[Ffmpeg-devel-irc] ffmpeg.log.20171130

burek burek021 at gmail.com
Fri Dec 1 03:05:02 EET 2017


[00:00:25 CET] <TheRock> Hello, i remember clonecd from 2004
[00:00:29 CET] <TheRock> ffmpeg is almost as good as clonecd
[00:00:58 CET] <durandal_1707> clonecd is?
[00:01:22 CET] <alexpigment> I really appreciated this bit of randomness :)
[00:01:29 CET] <fella> durandal_1707: ripping copyrighted DVDs
[00:01:44 CET] <alexpigment> Hello, i remember audiograbber from 2002
[00:01:49 CET] <alexpigment> ffmpeg is almost as good as audiograbber
[00:02:44 CET] <durandal_1707> fella: thats clonedvd :)
[00:03:17 CET] <fella> durandal_1707: see? i can't remember ;)
[00:03:18 CET] <alexpigment> pretty sure it's the same company
[00:03:33 CET] <alexpigment> they also have clonebd
[00:04:41 CET] <durandal_1707> ok, ffmpeg have nothing with ripping, it just have cdparanoia wrapper
[01:08:33 CET] <lightbulb6> hello, how can i make the "crop" filter update the "n", "pos" and "t" variables when the source is a single png image? please see: https://bpaste.net/show/54eeaa31a2b6
[01:09:06 CET] <lightbulb6> here, the text.png file is a simple 640x360 pixels image with some sample text
[01:09:35 CET] <lightbulb6> and for some reason, the "n" variable stays at 0
[01:10:14 CET] <lightbulb6> i expected it to increase with each frame, so that with each frame, more of the text.png can be seen
[01:11:14 CET] <lightbulb6> if i try to use the "pos" or "t" variables, ffmpeg complains "Error when evaluating the expression 't+100'"
[01:31:40 CET] <therage3> what exactly does '-tune animation' do? the docs aren't really clear about how it optimizes cartoons and animation
[01:38:18 CET] <kinkinkijkin> anybody here use vapoursynth?
[01:38:33 CET] <c_14> therage3: ask in #x264
[01:38:52 CET] <therage3> c_14: cheers, let's see what they say
[01:38:55 CET] <c_14> tunes are business of the encoder, not ffmpeg itself
[01:39:07 CET] <therage3> i see
[02:00:22 CET] <yukiup> is there a channel for hevc_nvenc?
[02:00:47 CET] <yukiup> for making the quality of the video higher
[02:12:37 CET] <jfmcarreira> heyy guys
[02:12:48 CET] <jfmcarreira> the new avcodec api works on all codecs?
[02:25:12 CET] <julianc> Hi, I'm re-encoding an mp4 using an HTTP resource as an input. The input's MOOV atom is at the end of the file and is large enough that I want to avoid downloading the MP4 in advance. FFMPEG is getting bogged down making 1000s of HTTP range requests while seeking the MOOV atom. Is there a way to tweak FFMPEG's MOOV atom seeking behaviour so that it makes larger range requests from the HTTP source? I've looked at https://ffmpeg.org/f
[02:25:24 CET] <julianc> and couldn't find anything helpful.
[07:37:41 CET] <Guest72929> >>> youtube.com/watch?v=X8qL6cAI0zQ
[09:18:27 CET] <kode54> Lavc56.12.101 libvorbis <- was this known to write malformed vorbis streams?
[09:19:05 CET] <kode54> I have some game-bundled .ogg files that have sequences of two or three consecutive packets with malformed offsets in their granulepos
[09:19:34 CET] <kode54> two or three packets that all have the same reported dependent timestamp, in the future
[09:20:09 CET] <kode54> https://hydrogenaud.io/index.php/topic,115032.msg948508.html#msg948508 <- log here
[09:20:35 CET] <kode54> my bad
[09:20:46 CET] <kode54> "previous" is in the future, and the same value on two or three packets in a row
[11:39:10 CET] <Kristjan> My name is Kristjan. I am in urgent need for money. People, please send me money with min sum of 50 $. IBAN: EE671010010225901016 (SEB Estonia). Name: Kristjan Robam
[11:41:35 CET] <pomaranc> Kristjan: do you accept bitcoin?
[11:42:15 CET] <Nacht> Yeah man, go with your time. Include a wallet as well
[11:42:49 CET] <fsphil> I am in urgent need of breakfast. Please send me toast
[13:59:27 CET] <dradakovic> So i followed the ffmpeg compile guide on https://trac.ffmpeg.org/wiki/CompilationGuide/Centos. I followed it in a nutshell step by stem. But when i get to ./configure of the ffmpeg itself with enabled encoders i get the ERROR: libmp3lame >= 3.98.3 not found
[14:00:37 CET] <dradakovic> It is like the pkg-config path is not known but i did specify it
[14:02:01 CET] <JEEB> most likely because LAME has a bad .pc file
[14:02:10 CET] <JEEB> check the ffbuild/config.log file for what's missing
[14:06:13 CET] <dradakovic> may i upload the log to be checked?
[14:06:35 CET] <sfan5> yes just pastebin it
[14:06:57 CET] <dradakovic> ah yes
[14:12:37 CET] <dradakovic> https://pastebin.com/H9hxwLzK
[14:15:00 CET] <sfan5> oh yeah someoene had the same issue last week
[14:15:30 CET] <sfan5> add --extra-libs="-lm" to the ffmpeg configure
[14:15:50 CET] <dradakovic> ohh i see. Let me try this immediately
[14:28:59 CET] <dradakovic> i think it worked. installing now. Ive had a similar error for a few other libraries but i removed the --enable from them. It passed libmp3lame tho which it matters
[14:29:18 CET] <dradakovic> i will report the outcome
[14:29:27 CET] <dradakovic> ty very much for now
[14:30:07 CET] <dradakovic> good bunch of people here
[14:35:06 CET] <dradakovic> Damn it works!
[14:35:09 CET] <dradakovic> Ty so much
[14:42:50 CET] <dradakovic> Maybe someone who is maintaining that guide should add that line to it. It would probably help a bunch of us noobs
[15:40:55 CET] <tyng> does libavcodec plan to support its own threaded av1 decoder?
[15:42:55 CET] <DHE> is there a decoder available today?
[15:43:00 CET] <DHE> (or even an encoder)0000000
[15:43:04 CET] <DHE> gah, stupid vnc
[15:43:13 CET] <durandal_1707> no today
[15:45:56 CET] <tyng> there is a reference codec under development
[15:46:13 CET] <tyng> but as libvpx it is not meant to be efficient
[15:46:22 CET] <JEEB> it's libaom now
[15:46:30 CET] <JEEB> libvpx is for VP8/9
[16:12:39 CET] <Fenrirthviti> Hmm, anyone done stress testing with nvenc on a 10-series GPU? Trying to find what FPS it should be hitting at 4k encoding. NVIDIA docs only list 1080p values (261fps) that I can find
[16:13:03 CET] <fx159159> Hey, how do I account for pixel stride of a picture when using av_image_copy_to_buffer?
[16:13:04 CET] <fx159159> When pixel stride is one and using row strides as linesizes my picture has the correct colors, on a device with pixel stride 2 the colors are off, only the base image is correctly visible
[16:13:26 CET] <DHE> Fenrirthviti: roughly speaking the limit is pixels/second. so you could expect 1/4 that number for 4k
[16:16:16 CET] <Fenrirthviti> hmm, ok.
[16:19:06 CET] <Guest90028> hello, i am trying to use ffmpeg to save udp stream to check for decoding errors and such. I am using the following command: https://pastebin.com/9qYHeCPs allthought i have a hunch that ffmpeg demuxes and remuxus the stream before it is saved which in the end possibly masking errors
[16:20:43 CET] <DHE> unfortunately yes. also the mpegts demuxer generates alerts for continuity errors at the debug level making them rather useless
[16:21:01 CET] <DHE> I recommend saving it using something else. considering the format, I'd just use netcat in UDP mode
[16:22:09 CET] <Guest90028> i have tried netcat, allthought there isnt a way for specify how long i should record for, and when i kill the process i am getting an issue where last key frame is incomplete and fails to decode
[16:23:02 CET] <Guest90028> unless there is an easy way to remove everything after last key frame
[16:23:33 CET] <DHE> if you're looking for packetloss, wireshark might be able to help you more here. though 10 minutes is a lot
[16:24:44 CET] <Guest90028> I need roughly 5 min, the example above was more of a longer test with ffmpeg
[16:25:58 CET] <Guest90028> but i guess there is no way to disable demuxing/remuxing in ffmpeg
[16:28:55 CET] <DHE> well, how about: ffmpeg [input] -c:v rawvideo -c:a pcm_s8 -f null /dev/null   # and just watch for errors from the video/audio decoders
[16:29:20 CET] <DHE> doesn't save it, but does check for errors in realtime?
[16:37:12 CET] <therage3> are there any flags/parameters for detelecine for the x264 encoder? i know there's options for Handbrake, but i can't seem to find anything for ffmpeg
[16:37:39 CET] <Guest90028> hmm, thats possible
[16:38:21 CET] <therage3> were you talking to DHE or to me?
[16:38:52 CET] <Guest90028> DHE
[16:39:06 CET] <therage3> oh
[16:44:52 CET] <Guest90028> quick question, i am trying to figure out what does "co located POCs unavailable" mean?
[16:54:38 CET] <alexpigment> therage3: https://ffmpeg.org/ffmpeg-filters.html#detelecine
[16:56:49 CET] <alexpigment> to be clear, this isn't for the x264 encoder; this is an ffmpeg filter that occurs before x264 in the chain
[17:22:23 CET] <therage3> alexpigment: i see, thanks
[17:22:48 CET] <alexpigment> therage3: if you are detelecining NTSC content, the defaults are usually fine
[17:23:39 CET] <therage3> alexpigment: that's exactly what i want to detelecine, yeah. i got this 2007 animated NTSC DVD, and yesterday when i was comparing quality of frames of a raw mkv with a transcoded one, i noticed the telecined frames
[17:25:01 CET] <alexpigment> therage3: yeah, it should end up as 23.976 (24000/1001) after detelecining then
[17:25:49 CET] <therage3> alexpigment: i see. i told a friend of mine who also happens to be a fan of this show, and he was surprised the publisher released this DVD as is
[17:25:57 CET] <therage3> i'm not sure if this was standard at the time
[17:26:11 CET] <alexpigment> yeah, telecined content is very common
[17:26:17 CET] <therage3> i see
[17:26:22 CET] <alexpigment> the thing is, there is no true 24fps standard for DVD
[17:26:32 CET] <alexpigment> it was always 30fps for NTSC and 25fps for PAL
[17:26:55 CET] <alexpigment> so they later started adding "pulldown flags" so that the player would play 24p as 60i
[17:27:35 CET] <alexpigment> still, many of the broadcast masters are already in 60i, so to make a true 24fps dvd, it would mean that the studio would have to go back to film masters or detelecine it themselves
[17:27:42 CET] <therage3> i see, i guess these flags were somehow embedded into the mpeg-2 stream so that the dvd player knew to do it
[17:27:51 CET] <alexpigment> when it's way simpler to just digitize the 60i tape masters
[17:27:57 CET] <therage3> o see
[17:28:02 CET] <therage3> i see*
[17:29:29 CET] <alexpigment> honestly, I don't really even mind keeping things as 60i so I don't bother detelecining
[17:29:44 CET] <alexpigment> maybe if I ever upgrade from my 60hz TV, i'll see the light ;)
[17:29:50 CET] <therage3> i don't mind it either that much, but i did kind of notice it in the show whe nwatching it
[17:29:59 CET] <therage3> but it was patently obvious when looking at static frames
[17:30:02 CET] <alexpigment> what, exactly, did you notice?
[17:30:09 CET] <therage3> let me give you a visual example
[17:30:37 CET] <alexpigment> 3:2 pulldown usually isn't noticeable on a player that's doing its job correctly
[17:30:55 CET] <therage3> https://imgur.com/a/npSUR
[17:31:13 CET] <therage3> the first frame is what maybe 3 frames look like, followed by two good ones, like the second frame
[17:31:19 CET] <therage3> that's the pattern i observed
[17:31:34 CET] <therage3> then i did googling, and what came up was "hard telecine"
[17:31:35 CET] <alexpigment> hmmm
[17:31:47 CET] <alexpigment> is your player actually deinterlacing to 60i?
[17:31:51 CET] <alexpigment> or is it doing a blend?
[17:32:02 CET] <therage3> not sure what the actual player is doing, I use vlc to play the DVD
[17:32:06 CET] <alexpigment> ahhhh
[17:32:09 CET] <alexpigment> there's your problem
[17:32:09 CET] <therage3> but these are apparently stills from the show
[17:32:21 CET] <therage3> no, no... these stills were extracted from raw mkv's
[17:32:21 CET] <alexpigment> VLC is an awful player and I wish they would get their shit together
[17:32:24 CET] <therage3> before you say anything
[17:32:24 CET] <alexpigment> right
[17:32:31 CET] <therage3> VLC didn't pull these
[17:32:32 CET] <alexpigment> but the stills were from vlc, right?
[17:32:34 CET] <alexpigment> ok
[17:32:42 CET] <therage3> no, I got them from ...... ffmpeg
[17:32:42 CET] <therage3> lol
[17:33:01 CET] <alexpigment> try getting the stills from ffmpeg with -vf yadif=1
[17:33:09 CET] <therage3> what does that do
[17:33:15 CET] <alexpigment> it deinterlaces to 60p
[17:33:27 CET] <alexpigment> which is what VLC would do if it were worth using at all
[17:33:35 CET] <therage3> i see, so it doesn't get the raw but applies a deinterlacing filter first
[17:33:48 CET] <alexpigment> well, the thing is that ffmpeg is getting a "frame"
[17:33:52 CET] <alexpigment> and there are two fields to a frame
[17:33:53 CET] <therage3> yes
[17:33:57 CET] <therage3> aha
[17:34:13 CET] <alexpigment> so if the fields are from different time stamps (they technically always are), you get a blended frame of two different time stamps
[17:34:48 CET] <therage3> right, i see
[17:34:55 CET] <alexpigment> by using -vf yadif=1, you're saying "make every field a whole frame" rather than blend fields
[17:35:01 CET] <therage3> so a frame isn't the most basic unit, but a field is
[17:35:08 CET] <alexpigment> for interlaced content, yes
[17:35:13 CET] <therage3> i see, interesting
[17:35:20 CET] <alexpigment> which is why there's so much confusion around interlaced content
[17:35:28 CET] <therage3> apparently o.O
[17:35:32 CET] <therage3> i'm learning so much lately
[17:35:34 CET] <alexpigment> on top of that, some players don't deinterlace by default, which is frankly stupid
[17:35:51 CET] <therage3> ok let me run that deinterlacing frame yank command
[17:35:54 CET] <therage3> and show you the results
[17:35:55 CET] <therage3> hold on
[17:37:35 CET] <alexpigment> k, now to be clear, if it's actually telecined, you'll still get duplicate frames, but they won't be blended. detelecining can technically reconstruct a true 23.976fps video from the 29.97i source. it uses it by blending certain fields in a particular pattern
[17:38:18 CET] <therage3> alexpigment: hah, yes, sure enough now all frames are clear, but there's a bunch that are appaerently the exact same thing
[17:38:28 CET] <alexpigment> i wasn't advocating that you don't detelecine - most people do that - i was just saying that I don't notice the pulldown enough to worry about it. i leave everything as 60i
[17:39:56 CET] <alexpigment> therage3: now about VLC. if you absolutely must use that horrible piece of garbage, go to Preferences > Video and set the deinterlacing mode to Automatic and the mdoe to Yadif (2x)
[17:40:19 CET] <therage3> alexpigment: i've been meaning to ditch VLC for a while now
[17:40:44 CET] <alexpigment> yadif(2x) is the best deinterlacing mode they have, which is passable for SD, but causes performance problems on high def content on many computers
[17:41:10 CET] <therage3> alexpigment: when i went to their channel here, and merely asked if gapless playback is supported (because i couldnt figure out how to get it running), the dev told me to do it myself because it wasnt going to happen
[17:41:20 CET] <therage3> this gives me yet another reason to ditch VLC
[17:41:29 CET] <alexpigment> yep
[17:41:46 CET] <therage3> i dont understand why it doesn't support gapless playback in this day and age
[17:41:51 CET] <therage3> it completely ruins some music albums
[17:42:07 CET] <alexpigment> well, it's technically a little advanced, but yes, other people do it
[17:42:09 CET] <alexpigment> brb
[17:42:13 CET] <therage3> ok
[19:08:00 CET] <wasilewski> hi guys, i have a text file with about 20 .m3u8 audio stream urls that I would like to download.  I have tried downloading a few with FFmpeg using the -i command and it works.  I'm wondering if I can use FFmpeg to download all the urls in the text file someway so I don't have to input them all individually?
[19:09:06 CET] <DHE> that's not really how ffmpeg works. a list of m3u8 files isn't something ffmpeg handles natively (or at least, properly)
[19:09:40 CET] <wasilewski> ok, thanks.....so I just have to enter them all in individually?
[19:10:42 CET] <therage3> you can use some script to automate it, i guess
[19:11:43 CET] <wasilewski> yeah
[19:13:07 CET] <wasilewski> just one more thing - am i getting the best quality by just entering ffmpeg -i "http://etc.etc" "file.mp3". ?
[19:13:53 CET] <sfan5> if those audio streams are already mp3 you can add -c copy to avoid any quality loss
[19:14:17 CET] <sfan5> if not i would suggest saving them in the original format if possible
[19:17:48 CET] <wasilewski> sfan5:  i'm not sure if they are in mp3, it all comes from the m3u8 playlist....sorry if im' being stupid here
[19:18:16 CET] <sfan5> running ffprobe on one will tell you the codec in use
[19:18:21 CET] <wasilewski> would I add the -c at the start of the command? such as ffmpeg -i -c "http://etc.etc" "file.mp3". ?
[19:18:28 CET] <wasilewski> oh cool - i didn't know that thanks
[19:18:57 CET] <wasilewski> actually i'd probably put the -c after the url part.....
[19:19:08 CET] <sfan5> no, ffmpeg -i http://url -c copy output.mp3
[19:19:13 CET] <wasilewski> oh cool
[19:19:15 CET] <wasilewski> thanks
[19:19:55 CET] <wasilewski> i will use the ffprobe command too
[19:20:17 CET] <wasilewski> thanks again guys
[19:31:55 CET] <DHE> that only works if the input is mp3 itself, which seems unlikely. officially HLS only supports AAC (and maybe in newer versions, other codecs)
[19:39:36 CET] <wasilewski> OK thanks
[19:40:00 CET] <wasilewski> i will take a look later
[19:40:06 CET] <wasilewski> gotta finish making dinner
[19:40:08 CET] <wasilewski> thanks again folks
[19:40:10 CET] <wasilewski> all the best
[20:58:53 CET] <therage3> is it normal for a 30fps hard telecined NTSC DVD movie to still exhibit some telecined frames after having the default detelecine filter applied to it (when it is now 24fps)?
[20:59:42 CET] <JEEB> then whatever filter didn't do the right job, or your source has more issues
[21:00:04 CET] <therage3> that's possible, i'm analyzing the source, frame by frame, to see if it's not the standard pattern
[21:00:30 CET] <JEEB> inverse telecine has the filters fieldmatch and decimate, btw
[21:00:36 CET] <JEEB> those were ported from vapoursynth ages ago
[21:00:53 CET] <JEEB> fieldmatch does field matching and decimation does the removal of duplicates
[21:01:09 CET] <therage3> i see
[21:03:44 CET] <therage3> what's the syntax for the filter? i used ' -filter:v "detelecine" ', and i guess that uses the default parameters?
[21:04:14 CET] <sfan5> it does
[21:04:24 CET] <sfan5> reference for those is here btw http://ffmpeg.org/ffmpeg-filters.html#detelecine
[21:04:26 CET] <therage3> huh, interesting
[21:04:41 CET] <JEEB> https://www.ffmpeg.org/ffmpeg-all.html#fieldmatch
[21:04:55 CET] <therage3> it's bizarre how it gets rid of the telecined frames in certain sequences of the movie, but not all. and the fps does go from 30 to 24
[23:03:02 CET] <alexpigment> therage3: out of curiosity, is the fps exactly 24 or is it 23.976 (24000/1001)?
[23:03:34 CET] <therage3> alexpigment: the latter (sorry if I didn't specify; I know there's some specification stuff there with MHz stuff that makes it use that odd number)
[23:03:44 CET] <alexpigment> that's fine, just wanted to check
[23:04:01 CET] <alexpigment> because ntsc telecined content would have duplicate frames every now and then if the frame rate ended up at exactly 24
[23:04:12 CET] <therage3> alexpigment: what I think is happening is that this video that I have isn't telecined in a consistent pattern
[23:04:15 CET] <alexpigment> just figured i'd ask that. i don't really have any more info on what's going on
[23:04:21 CET] <therage3> which throws the filter off and makes it miss some stuff
[23:04:25 CET] <alexpigment> ah
[23:05:05 CET] <alexpigment> well, that's a scenario where I would either leave it as interlaced (force x264 to do MBAFF interlacing), or deinterlace to 60p
[23:05:24 CET] <therage3> i just let VLC or whatever take care of it now
[23:05:33 CET] <alexpigment> k
[23:05:48 CET] <therage3> i have absolutely no idea how to deal with variably telecined footage
[23:05:57 CET] <alexpigment> neither do i ;)
[23:06:06 CET] <alexpigment> fortunately i've never dealt with that
[23:06:08 CET] <JEEB> if it's actually telecined and it's just the pattern that changes then TIVTC/fieldmatch+decimate should do it
[23:06:17 CET] <JEEB> too bad lavfi preview is not really a thing
[23:06:25 CET] <alexpigment> JEEB knows what he's talking about, so i'd trust him
[23:06:26 CET] <JEEB> (although you can test it out with something like mpv)
[23:06:46 CET] <JEEB> if it's not telecined and mixed interlacism and telecine, you will have to handle it in parts
[23:06:47 CET] <therage3> JEEB: hm i see
[23:06:54 CET] <therage3> oh dear Lord
[23:07:00 CET] <therage3> this is getting unnecessarily complicated
[23:07:03 CET] <alexpigment> yeah
[23:07:14 CET] <JEEB> therage3: you hit TIVTC+decimate at it and see if it looks skookem
[23:07:19 CET] <alexpigment> the credits are often true interlaced on otherwise-telecined content
[23:07:26 CET] <therage3> JEEB: i see
[23:07:35 CET] <JEEB> if it doesn't, time to check the fields proper
[23:07:42 CET] <JEEB> vapoursynth or avisynth recommended
[23:07:49 CET] <JEEB> they have the best preview things
[23:07:54 CET] <JEEB> with vapoursynth editor and AvsPmod
[23:07:57 CET] <therage3> oh snap avisynth ;D
[23:08:02 CET] <therage3> i once used it and my brain blew up
[23:08:46 CET] <JEEB> people tend to overcomplicate their crap
[23:09:33 CET] <therage3> i'll try that, hopefully i don't have to tweak any fields
[23:09:52 CET] <JEEB> try a check-up run with fieldmatch and decimate
[23:10:00 CET] <JEEB> that is supposed to be a dynamic IVTC thing in lavfi
[23:10:00 CET] <therage3> good call
[23:10:13 CET] <JEEB> and it's supposed to be a port of the LGPL IVTC filter in vapoursynth
[23:10:23 CET] <therage3> interesting, iirc VLC also uses that
[23:10:25 CET] <JEEB> although I think the latter got some continued development after the lavfi filter got made
[23:28:11 CET] <Cracki_> mjpeg is listed under *audio* codecs... https://ffmpeg.org/ffmpeg-codecs.html#mjpeg
[23:29:28 CET] <jfmcarreira> Heyy guys
[23:29:43 CET] <jfmcarreira> i am using this code to decode a stream using ffmpeg:  http://dpaste.com/1CM1RNT
[23:30:00 CET] <jfmcarreira> but i am always getting EOF errors from av_read_frame. even on first try
[23:33:57 CET] <zyme> if the current version is 3.4,  when was "ffmpeg version n2.8.11" and whats the n stand for?
[23:37:22 CET] <SortaCore> jfm: I suspect ffmpeg is in release mode?
[23:37:32 CET] <SortaCore> and debugging works best with full source code
[23:37:43 CET] <SortaCore> the PDB files may be stripped
[23:38:32 CET] <jfmcarreira> SortaCore: i know. i got ffmpeg. build it from source (using some kind of debug option in configure) then linking my app against the ffmpeg i just builed and installed in my home folder
[23:40:44 CET] <jfmcarreira> SortaCore: ./configure --prefix=/home/joaoc/.libraries --enable-debug=3 --disable-optimizations --disable-static --enable-shared --disable-stripping --disable-mmx --disable-ssse3 --extra-cflags
[23:41:32 CET] <SortaCore> should be ok then
[23:41:39 CET] <SortaCore> did you try Step Into?
[23:42:01 CET] <SortaCore> put a breakpoint just before a ffmpeg function call, then use step into command on your ide
[23:42:17 CET] <jfmcarreira> SortaCore: idk if i tried this config yet. i tried a different one with static libs
[23:42:51 CET] <jfmcarreira> `.rodata' can not be used when making a shared object; recompile with -fPIC
[23:42:56 CET] <jfmcarreira> how do i enable fpic?
[23:46:09 CET] <zash> Add it to CFLAGS
[23:47:11 CET] <durandal_1707> therage3: detelecine filter works only for fixed patterns
[23:47:29 CET] <therage3> durandal_1707: i see
[23:47:35 CET] <therage3> durandal_1707: that then does explain that
[23:47:36 CET] <JEEB> which is why I recommended the fieldmatch+decimate combo
[23:47:40 CET] <therage3> ^
[23:47:51 CET] <therage3> i'll have to do some fiddling with that as JEEB suggested
[23:48:17 CET] <therage3> but thanks, that does confirm why the plain vanilla defauly with detelecine didn't work fully
[23:48:21 CET] <therage3> default*
[23:49:05 CET] <durandal_1707> there should be telecine detection filter which detect patterns across stream
[23:49:15 CET] <alexpigment> ^^^
[23:50:40 CET] <therage3> yeah
[23:54:13 CET] <TheRock> hi guys, is there  filter to detect if the video is sex movie
[23:55:17 CET] Action: therage3 raises an eyebrow
[00:00:00 CET] --- Fri Dec  1 2017


More information about the Ffmpeg-devel-irc mailing list