[Ffmpeg-devel-irc] ffmpeg-devel.log.20150306

burek burek021 at gmail.com
Sat Mar 7 02:05:02 CET 2015


[01:07:53 CET] <cone-620> ffmpeg 03Carl Eugen Hoyos 07master:23e483269d21: lavd/avfoundation: Add support for 16bit integer input.
[01:11:54 CET] <cone-620> ffmpeg 03John Robinson 07master:98d19ca8f189: lavd/avfoundation: Add support for 24 and 32bit integer input.
[01:26:24 CET] <cone-620> ffmpeg 03Himangi Saraogi 07master:88b160a457e4: avfilter: Return more meaningful error codes
[01:26:25 CET] <cone-620> ffmpeg 03Michael Niedermayer 07master:f455c8fdef0a: Merge commit '88b160a457e491cb9e014630ef5387ad3500258e'
[03:59:41 CET] <cone-620> ffmpeg 03Andreas Cadhalpun 07master:c089e720c1b7: webp: ensure that each transform is only used once
[09:41:42 CET] <ninten> hi all, i want to know from where i can find MPEG-4 ALS conformance files ?? is there any specific link on net or is it there on source code !! 
[11:15:58 CET] <ubitux> https://vimeo.com/112149609 pretty nice video encoding stress test
[11:17:01 CET] <nevcairiel> bewcause of the fake analog artifacts?
[11:17:12 CET] <ubitux> i guess
[11:17:46 CET] <ubitux> there are all kind of fake crazy glitches
[11:28:01 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:dbd6ba24a870: MAINTAINERS: Remove 1.2 from the maintained releases
[11:32:46 CET] <alasin> @michaelni: Hi! I am interested in implementing the FFv1 P Frame Support project. I don't have much background on video encoding and would be really grateful if you could help me get started on it.
[11:41:05 CET] <michaelni> alasin, see libavcodec/ffv1*.c and https://github.com/FFmpeg/FFV1
[11:56:21 CET] <cone-878> ffmpeg 03Martin Storsjö 07master:25c29d32835f: libopenh264enc: Add support for building with OpenH264 1.4
[11:56:22 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:39becc949c04: Merge commit '25c29d32835f38cdc5f0c84fa27dfc489a228770'
[12:13:34 CET] <cone-878> ffmpeg 03Timo Rothenpieler 07master:6cf7f30655e9: dashenc: Simplify code by using a local variable
[12:13:35 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:c3f64a223380: Merge commit '6cf7f30655e95e27fd0394b5a80970d6f9517015'
[12:28:35 CET] <cone-878> ffmpeg 03Timo Rothenpieler 07master:5aef535a6435: dashenc: Update extradata for mov muxer
[12:28:36 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:b32cae1cc24f: Merge commit '5aef535a64350b7bc06c5bdda8c26c9efec9443b'
[13:59:36 CET] <cone-878> ffmpeg 03Claudio Freire 07master:6dbbb981b573: AAC: Add support for 7350Hz sampling rates, no error on too hight bitrate.
[14:37:44 CET] <cone-878> ffmpeg 03James Cowgill 07master:a99de50ec767: mips/aacdec: remove uses of mips32r2 specific ext instructions
[14:37:45 CET] <cone-878> ffmpeg 03James Cowgill 07master:eae13eae9d64: configure, mips: remove MIPS32R2, merging it with MIPSFPU
[15:14:31 CET] <cone-878> ffmpeg 03James Cowgill 07master:157d6f0d5b95: mips: port optimizations to mips n64
[15:14:33 CET] <cone-878> ffmpeg 03James Cowgill 07master:d6f5b3eee3ab: changelog: add mips 64-bit port
[15:29:49 CET] <cone-878> ffmpeg 03Srikanth G 07master:0f2359b86926: avutil/opencl: is_compiled flag not being cleared in av_opencl_uninit
[15:42:17 CET] Action: ubitux wonders if it will be always OK to access AVFrame even after the codec and format is closed
[15:42:26 CET] <ubitux> (typically in case of hwaccel)
[15:45:10 CET] <nevcairiel> the user app controls the hwaccel surface allocation usually
[15:45:22 CET] <nevcairiel> so it should know
[15:45:33 CET] <ubitux> what about through a device?
[15:48:38 CET] <ubitux> (also, no codec is going to keep a buffer in its context and hack the buf of the frame to point to it?)
[15:49:42 CET] <nevcairiel> the whole point of refcounting is that this is not something thats going to happen
[15:49:59 CET] <nevcairiel> or well, if it did, then the buf would only be free'ed once every last ref to it is free'ed, or its a bug
[15:50:10 CET] <wm4> * ubitux wonders if it will be always OK to access AVFrame even after the codec and format is closed <- yes most definitely
[15:50:21 CET] <wm4> don't even think about breaking this
[15:50:29 CET] <ubitux> i'm not thinking of breaking it
[15:50:40 CET] Action: wm4 puts back the gun
[15:50:41 CET] <ubitux> i'm wondering if i can do this, as a user
[15:51:05 CET] <ubitux> (i know it's surreal, but dev sometimes use the api outside of the project :p)
[15:53:34 CET] <wm4> so the modern AVFrame is just data, it has nothing to do with the codec it comes from
[15:54:55 CET] <nevcairiel> any data it refs, it should also hold a ref to
[15:55:01 CET] <nevcairiel> like, frames come from a frame pool
[15:55:07 CET] <nevcairiel> but the frame itself holds a ref to the pool
[15:55:20 CET] <nevcairiel> so even if the codec is closed, the pool lives on until all its frames are destroyed
[15:55:23 CET] <wm4> the pool can also be destroyed before the frame
[15:55:25 CET] <wm4> yep
[15:56:46 CET] Action: Daemon404 is surprised anyone can even use this api correctly
[15:56:49 CET] <Daemon404> sounds very complicated.
[15:56:56 CET] <nevcairiel> its really not
[15:57:14 CET] <nevcairiel> from the users point of view, you just unref/free the avframe when you are done
[15:57:21 CET] <Daemon404> ic
[15:57:25 CET] <nevcairiel> you just have to trust that it remains valid until then
[15:57:31 CET] <Daemon404> well it only has 3 users afaict anyway :P
[15:57:32 CET] <nevcairiel> but ubitux didn't have trust
[15:58:12 CET] <nevcairiel> i really like the new ref counting things, it made so many things easier for me
[15:58:21 CET] <Daemon404> aye
[15:58:23 CET] <Daemon404> me too
[15:58:33 CET] <Daemon404> (unless you meant for hwaccel)
[15:58:39 CET] <nevcairiel> both
[15:58:45 CET] <nevcairiel> normal and hwaccel
[15:59:42 CET] <Daemon404> the naye
[15:59:46 CET] <Daemon404> then aye*
[16:08:04 CET] <Daemon404> i have an idea for a gsoc, but it touches the horrible libavdevice
[16:08:18 CET] <nevcairiel> inappropriately so?
[16:08:25 CET] <Daemon404> wonder if its at all useful to support proper screen capture on windows (not gdi+ shit)
[16:08:28 CET] <Daemon404> the dx11 stuff
[16:08:33 CET] <Daemon404> like vdub has
[16:08:34 CET] <nevcairiel> i wrote that once
[16:08:36 CET] <nevcairiel> its not very hard
[16:08:52 CET] <Daemon404> well the libavdevice cruft might be
[16:08:53 CET] <Daemon404> lol
[16:09:00 CET] <nevcairiel> hardly a couple month filling work
[16:09:03 CET] <nevcairiel> took me a couple hours
[16:09:06 CET] <nevcairiel> but then, i'm me
[16:09:24 CET] <BtbN> propper screen capture on windows only work on win8+
[16:09:25 CET] <Daemon404> lots of the tasks would take a couple hours
[16:09:29 CET] <Daemon404> BtbN, win7+
[16:09:30 CET] <nevcairiel> i assume you mean the win8+ DesktopDuplication  api?
[16:09:33 CET] <BtbN> Nope, win8+
[16:09:39 CET] <BtbN> Desktop Duplication was added there
[16:09:45 CET] <nevcairiel> that one works quite nicely
[16:09:53 CET] <nevcairiel> thats what i implemented
[16:09:56 CET] <nevcairiel> fast and reliable, too
[16:10:06 CET] <nevcairiel> it just steals the image from the desktop compositor
[16:10:10 CET] <BtbN> on win7 you have to write a mirror driver to achive the same, it's horrible.
[16:10:24 CET] <Daemon404> http://www.virtualdub.org/blog/pivot/entry.php?id=356
[16:10:26 CET] <Daemon404> youre right
[16:10:27 CET] <Daemon404> 8+
[16:10:35 CET] <nevcairiel> anyway the implementation is easy
[16:10:41 CET] <BtbN> On win7 you can hook into dwm.exe, and do hooked screen capture
[16:10:50 CET] <BtbN> but adding that kind of capture to ffmpeg would be way overkill
[16:10:56 CET] <nevcairiel> my problem with libavdevice is that its C, not C++, and I bet you that mingw-w64 lacks some of the APIs
[16:11:11 CET] <BtbN> yes, C directshow is horrible
[16:11:18 CET] <nevcairiel> the C wrappers for the C++ DX APIs are terrible
[16:11:20 CET] <Daemon404> isnt it COM crap?
[16:11:26 CET] <BtbN> Yes, it's COM stuff in C
[16:11:50 CET] <BtbN> Not that COM is great in C++, but definitely better than in C
[16:12:04 CET] <Daemon404> could be worse. could be 
[16:12:05 CET] <Daemon404> MF
[16:12:13 CET] <nevcairiel> MF is actually pretty nice
[16:12:18 CET] <wm4> COM in C is only horrible if you need to implement COM interfaces
[16:12:25 CET] <wm4> other than that it's merely very verbose
[16:12:29 CET] <Daemon404> nevcairiel, iirc wbs said he needed liek 2000 lines of boilerplate
[16:12:31 CET] <Daemon404> to do anything
[16:12:32 CET] <BtbN> What you have to do for dshow capture, wm4 
[16:12:43 CET] <BtbN> You have to implement a filter
[16:12:52 CET] <nevcairiel> Daemon404: and directshow is different how? =P
[16:13:01 CET] <Daemon404> "pretty nice" is not teh word id use
[16:13:02 CET] <Daemon404> for that
[16:13:07 CET] <Daemon404> words even.
[16:13:13 CET] <nevcairiel> coming from directshow, its much nicer!
[16:13:18 CET] <Daemon404> lul
[16:15:17 CET] <nevcairiel> in any case, if someone really wanted dxgi desktop duplication capture in avdevice, i could even write it, if someone asks nicely
[16:15:29 CET] <nevcairiel> just need to port my old code to C <.<
[16:16:45 CET] <Daemon404> lol
[16:17:20 CET] <Daemon404> not so easy if mingw lacks the APIs
[16:17:25 CET] <nevcairiel> yeah i dont know
[16:17:32 CET] <nevcairiel> never tried to do dx11/dxgi with mingw
[16:17:44 CET] <nevcairiel> could just build it for msvc and add configure checks, let someone else figure out mingw
[16:17:50 CET] <nevcairiel> same I did with hevc hwaccel =p
[16:17:53 CET] <Daemon404> :D
[16:18:07 CET] <Daemon404> but i thought you used mingw builds in your binary distrib
[16:18:10 CET] <Daemon404> for SPEEDUMS
[16:18:26 CET] <nevcairiel> its not SPEEDUMS, its actually a significant difference
[16:18:34 CET] <Daemon404> i know
[16:18:35 CET] <nevcairiel> but I just copy-pastad the MS header
[16:18:43 CET] <Daemon404> for shame
[16:18:47 CET] <nevcairiel> sue me
[16:18:49 CET] <Daemon404> mingw is supposed to copypasta msdn
[16:18:51 CET] <Daemon404> ;)
[16:19:10 CET] <nevcairiel> http://git.1f0.de/gitweb?p=ffmpeg.git;a=commitdiff;h=901bc25983fe119637f73c0467731c1a96c49ad7;js=1
[16:19:11 CET] <nevcairiel> evil me!
[16:19:32 CET] <Daemon404> i think that copyright year is a bit off
[16:19:39 CET] <nevcairiel> its from their header like this
[16:19:42 CET] <nevcairiel> dont blame me
[16:20:06 CET] <ramiro> Daemon404: talk to the mingw-w64 guys about implementing the APIs you want
[16:20:12 CET] <ramiro> they're quite nice
[16:20:16 CET] <Daemon404> i know
[16:20:23 CET] <Daemon404> i just /care myself
[16:20:27 CET] <Daemon404> i dont screen cap usually
[16:20:35 CET] <nevcairiel> and usually the answer is "after the next release", so in about 6 to 12 months =P
[16:20:39 CET] <ramiro> the C wrapper for directshow is indeed horrible
[16:21:01 CET] <Daemon404> im surprised there is even a c wrapper.
[16:21:27 CET] <ubitux> we don't have anything to capture screen on windows?
[16:21:33 CET] <ubitux> we have 2 for linux and 1 for OSX afaik
[16:21:34 CET] <nevcairiel> we do, but only the crappy way
[16:21:38 CET] <Daemon404> we have gdi+ crap iirc
[16:21:44 CET] <nevcairiel> its extremely slow
[16:21:47 CET] <kepstin-laptop> we do, i cleaned up the gdigrab to get it in a while back
[16:21:55 CET] <kepstin-laptop> it works, but that's about it :)
[16:21:57 CET] <ubitux> ok
[16:22:20 CET] <nevcairiel> desktop  duplication is fast and reliable, its actually a nice  API, with the limit that its Windows 8 and up only
[16:22:50 CET] <kepstin-laptop> I found it hilarious in testing that gdigrab was faster in wine on linux than in real windows.
[16:23:20 CET] <ubitux> :D
[16:24:31 CET] <Daemon404> its not even usable IMO
[16:24:31 CET] <Daemon404> nto for anything sane
[16:28:25 CET] <klaxa> on the gsoc site Nicolas George (Cigaes) is noted as a mentor and his only contact info is irc, i haven't seen him yet and he doesn't even show up in my logs. would it be okay to send him an email?
[16:29:27 CET] <ubitux> yes better contact him through mail
[16:29:32 CET] <ubitux> he isn't very present on irc
[16:29:54 CET] <klaxa> i guess the email he uses on the ffmpeg-devel list is fine?
[16:30:08 CET] <klaxa> it is not listed on the page, that is why i ask
[16:30:18 CET] <nevcairiel> yeah that should be fine
[16:30:26 CET] <klaxa> okay thanks!
[16:56:45 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:4034e146baf6: Changelog: rename next to 2.6
[16:56:52 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:34300866a8c4: doc/APIchanges: Add 2.6 cut marker
[16:56:59 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:0b48ddc5f230: doc/APIchanges: Fill in the xx
[17:42:03 CET] <cone-878> ffmpeg 03Michael Niedermayer 07release/2.6:HEAD: doc/APIchanges: Fill in the xx
[17:51:11 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:0bcb6ac15069: Add 2.6 to maintained releases
[17:51:29 CET] <cone-878> ffmpeg 03Michael Niedermayer 07release/2.6:0d4549c2d688: Add 2.6 to maintained releases
[18:02:08 CET] <ninten> hi all i want to use mp4alsRM23  decoder but there isn't any documentation regarding how to use it, i searched on net also i could not found any so if any one has used it before or have any docmentation regarding this !! 
[18:02:27 CET] <ninten> plz help !!
[18:37:28 CET] <michaelni> ninten, i see a readme.txt in mp4alsRM23
[18:38:04 CET] <michaelni> also "mp4alsRM23 -h" prints a help text that explains use
[19:44:12 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:b4d89139fbaf: Changelog: Add back "version next"
[20:20:21 CET] <wm4> can we use some gsoc slaves for subtitles? I still have support for this as feature request: http://pastebin.com/6kKpZvGr
[20:20:54 CET] <j-b> smi
[20:20:55 CET] <j-b> lol
[20:23:18 CET] <wm4> better than webvtt
[20:23:33 CET] <j-b> +1
[20:26:10 CET] <ubitux> wm4: what doesn't work with ffmpeg?
[20:26:31 CET] <wm4> I think this was about font color tags
[20:26:48 CET] <ubitux> that's all?
[20:26:57 CET] <ubitux> rest is ok?
[20:27:19 CET] <wm4> another one http://pastebin.com/mhx3sSdk
[20:27:22 CET] <wm4> well I don't know
[20:27:33 CET] <wm4> but apparently
[20:28:53 CET] <ubitux> going to check this
[20:29:32 CET] <ubitux> wm4: can you give me the original filenames btw? i'm lacking inspiration
[20:30:02 CET] <wm4> they were posted as these pastebins
[20:30:06 CET] <wm4> (look at the dates)
[20:30:06 CET] <ubitux> ok
[20:31:07 CET] <ubitux> i'll go for russian-color.smi and ©D ´ ü8”.smi then
[20:31:22 CET] <wm4> the other one is korean
[20:31:35 CET] <ubitux> yeah i recognized the smiley characters
[20:31:55 CET] <ubitux> but i think using the <title> is appropriate :p
[20:40:20 CET] <cone-878> ffmpeg 03Clément BSsch 07master:70082a1e533d: avcodec/samidec: make sure to properly restore parsing context after a tag
[20:41:55 CET] <jamrial> ubitux: maybe backport it to 2.6 branch before the release is made?
[20:42:13 CET] <ubitux> yeah
[20:42:51 CET] <cone-878> ffmpeg 03Clément BSsch 07release/2.6:1dab67b647ef: avcodec/samidec: make sure to properly restore parsing context after a tag
[20:44:05 CET] <rcombs> if you want to put someone on charenc detection, my patch for that is sitting in the ML and could be a starting point
[20:44:43 CET] <ubitux> wm4: i'll add the color support later, but converting the two samples to srt looks mostly OK
[20:45:07 CET] <ubitux> there are some trailing whitespaces but the rest looks fine
[20:45:21 CET] <ubitux> (i just had a glitch with the <BR>)
[20:45:29 CET] <wm4> that's nice
[20:45:30 CET] <ubitux> (which should be solved)
[21:07:37 CET] <cone-878> ffmpeg 03Martin Storsjö 07master:9731cf400137: movenc: Keep writing zero-entry stts atoms as intended
[21:07:38 CET] <cone-878> ffmpeg 03Michael Niedermayer 07master:1468ff49dfde: Merge commit '9731cf4001377fa2f75c279072cc2b0cbd57bf8e'
[21:07:44 CET] <ubitux> anyone to re-read the latest release note before i push please? http://b.pkh.me/RELEASE_NOTES
[21:08:31 CET] <ubitux> my english is as usual barely decent, so if you have better wording etc, please suggest
[21:09:25 CET] <wm4> what is a Grothendieck? oh, some old guy
[21:09:44 CET] <ubitux> a mathematician that died recently
[21:09:55 CET] <ubitux> apparently a brillant one from the last article i read about him
[21:17:27 CET] <wm4> "and criticized what he saw as the declining ethics of the scientific community, characterized by outright scientific theft that, according to him, had become commonplace and tolerated. "
[21:17:31 CET] <wm4> what does he mean by this?
[21:20:23 CET] <ubitux> he was kind of a radical, pacifist, ecologist, kind of influenced by the hippy movement
[21:21:00 CET] <ubitux> apparently one day he realized https://en.wikipedia.org/wiki/Institut_des_Hautes_%C3%89tudes_Scientifiques was partially financed by some random governement thing
[21:21:08 CET] <ubitux> and left it
[21:21:15 CET] <ubitux> anyway, just a random extremist
[21:21:20 CET] <wm4> heh
[21:21:21 CET] <ubitux> he could have been a ffmpeg developer
[21:21:23 CET] <ubitux> ;)
[21:22:22 CET] <ubitux> let me check what he wrote about the "thieves" thing
[21:23:22 CET] <ubitux> ok so he was offered a price he refused because he has enough money
[21:24:33 CET] <ubitux> he was condemning that most scientifist getting such price always had a high social status with all the perversity associated with
[21:25:12 CET] <ubitux> ah and indeed he was condeming the theft of ideas 
[21:25:22 CET] <ubitux> which was apparently kind of a "general rule"
[21:25:29 CET] <ubitux> tolerated by everyone
[21:25:56 CET] <ubitux> he doesn't say "ideas" sorry
[21:26:06 CET] <wm4> but?
[21:26:40 CET] <ubitux> let me try a translation
[21:27:59 CET] <jamrial> ubitux: heh, no more "supreme leader"? :P
[21:29:34 CET] <ubitux> wm4: http://pastie.org/10005830
[21:30:05 CET] <wm4> thanks
[21:30:12 CET] <wm4> so basically scientists are being assholes
[21:30:29 CET] <ubitux> jamrial: yeah, Michael didn't want to be on https://en.wikipedia.org/wiki/List_of_leaders_of_North_Korea :(
[21:31:28 CET] <wm4> ubitux: anyway, in general I don't like the "casual marketing speech" tone of the announcement, but maybe that's just me
[21:32:03 CET] <ubitux> i'll keep that in mind
[21:32:08 CET] <ubitux> we're just trying some stuff
[21:32:15 CET] <ubitux> like, marketting somehow indeed
[21:32:46 CET] <ubitux> some people can obviously propose a release notes
[21:33:00 CET] <ubitux> i mean, i wanted to do this one because i like a lot of these changes
[21:33:32 CET] <ubitux> but yeah, i'm a terrible commercial
[21:33:37 CET] <ubitux> that's not my job you know ;)
[21:36:15 CET] <ubitux> anyway, i'll push this in about 1 hour or so
[21:36:22 CET] <ubitux> so if a native english speaker is around&
[21:36:51 CET] <wm4> "libmpcodecs (MPlayer filters) wrapper is finally dead."
[21:36:57 CET] <wm4> missing "the"?
[21:37:23 CET] <Compn> what
[21:37:27 CET] <Compn> +the
[21:37:37 CET] <ubitux> wm4: fixed
[21:42:37 CET] <selsta> okay so i am thinking about implementing support for sample-aes encryption in hls files (even though no website is using it, but it's part of the spec). to start it would be great if someone could tell me where encrypted segments currently get decrypted in ffmpeg.
[21:43:18 CET] <selsta> i've read aes.c / crypto.c but i can't find the decryption function is getting called
[21:43:41 CET] <selsta> does libavformat even decrypt segments or should i look somewhere else?
[21:44:39 CET] <cone-878> ffmpeg 03Andreas Cadhalpun 07master:482c86f2319c: fix spelling errors
[21:57:07 CET] <cone-878> ffmpeg 03Peter Tissen 07master:42493843dc6f: configure: Make compilable with VS2015
[22:07:10 CET] <ubitux> selsta: av_aes_crypt()?
[22:07:15 CET] <ubitux> i see a decrypt flag
[22:08:08 CET] <selsta> ubitux: i've seen this function, but i can't find where it gets called
[22:11:07 CET] <ubitux> selsta: so crypto protocol? libavformat/crypto.c
[22:11:24 CET] <ubitux> i see a call to it in crypto read
[22:12:08 CET] <ubitux> the hls demuxer is prefixing with a "crypto" to trigger it
[22:12:10 CET] <ubitux> i suppose
[22:12:12 CET] <selsta> ubitux: i'll look for files using the CryptoContext, thanks
[22:14:23 CET] <wm4> if ubitux is right, this is a runtime mechanism
[22:14:28 CET] <wm4> at least the way hls uses it
[22:14:59 CET] <selsta>             snprintf(url, sizeof(url), "crypto+%s", seg->url); 
[22:15:23 CET] <ubitux> so evil :)
[22:15:25 CET] <selsta> what exactly does this do? set a crypto-url?
[22:16:20 CET] <ubitux> i don't know that code very well, but if let's say you get a "crypto+http://..." it will go through crypto protocol, then http protocol
[22:16:26 CET] <ubitux> (or maybe the other way around?)
[22:16:44 CET] <ubitux> but just a guess
[22:17:30 CET] <selsta> hmm so crypto.c is probably the file i have to understand
[22:18:57 CET] <wm4> they're chained kind of like unix pipes
[22:19:14 CET] <wm4> except instead of pipes it goes through avio
[22:19:34 CET] <ubitux> the hls demuxer is transmitting the key/iv through av_opt_set(); that will end up in CryptoContext
[22:19:42 CET] <ubitux> through the AVOption mechanism
[22:20:18 CET] <ubitux> so what's this sample-aes thing?
[22:20:36 CET] <ubitux> do we have everything needed already?
[22:20:58 CET] <selsta> ubitux: kinda, the decryption is a little bit more complicated than normal aes-128 encryption
[22:21:36 CET] <wm4> yet it's called "simple"?
[22:21:54 CET] <selsta> it's called sample because only parts of the video are encrypted
[22:22:25 CET] <wm4> sounds complicated
[22:22:29 CET] <selsta> https://developer.apple.com/library/iad/documentation/AudioVideo/Conceptual/HLS_Sample_Encryption/Introduction/Introduction.html#//apple_ref/doc/uid/TP40012862-CH1-SW1
[22:22:38 CET] <ubitux> wm4: sample ` simple :p
[22:23:11 CET] <selsta> wm4: yeah i will probably fail to implement it
[22:23:39 CET] <wm4> my eyes are failing
[22:24:16 CET] <ubitux> lol @ "Made-Seriously-Happy:" tag
[22:27:00 CET] <cone-878> ffmpeg 03Andreas Cadhalpun 07release/2.6:4e2cab5a79f4: fix spelling errors
[22:28:31 CET] <selsta> i just realised there is no way i can implement this
[22:28:51 CET] <ubitux> already giving up? ;)
[22:30:18 CET] <selsta> it's probably not worth the effort because i've only seen a demo stream using this
[22:30:23 CET] <ubitux> the link in 42493843dc6f990ebab6b82df2e9cfab0e019b80 is refering to VS2014, is this on purpose?
[22:31:49 CET] <selsta> the only "cool" thing about it would be that it would be the first open source implementation, i've only seen proprietary players that can play it
[22:32:52 CET] <cone-878> ffmpeg 03Clément BSsch 07release/2.6:3dc88e1fd6b8: Add release notes
[22:33:11 CET] <Plorkyeran> VS 14 == VS 2015
[22:33:32 CET] <ubitux> ah... the link looked old
[22:54:40 CET] <cone-878> ffmpeg 03Peter Tissen 07master:79f83523f771: configure: Make compilable with VS2015 (missing hunk)
[23:46:22 CET] <wm4> is there a barrel distortion filter in lavfi?
[00:00:00 CET] --- Sat Mar  7 2015


More information about the Ffmpeg-devel-irc mailing list