[Ffmpeg-devel-irc] ffmpeg-devel.log.20140304

[burek]

[00:12] <BBB> kurosu_: I'll have a look after dinner; I still think the initial shift after unpack is odd but I only glanced through it very quickly
[00:22] <cone-430> ffmpeg.git 03Andrew Kelley 07release/2.2:ba21499648bb: lavfi: add compand audio filter
[00:22] <cone-430> ffmpeg.git 03Anton Khirnov 07release/2.2:15ae305007c0: af_compand: add a dependency on strtok_r
[00:22] <cone-430> ffmpeg.git 03Hendrik Leppkes 07release/2.2:031d3b66c2ea: latm: Always reconfigure if no extradata was set previously
[00:22] <cone-430> ffmpeg.git 03Anton Khirnov 07release/2.2:00d5ff64315a: af_compand: replace strtok_r() with av_get_token()
[00:22] <cone-430> ffmpeg.git 03Keiji Costantini 07release/2.2:5df52b0131d3: ituh263: reject b-frame with pp_time = 0
[00:22] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.2:1af235f6b344: Merge remote-tracking branch 'qatar/release/10' into release/2.2
[01:19] <cone-430> ffmpeg.git 03Dale Curtis 07release/2.0:33aa2c5d6b22: h264: Clear ERContext.cur_pic when unref'ing current picture.
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:dfdeabadcaf1: cmdutils: update year
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:9847f02fafac: dnxhdenc: fix mb_rc size
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:7c17207ab9ac: avcodec/vmnc: Check  that rectangles are within the picture
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:ca5d6c615e50: avcodec/jpeg2000dec: fix error detection in pix_fmt_match()
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:d0d0924947a4: avcodec/takdec: always check bits_per_raw_sample
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:33c47d3976f7: avcodec/vc1: reset fcm/field_mode in non advanced header parsing
[01:19] <cone-430> ffmpeg.git 03Justin Ruggles 07release/2.0:0962c26b6ba8: samplefmt: avoid integer overflow in av_samples_get_buffer_size()
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:13ce3673684e: avcodec/wmalosslessdec: fix mclms_coeffs* array size
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:5e7e43c33ea4: avformat/mpegtsenc: Check data array size in mpegts_write_pmt()
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:67b943ad661e: avcodec/utils: set AVFrame format unconditional
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:bc1c8ec5e650: avcodec/msrle: use av_image_get_linesize() to calculate the linesize
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:d41f4e8dc82b: avcodec/ansi: fix integer overflow
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:f0ee0fcbfcb5: avcodec/snow: split block clipping checks
[01:19] <cone-430> ffmpeg.git 03Timothy Gu 07release/2.0:612ef09a1819: configure: use pkg-config to detect libbluray
[01:19] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:e939c7b7f533: avcodec/utvideoenc: fix slice_bits size
[01:19] <kierank> interesting that kodak are using ffmpeg
[01:27] <cone-430> ffmpeg.git 03Michael Niedermayer 07release/2.0:7de7bd4f563a: update for 2.0.4
[01:27] <Compn> yep thats pretty cool'
[01:41] <cone-430> ffmpeg.git 03wm4 07fatal: ambiguous argument 'refs/tags/n2.0.4': unknown revision or path not in the working tree.
[01:41] <cone-430> Use '--' to separate paths from revisions
[01:41] <cone-430> refs/tags/n2.0.4:HEAD: avformat/microdvd: export the declared frame rate
[01:47] <llogan> interesting that kodak still exists
[01:50] <Compn> in patents and trademarks only
[01:52] <cone-430> ffmpeg.git 03Anton Khirnov 07release/1.2:2256b2a3c1a6: pthread_frame: flush all threads on flush, not just the first one
[01:52] <cone-430> ffmpeg.git 03Anton Khirnov 07release/2.1:96603d5753c7: pthread_frame: flush all threads on flush, not just the first one
[03:01] <cone-430> ffmpeg.git 03Michael Niedermayer 07master:ec6d043f8e58: avcodec/h263dec: Workaround H263 end padding bug
[05:21] <cone-430> ffmpeg.git 03Lukasz Marek 07master:19b3a25048d9: lavu/log: add device categories
[05:21] <cone-430> ffmpeg.git 03Lukasz Marek 07master:f607767d93b7: lavd: add categories to device implementations
[05:22] <cone-430> ffmpeg.git 03Lukasz Marek 07master:06fed44051d0: lavd/avdevice: add device iterators
[05:22] <cone-430> ffmpeg.git 03Michael Niedermayer 07master:f34cceb1fc58: Merge remote-tracking branch 'lukaszmluki/master'
[11:24] <ubitux> http://joco.name/2014/03/02/all-rgb-colors-in-one-image/ fun
[12:50] <cone-645> ffmpeg.git 03Vittorio Giovara 07master:8accddeb5806: vf_frei0r: adjust error messages
[12:50] <cone-645> ffmpeg.git 03Michael Niedermayer 07master:2b0ff8d6424f: Merge commit '8accddeb5806cd98e3803b4ddf2a2ef576d0e4d9'
[13:18] <cone-645> ffmpeg.git 03Vittorio Giovara 07master:4e0be9c86f20: vf_frei0r: prevent a segfault when filter parameters are not set
[13:18] <cone-645> ffmpeg.git 03Michael Niedermayer 07master:bfe03205a86c: Merge commit '4e0be9c86f2003dd3ba2eb8fad01e8d0e538075e'
[13:38] <cone-645> ffmpeg.git 03Vittorio Giovara 07master:61b323ce7c7c: vf_frei0r: refactor library loading from env variable
[13:38] <cone-645> ffmpeg.git 03Michael Niedermayer 07master:d63b032d64e9: Merge commit '61b323ce7c7cdc101eadfd7de2203922b8a39e8d'
[13:53] <plepere> BBB : thanks for the feedback. it's very helpful
[13:53] <BBB> lots of code still to go ...
[13:53] <BBB> that's a lot of code
[13:56] <BBB> oh labels are loop-internal also
[13:56] <BBB> so LOOP_INIT doesn't need an argument
[13:57] <BBB> just make it .loop: everywhere
[13:57] <BBB> it'll work
[13:57] <plepere> *mind blown*
[13:58] <JEEB> :)
[14:02] <plepere> well seems like I've got still a lot of work. I'm impressed by how rigorous the code can and must be.
[14:02] <JEEB> well, after you get into it once, it should become easier to keep stuff like that :)
[14:03] <plepere> yes.
[14:04] <ubitux> plepere: often with the asm, once it's written it's not much touched anymore, so better have it done well for at first
[14:05] <cone-645> ffmpeg.git 03Vittorio Giovara 07master:5397386effba: mathops: move macro to the only place it is used
[14:05] <cone-645> ffmpeg.git 03Michael Niedermayer 07master:8cc581fbf77a: Merge commit '5397386effba2e53e4ff82852a86f6be4d59e9c1'
[14:05] <plepere> yes, I understand, but I'm impressed how it's always possible to have an always cleaner code.
[14:12] <BBB> yeah, basically what ubitux says, it's unlikely we'll touch or refactor this in major ways in the future, so better do it right at once
[14:13] <BBB> otherwise in the future, it's triple the effort (redo x86 asm, redo arm/neon asm, redo ppc asm, redo mips asm, redo c, ...)
[14:16] <plepere> BBB : in which file is the C wrapper for the W > 16 ?
[14:19] <ubitux> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=732159 interesting thread about mplayer
[14:19] <ubitux> i like how the argument that mplayer2 is not maintained anymore is completely ignored
[14:20] <ubitux> as well as the fact that mplayer works with ffmpeg but not libav is an argument to drop mplayer and not (fix ?) libav
[14:22] <ubitux> "Shall we go ahead with the removal now?" wth
[14:23] <ubitux> how can he ignore all the arguments stated?
[14:23] <nevcairiel> the debian multimedia people seem all to be  silly people
[14:23] <ubitux> nevcairiel: no
[14:24] <ubitux> that's the security guy moritz who is blocking the ifp for ffmpeg
[14:24] <ubitux> and some libav folks
[14:24] <nevcairiel> s/multimedia//
[14:24] <ubitux> the "debian people" are actually stating arguments
[14:24] <ubitux> but they are just totally ignored
[14:24] <ubitux> wth is this FUD
[14:25] <nevcairiel> how can one person even do that, especially if he is not the packaging guy
[14:25] <ubitux> security team has a lot of power it seems.
[14:26] <nevcairiel> some days i'm even happier to never deal with linux distributions
[14:34] <ubitux> "In lack of any *constructive* comments about this, I would say yes, let's remove them." come on siretart, the fact that mplayer2 is less maintained than mplayer (and will break down with recent libav) is not a constructive argument?
[14:36] <nevcairiel> it doesnt even work with latest libav anymore? that'll haunt  him again for the libav10 release :p
[14:36] <siretart> ubitux: how does this help with still nobody willing to update the package?
[14:37] <ubitux> it says that dropping it will cause a lot of trouble
[14:37] <JEEB> nevcairiel, I think he has already poked it
[14:37] <JEEB> since the libav 10 transition is already under way
[14:38] <siretart> ubitux: at least, mplayer2 doesn't post lovely comments such as http://lists.mplayerhq.hu/pipermail/mplayer-users/2014-March/087255.html - not very encouraging from my POV
[14:39] <ubitux> well mplayer2 is dead so obviously you won't see any comment
[14:39] <nevcairiel> if it doesnt post any comments anymore, how is that any better :D
[14:39] <thardin> why are there intentionally broken things being kept around?
[14:41] <siretart> thardin: because you don't fix them
[14:41] <ubitux> the user is always the responsible ;)
[14:42] <thardin> speaking of which, I should figure out a nice-ish way to fix that mxf bug
[14:42] <thardin> probably just taking that libmxf source file as-is and poking at it so it compiles nicely
[14:43] <thardin> since I want to be able to keep it up-to-date as painlessly as possible
[14:44] <thardin> "the more code I have the poorer I am" and so on
[14:44] <JEEB> esp. if it's stuff you have to maintain >_>
[14:45] <thardin> yeah, esp. since gsoc doesn't seem to be happening
[14:45] <thardin> can we use git-subtree?
[14:45] <thardin> iirc that can be used to track a specific file in a specific repo
[14:49] <nevcairiel> doubtful thats going to fly really well
[14:52] <thardin> ill just call ppl names
[14:53] <thardin> anyway, that's in the future
[15:15] <j-b> any mplayer devs around?
[15:15] <j-b> or mp*
[15:16] <ubitux> i have write access to mplayer but i'm not sure that's what you need :p
[15:16] <cone-645> ffmpeg.git 03Vittorio Giovara 07master:973dc4e8d4ae: put_bits: Remove unused includes
[15:16] <cone-645> ffmpeg.git 03Michael Niedermayer 07master:efc4439c894b: Merge remote-tracking branch 'qatar/master'
[15:18] <kierank> thardin: i will be seeing the libmxf author next week if there's anything you want
[15:20] <j-b> ubitux: do you support MAP files?
[15:22] <ubitux> j-b: doesn't ring any bell; what's that? (hard to google such name)
[15:29] <j-b> ubitux: that's a good answer then
[15:29] <ubitux> i don't see anything in multimedia wiki or whatever
[15:48] <thardin> kierank: cool
[15:48] <thardin> I want to be able to have that EssenceContainer -> WrappingKind code as its own module
[15:49] <thardin> though it probably already works that way - haven't checked
[16:08] <cone-645> ffmpeg.git 03olugbone 07master:b2880a3de9e3: avformat/mpegtsenc: fix HEVC MPEG TS stream type
[16:12] <ubitux> funman: btw, 2.1.4 has av_alloc_vdpaucontext()
[16:13] <ubitux> i'm not sure to understand what exactly is the issue all this flaming is about though
[16:14] <ubitux> but maybe we should backport version added later later to that branch?
[16:15] <JEEB> hmm regarding that HEVC in MPEG-TS thing, did someone finally get access to the latest document version and thus made sure that the ID is correct?
[16:15] <kierank> yeah i was going to say
[16:15] <kierank> i've seen so many stream types
[16:15] <ubitux> funman: also, remi is talking about bugs in the old vdpau code; something we could/should backport from master?
[16:33] <funman> ubitux: i do not trust any word of him
[16:34] <funman> ubitux: also if you backport new API, how will we detect it with version numbers?
[16:34] <nevcairiel> carefully?
[16:36] <ubitux> by bumping minor but well..
[16:38] <nevcairiel> thats one of the reasons why new things shouldnt be ported to a stable release branch :p
[16:39] <ubitux> yeah i guess
[17:01] <b_jonas> for some reason, the png codec for ffmpeg creates very large images and seems to have no documented settings
[17:01] <b_jonas> I'll have to use imageMagick to create sane pngs instead
[17:02] <nevcairiel> it doesnt try to optimize the pngs at all, like many image tools do, it just takes the input image and compresses it with zlib
[17:02] <nevcairiel> you should be able to set the compression level however
[17:03] <nevcairiel> -compression_level 0..9 i guess
[17:04] <nevcairiel> default should be 6
[17:10] <b_jonas> nevcairiel: -compression_level ? let me check the docs
[17:11] <b_jonas> nevcairiel: I don't see that option mentioned in the docs anywhere. sad
[17:11] <b_jonas> but I can still try 
[17:13] <b_jonas> (imagemagick has three complicated options for png compression, and I only partly understand how they work)
[17:30] <Zeranoe> Can anyone explain why -ac 2 doesn't automatically downmix 5.1 to 2 channel?
[17:30] <nevcairiel> what else does it do
[17:33] <Zeranoe> it drops LFE http://superuser.com/questions/723731/downmix-drops-low-frequency-channel
[17:33] <nevcairiel> thats part of the down mix
[17:33] <JEEB> the specs say that you should drop it
[17:34] <JEEB> s/should/shall/
[17:34] <nevcairiel> its recommended to drop the lfe when down mixing surround to stereo
[17:34] <Zeranoe> I'm confused... Why would you do that
[17:35] <nevcairiel> to avoid overflowing the audio channels
[17:35] <Zeranoe> Don't you lose quite a bit by doing that though
[17:36] <nevcairiel> a normal 5.1->stereo downmix already puts 3 channels into one, adding LFE adds a 4th
[17:36] <Zeranoe> I ended up running pan=stereo: FL < FL + 0.5*FC + 0.6*BL + 0.6*SL : FR < FR + 0.5*FC + 0.6*BR + 0.6*SR
[17:36] <nevcairiel> that results in either overflows and clipping, or in a much lower base volume
[17:36] <nevcairiel> especially because you would need to boost LFE by 6dB to preserve the audio signal
[17:37] <nevcairiel> Zeranoe: there is still no LFE in that
[17:40] <Zeranoe> oh
[17:41] <Zeranoe> So how is that pan filter different from -ac 2? Or is it not?
[17:41] <nevcairiel> not sure how pan works
[17:41] <nevcairiel> but it lets you specify the mixing factors
[17:42] <nevcairiel> and if you want LFE in there, you need to add a "+ LFE" somewhere :p
[17:42] <nevcairiel> -ac is fully automatic
[17:42] <nevcairiel> not sure you can overwrite the mixing factors with just -ac
[17:43] <J_Darnley> In case you don't know: the LFE channel is not a "bass" channel.
[17:44] <Zeranoe> But technically I shouldn't want LFE when downmixing to stereo?
[17:44] <nevcairiel> Dolby suggests not to keep it
[17:45] <Zeranoe> J_Darnley: It's not? "(LFE) channel is the name of an audio track specifically intended for deep, low-pitched sounds..."
[17:48] <J_Darnley> Low Frequency Effects
[17:48] <J_Darnley> effects being the key word
[17:48] <J_Darnley> supplemental sounds for explosions ad the like
[17:48] <Mavrik> hmm, isn't that just the channel that gets routed to subwoofer on most systems?
[17:49] <nevcairiel> it is
[17:49] <nevcairiel> but many audio systems also low pass the other channels and redirect that
[17:50] <Mavrik> mhm, makes sense
[17:51] <Case> it's used for bass at least in music files I have
[17:51] <Case> and it sounds much worse without it. I personally don't understand the weird recommendation to leave it out from downmix
[17:52] <nevcairiel> surround music is rather rare
[17:52] <nevcairiel> music is also a bit different
[18:04] <mateo`> Hello there ! Would the ffmpeg project accept a patch which introduce linking to a totally closed library ?
[18:04] <j-b> No
[18:05] <ubitux> really?
[18:06] <mateo`> j-b: why ?
[18:07] <j-b> mateo`: dunno. :)
[18:07] <j-b> mateo`: it just seemed not really the open source way
[18:08] <ubitux> we have a --enable-nonfree flag for similar things
[18:08] <Zeranoe> mateo`: Your free to branch and patch locally though
[18:08] <ubitux> though, the source are open for most/all
[18:09] <ubitux> not sure about decklink for instance
[18:10] <Zeranoe> ubitux: On that topic, do you know how decklink works? Can the SDK be cross compiled and the ffmpeg work with a decklink device on win?
[18:11] <Zeranoe> I've been asked for it in the Windows builds, and I think it would be helpful. I don't know if it can be cross compiled and work through
[18:12] <j-b> Decklink works fine on Windows using DShow
[18:15] <Zeranoe> How is it working in FFmpeg though? Though DShow filters?
[18:46] <mateo`> so in the end, is it something that can be accepted or not by upstream ?
[19:00] <ubitux> Zeranoe: absolutely no idea
[19:05] <ubitux> lol @ 61b323ce7c ; the code wasn't even compiled
[19:05] <ubitux> nice review team work
[19:07] <mateo`> ubitux: ?
[19:07] <ubitux> look at the end of the commit diff
[19:08] <mateo`> what the ...
[19:10] <Compn> mateo` : patches for ffmpeg are always welcome, please post to ffmpeg-devel list
[19:11] <Compn> dont listen to j-b trolling, hes a vlc fanboy :P
[19:11] <ubitux> Compn: well if it's not going to be applied...
[19:11] <ubitux> Compn: such question looks relevant
[19:11] <Compn> ubitux : users like to have patches , even if not applied
[19:11] <ubitux> Compn: i don't like encouraging people to waste time
[19:11] <ubitux> you seem to forget the time/effort cost of writing code
[19:12] <mateo`> Compn: i'd like to know if it's gonna be accepted before work is being done, what's why I ask
[19:12] <Compn> ubitux : you know there are people who have patches that wont be applied to ffmpeg, they are still useful to people even if not included here
[19:12] <Compn> mateo` : well what feature are you thinking about working on ?
[19:13] <Compn> you said a patch for a closed lib, what lib ?
[19:13] <mateo`> Compn: closed lib, a dolby one
[19:14] <j-b> what for?
[19:14] <Compn> mateo` : encoder / decoder of dts-ma-hd ?
[19:14] <Compn> lol, now that j-b sees its a patch to piss off dolby, now hes interested :)
[19:14] <mateo`> decoding, it's part of the dolby sdk i think, i don't have much detaisl
[19:15] <j-b> Compn: clearly not :)
[19:16] <Compn> mateo` : it sounds good, bu i dont particularly know what feature could be used. ask michaelni , he can say wether or not it will be applied ...
[19:17] <j-b> Compn: however, I fail to see what codecs are in this sdk that FFMpeg cannot decode
[19:18] <mateo`> j-b: a matter of license maybe, can't tell much more
[20:22] <kierank> i doubt dolby will be happy with you adding things to ffmpeg
[21:17] <jhoffing> Hello, where can I make a feature request for ffmpeg's libfaac codec? 
[21:18] <Zeranoe> jhoffing: faac is a little old, what feature did you want
[21:19] <jhoffing> I'd like to have an option to choose the mpeg version for encoding an audio file into AAC
[21:19] <jhoffing> right now it's hard-coded to MPEG4. but i'd like to be able to make it MPEG2
[21:20] <jhoffing> Is there another codec I can use instead that lets me have this option? 
[21:21] <jhoffing> I believe libfdk_aac wraps an audio file in an MPEG-4 container by default as well 
[21:21] <Zeranoe> jhoffing: What are you trying to do?
[21:23] <jhoffing> Zeranoe: I'm trying to insert an audio file in front of a live audio stream. The audio stream is being pumped out with an MPEG-2 version  that I cannot change, and so the audio file I need to position in front of the stream has to have the same exact MPEG-2 container format
[21:24] <jhoffing> Otherwise the audio file data will play, but then the stream will be corrupt and fail to transition into the live audio stream
[21:27] <jhoffing> Zeranoe: Let me know if that doesn't make sense. Essentially, I'm trying to pump a 'preroll' audio buffer in front of a livestream, and its MPEG version has to match exactly that of the livestream's
[21:29] <Zeranoe> I've never done anything like that before, so I'm not sure
[21:29] <jhoffing> Zeranoe: ok. Looks like I have some help in #ffmpeg i'll continue discussion in there, thanks!
[22:00] <cone-645> ffmpeg.git 03Vittorio Giovara 07master:a18ef7a76c73: doc: fix a couple of typos in frame.h
[22:00] <cone-645> ffmpeg.git 03Michael Niedermayer 07master:4a59994e0317: Merge commit 'a18ef7a76c735bcf78ed4825e33ad7f9f6f77a54'
[22:09] <cone-645> ffmpeg.git 03Vittorio Giovara 07master:4c41a7a1798d: vf_frei0r: fix missing end of line character
[22:09] <cone-645> ffmpeg.git 03Michael Niedermayer 07master:adaba6a108aa: Merge commit '4c41a7a1798dd6c60e40f79be12faa98a9347151'
[22:38] <cone-645> ffmpeg.git 03Keiji Costantini 07master:3407172b36a9: avienc: K&R formatting cosmetics
[22:38] <cone-645> ffmpeg.git 03Michael Niedermayer 07master:309171bea9cd: Merge commit '3407172b36a93cec410c6a287dbeec0f41257512'
[23:14] <cone-645> ffmpeg.git 03James Almer 07master:206167a295a5: x86/synth_filter: add missing HAVE_YASM guard
[23:14] <cone-645> ffmpeg.git 03Andrey Utkin 07master:bc764d786fe6: Add "prefer_tcp" flag to "rtsp_flags"
[00:00] --- Wed Mar  5 2014