[Ffmpeg-devel-irc] ffmpeg-devel.log.20150914

burek burek021 at gmail.com
Tue Sep 15 02:05:03 CEST 2015


[09:00:11 CEST] <satiender> Hi , Can any body help me , I want use ffmpeg API in my C code 
[09:06:21 CEST] <fritsch> satiender: http://dranger.com/ffmpeg/tutorial01.html
[09:06:22 CEST] <fritsch> good start
[09:06:38 CEST] <fritsch> another example: https://www.ffmpeg.org/doxygen/0.6/output-example_8c-source.html
[09:11:55 CEST] <satiender> fritsch : I already see that tutorial , but when I compile the tutorial1.c with gcc in linux there have multiple undefined errors which are following :
[09:11:58 CEST] <satiender> exmp1.c:(.text+0x166): undefined reference to `av_register_all'
[09:11:59 CEST] <satiender> exmp1.c:(.text+0x18d): undefined reference to `avformat_open_input'
[09:11:59 CEST] <satiender> exmp1.c:(.text+0x1af): undefined reference to `avformat_find_stream_info'
[09:11:59 CEST] <satiender> exmp1.c:(.text+0x1e4): undefined reference to `av_dump_format'
[09:11:59 CEST] <satiender> exmp1.c:(.text+0x299): undefined reference to `avcodec_find_decoder'
[09:11:59 CEST] <satiender> exmp1.c:(.text+0x2d8): undefined reference to `avcodec_alloc_context3'
[09:12:16 CEST] <satiender> what I can do 
[09:12:20 CEST] <fritsch> yeah, you obviously need to link correctly
[09:12:52 CEST] <satiender> how I can link ffmpeg library at compile time 
[09:13:01 CEST] <satiender> please help 
[09:13:31 CEST] <satiender> my ffmpeg library have in /usr/lib/
[09:13:55 CEST] <satiender> my ffmpeg header files have in /usr/include 
[09:24:51 CEST] <durandal_1707> satiender: wrong channel, also you can ask on libav ffmpeg mailing list
[09:25:25 CEST] <satiender> durandal_1707 : ok sir ! 
[13:49:18 CEST] <cone-326> ffmpeg 03Muhammad Faiz 07master:145c5a99783c: avfilter/avf_showcqt: change fft left-right separation
[14:38:51 CEST] <cone-326> ffmpeg 03Agatha Hu 07master:49046580ce62: avcodec/nvenc: Optimize nvenc parameters
[14:38:52 CEST] <cone-326> ffmpeg 03Agatha Hu 07master:a8bedd56aaee: avcodec/nvenc: merge compute initialQP sections
[15:10:54 CEST] <cbsrobot_> ubitux: in your sc.py it looks like the K values has to do with saturation - at least there seems to be some sort of correlation between K and S (from HSV)
[15:11:14 CEST] <ubitux> K is black of CMYK
[15:13:36 CEST] <cbsrobot_> yeah I know
[15:16:07 CEST] <ubitux> but anyway, it's just a basic line; i need to extract every slope and try to correlate them with all the other parameters
[15:17:05 CEST] <ubitux> the clamping i have seems already ok, so i guess i just need to find that slope factor
[15:19:26 CEST] <BBB> j-b: ping
[15:46:08 CEST] <cone-326> ffmpeg 03Nedeljko Babic 07master:b65ffa316e37: avcodec/mips/aaccoder_mips: Sync with generic aaccoder file.
[15:48:53 CEST] <Daemon404> lul
[15:51:23 CEST] <atomnuker> this 100% overlaps with what Claudio's doing
[15:51:47 CEST] <Daemon404> atomnuker, i'd argue claudio's is far more important
[15:51:49 CEST] <Daemon404> >mips
[15:52:03 CEST] <cbsrobot_> ubitux: hmm nope, even converted to cmyk, hsv or hls the K variation makes not much sense
[15:52:27 CEST] <Daemon404> ^ ticket above; we do it via api with one request
[15:52:33 CEST] <Daemon404> special-casing for images
[15:52:34 CEST] <Daemon404> :D
[16:12:43 CEST] <nevcairiel> as long as he only touches mips decoder files, I would just ignore these things, and just disable it again when needed due to main source changes =p
[16:12:56 CEST] <Daemon404> indeed
[16:19:50 CEST] <atomnuker> well, just talked with Claudio
[16:20:46 CEST] <atomnuker> the reason why he hasn't pushed his changes is gcc cross compiling to MIPS with opensuse's build system segfaults
[16:21:27 CEST] <nevcairiel> tell him to screw MIPS and let the mips guys figure it out
[16:21:46 CEST] <nevcairiel> we're writing generic C here
[16:21:46 CEST] <JEEB> agreed
[16:22:10 CEST] <nevcairiel> if they need "optimized C" for their arch, they should maintain i t
[16:44:46 CEST] <atomnuker> klaussfreire: "Well, if I can't solve it tonight I'll do as nevcairiel says"
[16:45:02 CEST] <nevcairiel> yay
[16:45:16 CEST] <nevcairiel> cant wait for those patches to land as well
[16:45:21 CEST] <nevcairiel> new and improved aac encoder hooray
[16:45:50 CEST] <ubitux> :)
[16:45:51 CEST] <Daemon404> \o
[16:48:36 CEST] <atomnuker> you don't have to wait long for my patches though
[16:48:59 CEST] <atomnuker> going to merge the LTP extension and the profile stuff tonight
[16:52:06 CEST] <Daemon404> atomnuker, cant wait till we reach the "let HA loose on it" point ;)
[16:53:51 CEST] <ubitux> the drop of the experimental flag will be the rampage signal
[16:53:59 CEST] <j-b> BBB: pong
[16:54:23 CEST] <BBB> still awaiting your slides and comments on mine (see privmsg last week)
[16:54:33 CEST] <Daemon404> j-b was busy being homeless
[16:54:38 CEST] <j-b> yes :)
[16:54:51 CEST] <j-b> and busy organizing VDD mess and people registering late
[16:54:54 CEST] <BBB> hm& homeless in france?
[16:54:59 CEST] <j-b> in Amsterdam
[16:54:59 CEST] <Daemon404> netherlands
[16:55:15 CEST] <j-b> that was a nice stupid story.
[16:55:28 CEST] <BBB> I hope you smoked some good stuff while you were out there
[16:55:30 CEST] <BBB> ibc?
[16:55:33 CEST] <j-b> yes
[16:55:40 CEST] <BBB> tres bien
[16:56:05 CEST] <j-b> BBB: don't worry, I'm on it. And I need it before VDD
[16:56:14 CEST] Action: Daemon404 owes j-b and/or kierank 16 euro now
[16:56:19 CEST] <Daemon404> or so.
[16:56:32 CEST] <j-b> for the beer?
[16:56:38 CEST] <Daemon404> yes
[16:56:46 CEST] <j-b> Daemon404: Julien paid everything, including kieran's beer
[16:56:54 CEST] <Daemon404> O_o
[16:57:53 CEST] <j-b> so you owe it to him
[16:58:15 CEST] <Daemon404> all right then
[16:58:36 CEST] <j-b> even though he will refuse your money.
[17:00:43 CEST] <BBB> I wish there was a way for me to figure out how to get from airport to hotel without begging for directions in this irc channel
[17:01:26 CEST] <j-b> BBB: from CDG?
[17:01:30 CEST] <j-b> BBB: it will be on the wiki
[17:01:30 CEST] <BBB> yeah
[17:01:38 CEST] <BBB> porte de clichy line 13?
[17:02:51 CEST] <j-b> https://wiki.videolan.org/VDD15/#Transportation
[17:03:56 CEST] <BBB> ah the bottom
[17:03:58 CEST] <BBB> ty
[17:04:02 CEST] <Daemon404> my mobile provider gvies me data, so google maps for me
[17:04:05 CEST] Action: Daemon404 lazy
[17:04:52 CEST] <BBB> mine too
[17:04:56 CEST] <BBB> but you never know if it works
[17:05:05 CEST] <BBB> so it takes quite a while to get to the conference venue, does it
[17:05:25 CEST] <BBB> 13, 2, E
[17:05:42 CEST] <j-b> from the hotel?
[17:05:58 CEST] <BBB> yes
[17:05:59 CEST] <BBB> ?
[17:06:20 CEST] <j-b> no.
[17:06:32 CEST] <j-b> hotel <-> conference is 4 stops of line 13
[17:06:42 CEST] <BBB> it doesnt say that on the wiki :-p
[17:07:30 CEST] <JEEB> this discussion reminds me trying to look up a route from heathrow to cambridge with google maps
[17:07:51 CEST] <JEEB> the alternatives were all pants-on-head retarded
[17:10:18 CEST] <BBB> ah I see, I get out at liege
[17:10:19 CEST] <BBB> ok
[17:10:27 CEST] <BBB> now I only need to print 100 google maps
[17:10:32 CEST] <BBB> so I dont forget any of this
[17:11:06 CEST] <JEEB> :)
[17:12:40 CEST] <cone-326> ffmpeg 03Alex Agranovsky 07master:0572bd1a62c0: mpjpeg: probe should require same constraints as packet reader - both proper content-type and content-size must be present
[17:13:32 CEST] <BBB> do regular dutch power adapters fit in france?
[17:13:58 CEST] <Daemon404> yes
[17:14:01 CEST] <BBB> \o/
[17:14:04 CEST] <BBB> one problem less
[17:14:17 CEST] <iive> i though all europe uses same power adapters...
[17:14:17 CEST] <BBB> and therell be ATMs that understand american-style debit cards for cash, right?
[17:14:25 CEST] <JEEB> iive: except britain
[17:14:34 CEST] <JEEB> BBB: do you have chip?
[17:14:38 CEST] <iive> yeh... they are island ;)
[17:14:38 CEST] <JEEB> or just magnetic
[17:14:46 CEST] <BBB> iive: I wouldnt know, I havent lived there for ages& I went to france in the 90s and it was a pain
[17:15:00 CEST] <BBB> JEEB: dunno? magnetic I guess, it swipes
[17:15:15 CEST] <Daemon404> [16:14] < JEEB> iive: except britain <-- and ireland.
[17:15:25 CEST] <BBB> Hopefully, there will be Wi-Fi.
[17:15:26 CEST] <BBB> ...
[17:15:37 CEST] <JEEB> BBB: only swipe or does it have the chip? :P
[17:15:45 CEST] <BBB> JEEB: I dont know :-p
[17:15:47 CEST] <JEEB> ok
[17:15:48 CEST] <BBB> how do I find out
[17:15:53 CEST] <JEEB> well, by looking at the card?
[17:15:54 CEST] <Daemon404> look at front of card for a silver chip
[17:16:01 CEST] <JEEB> silver or golden
[17:16:01 CEST] <iive> the chip is visible on the card
[17:16:04 CEST] <JEEB> it should be visible
[17:16:14 CEST] <Daemon404> http://content.linkoffers.net/SharedImages/Products/70/605712.gif?c22584 <-- chip
[17:16:15 CEST] <iive> like smartcard chip
[17:16:30 CEST] <iive> well, it is smart card chip.
[17:16:34 CEST] <BBB> my credit card has a chip
[17:16:36 CEST] <BBB> my debit card doesn't
[17:16:50 CEST] <JEEB> the one with chip most probably will work
[17:16:59 CEST] <Daemon404> if you know your pin.
[17:17:02 CEST] <JEEB> sure
[17:17:07 CEST] <JEEB> I'm expecting people to remember that 8)
[17:17:14 CEST] <BBB> credit cards dont have a pin
[17:17:16 CEST] <Daemon404> considerng you dont use it in the US, he might now
[17:17:16 CEST] <BBB> only debit cards do
[17:17:22 CEST] <Daemon404> BBB, chip needs a pin
[17:17:26 CEST] <Daemon404> phone your CC provider
[17:17:27 CEST] <JEEB> ^
[17:17:30 CEST] <BBB> &
[17:17:31 CEST] <BBB> pain
[17:17:35 CEST] <Plorkyeran> chip+signature does exist
[17:17:39 CEST] <Plorkyeran> it's super dumb
[17:17:41 CEST] <JEEB> well yes
[17:17:44 CEST] <JEEB> I've seen that in japan
[17:17:44 CEST] <Daemon404> never seen it
[17:17:50 CEST] <Daemon404> and i doubt it is well supported
[17:17:50 CEST] <JEEB> "I don't remember my PIN" "Please sign"
[17:17:56 CEST] <JEEB> yeah, europe probably not gonna fly
[17:18:03 CEST] <Plorkyeran> I haven't had any issues yet in denmark
[17:18:06 CEST] <JEEB> oh
[17:18:08 CEST] <JEEB> surprising
[17:18:29 CEST] <BBB> Plorkyeran: you american?
[17:18:33 CEST] <Plorkyeran> yes
[17:18:38 CEST] <BBB> goodie!
[17:18:41 CEST] <j-b> American CC work fine in France
[17:18:54 CEST] <BBB> but does debit card work to get cash?
[17:18:59 CEST] <BBB> in a regular ATM
[17:19:00 CEST] <j-b> Debit, I doubt it
[17:19:06 CEST] <BBB> bank card
[17:19:13 CEST] <BBB> you know, to take cash out of your checking account
[17:19:21 CEST] <j-b> No I don't.
[17:19:22 CEST] <iive> even here all ATM have visa and mastercard logos, so I do expect them to work.
[17:19:30 CEST] <Daemon404> j-b, it will if it has a chip (but his doesnt)
[17:19:40 CEST] <j-b> I have one card only, it works as debit/credit with a chip anywhere in the world.
[17:19:40 CEST] <Daemon404> also he MAY be able to use the travelex atms
[17:19:54 CEST] <BBB> what could possibly go wrong"
[17:20:14 CEST] <JEEB> just have both cards just in case
[17:20:14 CEST] Action: Daemon404 has CC cards to earn points
[17:20:18 CEST] <Daemon404> er CCs
[17:20:20 CEST] <j-b> And I never have to care whether it's credit/debit/saving or whatever stupidity US bank has.
[17:20:43 CEST] <iive> it might be good idea to call the bank and ask them in advance. It would also let them know that you are going to travel.
[17:20:55 CEST] <Daemon404> i also have one CC that has no foriegn transaction fees
[17:21:00 CEST] <Daemon404> which comes in useful.
[17:21:00 CEST] <iive> sudden change in continent would probably make them call you anyway.
[17:22:24 CEST] <BBB> ok so according to most travel sites, atm cards dont need chips
[17:22:27 CEST] <BBB> (for cash/atm)
[17:22:34 CEST] <BBB> only point of sale stuff cc needs chip
[17:22:41 CEST] <BBB> (and I dont have a pin, so I wont be using them)
[17:22:43 CEST] <JEEB> yaeh, I've had a US debit card work in .fi
[17:22:50 CEST] <JEEB> in an ATM
[17:22:55 CEST] <BBB> I guess you guys are paying for my beer
[17:23:01 CEST] <BBB> :D
[17:23:01 CEST] <JEEB> (citybank IIRC)
[17:23:04 CEST] <Daemon404> you have to phone your bank to tell them youre travelling anywya
[17:23:10 CEST] <Daemon404> may as well set up a pin for your CC at the same time
[17:23:13 CEST] <BBB> phone?
[17:23:16 CEST] <BBB> website
[17:23:20 CEST] <BBB> Ive never called my bank
[17:23:21 CEST] <Plorkyeran> >phone
[17:23:23 CEST] <Plorkyeran> what is that
[17:23:24 CEST] <Daemon404> depends on the bank
[17:23:51 CEST] <Daemon404> i just assume most US banks are behind the rest of the world
[17:23:55 CEST] <Daemon404> it's usually a valdi assumption
[17:24:04 CEST] <BBB> theyre just heading in a different direction
[17:24:10 CEST] <BBB> just like most eu countries individually
[17:24:18 CEST] <BBB> :D
[17:24:27 CEST] <Daemon404> i dunno, i can do a bank transfer in 2 hrs or less here
[17:24:32 CEST] <Daemon404> its pretty nice
[17:24:34 CEST] <Daemon404> usually it is instant
[17:25:55 CEST] <iive> Plorkyeran: It is an old feature that give the name of your smart-phone. The one you are chatting from now :P
[17:27:38 CEST] <cone-326> ffmpeg 03Alex Agranovsky 07master:1de21215d467: mpjpegde: trim header name/value of MIME headers
[17:31:05 CEST] <ubitux> michaelni: please tell the guy to fix his style next time
[17:31:14 CEST] <ubitux> he's making a mess out of the file
[17:31:34 CEST] <ubitux> (mjpegdec)
[17:34:59 CEST] <michaelni> ubitux, forwarded your comment to the author & ML
[17:35:15 CEST] <ubitux> thx
[17:35:34 CEST] <michaelni> np, iam terrible in noticing these things unless formatting is really really bad
[17:36:11 CEST] <ubitux> wellitslackingsomelittlespaceabit
[17:40:04 CEST] <ubitux> :)
[17:44:08 CEST] <cone-326> ffmpeg 03Clément BSsch 07master:c8370a17f57f: avformat: fix style after recent commits
[17:54:52 CEST] <cone-326> ffmpeg 03Timothy Gu 07master:ca00dda2167f: aaccoder_mips: Fix indentation
[17:59:10 CEST] <pa> hi, has anyone here tried to debug ffmpeg with visual studio?
[17:59:49 CEST] <Daemon404> several of us have / do
[18:00:20 CEST] <pa> Daemon404, do you attach the debugger to the dll after you start some player? or what is the best way to do it?
[18:01:03 CEST] <Daemon404> i do it like i'd do with anything?
[18:01:12 CEST] <Daemon404> what makes ffmpeg special?
[18:01:18 CEST] <pa> well it's a dll
[18:01:28 CEST] <Daemon404> is your question 
[18:01:33 CEST] <Daemon404> "how do i debug a dll"?
[18:01:36 CEST] <Daemon404> thats hardly ffmpeg specific
[18:02:10 CEST] <pa> that's true.. i'll check better on SO :)
[18:03:39 CEST] <nevcairiel> the answer obviously is you debug the player its running in
[18:03:43 CEST] <nevcairiel> or you debug ffmpeg.exe
[18:03:45 CEST] <nevcairiel> which isnt a dll :P
[18:05:17 CEST] <pa> thanks :)
[18:22:18 CEST] <Daemon404> -0.0 is the best float
[18:23:13 CEST] <pa> is anybody using siliconandlithium tool to generate visual studio solution project files?
[18:23:41 CEST] <Daemon404> i dont see why
[18:23:46 CEST] <Daemon404> the build system supports msvc as-is.
[18:23:51 CEST] <Daemon404> people use that
[18:24:05 CEST] <pa> well i mean to load the project inside visual studio
[18:24:51 CEST] <pa> and be able to modify it from there
[18:25:00 CEST] <Daemon404> nobody here works on ffmpeg inside the MSVS gui
[18:25:01 CEST] <nevcairiel> i just use an exe project, no mess and no fuss
[18:25:26 CEST] <Daemon404> (except for debug)
[18:25:28 CEST] <nevcairiel> and yeah only for debugging
[18:25:48 CEST] <pa> i see
[18:27:05 CEST] <pa> also, are you aware of some repo containing therein as many deps as possible?
[18:27:24 CEST] <pa> unofficial i mean
[18:27:35 CEST] <nevcairiel> why would you do that
[18:28:09 CEST] <pa> well to simplify the build, i guess
[18:28:10 CEST] <j-b> For fun
[18:46:04 CEST] <pa> ok i think i dont manage. but technically i could also use eclipse on linux to step through the libavformat/demuxer, right?
[18:51:04 CEST] <durandal_1707> who will adopt loop from libav they appear to abandoned it
[20:05:13 CEST] <cone-326> ffmpeg 03Vittorio Giovara 07master:65b96aba283b: wmalossless: Warn when decoding 24 bit depth
[21:06:12 CEST] <BBB> whooooooo wants to review my last vp9 patch
[21:11:13 CEST] <BBB> (its sort of a milestone)
[21:18:29 CEST] <Daemon404> a buggy milestone?
[21:18:34 CEST] <Daemon404> (lubvpx)
[21:23:39 CEST] <BBB> Daemon404: Dude, its the Golden Standard[TM]
[21:23:46 CEST] <BBB> sorry
[21:23:50 CEST] <BBB> The Golden Standard[TM]
[21:23:53 CEST] <BBB> lets get my caps right
[21:24:00 CEST] <Daemon404> it's a pun on golden frame
[21:24:06 CEST] <Daemon404> is taht*
[21:24:09 CEST] <Daemon404> that**
[21:24:12 CEST] <BBB> was it funny?
[21:24:34 CEST] <Daemon404> i see where this is going
[21:24:39 CEST] <Daemon404> 'yes'.
[21:24:45 CEST] <BBB> although I guess factually speaking vp9 doesnt really have golden frames anymore
[21:24:53 CEST] <BBB> the encoder still does, but the decoder doesnt
[21:24:58 CEST] <BBB> (dont ask)
[21:25:14 CEST] <Daemon404> sounds like a weird implementation detail to support having both codecs in the encoder
[21:25:16 CEST] <BBB> anyway, yes, it reproduces libvpx on random garbage now
[21:26:08 CEST] <BBB> well, so, my understanding (Im a little out of date w.r.t. now) is that the decoder supports 8 ref frames, of which 3 are active, but the encoder just uses 3 and they are in the same order/setting as golden/altref/last in vp8
[21:26:46 CEST] <Daemon404> O_o
[21:28:14 CEST] Action: Daemon404 checks in for his paris flight (i dont know why the airline lets you check in 3 days before)
[21:29:32 CEST] <BBB> so you can buy upgrades to business class
[21:29:36 CEST] <BBB> and rent hotels
[21:29:38 CEST] <BBB> and cars
[21:29:40 CEST] <BBB> and other stuff
[21:29:49 CEST] <Daemon404> there is no such thing as business class on germanwings
[21:30:19 CEST] <Daemon404> but yes, they ask you to upgrade
[21:30:48 CEST] <BBB> toldya
[21:31:13 CEST] <fritsch> Daemon404: lol
[21:31:26 CEST] <fritsch> hopefully you are < 1.80 ...
[21:31:52 CEST] <Daemon404> i assume you mean meters
[21:32:13 CEST] <Daemon404> (i have flown germanwings before)
[21:32:33 CEST] <fritsch> yeah meters
[21:32:33 CEST] <Daemon404> and im just shy of 1.8 ;)
[21:32:36 CEST] <fritsch> hehe
[21:32:38 CEST] <Daemon404> ~1.77
[21:32:54 CEST] <fritsch> last time i was booked to use them ... it sucked
[21:33:01 CEST] <Daemon404> theyre not bad
[21:33:10 CEST] <Daemon404> theyre miles better than a real low cost carrier liek ryanair
[21:33:12 CEST] <fritsch> had to take the plane to Mallorca
[21:33:17 CEST] <fritsch> business trip
[21:33:24 CEST] <fritsch> between a whole lot of those clapping touris
[21:33:45 CEST] <Daemon404> oh. a tourist party plane. my condolences.
[21:33:53 CEST] <Daemon404> thats why i dont fly e.g. easyjet
[21:33:57 CEST] <fritsch> my institute books the cheapest they find
[21:34:02 CEST] <Daemon404> stag dos on every flight on easyjet
[21:34:02 CEST] <fritsch> and if the conference is in summer
[21:34:04 CEST] <Daemon404> without fail.
[21:34:13 CEST] <Daemon404> ic
[21:34:25 CEST] <fritsch> ah, it was to antalya ... not mallorca, but same people
[21:34:30 CEST] <Daemon404> if i fly on business, i can book any econo fair within  reason on company dime
[21:34:51 CEST] <fritsch> we are just allowed to drive 1st class deutsche bahn
[21:35:00 CEST] <fritsch> but that's allover the same in public sector I think
[21:35:04 CEST] <Daemon404> (there is a separate channel for FOSS media guys to discuss travel; we all frequent flyertalk)
[21:35:16 CEST] <fritsch> hehe
[21:35:20 CEST] <fritsch> yeah sorry for offtopic
[21:35:27 CEST] <Daemon404> doubt it matters
[21:35:33 CEST] <Daemon404> we only interrupted VP9 trolling
[21:35:39 CEST] <Daemon404> Very Important
[21:36:16 CEST] <fritsch> that last thing about vp9 remembered me of amd's xvba
[21:36:29 CEST] <fritsch> decoding h264 level4.1 at 4 reframes worked
[21:36:38 CEST] <fritsch> for the 5.1 you needed to hack the "catalyst registry"
[21:36:38 CEST] <Daemon404> i had to google that
[21:36:42 CEST] <Daemon404> shows you how widespread it was...
[21:36:46 CEST] <BBB> Daemon404: you dstill dont like vp9? :(
[21:36:52 CEST] <BBB> I thought you just disliked libvpx
[21:36:55 CEST] <Daemon404> BBB, i dont dislike vp9
[21:36:59 CEST] <Daemon404> i dislike the way google markets it
[21:37:05 CEST] <cone-326> ffmpeg 03Michael Niedermayer 07master:8b47e106514b: avformat/hlsenc: Fix the method command line parameter
[21:37:16 CEST] <BBB> so hypothetically, if we iamgined on2/google/libvpx didnt exist
[21:37:20 CEST] <BBB> would you be ok with vp9?
[21:37:24 CEST] <BBB> *imagined
[21:37:40 CEST] <Daemon404> if there was a reasonable encoder that allowed decent gaisn over h264
[21:37:47 CEST] <Daemon404> now with supprt comign e.g. Edge
[21:37:53 CEST] <Daemon404> yeah, probably.
[21:38:01 CEST] <Daemon404> (and a spec obv)
[21:38:03 CEST] <fritsch> vaapi will support vp8 encode and vp9 decode in the future
[21:38:07 CEST] <BBB> there is a spec
[21:38:12 CEST] <fritsch> "thanks to google"
[21:38:12 CEST] <BBB> you just need to pay google for it, or so
[21:38:13 CEST] <BBB> :-p
[21:38:20 CEST] <Daemon404> BBB, funman saw it under NDA
[21:38:22 CEST] <BBB> and sign a nda
[21:38:23 CEST] <Daemon404> its a 'bitstream guide'
[21:38:33 CEST] <BBB> I dont know, I dont have it
[21:38:33 CEST] <BBB> sorry
[21:38:39 CEST] <Daemon404> lol
[21:38:48 CEST] <BBB> I dont have time to write one either
[21:38:51 CEST] <nevcairiel> why does google hide their spec like that
[21:38:52 CEST] <BBB> ffvp9 took enough of my time
[21:38:56 CEST] <nevcairiel> i thought they wanted a open format
[21:39:04 CEST] <fritsch> yeah, thought the same
[21:39:05 CEST] <BBB> if you want me to write a spec, pay me :-p
[21:39:30 CEST] <BBB> Im probably not even the right person to write one
[21:39:40 CEST] <BBB> maybe that guy writing a ffv1 spec can write a vp9 spec?
[21:40:01 CEST] <Daemon404> ffv1 spec is tax funded, i think?
[21:40:07 CEST] <Daemon404> it's a preservation thing
[21:40:08 CEST] <Daemon404> for archives
[21:40:42 CEST] <nevcairiel>  a spec for vp9 would only be worht something if vpx actually obeys it, but since vpx is the spec including its weird and broken behaviors, why would anyone bother
[21:41:04 CEST] <BBB> ffvp9 exactly reproduces libvpx
[21:41:11 CEST] <BBB> (including bugs)
[21:41:18 CEST] <BBB> and serves as sweet documentation of these bugs
[21:41:24 CEST] <BBB> so & I dont think a spec is impossible
[21:41:30 CEST] <BBB> (and yes I filed bugs for each bug)
[21:41:30 CEST] <nevcairiel> well, i would argue its not exact, but closer everytime you send a patch =p
[21:41:43 CEST] <BBB> I think its actually exact now
[21:42:54 CEST] <BBB> or so Im being told
[21:43:17 CEST] <Daemon404> we now decode /dev/urandom like libvpx does
[21:43:28 CEST] <BBB> as long as /dev/random is valid input, yes
[21:43:54 CEST] <Daemon404> eventually it will be
[21:44:03 CEST] <BBB> I guess
[21:44:11 CEST] <BBB> anyway, yes, we decode /dev/urandom like libvpx, essentially
[21:44:21 CEST] <BBB> theres a few buts but thats basically what it means
[21:44:29 CEST] <BBB> and were still faster
[21:44:51 CEST] <Daemon404> i thought chromes reason for using vpx was error resiliance
[21:45:49 CEST] <genio> I was asking a question on #ffmpeg, but I think it would probably be best in here as it will likely require code changes.  should I repeat it here or are you guys there, too?
[21:46:17 CEST] <nevcairiel> you should only ask here if you are actually going to make said code changes yourself
[21:46:37 CEST] <BBB> Daemon404: only for cases where that matters (lik webrtc)
[21:46:58 CEST] <Daemon404> i think they care a bunch about that
[21:46:58 CEST] <BBB> Daemon404: for vp8 in html5/mediasource, chrome used ffvp8, and my understanding is that they were looking to use ffvp9 also
[21:47:07 CEST] <Daemon404> iirc peloverde has spent a lot of time working on screensharing via webrtc
[21:47:10 CEST] <BBB> I dont know if they finished it
[21:47:11 CEST] <Daemon404> (and its concave rd curve)
[21:47:25 CEST] <genio> nevcairiel: I doubt I have the knowledge set required to do so.
[21:47:36 CEST] <BBB> right but thats for vp10
[21:47:40 CEST] <BBB> vp10 is still fair bits away
[21:47:45 CEST] <Daemon404> in theory
[21:49:53 CEST] <genio> https://www.cwhitener.com/2015-09-14t12.47.26.avi is a crappy video with many errors (I can't change how they're made -IPCams).  VLC seems to play the original AVI fine, but when I try to copy it to an MP4, there are so many errors that the resulting video can't be played.    fmpeg -i crap.avi -an -vcodec copy -movflags faststart output.mp4
[21:50:54 CEST] <Daemon404> this should be filed as a bug
[22:00:10 CEST] <llogan> michaelni: i want to update the github mirror description to mention that patches should be sent to the mailing list instead of sending merge requests, but, AFAICT, i don't have the pemissions to do so
[22:00:24 CEST] <BtbN> Why not just disable pull requests?
[22:00:42 CEST] <llogan> because i'm stupid
[22:01:09 CEST] <BtbN> Hm, is it even possible to disable them?
[22:01:17 CEST] <llogan> i assumed you knew
[22:01:24 CEST] <llogan> because i didn't see a way
[22:01:36 CEST] <BtbN> I'm sure i saw repos without them
[22:01:43 CEST] <BtbN> But i can't find an obvious way to do so.
[22:02:16 CEST] <BBB> http://stackoverflow.com/questions/27957454/preventing-pull-request-in-github
[22:02:55 CEST] <BBB> so apparently theres no way
[23:15:26 CEST] <cone-326> ffmpeg 03Carl Eugen Hoyos 07master:a85f2cdfbff1: lavc: Fix standalone compilation of the avi demuxer.
[00:00:00 CEST] --- Tue Sep 15 2015


More information about the Ffmpeg-devel-irc mailing list