_ashkan> kepstin: video's framerate is 25 too, so they match. [23:13:05 CET] <paule32> hello, need help [23:13:07 CET] <paule32> https://pastebin.com/5nPnG5ir [23:13:27 CET] <paule32> i would like stream webm or mp3 [23:13:38 CET] <paule32> firefox does not support mp4 [23:13:43 CET] <paule32> that the problem [23:13:53 CET] <paule32> but i think mp3 does support [23:14:49 CET] <notashark> is it normal for ffmpeg to mangle variable values without assignment? [23:15:17 CET] <paule32> me? [23:15:52 CET] <notashark> anyone [23:16:12 CET] <notashark> paule32: I can't help, sorry. [23:16:16 CET] <paule32> "-video_size" is the same as "-s [23:16:45 CET] <paule32> notashark: what do you stream? [23:17:00 CET] <notashark> I don't stream anything? [23:17:18 CET] <paule32> transcode? [23:18:22 CET] <notashark> my use of ffmpeg is just basic conversion; like, wav to flac. I don't use any of the really complicated parts [23:18:38 CET] <paule32> ok [23:24:28 CET] <furq> notashark: i don't know what you're actually asking [23:33:10 CET] <notashark> furq: i have a short (~30loc) script to batch process files. commenting out the ffmpeg line lots the file names get printed just fine. if I uncomment ffmpeg, they are mangled; this is before and after ffmpeg is called [23:33:25 CET] <notashark> I can only think it's the way the shell is parsing it before even running it, but idk [23:33:38 CET] <furq> probably some kind of quoting nightmare [23:33:41 CET] <furq> but pastebin the script anyway [23:38:46 CET] <notashark> https://transfer.sh/kg04F/ffmpeg.sh [23:58:48 CET] <sfan5> doesn't ${TRACK/\.original/} make use of a bash-only feature? [23:58:51 CET] <sfan5> since you're using /bin/sh [00:00:00 CET] --- Mon Feb 12 2018