[Ffmpeg-devel-irc] ffmpeg.log.20180428

Sun Apr 29 03:05:01 EEST 2018

[00:11:40 CEST] <rindolf> furq: ping
[00:21:22 CEST] <rindolf> Lite: <rindolf> furq: see http://www.shlomifish.org/Files/files/text/f.txt ==> do you know how to fix it?
[00:24:00 CEST] <Lite> Unfortunely not. I am facing such not so much crutucal 'errors' connected with DTS but never in real faced that output file was incrorrect/wrong
[00:24:36 CEST] <Lite> I have a question too
[00:24:57 CEST] <Lite> to these guys who knows my issue I provided yesterday in detail
[00:25:13 CEST] <Lite> with HTTPS URLs
[00:25:45 CEST] <Lite> I just deleted all cryptography software but that's not affected
[00:25:47 CEST] <Lite> but
[00:26:47 CEST] <Lite> there switched on totally all possible to switch on, you know, Windows 10 OS components
[00:26:59 CEST] <Lite> They included OpenSSL
[00:27:49 CEST] <Lite> Could it be faced some, you know incompatibility between latest Nightly builds and OpenSSL?
[00:28:15 CEST] <Lite> Which was not before released version 02.11.2017?
[00:28:20 CEST] <Lite> ???
[00:36:18 CEST] <kepstin> you're talking about openssh?
[00:36:23 CEST] <kepstin> completely unrelated
[00:36:29 CEST] <Lite> Ye, ye
[00:36:43 CEST] <Lite> should be ssh
[00:36:44 CEST] <kepstin> (the version of openssh shipped in the windows components doesn't use openssl, it uses windows builtin crypto)
[00:36:55 CEST] <Lite> wait no
[00:36:56 CEST] <Lite> SSL
[00:36:59 CEST] <Lite> OpenSSL
[00:37:43 CEST] <Lite> Or SSH :D
[00:37:46 CEST] <Lite> dunno
[00:37:49 CEST] <Lite> :'C
[00:38:30 CEST] <Lite> But SSL installed at my PC too definately
[00:51:30 CEST] <rindolf> good night, all
[00:54:25 CEST] <Guma> I am looking for advise and direction for my hobby project.  I have application that I am reading images from IP camera. I have few of them to experiment and they provide different formats. Some are in raw format. What I would like to do is write service in C++ that I can connect with web browser and see image stream from remote camera. Since such cameras might be accessed over 4G LTE, speeds might very. Is anyone here who has done this can explain some
[00:54:25 CEST] <Guma>  general directions where to start? Maybe some sample code?
[00:55:25 CEST] <Mavrik> Hmm, for starters, you probably don't need to write the service yourself at all. Would you consider a more script oriented approach? :)
[00:57:25 CEST] <Guma> Mavrik I might. What scripting are you talking about? But I would prefer C++ since it is my language of choice
[00:57:49 CEST] <Mavrik> Guma, basically for your use-case you probably don't have to write code at all
[00:58:04 CEST] <Mavrik> You can just use ffmpeg directly to transcode streams and pass them on to something like nginx-rtmp-module
[00:58:21 CEST] <Mavrik> Which acts as a streaming server for web and then use video.js or something to show it in the browser
[00:58:46 CEST] <Mavrik> If you know nothing about how digital video is structured you'll have a pretty hard time with ffmpeg APIs
[00:59:07 CEST] <Guma> Mavrik So I have already C++ library that is reading frames from camera (USB or IP) so how could I feed such frame directly to ffmpeg?
[00:59:35 CEST] <Guma> Mavrik I do not might learning API. I got time :)
[00:59:46 CEST] <Lite> Is there anybody who uses Fiddler?
[01:00:08 CEST] <Guma> I just want to know where to start to be as much as efficient with my learning time
[01:01:02 CEST] <Mavrik> Guma, ad
[01:01:04 CEST] <Mavrik> *ah
[01:01:06 CEST] <Guma> I see there is ffserver. This sounds something like I would like to do
[01:01:13 CEST] <Mavrik> Guma, yeah, you'd have to look at libav API
[01:01:18 CEST] <Mavrik> ffserver is dead, don't touch that :P
[01:01:21 CEST] <Guma> I am doing some processing in OpenCV first then publish
[01:01:42 CEST] <Mavrik> basically you'll have to use libavcodec to encode frames into a video stream
[01:01:48 CEST] <Mavrik> and then libavformat to mux it into a container
[01:01:58 CEST] <Mavrik> then send it to nginx-rtmp-module or equivalent streaming server
[01:02:10 CEST] <Mavrik> (ffmpeg / libavformat knows how to do that for you)
[01:02:31 CEST] <Mavrik> I'd try first setting up a pipeline with plain ffmpeg and static file so you get the web part working
[01:02:37 CEST] <Mavrik> And then start writing C++ :)
[01:02:39 CEST] <Lite> Just strange...after I installed Cryptographic software and setted ip how it should working in 2017 (I used also Fiddler before and didn't see such request), it starts to show some csp/ocsp requests
[01:02:51 CEST] <Mavrik> Guma, there's an example folder in ffmpeg source :)
[01:03:01 CEST] <Guma> Mavrik do you know of any decent sample tutorial code so I can start go over and read API docs?
[01:03:15 CEST] <Guma> You just did beat me with answer
[01:03:21 CEST] <Lite> user agent there in request headers shows as 'Microsoft-CryptoAPI/10.0'
[01:03:28 CEST] <Mavrik> Sadly the documentation isn't that stellar
[01:03:46 CEST] <Guma> Ok I will go download the source and look there.
[01:03:49 CEST] <Mavrik> Guma, this helped me a lot as well: https://www.ffmpeg.org/doxygen/3.4/index.html
[01:03:58 CEST] <Lite> It still continue. Even after uninstalling all cryptographic software from machine
[01:03:59 CEST] <Guma> So if I have any question is this best channel to ask?
[01:04:13 CEST] <Mavrik> It's as good as it gets I think :P
[01:04:52 CEST] <Lite> Could this sh*t 'Microsoft-CryptoAPI/10.0' affect my issue????! Untill I won't switch it off?
[01:05:08 CEST] <Lite> kepstin?
[01:05:11 CEST] <Guma> Sweet. Thank you. Will start there
[01:08:41 CEST] <Guma> Mavrik I was wondering if I build all libs from source at some point for specific hardware is there any speed gain? I am on Ubuntu 16.04. Is the latest packages good or there is a benefit to use latest and build from source.
[01:09:15 CEST] <Mavrik> Hmm, the bigger concern is that 16.04 probably has stone age ffmpeg in repos
[01:09:35 CEST] <Mavrik> Performance won't differ much because pretty much all hot codepaths are handtuned ASM for popular codecs
[01:09:41 CEST] <Mavrik> Which is loaded according to CPU capabilities
[01:09:53 CEST] <Guma> Ok so building from source will be first step
[01:09:55 CEST] <Mavrik> So you'd only be optimizing most of the code that doesn't really do any CPU work :P
[01:10:05 CEST] <Mavrik> But yeah, I'd suggest you build from source
[01:10:18 CEST] <Mavrik> And while developing link against ffmpeg with debug symbols enabled and optimizations disabled
[01:10:23 CEST] <Mavrik> So you can use gdb to step through ffmpeg code as well
[01:10:25 CEST] <Guma> I was thinking like enabling SSE4 or anything else specific to my hardware.
[01:10:58 CEST] <Guma> I might worry about this later but I am just asking for future.
[01:11:00 CEST] <Mavrik> As I said, the places where that helps already have AVX etc. code written
[01:11:07 CEST] <Mavrik> And will use it if they detect you have a CPU capable of that.
[01:11:12 CEST] <Mavrik> Other places aren't doing anything CPU intensive :)
[01:12:43 CEST] <Guma> ok. Thank you for you help and time. Talk soon
[04:09:44 CEST] <oerg866> Problem: -f concat with multiple VOB files with MPEG2 + PCM audio streams, with -c copy into a new VOB or MPG files, audio is unplayable. MPC-HC detects it as "MPEG audio 11025 hz" instead of PCM.
[04:10:45 CEST] <oerg866> https://i.snag.gy/G8kmn3.jpg https://i.snag.gy/hXJ6OR.jpg
[04:11:54 CEST] <oerg866> command: ffmpeg -f concat -safe 0 -i list.txt -codec copy output.mpg (or vob)
[04:22:01 CEST] <oerg866> "
[04:22:01 CEST] <oerg866> At this time, ffmpeg does not write the header for LPCM streams in MPEG-1/2 when streamcopying."
[04:22:05 CEST] <oerg866> is this a bug????
[10:29:18 CEST] <rindolf> hi all
[10:41:00 CEST] <rindolf> furq: hi! See http://www.shlomifish.org/Files/files/text/f.txt
[10:48:10 CEST] <rindolf> Hi all! I have an issue with overlaying audio and video - see https://github.com/shlomif/pysolfc-black-hole-solver--screencast and http://www.shlomifish.org/Files/files/text/f.txt
[11:22:27 CEST] <ChocolateArmpits> rindolf, have you tried another version?
[11:23:13 CEST] <rindolf> ChocolateArmpits: what for?
[11:23:21 CEST] <ChocolateArmpits> so see if that works
[11:24:07 CEST] <rindolf> ChocolateArmpits: no
[11:24:31 CEST] <rindolf> ChocolateArmpits: ok, let me build 3.x.y
[11:39:29 CEST] <helmo_> hello there, sorry for probably newbie question, is it normal to get THIS.. "WARNING: Building with deprecated library libavresample"  and if not how to solve it? am i missing something?
[11:40:13 CEST] <helmo_> its appeared now when trying v 4.0
[11:41:01 CEST] <helmo_> and obviously fails the build any hintt?
[11:41:03 CEST] <durandal_1707> helmo_: do not build it and do not use it
[11:42:48 CEST] <helmo_> @durandal_1707  ???
[11:50:02 CEST] <rindolf> ChocolateArmpits: same problem with 3.4.2
[11:51:09 CEST] <ChocolateArmpits> rindolf, have you tried omiting parameters like -shortest or the timestamp filter ?
[11:51:13 CEST] <helmo_>  3.4.2 for me builds just fine its only v 4 which seem have this issue
[11:51:59 CEST] <rindolf> ChocolateArmpits: which timestamp filter?
[11:52:11 CEST] <ChocolateArmpits> asetpts=PTS-STARTPTS
[11:52:58 CEST] <ChocolateArmpits> I don't see why you would need to create timestamps like this
[11:53:04 CEST] <durandal_1707> helmo_: post your full configure line
[11:55:31 CEST] <helmo_> @durandal_1707 https://pastebin.com/fi9MtKdn
[11:55:32 CEST] <rindolf> ChocolateArmpits: https://github.com/shlomif/pysolfc-black-hole-solver--screencast/blob/master/old-mix.bash - this is with ffmpeg 3.4.2
[11:57:20 CEST] <ChocolateArmpits> rindolf, and this fails too right ?
[11:57:40 CEST] <rindolf> ChocolateArmpits: i get a zero length file
[12:04:07 CEST] <ChocolateArmpits> rindolf, well best I can suggest is to rewrap each file separately to see if there's an offender
[12:04:25 CEST] <ChocolateArmpits> or remove shortest and set the target length for the audio input directly
[12:04:55 CEST] <ChocolateArmpits> via the duration input option -t
[12:04:58 CEST] <rindolf> ChocolateArmpits: wo -shortest i only get thhe first frame of the video
[12:05:18 CEST] <rindolf> with the full audio
[12:05:47 CEST] <ChocolateArmpits> try rewrapping that one separately
[12:06:30 CEST] <rindolf> ChocolateArmpits: what is rewrapping?
[12:06:42 CEST] <ChocolateArmpits> remuxing lol
[12:07:01 CEST] <rindolf> ChocolateArmpits: what?
[12:07:17 CEST] <ChocolateArmpits> pick the ogv file and do a -c copy just with it
[12:07:38 CEST] <ChocolateArmpits> see if it plays back afterwards
[12:10:27 CEST] <rindolf> ChocolateArmpits: it has no video
[12:11:30 CEST] <durandal_1707> helmo_: remove this line: --enable-avresample
[12:11:37 CEST] <rindolf> ChocolateArmpits: can you test it yourself? See https://github.com/shlomif/pysolfc-black-hole-solver--screencast
[12:12:39 CEST] <ChocolateArmpits> rindolf, ok I'll take a look
[12:12:43 CEST] <helmo_> @durandal_1707  thx will try
[12:12:49 CEST] <rindolf> ChocolateArmpits: thanks
[12:17:42 CEST] <ChocolateArmpits> rindolf, the timestamps for the ogv file are all over the place
[12:18:50 CEST] <ChocolateArmpits> in the stream audio they go backwards somehow, the video stream has them start at some ridiculous value then suddenly jump to normal
[12:19:31 CEST] <ChocolateArmpits> I think you'll have to transcode the video
[12:19:42 CEST] <ChocolateArmpits> to set timestamps properly
[12:21:47 CEST] <helmo_> @durandal_1707 nope even omitting --enable-avresample it fails
[12:22:09 CEST] <helmo_> @durandal_1707 thanks anyway
[12:22:22 CEST] <rindolf> ChocolateArmpits: ah
[12:24:22 CEST] <oerg866> Question: Is there a way to make -shortest consider only audio and video? I'm also copying a subtitle stream and since that only displays something at the start of the file the resulting output is as long as that subtitle XD
[12:26:17 CEST] <durandal_1707> helmo_: that add --disable-avresample
[12:26:33 CEST] <durandal_1707> and it warnings are not errors
[12:26:55 CEST] <ChocolateArmpits> rindolf, you can check timestamps using this: ffprobe -select_streams a -show_packets -print_format json=c=1 file.ogv
[12:27:06 CEST] <ChocolateArmpits> where select_streams will select the stream type
[12:27:35 CEST] <durandal_1707> helmo_: pastebin ffbuild/config.log
[12:27:47 CEST] <rindolf> ChocolateArmpits: hi
[12:28:29 CEST] <rindolf> ChocolateArmpits: this is the original file - i want to extract a duration from it - http://www.shlomifish.org/Files/files/video/out.ogv
[12:29:42 CEST] <rindolf> ChocolateArmpits: i tried "ffmpeg -ss 2.0 -to 1:27 -i out.ogv -c copy out2.ogv " but it yielded the bad file
[12:29:57 CEST] <ChocolateArmpits> rindolf, you need to transcode the file for it to work
[12:30:04 CEST] <ChocolateArmpits> as I said before
[12:31:32 CEST] <helmo_> @durandal_1707 rebuilding...  i'll post the log as soon its done
[12:32:11 CEST] <helmo_> trashed the other before reading.. :(
[12:33:27 CEST] <ChocolateArmpits> rindolf, ffmpeg can't fix timestamps without reencoding the file
[12:34:24 CEST] <rindolf> ChocolateArmpits: ah
[12:35:24 CEST] <rindolf> ChocolateArmpits: wouldn't that be lossy?
[12:36:40 CEST] <ChocolateArmpits> well of course it is, but the video content is static, don't think it's going to hurt by any margin
[12:37:10 CEST] <helmo_> @durandal_1707 log too long for non pro user, hope you dont mind i made a zip here sorry for the hassle... https://workupload.com/file/rHyQ6gw
[12:37:29 CEST] <helmo_> anyway didnt work either
[12:37:30 CEST] <rindolf> ChocolateArmpits: wo -c copy, the result was bad
[12:37:37 CEST] <ChocolateArmpits> rindolf, I already told you why
[12:38:06 CEST] <rindolf> ChocolateArmpits: without -c copy, the result was bad
[12:38:23 CEST] <ChocolateArmpits> how bad ?
[12:38:35 CEST] <rindolf> ChocolateArmpits: very low quality
[12:38:46 CEST] <ChocolateArmpits> rindolf, so set the bitrate
[12:38:52 CEST] <rindolf> ChocolateArmpits: ok
[12:39:11 CEST] <ChocolateArmpits> your current video has 1mbps listed, you can try specifying the same
[12:42:42 CEST] <durandal_1707> helmo_: make clean && make distclean ??
[12:43:25 CEST] <helmo_> @durandal_1707 same
[12:46:34 CEST] <durandal_1707> helmo_: why --as=yasm ?
[12:49:15 CEST] <helmo_> @durandal_1707 always used as since v3 in osx it wouldn't work without it
[12:50:20 CEST] <helmo_> @durandal_1707 was some issue with xcode
[12:50:36 CEST] <durandal_1707> --as=clang should be
[12:51:03 CEST] <durandal_1707> looks like your build environment is broken
[12:52:03 CEST] <helmo_> @durandal_1707 curious thing is 3.4.2 still builds..
[13:20:36 CEST] <rindolf> durandal_1707: hi, how can I trim the duration of this file while preserving quality? http://www.shlomifish.org/Files/files/video/out.ogv
[13:22:40 CEST] <rindolf> durandal_1707: i tried ffmpeg -ss 2.0 -to 27.0 -i out.ogv -b:v 1M -bufsize 1M -crf 18 -preset veryfast out5.ogv --- there is a loss of quality
[13:24:47 CEST] <rindolf> anyone?
[13:26:42 CEST] <durandal_1707> nobody sane uses ogv
[13:36:32 CEST] <rindolf> durandal_1707: ok, that is what http://recordmydesktop.sourceforge.net/about.php gave me
[13:49:03 CEST] <rindolf> durandal_1707: apparently there is http://www.maartenbaert.be/simplescreenrecorder/ which gives an .mkv file - is that better?
[13:52:30 CEST] <rindolf> anyone?
[13:52:43 CEST] <durandal_1707> dunno, never tried, just know that ogv support is flaky in ffmpeg
[13:52:54 CEST] <rindolf> durandal_1707: ah
[14:12:17 CEST] <rindolf> durandal_1707: yay! Working now with .mkv - thanks
[14:13:11 CEST] <JEEB> -39
[14:36:32 CEST] <rindolf> durandal_1707: here is the final result - http://www.shlomifish.org/Files/files/video/pysolfc-bhs-with-audio.mkv
[14:37:39 CEST] <durandal_1707> this channel is logged, so do not post important stuff
[14:38:41 CEST] <rindolf> durandal_1707: yes, i know
[14:38:57 CEST] <rindolf> durandal_1707: it is public
[15:20:02 CEST] <rindolf> yay! it is now live at https://www.youtube.com/watch?v=t_cuDHj5V2k&feature=youtu.be
[16:56:45 CEST] <oerg866> hi I have an mpeg2 stream that changes aspect ratio in the middle for some reason from 4:3 to 16:9, can I fix the file to.... not do that? I tried DVDpatcher to fix the headers but that doesn't work
[17:14:29 CEST] <Mavrik> oerg866: that's probably because SAR changes
[17:14:35 CEST] <Mavrik> It's standard in broadcast TV
[17:14:46 CEST] <Mavrik> I guess you could use a setsar= filter to lock the video to a single SAR
[17:14:55 CEST] <oerg866> wouldn't that require me to re-encode
[17:15:04 CEST] <JEEB> if the content is actually of a specific SAR
[17:15:12 CEST] <JEEB> you could rewrite the stream SAR with the bitstream filter
[17:15:21 CEST] <JEEB> there's one recently added for MPEG-2 headers
[17:15:28 CEST] <JEEB> *MPEG-2 Video
[17:15:28 CEST] <Mavrik> oerg866: by MPEG2 i'm assuming you have a MPEG-2TS file captured from a broadcast TV? :)
[17:15:41 CEST] <oerg866> not TS, it was in VOB format
[17:15:45 CEST] <Mavrik> because TS signals those changes as a packet in the stream so I guess it's possible to drop it
[17:15:49 CEST] <oerg866> just to make things clear, this is how it starts: https://i.snag.gy/356Dps.jpg
[17:16:16 CEST] <oerg866> and then it goes to this for no reason: https://i.snag.gy/NASewi.jpg
[17:16:17 CEST] <Mavrik> Yeah, VOB is structurally not far away from TS.
[17:16:30 CEST] <JEEB> it's PS
[17:16:39 CEST] <JEEB> program stream VS transport stream
[17:17:01 CEST] <JEEB> oerg866: if you are sure the content is actually of one aspect ratio
[17:17:06 CEST] <JEEB> cbs_mpeg2 is probably what you want
[17:17:11 CEST] <oerg866> yes, it is all 4:3
[17:17:20 CEST] <oerg866> the video is from 1984 and remained unchanged,it's just the stream that's messed up
[17:17:57 CEST] <JEEB> too bad I don't remember if or how you can do the header modification with ffmpeg.c
[17:18:32 CEST] <oerg866> i have tried a couple patcher programs now that supposedly fix this but no luck so far
[17:18:59 CEST] <JEEB> the bit stream filter is seemingly there now in FFmpeg's libraries but I've just never used it :P
[17:19:04 CEST] <JEEB> (through the command line)
[17:19:20 CEST] <oerg866> interestingly, if I use mpgpatcher to force 1:1 aspect ratio
[17:19:36 CEST] <oerg866> then it is played as 1:1 throughout by both windows 10's media player and MPC-HC
[17:19:47 CEST] <oerg866> but if i patch it back to 4:3 it switches to 16:9 in the middle as usual
[17:19:52 CEST] <oerg866> ?!?!?!
[17:21:01 CEST] <oerg866> ah actually, windows' thingy gets confused, drops frames and changes aspect ratio to 16:9 and then 1:1 again shortly after
[17:21:11 CEST] <oerg866> this is getting very fishy I think the file might just be unlucky :P
[17:21:50 CEST] <oerg866> I can reduce the file to the offending parts and upload somewhere if someone more knowledgeable wants to try his luck :-)
[17:23:31 CEST] <oerg866> https://drive.google.com/open?id=1dV_OZTWAENVV56ZdrNWg0RgZX4DzoZy1
[17:23:52 CEST] <oerg866> I'm out of ideas now :-)
[17:26:55 CEST] <Mavrik> Hmm, you're probably just confusing the video player with non-standard AR
[17:27:06 CEST] <Mavrik> And most likely mpegpatcher isn't doing anything hence why you're still seeing the switch :P
[17:27:17 CEST] <Mavrik> It should scan the stream and throw away all the AR change packets
[17:28:03 CEST] <oerg866> mpegpatcher is just patching sequence headers
[17:28:15 CEST] <oerg866> if theres a program to ditch the packets let me know :S
[17:29:19 CEST] <oerg866> interestingly if I remux into mkv the switch doesn't happen
[17:29:29 CEST] <oerg866> but if i mux the mkv back to mpg or vob it happens again
[17:29:31 CEST] <oerg866> so weird
[17:29:55 CEST] <oerg866> according to mpegpatcher all headers now say 4:3 but if theres some packets inbetween i dunno
[17:36:07 CEST] <oerg866> I would just try and find another recording of this but this is the only PAL/25fps version of this clip I could find, all dvd releases are 480i :(
[17:45:07 CEST] <Mavrik> Yes, that's because mkv doesn't support AR switching.
[17:45:37 CEST] <Mavrik> (Which kinda surprises me.=
[18:26:26 CEST] <JEEB> Mavrik: I think it has more to do with "do you trust container vs contained format"
[18:26:46 CEST] <JEEB> and while matroska in theory supports multiple extradata blocks, those aren't used
[18:27:06 CEST] <JEEB> and usually when you mux into matroska extradata/init data packets get dropped
[18:27:13 CEST] <JEEB> I mean, in-band
[18:27:20 CEST] <JEEB> so you lose the update even if it was there
[18:32:25 CEST] <Mavrik> mhm
[18:32:44 CEST] <JEEB> of course some muxers will not filter those out
[18:32:45 CEST] <Mavrik> I'm surprised because I managed to mux some of H.264 into FLV and the AR change code stayed
[18:32:57 CEST] <JEEB> yea, FLV does in-band
[18:33:02 CEST] <JEEB> as far as I know
[18:33:08 CEST] <Mavrik> And it's surprising in a way that I'd expect MKV to keep those packets as well.
[18:33:19 CEST] <Mavrik> Although the whole 4:3 / 16:9 switch is now kinda obsolete.
[18:33:20 CEST] <JEEB> or at least I don't know of muxers that aproactively "clear" those init packets out
[18:33:34 CEST] <JEEB> with matroska or mp4 I wouldn't be surprised
[18:33:43 CEST] <JEEB> if various muxers were "helpful"
[18:33:51 CEST] <JEEB> and removed "duplicated" init packets
[18:35:12 CEST] <Mavrik> Yeah, especially since MP4 has all of that info in the header.
[22:30:32 CEST] <v1v1> Simple command ( cat /home/ec2-user/FILES/*.jpg* | /home/ec2-user/ffmpeg/ffmpeg -f image2pipe -i /dev/stdin /home/ec2-user/OUT.MP4 ) failing with warning ( [image2pipe @ 0x4f45bc0] Could not find codec parameters for stream 0 (Video: none, none): unknown codec Consider increasing the value for the 'analyzeduration' and 'probesize' options ) and error ( Output file #0 does not contain any stream ). Any recommendations?
[22:31:00 CEST] <v1v1> Full output pasted at: https://pastebin.com/2tChqvvq
[22:34:14 CEST] <furq> why not just use -pattern_type glob -i *.jpg
[22:45:50 CEST] <v1v1> furq: thanks, giving that approach a try
[00:00:00 CEST] --- Sun Apr 29 2018