[Ffmpeg-devel-irc] ffmpeg.log.20180422

Mon Apr 23 03:05:02 EEST 2018

[01:16:33 CEST] <alone-y> why drawtext have some halo around text?
[01:18:08 CEST] <klaxa> bad font? bad encoding?
[01:18:13 CEST] <klaxa> could be a number of things
[01:18:22 CEST] <alone-y> standart tahoma
[01:18:32 CEST] <klaxa> screenshot would be good i think
[01:18:35 CEST] <alone-y> not, codec lossless (zmbv)
[01:19:09 CEST] <klaxa> and ffmpeg call as well
[01:19:11 CEST] <klaxa> + output
[01:19:20 CEST] <klaxa> (on pastebin or similar)
[01:20:22 CEST] <alone-y> klaxa, how can i draw SAME size of text and font? i know my font tahoma 8.
[01:21:08 CEST] <alone-y> https://pastebin.com/bszW4nF8
[01:22:39 CEST] <klaxa> maybe use 16 instead of 15 then to scale by 2 instead of 1.something
[01:38:17 CEST] <alone-y> where is scale *
[01:38:20 CEST] <alone-y> ?
[01:39:15 CEST] <alone-y> look
[02:03:55 CEST] <alone-y> if u will print WHITE text its anyway will be with grayscale halo
[02:04:00 CEST] <alone-y> as antialiasing
[02:04:09 CEST] <alone-y> text, clearfonts etc
[02:04:24 CEST] <alone-y> how can i disabling it
[02:04:25 CEST] <alone-y> ?
[02:08:58 CEST] <klaxa> maybe see https://ffmpeg.org/ffmpeg-filters.html#drawtext-1 and the ft_load_flags?
[02:10:03 CEST] <alone-y> THANKYOU
[02:10:21 CEST] <alone-y> i have read it already but not guessed
[02:10:39 CEST] <alone-y> i hope with windows same?
[02:10:41 CEST] <alone-y> dont know.
[02:10:44 CEST] <klaxa> well i'm just throwing around ideas
[02:11:14 CEST] <alone-y> ft_load_flags=monochrome
[02:11:15 CEST] <alone-y> ?
[02:22:14 CEST] <wfbarksdale> a co-worker and I are experiencing an issue with getting video encoded with ffmpeg to play in quick time? does anyone know why this happens? we posted and SO question here: https://stackoverflow.com/questions/49961326/c-ffmpeg-video-missing-frames-and-wont-play-in-quicktime
[02:25:12 CEST] <furq> wfbarksdale: pastebin the ffprobe -show_streams output
[02:25:46 CEST] <alone-y> klaxa, all options still with anti-aliasing text
[02:25:47 CEST] <alone-y> ;(
[02:25:55 CEST] <alone-y> furq hello
[02:26:53 CEST] <furq> alone-y: i've never had that issue so i wouldn't know where to start
[02:27:48 CEST] <alone-y> you mean impossibility to put NO-anti-aliasing text?
[02:28:05 CEST] <furq> i have no idea
[02:28:08 CEST] <furq> i've never wanted to do that
[02:28:33 CEST] <alone-y> well, just in black-white mode fonts with antialiasing
[02:28:40 CEST] <furq> you could maybe do it with a fontconfig fonts.conf but idk how that would work on windows
[02:28:44 CEST] <alone-y> not same as in color mode ;)
[02:28:51 CEST] <alone-y> yesyes
[02:29:07 CEST] <alone-y> well,i have set full NO antialias
[02:30:24 CEST] <alone-y> thank you again for bw mode!
[02:30:27 CEST] <alone-y> its cool
[02:30:44 CEST] <wfbarksdale> furq: https://pastebin.com/Sr7GBS0c
[02:33:02 CEST] <furq> i don't see anything particularly unusual there other than the framerate
[02:40:11 CEST] <alone-y> farq i guess it's not possible
[02:40:11 CEST] <alone-y> https://stackoverflow.com/questions/45904104/rendering-pixel-perfect-subtitles-for-small-videos
[02:43:24 CEST] <alone-y> farq!!!!
[02:43:27 CEST] <alone-y> it's seems to me
[02:43:31 CEST] <alone-y> i found!!!
[02:43:47 CEST] <alone-y> i have combined filters with +
[02:43:56 CEST] <alone-y> and some combination is working.
[02:43:59 CEST] <alone-y> just a moment!!!
[02:53:04 CEST] <alone-y> ft_load_flags=render+monochrome
[02:53:06 CEST] <alone-y> this one!
[02:53:17 CEST] <alone-y> only together.
[02:53:45 CEST] <alone-y> <klaxa thank you very much!!!!
[10:27:04 CEST] <acresearch> people i am using this command to convert an audio file into a video file displaying the logo as background with a wave form,    everything works except that the image is no used and i get a black background           ffmpeg -i 077-ilm-fm-16-02-2018.mp3 -i Logo.jpg -filter_complex "showwaves=s=1920x1080:mode=cline:colors=blue" test.avi       anyone can help me?
[11:08:43 CEST] <Yukkuri> hi, i am trying to put h264 stream inside the fragmented contianer, but not sure how individual NAL units are related to MP4 samples. Question in detail on SO: https://stackoverflow.com/questions/49964268/how-h264-nal-units-to-be-corectly-sampled-inside-fragmented-mp4
[14:19:17 CEST] <acresearch1> people i am using this command to convert an audio file into a video file displaying the logo as background with a wave form,    everything works except that the image is no used and i get a black background           ffmpeg -i 077-ilm-fm-16-02-2018.mp3 -i Logo.jpg -filter_complex "showwaves=s=1920x1080:mode=cline:colors=blue" test.avi       anyone can help me?
[15:09:56 CEST] <Nethe> Hi @all. I am having a little issue with FFMPEG transcoding on my Nvidia P5000 (there are 4 in the server). The ffprobe of input: https://imgur.com/zuxYSey ... FFMPEG compiled with params: https://imgur.com/iNDs01l  And The error + command how I started transcoder: https://pastebin.blesmrt.net/ppyylwrwv/tcwejz .  The input source is in Brazil country. Can you please help me to check what is wrong?
[15:27:21 CEST] <Nethe_2> register Nethe Heslo123!! matyno at gmail.com
[15:27:52 CEST] <azarus> great job Nethe_2
[15:28:12 CEST] <Nethe_2> :D Just installed a client on my laptop :) So... :D
[15:28:18 CEST] <azarus> :D
[15:28:55 CEST] <Nethe> better :)
[15:29:12 CEST] <Nethe> azarus: do you have an idea about my issue? :/
[15:29:37 CEST] <kepstin> Nethe: looks like a mismatch between the installed nvidia driver version and the one the ffmpeg was compiled for
[15:29:38 CEST] <azarus> I do not.
[15:29:51 CEST] <kepstin> Nethe: make sure you're using the latest nvidia drivers *and* latest ffmpeg
[15:29:57 CEST] <kepstin> (what ffmpeg version are you using?)
[15:30:03 CEST] <Nethe> I am using same versions as on another servers
[15:30:06 CEST] <Nethe> where it is working
[15:30:12 CEST] <Nethe> ah
[15:30:15 CEST] <acresearch1> people i am using this command to convert an audio file into a video file displaying the logo as background with a wave form,    everything works except that the image is no used and i get a black background           ffmpeg -i 077-ilm-fm-16-02-2018.mp3 -i Logo.jpg -filter_complex "showwaves=s=1920x1080:mode=cline:colors=blue" test.avi       anyone can help me?
[15:30:46 CEST] <Nethe> kepstin: ffmpeg version N-90793-g30940be
[15:31:24 CEST] <Nethe> kepstin: This version we have on our testing server with same configuration and it is working well there... Same as cuda 9.1
[15:31:40 CEST] <kepstin> hmm, git master build from a few weeks ago, should be fine
[15:32:01 CEST] <kepstin> er, from a few days ago*
[15:32:08 CEST] <Nethe> :)
[15:32:47 CEST] <Nethe> Should I try to reinstall nvidia cuda and try to compile ffmpeg again? Do you think it could help/
[15:32:49 CEST] <kepstin> are you using the exact same ffmpeg binary as on the other machines?
[15:32:59 CEST] <Nethe> yes
[15:33:30 CEST] <ziloxiju> where is the windows version download link ? :)
[15:33:40 CEST] <Nethe> https://imgur.com/a/Trn2tqB
[15:33:54 CEST] <kepstin> ziloxiju: the ffmpeg download page has links to some unofficial windows builds
[15:34:08 CEST] <Nethe> kepstin: jones is testing and the second one is transcoder
[15:35:35 CEST] <Nethe> There is only one difference. Testing machine has 1 GPU, but transcoder has 4
[15:35:47 CEST] <kepstin> Nethe: hmm. I haven't done much nvenc stuff. One thing to note is that current ffmpeg versions want https://git.videolan.org/?p=ffmpeg/nv-codec-headers.git installed to get the headers.
[15:36:02 CEST] <ziloxiju> !n
[15:37:46 CEST] <Nethe> yes, those libraries have been opmiled already ... will try to do something else and try to google more
[15:37:57 CEST] <ziloxiju> Thanks you !
[16:01:25 CEST] <acresearch1> people anyone knows how to add a background to waveform in ffmpeg?    the commands in the firums seem to be old
[16:06:17 CEST] <furq> acresearch1: "showwaves=s=320x240[wv];[0:v][wv]overlay"
[16:06:28 CEST] <furq> change s to match the video dimensions
[16:09:20 CEST] <acresearch1> furq: i get a black screen, no waveform and no image
[16:09:30 CEST] <acresearch1> ffmpeg -y -i 077-ilm-fm-16-02-2018.mp3 -i Logo.jpg -filter_complex "showwaves=s=3000x3000[wv];[0:v][wv]overlay" test.mkv
[16:09:55 CEST] <furq> change 0:v to 1:v
[16:11:50 CEST] <Nethe> kepstin: I found something. The FFMPEG don't like nvidia headers version probably
[16:11:57 CEST] <Nethe> It is still showing [h264_nvenc @ 0x55f0b4dafd00] Cannot get the preset configuration: invalid version (15)
[16:12:44 CEST] <acresearch1> furq: hmmm ok i have a background image but no waveform
[16:18:11 CEST] <acresearch1> furq: its ok i got it,,, thanks :-)
[16:20:30 CEST] <IchGucksLive> hi all im from germany what is a good bitrate for YT 720p
[16:21:01 CEST] <IchGucksLive> the 7min Vido has 250MB i woudt like to shorten it
[16:22:40 CEST] <IchGucksLive> ffmpeg -i in.MOV -acodec libfdk_aac -ab 90k -vcodec libx264  -preset fast -crf 28 out.mp4
[16:23:08 CEST] <IchGucksLive> can i get a constant bitrate at 1500 aroun
[16:23:38 CEST] <IchGucksLive> it showes 800
[16:25:06 CEST] <IchGucksLive> ffmpeg 3.2
[16:29:58 CEST] <IchGucksLive> https://gist.github.com/mikoim/27e4e0dc64e384adbcb91ff10a2d3678
[16:30:04 CEST] <IchGucksLive> i will try this
[16:30:21 CEST] <sfan5> crf 28 is already fairly high, and 90k very low for aac-lc
[16:31:59 CEST] <furq> IchGucksLive: if you're uploading this to youtube then just upload the source
[16:33:52 CEST] <IchGucksLive> furq,  on ca low I-net connection not a good idee
[16:34:01 CEST] <IchGucksLive> as well is im limited in traffic
[16:34:24 CEST] <IchGucksLive> ffmpeg -i input -c:v libx264 -preset slow -profile:v high -crf 18 -coder 1 -pix_fmt yuv420p -movflags +faststart -g 30 -bf 2 -c:a aac -b:a 384k -profile:a aac_low output
[16:34:42 CEST] <sfan5> that command will probably even increase the size of your file
[16:34:54 CEST] <sfan5> just upload the original, even if it's painful to wait
[16:35:24 CEST] <IchGucksLive> ok isee
[16:35:33 CEST] <IchGucksLive> now down to 55MB
[16:35:57 CEST] <IchGucksLive> i will go for this
[16:50:39 CEST] <IchGucksLive> uploaded https://youtu.be/PXYliAOGsMM
[17:21:37 CEST] <romano> Is there a way to overlay a video beginning in a specific frame?
[18:24:49 CEST] <shfil> Is someone who is familiar with AVIOContext online?
[18:31:22 CEST] <klaxa> shfil: just post your question
[18:33:53 CEST] <shfil> how long AVIOContext's buffor should be allocated? I was thinking that only for decoding it's needed, but if I do this, there's nullptr dereference.
[18:33:59 CEST] <shfil> https://github.com/rwengine/openrw/pull/420/files
[18:34:56 CEST] <shfil> uncommenting lines in SDTloader's dtor causes crash.
[18:38:23 CEST] <klaxa> https://www.ffmpeg.org/doxygen/trunk/avio_8h.html#a853f5149136a27ffba3207d8520172a5 suggests 4kb for buffer size, but when using avio_open2() i don't think you should need to allocate it yourself?
[18:46:12 CEST] <shfil> 1) yeah I should increase size of buffer 2) I think I should allocate on my own it's like constructing AVFormatContext manually(I can be wrong).
[18:47:49 CEST] <shfil> ops it's already 4KB.
[18:50:16 CEST] <klaxa> dunno man
[18:50:16 CEST] <klaxa> >formatContext = loader.loadSound(index);
[18:50:28 CEST] <klaxa> what does that do?
[18:50:56 CEST] <klaxa> oh there we go
[18:52:42 CEST] <klaxa> where does it crash?
[18:57:03 CEST] <shfil> I'll rebuild it once again and tell in detail.
[18:59:10 CEST] <shfil> it crashes in dtor, debugger says something like _int_free. hmmm
[19:01:07 CEST] <klaxa> a stacktrace would help
[19:05:12 CEST] <shfil> (rebuilding in progress)
[19:12:51 CEST] <shfil> it's line : "av_free(ioBuffer);"
[19:14:42 CEST] <shfil> maybe I unnecessary think about this buffer and it's already handled by "avformat_close_input(&formatContext);"
[19:15:08 CEST] <klaxa> yes, don't av_free() the buffer you assigned, free AVIOContext.buffer
[19:21:34 CEST] <shfil> btw we use "AVCodecContext *codec" in project, it's deprecated, I was't able to find good examples with codecpar struct. Do you know  some of them?
[19:24:54 CEST] <klaxa> i use it in my code, recently cleaned up AVCodecContext* usage as well, one sec
[19:26:13 CEST] <klaxa> https://gist.github.com/klaxa/7f26c2e5137e99ff5883aaeb52d7a248
[19:26:24 CEST] <klaxa> although i think you need decoding?
[19:26:27 CEST] <klaxa> i just need demuxing
[19:32:58 CEST] <shfil> thanks it's good start
[19:34:53 CEST] <shfil> ah, last question, how AV_VERSION_INT works?
[19:35:12 CEST] <klaxa> dunno, never heard of that
[19:38:11 CEST] <shfil> I think about lts like trusty, doesn't it contain old ffmpeg without new interface?  So should I use precompiler directive for keeping backwards compatibility?
[19:41:21 CEST] <JEEB> trusty is really old at this point, and doesn't even contain FFmpeg IIRC
[19:42:11 CEST] <klaxa> ditched ubuntu after 10.04, not really up-to-date with their latest pokemons
[19:47:33 CEST] <JEEB> 16.04 has 2.8, 18.04 has 3.4
[19:50:17 CEST] <shfil> yeah, avcodec_send_packet was introduced in branch 3.x.
[21:31:22 CEST] <Nethe> kepstin: lol :D Imagine that the same situation already happened on the testing server :D Absolutely don't know why
[22:03:00 CEST] <Nethe> kepstin: i fixed it. nv-headers was too new. I used older version of them and it is working.
[22:06:20 CEST] <alone-y> hello, any 1 know how to add spaces after text with drawtext?
[22:06:36 CEST] <alone-y> i need to replace some characters on screen without symbols..
[22:32:58 CEST] <alone-y> how can i add spaces AFTER text in drawtext?
[22:33:13 CEST] <alone-y> \    Hello World
[22:33:21 CEST] <alone-y> but after Hello World?
[22:33:30 CEST] <alone-y> not before..
[22:38:38 CEST] <alone-y> dont worry, thank you
[00:00:00 CEST] --- Mon Apr 23 2018