[Ffmpeg-devel-irc] ffmpeg.log.20190423

[burek]

[04:18:08 CEST] <TheBest_F-22> Does anyone know where to get the Latest FFmpeg "Stable" (4.1.3) for Windows (pre-built)?
[04:19:15 CEST] <TheBest_F-22> Zeranoe tends to no longer keep up-to-date with "Stable", for some Reason...
[04:25:49 CEST] <TheBest_F-22> Anyone knows if Kyle/Zeranoe simply gave up on Building "Stable" Releases?
[04:25:52 CEST] <TheBest_F-22>     (even though there are so few of them...)
[05:14:32 CEST] Action: TheBest_F-22 wonders if there's anyone listening...
[05:23:37 CEST] <elmofan> TheBest_F-22: those who read your question do not know the answer, or are witholding it from you.
[05:23:44 CEST] <elmofan> Perhaps you need to offer incentives.
[05:23:57 CEST] <elmofan> but i doubt that would help
[05:27:22 CEST] <TheBest_F-22> elmofan: Thanks.  I'm actually wondering if it's because Kyle/Zeranoe is the only one still bothering to do a Proper "trustworthy" Windows Build of FFmpeg from Source (as feature complete as possible on this Platform)...
[05:28:02 CEST] <elmofan> does he have an email addr?
[05:30:46 CEST] <TheBest_F-22> Maybe He (Zeranoe) doesn't think that the Changes from 4.1.1 to 4.1.3 are "Worth it"(?)...  (Or He just has an Automated Process for Git "Nightlies" that isn't capable of Processing "Releases"/"Stable"s...  Who knows...)
[05:31:09 CEST] <TheBest_F-22> He has a Forum, a Contact Page, and a Twitter account...
[05:31:29 CEST] <elmofan> i expect he knows better why he does or doesn't do things than I do
[05:31:40 CEST] <TheBest_F-22> Last time I gave him this thread  ->  twitter.com/TheBestF22/status/1098260668191510528
[05:32:02 CEST] <TheBest_F-22> He apparently hear it and it appeared a couple of days later.
[05:32:47 CEST] <TheBest_F-22> I just was hoping for him to get back to what he did before:  He was almost always only half a week behind a Stable Release...
[05:33:14 CEST] <TheBest_F-22> (but normally, just a single day, I think...)
[05:37:11 CEST] <TheBest_F-22> I just read through "https://git.ffmpeg.org/gitweb/ffmpeg.git/blob/4154f8967820ca734a77ce91bb590cd77649dee8:/Changelog" and there seems to be a lot of useful "Fixes"...
[05:38:05 CEST] <TheBest_F-22> it's weird that Termux for Android (on arm64) got the Stable update before Windows...
[05:39:04 CEST] <TheBest_F-22> (BTW:  Zeranoe seems to be the only Windows Build Listed on the Official "http://ffmpeg.org/download.html#build-windows" Page...
[05:49:33 CEST] <TheBest_F-22> Maybe it has to do with what's written on the right side of said Official Download site...  (in the "Get the Sources" Section"
[05:52:13 CEST] <TheBest_F-22> Sometimes, it seems that FFmpeg developers themselves recommend the Latest Git Commit Snapshot over the "Stable" one that, hopefully, should be more Reliable and with less possible Security Vulnerabilities...
[05:56:01 CEST] <TheBest_F-22> Even Zeranoe's Own Website has "hover text" on the "Version" selection that Defaults and suggests the Latest "Git Version" over the "Stable" one...
[05:58:04 CEST] <TheBest_F-22> I'd Like to see all those Fixes (and Security Patches) on Windows too...  (though I doubt that I would encounter an exploit coming from Videos on Legal Streaming Sites, the Source of most of My FFmpeg use cases...)
[06:00:24 CEST] <TheBest_F-22> I just Noticed I've been at this for far too long...  Gotta go get some Dinner and some Sleep too...  -->  5 a.m. here in Lisbon, Portugal...
[06:00:36 CEST] <TheBest_F-22>   C Ya L8r then!.    ;P
[06:04:46 CEST] <TheBest_F-22> if someone sees Zeranoe around Here, Kindly ask Him to Build a "Stable" (4.1.3) Release of FFmpeg...
[06:05:32 CEST] <TheBest_F-22> K thx Bai.  (for now)
[07:48:31 CEST] <gour> ll
[07:50:01 CEST] <gour> morning
[07:51:00 CEST] <gour> our non-profit organization is recorring video from our public programs/lectures and here is the ffmpeg -i output: https://pastebin.com/vR5Sv2U6
[07:53:01 CEST] <gour> i need some recommendation how (to what) convert such recordings to make them smaller for more convenient upload to the server's public download area, but not to sacrifice quality too much? now ~1hr15min is ~1.7G
[08:28:54 CEST] <elmofan> gour: will the users want to view them in browser?
[08:29:53 CEST] <gour> elmofan: i belive that downloading them is preferred option, although we make a version for online viewing as well
[08:30:05 CEST] <gour> *we can make...
[08:30:43 CEST] <elmofan> i tend to prefer h264, but browsers don't generally seem to support it
[08:30:52 CEST] <elmofan> so that is one consideration
[08:31:07 CEST] <elmofan> i am no expert in encoding though, so I can not advise
[08:31:07 CEST] <furq> all browsers support h264
[08:31:24 CEST] <elmofan> is it the container then?  no .mp4?
[08:31:40 CEST] <furq> h264/aac in mp4 is the most widely supported thing there is
[08:32:03 CEST] <furq> gour: 1.7G for 75 minutes of 720p is already pretty small
[08:32:44 CEST] <elmofan> thank you for the correction furq
[08:33:01 CEST] <gour> furq: so, not much can be saved space-wise?
[08:33:17 CEST] <furq> probably not without downscaling or something like that
[08:33:25 CEST] <furq> it depends on the content though
[08:34:09 CEST] <gour> it's kind of "classical" presentation with light usage of flip-chart
[11:54:51 CEST] <seni> how does one read from an input audio stream? im getting an "invalid start code [..] pipe::invalid data found when proc input".the input format is mulaw in wav. im reading like `ffmpeg -y -loglevel error -acodec:a pcm_mulaw -f wav -i`
[14:37:50 CEST] <Ariyasu> i have a h264 video that plays back fine and a e-ac3 audio file that also plays back fine
[14:38:10 CEST] <Ariyasu> when i mux the two together, the output becomes very stuttery/jerky/choppy
[14:38:26 CEST] <Ariyasu> any ideas why ? or how i can prevent/fix it
[14:45:40 CEST] <elmofan> what are you using for playback Ariyasu
[14:45:53 CEST] <elmofan> try with -ao null
[14:48:10 CEST] <Ariyasu> mpc but it doesn't matter what media player i use
[14:48:17 CEST] <Ariyasu> i will give -ao null a try
[14:50:32 CEST] <Ariyasu> "Unrecognized option 'ao'."
[14:51:22 CEST] <elmofan> well it's a long shot anyway
[14:51:28 CEST] <lemourin> Is it possible to distinguish mp4 from mov using ffmpeg? Other than just looking at the extension.
[14:51:47 CEST] <Ariyasu> what was -ao null supposed to do?
[14:52:09 CEST] <elmofan> selects a null sink for your audio output
[14:52:36 CEST] <elmofan> to eliminate possibility of your player+system having audio driver issues
[14:52:46 CEST] <elmofan> that might get triggered by unusual files
[15:10:00 CEST] <Ariyasu> elmofan i see this error in output everytime i try to mux
[15:10:03 CEST] <Ariyasu> "[eac3 @ 0000000002c4e040] Estimating duration from bitrate, this may be inaccurate"
[15:10:18 CEST] <Ariyasu> you think this is the cause of the choppy playback?
[15:10:36 CEST] <elmofan> this is beyond my knowledge
[15:13:02 CEST] <ritsuka> lemourin: you can check the major_brand with ffprobe
[15:19:59 CEST] <lemourin> ritsuka: this won't work most of the time because i have mov files which don't have this metadata field set
[15:20:29 CEST] <lemourin> MediaInfo manages to differentiate those files somehow though
[15:20:36 CEST] <ritsuka> uh, major_brand is a required part of mov/mp4
[15:20:55 CEST] <lemourin> oh
[16:27:24 CEST] <Ariyasu> [matroska,webm @ 00000000005895c0] Element at 0x2687a58 ending at 0x268b7b6 exceeds containing master element ending at 0x2687a52
[16:27:30 CEST] <Ariyasu> what does this mean?
[17:15:59 CEST] <Adcock> I need help to reverse a video and the reversed portition at the end of actual video.
[17:16:07 CEST] <Adcock> Can this be done?
[17:16:40 CEST] <Adcock> I need help to reverse a video and then add reversed portition at the end of actual video.
[17:49:46 CEST] <Adcock> I figured it out guys.
[17:49:48 CEST] <Adcock> :D
[19:55:56 CEST] <dablitz> good afternoon channel I want to rtp stream audio and also recieve rtp audio at the same time. is there a command with ffmpeg where I can ffmepg and ffplay off 2 seperate ports simulataniously
[20:03:06 CEST] <dablitz> I am thinking, and please correct me -->  ffmpeg -re -f alsa -i hw,1:0 -acodec mulaw -f rtp rtp://xxx.xxx.xxx.xxx:1234 | ffplay rtp://xxx.xxx.xxx.xxx:4321
[20:12:00 CEST] <dablitz> I have to run out a get some batteries, i will be back in 30 minutes
[21:25:49 CEST] <dablitz> can anyone help
[21:49:17 CEST] <ChocolateArmpits> dablitz, yeah go ahead
[21:55:27 CEST] <DHE> he's doing rtp shenanigans
[21:57:39 CEST] <another> why don't you just start two processes separately?
[22:19:25 CEST] <dablitz> ChocolateArmpits: good afternoon channel I want to rtp stream audio and also recieve rtp audio at the same time. is there a command with ffmpeg where I can ffmepg and ffplay off 2 seperate ports simulataniously
[22:19:50 CEST] <dablitz> ChocolateArmpits: I am thinking, and please correct me -->  ffmpeg -re -f alsa -i hw,1:0 -acodec mulaw -f rtp rtp://xxx.xxx.xxx.xxx:1234 | ffplay rtp://xxx.xxx.xxx.xxx:4321
[22:20:17 CEST] <ChocolateArmpits> wew you don't want to use a pipe for that
[22:20:29 CEST] <dablitz> this is for my ham network and I don't know how
[22:22:34 CEST] <dablitz> I have 6 radio repeaters, and my home pc. I want to be able to key up my radio at home and rtp audio direct to all my repeaters at once
[22:22:37 CEST] <ChocolateArmpits> While it's possible to have both a sender and receiver in the same process via output mapping it would probably be much better launch two separate processes as another suggests
[22:22:53 CEST] <dablitz> ok
[22:23:03 CEST] <dablitz> is my ffmpeg string correct
[22:23:08 CEST] <ChocolateArmpits> It's not
[22:23:20 CEST] <dablitz> can i get a correction
[22:25:07 CEST] <ChocolateArmpits> Provided your output destination is correct
[22:25:09 CEST] <ChocolateArmpits> Maybe this: ffmpeg -re -f alsa -i hw,1:0 -i rtp://xxx.xxx.xxx.xxx:4321 -map 0 -acodec mulaw -f rtp rtp://xxx.xxx.xxx.xxx:1234 -map 1 -codec copy -f nut - | ffplay -sync ext -probesize 32 -f nut -
[22:25:25 CEST] <ChocolateArmpits> But because I don't know bash I can't suggest to how to start two processes at once :/
[22:25:57 CEST] <ChocolateArmpits> That sample should work but it's not an ideal case
[22:27:29 CEST] <dablitz> the reason for the ffplay is when my repeaters receive RF they will stream that to my home pc, hence the two
[22:29:46 CEST] <dablitz> so I need to send to each repeater and receive back from the repeater, I dont mind having 6 different instances of this running I just need to know the config is correct
[22:35:46 CEST] <pzich> ChocolateArmpits: are you proposing using something like a fifo instead of a pipe? Or a different approach entirely?
[22:38:24 CEST] <ChocolateArmpits> dablitz, by sending to the repeater do you mean sending anything or what you expect to get back?
[22:39:50 CEST] <dablitz> ChocolateArmpits: my radio is connected to my mic input on my computer, so when I key my mic it streams to the raspberry pi at the repeater, and when the repeater recieves RF from another radio it streams back to my pc and out soundcard into my radio and I hear it
[22:45:40 CEST] <ChocolateArmpits> dablitz, by radio you mean using it as a loudspeaker?
[22:45:58 CEST] <ChocolateArmpits> I'm trying to figure it one by one
[22:47:26 CEST] <dablitz> yes kindof. the radio on my desk will play as a loud speaker what it recieves, but will also tx via vox (inside the radio) on a simplex channel to my portable radio that I carry around town with me
[22:48:08 CEST] <dablitz> i just need to have the stream of unencrypted audio go back and forth
[23:10:39 CEST] <ChocolateArmpits> dablitz, do you want only one mic to engage all the repeaters?
[23:11:24 CEST] <dablitz> the raspberry pi on each repeater only has 1 mic
[23:12:49 CEST] <dablitz> all pi's are on a vpn and I was multicasting
[23:12:55 CEST] <dablitz> on the vpn
[23:13:30 CEST] <dablitz> that way I can use internal ip's and not have to worry about remotes having static public ip's
[23:39:38 CEST] <calamari> The AVI container has a multitude of drawbacks, but one nice thing is that if you only copy some of it, you can still have a playable file, which isn't the case with MKV for example. Does anyone know of other containers that can tolerate truncated containers?
[23:40:10 CEST] <cehoyos> All do iirc
[23:40:50 CEST] <calamari> cehoyos: not true.. MKV doesn't work if you don't have the end
[23:40:58 CEST] <cehoyos> That's news to me.
[23:41:31 CEST] <cehoyos> (With mov and friends, the part that can be at the end of the file but can as well be at the beginning is needed - if it is at the end, you cannot cut the file)
[23:41:48 CEST] <calamari> cehoyos: try ripping a blu-ray with MakeMKV. It's writing a mkv all along, but you can't play it until it finishes up
[23:42:14 CEST] <furq> that sounds like a makemkv problem
[23:42:16 CEST] <cehoyos> So your questions was actually: Which applications apart from MakeMKV write files so that they cannot be played while written?
[23:42:33 CEST] <calamari> LOL!
[23:42:43 CEST] <furq> anyway mpegts is the canonical unbreakable container format
[23:42:58 CEST] <cehoyos> You mean like mpeg-ps?
[23:43:14 CEST] <cehoyos> And without pat/pmt?
[23:43:16 CEST] <calamari> furq: I'll give it a try, thanks
[23:46:42 CEST] <calamari> I believe the problematic part of mp4 and mkv is called the MOOV atom, which has to be updated for the file not to be corrupted
[23:46:50 CEST] <furq> of mp4, sure
[23:46:56 CEST] <furq> there's no such thing in mkv
[23:48:22 CEST] <BtbN> There is, but it doesn't break the file to uselessness if it's missing. It's just not seekable.
[23:48:34 CEST] <cehoyos> as said, the moov atom can be at the start of the file, if it is at the start of the file, you can cut it and (still) play it.
[23:48:34 CEST] <furq> i meant it's not called a moov atom
[23:49:02 CEST] <calamari> BtbN: are there any containers besides AVI that don't have a moov atom or anything like it?
[23:49:10 CEST] <furq> mpegts
[23:49:10 CEST] <BtbN> mpegts
[23:49:11 CEST] <cehoyos> All except mov?
[23:49:12 CEST] <BtbN> flv
[23:49:24 CEST] <furq> flv as well but that's much more restrictive in terms of codecs
[23:49:30 CEST] <BtbN> mkv in streamable mode
[23:49:32 CEST] <furq> basically any container designed for streaming
[23:49:42 CEST] <calamari> alrighty, thanks! sounds like mpeg-ts is the answer
[23:49:47 CEST] <furq> bear in mind mpegts has no seek index
[23:49:53 CEST] <BtbN> mpeg-ts is pretty heavy in terms of overhead as well
[23:49:56 CEST] <BtbN> I'd just go for mkv
[23:50:00 CEST] <BtbN> for pretty much everything
[23:50:10 CEST] <calamari> mkv doesn't work, it's broken if you don't have the entire file
[23:50:15 CEST] <furq> yeah idk what makemkv is doing but ffmpeg's mkv output will work if it's truncated
[23:50:15 CEST] <cehoyos> No
[23:50:17 CEST] <BtbN> Unless you want it to play in a Browser, then you have the choice between mp4 and mp4
[23:50:42 CEST] <cehoyos> Your application is "broken" because it does not write mkv in the normal way where it can be played before being finished.
[23:50:53 CEST] <cehoyos> In any case, you can certainly cut your finished mkv file and it is still playable.
[23:51:03 CEST] <calamari> cehoyos: in this case I'm not using makemkv, I was just offering that as an example
[23:51:16 CEST] <furq> what player are you using
[23:51:19 CEST] <furq> maybe that's broken
[23:52:52 CEST] <calamari> furq: well with the partial mkvs, I tried vlc, mpv, a browser, an old mplayer.. it seemed universally broken
[23:53:09 CEST] <furq> weird
[23:53:27 CEST] <calamari> I realize mkv and mplayer are probably not really a different test
[23:53:43 CEST] <calamari> err mpv *
[23:54:02 CEST] <calamari> anyhow thanks everyone, I'll try it with mpeg-ts
[23:54:05 CEST] <cehoyos> They all work fine with cut mkv files
[23:54:41 CEST] <cehoyos> None of them can play invalid mkv file which is likely what your application produces if you don't allow it to run until the end.
[23:58:12 CEST] <calamari> can you seek in a partial mpeg-ts?
[23:58:27 CEST] <calamari> (you can in avi)
[23:58:41 CEST] <calamari> I can try it and see :)
[23:58:46 CEST] <BtbN> mpeg-ts is not seekable
[23:59:34 CEST] <calamari> bummer. maybe AVI is the only answer then
[00:00:00 CEST] --- Wed Apr 24 2019