[Ffmpeg-devel-irc] ffmpeg.log.20140609

[burek]

[00:16] <llogan> unlord: mpeg quant scale. i believe that includes flv1/h263/h263+/mpeg1/mpeg2/mpeg4/msmpeg*/mjpeg. probably h261 too, but i'm not sure.
[00:27] <unlord> llogan: we are doing a broad comparison of video coding formats to see how daala compares.
[00:28] <unlord> llogan: silverdev has been looking at h261 and h263
[00:29] <silverdev> why does the filesize changes when -v:q is above 31 for h263 but not h261?
[00:32] <silverdev> 36K	test1.mpg
[00:33] <silverdev> 32K	test2.mpg
[00:33] <silverdev> they whre made from the same video with -q:v set to 34 and 36
[00:55] <llogan> unlord: https://people.xiph.org/~xiphmont/demo/daala/daala-vp9summit-20140606.pdf
[01:05] <unlord> llogan: yes, I helped make some of those slides
[01:09] <unlord> llogan: I think what silverdev is saying is that he is seeing some odd behavior when he encodes with -f h263
[01:12] <llogan> i don't know why it does that, but who uses h263 anymore?
[01:13] <silverdev> I was told that -q:v scail for h263 is only supposted to be a value between 0 and 31 but It seems to keep compressing the outputs more at values above 31 instead of capping
[01:13] <unlord> llogan: there have been some pretty crazy proposals at the IETF regarding the MTI video codec for webrtc
[02:13] <needmorespeed1> Can anyone point to a very simple project that with ffmpeg compiling in QT Creator?
[03:29] <ac_slater> hey guys, I want to use ffmpeg (if possible) to packetize my h264 nal units for RTP. I'm digging though the sources and cant tell if this is a feature
[03:31] <ac_slater> ah nevermind libavformat/rtpenc_h264.c
[03:55] <ac_slater> sadly I live all of the libav* libs to be poorly documented with examples. Is there a solid references/book on these libs?
[04:33] <needmorespeed1> Anyone know of a simple example of QT with recent ffmpeg?
[04:48] <ac_slater> needmorespeed1: simple? What do you want it to do ?
[04:54] <needmorespeed1> just compile it to start
[04:54] <needmorespeed1> I was able to compile latest ffmpeg in Ubuntu
[04:55] <needmorespeed1> but when including the C with extern "C"
[04:55] <needmorespeed1> /home/vmuser/lib/ffmpeg/libavcodec/internal.h:200: error: undefined reference to `av_rescale_q'
[04:55] <ac_slater> needmorespeed1: usually you dont need to compile the library ... that's the point of dynamic libraries
[04:56] <ac_slater> ffmpeg is a big framework. You should get to know how to dynamically link. You're best bet is to link against the version that ubuntu has ... so you dont have make your users compile the version of ffmpeg you usedf
[04:56] <ac_slater> used *&
[04:56] <ac_slater> used **
[04:57] <ac_slater> you probably arent linking correctly. Or you compiled out that feature on accident
[05:01] <needmorespeed1> To config, I used
[05:01] <needmorespeed1> ./configure --prefix=$HOME/bin --enable-gpl --extra-libs=-ldl --enable-libx264 --extra-cflags="-I$HOME/lib/ffmpeg/include" --extra-ldflags="-L$HOME/lib/ffmpeg/lib"
[05:02] <ac_slater> needmorespeed1: like I said, dont compile it yourselkf
[05:02] <ac_slater> you probably just need to pull the ffmpeg-devel package
[05:02] <ac_slater> then you can link against the libs with using -lavcodec, stc
[05:03] <ac_slater> or use pkg-config
[05:04] <niao> I have a video I converted with ffmpeg that runs fine for me on Linux, but when I send to 2 different friends on windows, they report a black screen
[05:04] <niao> Is there a way to convert it to a windows friendly format?
[05:04] <ac_slater> niao: what format is it in?
[05:04] <niao> mp4
[05:04] <ac_slater> thats the container
[05:05] <ac_slater> what's the stream format?
[05:05] <niao> oops, my bad
[05:05] <niao> libx264?
[05:05] <ac_slater> sall good I'm pretty new to video stuff too
[05:05] <ac_slater> so it was encoded in h.264
[05:05] <ac_slater> into*
[05:06] <ac_slater> I would probably put into some mpeg2 format. Older versions of Windows didnt come with many decoders
[05:06] <niao> I already tried  -profile:v baseline but it gives me an error
[05:06] <niao> > maybe incorrect parameters such as bit_rate, rate, width or height
[05:06] <niao> I will try converting it to mpeg2
[05:06] <ac_slater> doubtful. But maybe someone here can weigh in
[05:10] <needmorespeed1> thx ac_slater, I'm trying again with more configure flags like --enable-avformat --enable-avutil
[05:11] <ac_slater> needmorespeed1: or link against the system libs ;)
[05:11] <ac_slater> needmorespeed1: are you also building QT
[05:11] <ac_slater> ?
[05:12] <needmorespeed1> not building QT
[05:12] <needmorespeed1> I should just go with the apt-get ffmeg you're saying
[05:15] <ac_slater> needmorespeed1: Most times,  programs link with libraries. Rarely (unless you're writing a closed-source application), do people include static builds of their dependencies.
[05:16] <ac_slater> you should always build against your target system. Unless of course you're doing cross-platform stuff... in that case, you'd probably build Qt and ffmpeg for linux and windows, etc
[05:21] <niao> I solved it by using -pix_fmt yuv420p
[05:21] <niao> Thanks ac_slater
[05:35] <needmorespeed1> ac_slater: I think my goal with compiling myself was using latest ffmpeg and x264 lib
[05:41] <barry>     Stream #0:0: Audio: mlp, 96000 Hz, stereo, s32
[05:41] <barry> how come ffprobe is showing as s32 when it's s24
[05:41] <barry> is this a bug
[06:07] <ac_slater> needmorespeed1: ah that makes sense then. But any newer version of ubuntu has recent ffmpeg stuff
[08:06] <excalibr> >Unknown input format: 'x11grab'
[08:06] <excalibr> ugh..why?
[08:19] <rahul_> Hii all
[08:21] <rahul_> I am trying to read tutorials. but as i am compiling them all i am getting errors undefined reference to `av_frame_alloc'
[08:22] <sacarasc> excalibr: Because you're using a static binary?
[08:32] <rahul_> please help on compiling tutorial ..
[08:38] <rahul_> i am trying this https://github.com/phamquy/FFmpeg-tutorial-samples#ffmpeg-tutorial-samples
[09:15] <rahul_> any one heere?
[11:13] <loa> hello, i am trying to convert
[11:14] <spaam> Nice :D
[11:16] <loa> Lol. Sorry, it was strange. I am trying to convert m4a to mp3 without loosing quality, i use such command: ffmpeg -v 5 -y -i Riverline\ -\ Reasons-X_GsbrqAJ28.m4a -acodec libmp3lame -ac 2 -ab 192k Riverline\ -\ Reasons-X_GsbrqAJ28.mp3
[11:16] <loa> but when job done i see that m4a have 192kbit/s but resulting file have 32kbit/s
[11:16] <loa> can somebody explain what i am doing wrong?
[11:26] <loa> here is that file https://dl.dropboxusercontent.com/u/25725476/Riverline%20-%20Reasons-X_GsbrqAJ28.m4a
[13:08] <brontosaurusrex> loa, i think its -b:a or something similar this days
[13:09] <loa> brontosaurusrex, i tried -q:a 2
[13:09] <loa> nothing changed
[13:09] <brontosaurusrex> loa, could be a samplerate problem, whats the original samplerate?
[13:10] <loa> brontosaurusrex, is it possible to download that file?
[13:10] <brontosaurusrex> yeah, hold on ...
[13:10] <loa> there is pritty good electronic music
[13:13] <loa> brontosaurusrex, i use such command mplayer -ao pcm:file=output.wav Riverline\ -\ Reasons-X_GsbrqAJ28.m4a && lame -h output.wav output.mp3 && rm output.wav
[13:13] <loa> and recieve 128kbit/s
[13:13] <loa> when i use -V option for lame i recieve 32kbit/s again :D
[13:14] <loa> and i can't notice if there is difference between two files when playing them
[13:16] <loa> brontosaurusrex, maybe it is normal situation? i heard earlier that bitrate can be dynamic
[13:17] <brontosaurusrex> no
[13:17] <brontosaurusrex> if you specify 192k, then it should be 192k
[13:18] <brontosaurusrex> -acodec libmp3lame -ab 192k < this is working for me on your sample
[13:18] <brontosaurusrex> this is with :  built on Mar 29 2014 05:48:25 with gcc 4.6 (Debian 4.6.3-1) git ffmpeg from gusari
[13:21] <loa> brontosaurusrex, how you check bitrate?
[13:21] <brontosaurusrex> loa, a: filesize, b: mediainfo
[13:22] <loa> brontosaurusrex, it is command line programs or what?
[13:22] <loa> i use mplayer on that file
[13:23] <loa> and i see such result https://dl.dropboxusercontent.com/u/25725476/screenshot-2014.06.09-15%3A22%3A48.png
[13:23] <loa> when i run such command mplayer -ao pcm:file=output.wav Riverline\ -\ Reasons-X_GsbrqAJ28.m4a && lame --preset insane output.wav output.mp3 && rm output.wav
[13:23] <loa> i recieve such result https://dl.dropboxusercontent.com/u/25725476/screenshot-2014.06.09-15%3A23%3A31.png
[13:23] <finemann> Hi, can I concatenate PCM WAV files using just cat(1) ?
[13:23] <Hello71> no, they have a header
[13:24] <Hello71> otherwise how would they encode sample rate, channels, etc
[13:24] <finemann> Hello71, I ask this because in a script on the ffmpeg wiki, there's a line where multiple FIFOs are concatenated using just cat.
[13:25] <finemann> http://trac.ffmpeg.org/wiki/How%20to%20concatenate%20%28join,%20merge%29%20media%20files
[13:25] <loa> finemann, it is ffmpeg command)
[13:25] <finemann> In particular this line: cat $all_a > $TMP/mcs_a_all &
[13:26] <Hello71> -f u16le
[13:27] <Hello71> that's not a wav file
[13:27] <finemann> Hello71, yes all of them are unsigned 16 bit LE.
[13:27] <finemann> Oh.
[13:27] <Hello71> that'll work fine then.
[13:27] <Hello71> *wav* files have a header
[13:27] <Hello71> that's the entire point of wav
[13:27] <finemann> That's new info. for me.  Thanks!
[13:27] <finemann> I mean, I thought WAV was just raw audio.
[13:27] <Hello71> obviously if you delete the header and store it in your program you'll be fine
[13:28] <Hello71> !g wikipedia wav
[13:45] <loa> brontosaurusrex, looks like mplayer don't show average bitrate for vbr
[13:46] <brontosaurusrex> loa, no idea
[14:06] <finemann> I just encoded a raw audio file with -f s16le and -ac pcm_s32le.  I'm confused.  How can I put a 32 bit audio file inside a file which uses 16 bit blocks?
[14:12] <finemann> Is it because, it's headerless audio and two blocks of the file would be considered as a single 32 bit block?
[17:00] <ethreal> Hello. Could someone please help me figure out the format of the "time = xyz" part of the ffmpeg vstats file?
[17:02] <ethreal> It's being output as a large decimal instead of the colon separated output I keep seeing online.
[17:14] <ethreal> Anyone?
[17:16] <Fjorgynn> welcome
[17:19] <veek> how do i merge an audio and a video file that i downloaded from youtube. It's a dash file ffh262 and a aac audio file
[17:19] <veek> avconv  -i "Intro to curves and vector functions-ruK0xBwUYdk.mp4" -i "Intro to curves and vector functions-ruK0xBwUYdk.m4a"  -c:v copy -c:a copy -strict experimental "Intro to curves and vector functions.mp4"
[17:19] <veek> when i play the output i just get audio
[17:20] <Fjorgynn> ffh262?
[17:20] <veek> err ffh264
[17:20] <Fjorgynn> why strict experimental?
[17:20] <veek> Video: h264 (Main)
[17:20] <Fjorgynn> should be ffmpeg/avconv -i test.mp4 -i test.m4a -c copy output.mp4
[17:21] <c_14> veek: a) pastebin your complete commandline and output, b) avconv support is in #libav
[17:21] <veek> ugh sry hang on that was for aac
[17:22] <ethreal> Does anyone know about the time values in the vstats file?
[17:22] <c_14> ethreal: that time is the end pts * the time base
[17:23] <ethreal> Oh. Uhm. End pts?
[17:24] <veek> c_14, http://pastebin.com/br0hA90u
[17:25] <ethreal> Could I convert it into seconds purely based on the output of the vstats file?
[17:27] <c_14> veek: according to the command and output, that should be working. Can you try downloading one of the static ffmpeg builds?
[17:27] Action: veek noddles
[17:31] <c_14> ethreal: I don't think so.
[17:31] <c_14> There's also a note in the source saying that the attribute that is used in that calculation is deprecated.
[17:33] <ethreal> I extracted a bunch of frames with changes in scene using this command: http://pastebin.com/emDX9QcY
[17:33] <ethreal> Is there no way for me to determine at what time it extracted it?
[17:34] <ethreal> By time I mean time relative to the video's start time
[17:40] <c_14> ethreal: The only option I know of would be to burn the timecode into the pictures.
[17:41] <veek> c_14 tried this, no luck ffmpeg -acodec copy -vcodec copy "Intro\ to\ curves\ and\ vector\ functions-ruK0xBwUYdk.mp4" -i "Intro to curves and vector functions-ruK0xBwUYdk.m4a"  Intro\ to\ curves\ and\ vector\ functions.mp4
[17:42] <c_14> the order of the arguments is wrong and you forgot the -i and you were double escaping, try ffmpeg -i Intro\ to\ curves\ and\ vector\ functions-ruK0xBwUYdk.mp4 -i Intro to curves and vector functions-ruK0xBwUYdk.m4a -codec copy Intro\ to\ curves\ and\ vector\ functions.mp4
[17:42] <c_14> eh, the second one needs quotes
[17:43] <c_14> or escaping
[17:44] <veek> okay -i -i -codec copy
[17:47] <veek> ah phew!
[17:47] <veek> c_14, thanks
[17:47] <Fjorgynn> I see
[18:50] <mleise> Is "-vf interlace" the correct way to turn a progressive video into an interlaced one for encoding with x264's bff=1 ?
[18:52] <mleise> Interlacing is the only way to play a video @ 50 (half-)frames on my gfx card with VDPAU (level 4.1).
[18:52] <c_14> I'm not sure why you want it, but -vf interlace should interlace the video.
[18:53] <mleise> And I figured that I probably don't need -flags +ilme+ildct, since I found no equivalent in the x264.
[18:53] <mleise> ...options
[18:53] <mleise> c_14: Everyone seems to hate interlacing, but isn't this MBAFF just the best of both worlds?
[18:54] <mleise> 25p content stays progressive, while 50p parts retain their temporal resolution
[18:54] <t4nk489> Dowloaded ffmpeg from git ffmpeg. While making ffmpeg on centos receive warning "have_fminf" is not defined then fatal error libavutil/libm.h:87 static declaration of 'fminf' follows non-static declaration. Should I just go back to older version at github?
[18:54] <mleise> when encoding at 25 FPS with interlacing
[18:55] <llogan> t4nk489: can you use git bisect to find out what commit may have caused the issue?
[18:56] <t4nk489> point me to where I should go
[18:58] <llogan> http://git-scm.com/book/en/Git-Tools-Debugging-with-Git#Binary-Search
[18:58] <t4nk489> OK
[18:58] <llogan> what was your ffmpeg configure?
[19:01] <c_14> mleise: If I'm reading this correctly, mbaff is better than just interlacing, but it shouldn't be better than progressive video if you're going to be displaying it on a progressive display.
[19:02] <mleise> c_14: Right! But by using half the amount of macroblocks per second 25 FPS interlaced can be played back on h264 level 4.1 hardware @ 50 FPS (after hardware assisted deinterlacing)
[19:03] <mleise> I tried both mpv and Media Player Classic Home Cinema, and the deinterlaced output looks surprisingly well through nvidia's video engine.
[19:04] <c_14> Even with video with lots of motion?
[19:04] <mleise> Well, I have mostly panning and landscape shots
[19:05] <llogan> t4nk489: i can't duplicate the issue using git head with "./configure --enable-gpl"
[19:05] <mleise> But I _do_ see the artifacts in the credits scroller where I used a serif font, with 1 pixel wide serifs
[19:05] <mleise> The serifs pop up, the disappear
[19:05] <mleise> *then
[19:08] <mleise> Naturally side-to-side motion should not be affected, but up-down motion is. I guess it depends heavily on the content. Let's see how the shaky shots while driving through the dips in death valley come out. There is lots of up-down motion in it :)
[19:08] <t4nk489> new to git bisect. used git bisect start, then bad, then good.  good reports long string with text 'was both good and bad'.
[19:09] <iive> mleise: it is extremely hard to deinterlace properly, especially on progressive displays.
[19:10] <iive> the interlace encoder flags work on libx264, x264 have tff/bff
[19:10] <t4nk489> Complaint appears during make. If I remove gpl from configure then complaint goes away during make?
[19:11] <llogan> i don't know. i was just stating that it works for me, and that was my configure
[19:13] <iive> t4nk489:  do the bisect with the options that break the build
[19:18] <mleise> live: Yes, I'm using -vf interlace for ffmpeg and the -bff=1 flag for x264. I tried 25p, but I found it hard to keep track of details during motion. The brain has a much easier time at 50 fps it seems.
[19:18] <mleise> btw, do you say 50i or 25i for 50 half-frames per second?
[19:24] <iive> half-frame is called field. when you talk about fps then it is always in frames per second. but 50i is not fps, so...
[19:24] <iive> i don't know.
[19:25] <iive> since marketing prefers bigger numbers i'd say that 50i is 25fps
[19:28] <u19809> hi all,
[19:28] <u19809> I need some help forwarding my camera to youtube live events (RTMP)
[19:29] <u19809> I can see my camera's output (using rtsp)
[19:29] <u19809> and I can send a fixed avi file to youtube using rtmp
[19:29] <u19809> but I am unable to bind them together
[19:32] <u19809> these are the commands http://pastebin.com/c9d5FtcB
[19:40] <u19809> fflogger : tricky since I test on a remote server using teamviewer and cut and past does not work :(  Hold on
[19:40] <llogan> fflogger is a bot, BTW.
[19:42] <u19809> do I look stupid
[19:43] <u19809> ok ... this should be it ... the command I try to forward rtsp to rtmp http://pastebin.com/jjcu2T9F
[19:51] <llogan> this output looks like it came from a different command than the one you provided in your first link
[19:53] <u19809> yes the commands I first sent were the ones that work, the one I sent last is the one I tried for forwarding the rtsp stream to the rtmp
[19:53] <u19809> I will pastebin the others too ... hold on
[20:01] <u19809> llogan : this is the command that sends a good stream to youtube http://pastebin.com/CTRcpYGw
[20:02] <u19809> llogan : for the input stream I just ran ffplay rtps://.... and the content showed
[20:08] <llogan> u19809: but your failing command and console output is missing.
[20:09] <u19809> damn ... my paste failed ... hold on
[20:11] <u19809> here it is ... http://pastebin.com/wuQAhjPH (typed over by hand though :( )
[20:13] <llogan> so this is the command that goes with http://pastebin.com/jjcu2T9F ?
[20:16] <llogan> also it appears you are using two different versions of ffmpeg. old 0.10.11 and 2.1.4
[20:20] <u19809> yes ...
[20:20] <u19809> hmmm need to check these versions ... did not check it
[20:22] <u19809> llogan : odd ... both are recent but the 0.11 is from ubuntu saucy and the 2.1 is from a fedora ???
[20:22] <u19809> both should be up to date
[20:22] Action: llogan doesn't know what is going on
[20:23] <llogan> does wuQAhjPH work if you omit the audio? does it work re-encode the streams?
[20:24] <u19809> no youtube requires an audio or else it refuses the stream
[20:25] <llogan> i'm not much of a streaming guy. and im not sure if pcm_s16le is something you can use
[20:26] <u19809> youtube seems to need an aac
[20:26] <u19809> so what encoding do you propose ?
[20:28] <llogan> according to your build at the very least you have to use: -acodec aac -strict experimental
[20:30] <u19809> yes I just tried that ... it seems to be streaming but oddly only at 0.1kbps which is way to low for a 1080 p stream ... I do not get youtube errors but also youtube does not seem to detect my stream ... unless I press CTRL-C and then it seesm youtube did receive something
[20:31] <llogan> are you stream copying or re-encoding?
[20:31] <llogan> (the video and audio)
[20:31] <u19809> video copy (it is already h264) audio reencoding as it is pcm_mulaw and should be aac
[20:32] <u19809> It looks perhaps that the video stream does not arrive in the flv. only the audio is present perhaps
[20:32] <u19809> When I captured the video raw to a file I got 800kbps which is normal I think
[20:33] <llogan> try re-encoding the video like you did in one of your other many pastes
[20:34] <u19809> I can but should I since it is already in the correct format
[20:34] <llogan> it might be a really high bitrate.
[20:38] <u19809> nope ... still about 1kbps ...
[20:40] <llogan> then i'm out of ideas and fairly ignorant of RTMP. try the ffmpeg-user mailing list. make sure to include your actual failing ffmpeg command and the complete console output.
[20:41] <u19809> ok thx anyway
[22:38] <communiications> Hi
[22:39] <communiications> i have a problem with joining mkv files together, which i explayned in the ffmpeg support forum, could someone get a look at it? http://ffmpeg.gusari.org/viewtopic.php?f=11&t=1507
[22:41] <c_14> pretty sure mkv doesn't support the concat protocol
[22:41] <c_14> use the filter or the demuxer
[22:41] <c_14> https://trac.ffmpeg.org/wiki/How%20to%20concatenate%20(join,%20merge)%20media%20files
[22:46] <tyler1> is it possible to get ffmpeg to read a sequence of images that are  _00001.png  _00002.png _00003.png  but then something like _00003_1.png
[22:47] <c_14> tyler1: you can use -pattern_type glob
[22:52] <communiications> ok, im trying the demuxxer and get the following error http://pastebin.com/hr5chsKJ
[22:52] <c_14> can you pastebin the whole command and the output?
[22:52] <tyler1> c_14:  is there any good documentation on that you could link me?
[22:53] <llogan> http://ffmpeg.org/ffmpeg-formats.html#image2-2
[22:53] <llogan> http://trac.ffmpeg.org/wiki/Create%20a%20video%20slideshow%20from%20images
[22:53] <tyler1> so something like %05d*.png would capture that
[22:54] <communiications> c_14: http://pastebin.com/cL073p24
[22:56] <c_14> try switching from -c copy to some video codec for testing purposes
[22:59] <communiications> when im not using copy it works fine http://pastebin.com/UPPGsS6g
[23:00] <communiications> (also says invalid dropping on line 38)
[23:01] <tyler1> c_14:  if I change to glob pattern type, it doesn't seem to open the files
[23:01] <c_14> I'm guessing there's something messed up with the timestamps in one or multiple of the files you're concatenating. Can you remux each of the files individually?
[23:01] <tyler1> can I not use %05d if I'm using glob pattern ?
[23:02] <c_14> tyler1: I think the use of both glob and % patterns was deprecated.
[23:02] <c_14> that would be glob_sequence, you can test if it works.
[23:02] <c_14> But then you have to escape the *
[23:03] <tyler1> hm I can't find the docs on glob_sequence
[23:05] <communiications> i remuxed every input file in mkvmerge, but that soesnt fixes it
[23:06] <c_14> tyler1: in the link llogan pasted, the first one
[23:07] <tyler1> ah I see thanks
[23:07] <tyler1> it says glob_sequence will be deprecated, not glob
[23:07] <c_14> communiications: then I have no idea, try concatenating them pairwise and see which one is giving issues (if it's only one)
[23:08] <c_14> tyler1: yes, and the pattern you were trying to use (%05d*.png) is a glob_sequence not a glob [and it should be %05d%*.png]
[23:09] <tyler1> ah I understand c_14, let me try that
[23:11] <tyler1> ok c_14 i'm getting closer, i have another error now
[23:13] <tyler1> nvmind , still gettin cannot open file,  with this source :  /temp/test%05d%*.png
[23:14] <tyler1> and I have test00000.png  test00001.png  test00002_1.png   etc.
[23:17] <c_14> try switching to a simple glob /temp/test*.png ?
[23:19] <tyler1> i think so, let me try again though
[23:22] <tyler1> yea i'm getting in stderr:
[23:22] <tyler1> undefined
[23:28] <tyler1> it's so strange, because if I ls the pattern, its showing me all the files
[23:37] <c_14> Does the process running ffmpeg have permissions to access the files?
[23:37] <c_14> s/process/user
[23:43] <tyler1> that undefined was a false alarm, nevermind
[23:46] <tyler1> ok so glob_sequence doesn't break anything, but it doesn't look like it's getting the frames that are like this 00002_1.png
[23:58] <tyler1> i can't even get pattern type glob to work with /temp/*.png
[23:59] <c_14> Is it throwing errors?
[23:59] <tyler1> stderr is empty
[23:59] <tyler1> let me try it command line
[00:00] --- Tue Jun 10 2014