michaelni changed the topic of #ffmpeg-devel to: Welcome to the FFmpeg development channel | Questions about using FFmpeg or developing with libav* libs should be asked in #ffmpeg | This channel is publicly logged | FFmpeg 6.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
gust82 has quit [Remote host closed the connection]
MrZeus_ has joined #ffmpeg-devel
Gramner has joined #ffmpeg-devel
navi has quit [Quit: WeeChat 4.0.4]
ocrete2 has quit [Server closed connection]
ocrete2 has joined #ffmpeg-devel
lexano has quit [Ping timeout: 268 seconds]
Daemon404 has left #ffmpeg-devel [#ffmpeg-devel]
MrZeus_ has quit [Ping timeout: 246 seconds]
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
rodgort has quit [Server closed connection]
rodgort has joined #ffmpeg-devel
thilo has quit [Ping timeout: 240 seconds]
thilo has joined #ffmpeg-devel
balling has quit [Remote host closed the connection]
balling has joined #ffmpeg-devel
balling has quit [Remote host closed the connection]
balling has joined #ffmpeg-devel
bilboed has quit [Server closed connection]
bilboed has joined #ffmpeg-devel
balling has quit [Ping timeout: 268 seconds]
wyatt8740 has quit [Remote host closed the connection]
wyatt8740 has joined #ffmpeg-devel
dellas has quit [Remote host closed the connection]
wyatt8740 has quit [Remote host closed the connection]
wyatt8740 has joined #ffmpeg-devel
wyatt8740 has quit [Ping timeout: 246 seconds]
jamrial has quit []
wyatt8740 has joined #ffmpeg-devel
wangbin has joined #ffmpeg-devel
mkver has quit [Ping timeout: 256 seconds]
lemourin has quit [Read error: Connection reset by peer]
lemourin5 has joined #ffmpeg-devel
lemourin5 is now known as lemourin
balling has joined #ffmpeg-devel
MisterMinister has quit [Ping timeout: 256 seconds]
bpmedley has quit [Server closed connection]
bpmedley has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Remote host closed the connection]
balling has joined #ffmpeg-devel
durandal_1707 has quit [Ping timeout: 268 seconds]
durandal_1707 has joined #ffmpeg-devel
<another|> Anyone else getting AI spam on their commit email?
balling has quit [Ping timeout: 268 seconds]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Ping timeout: 246 seconds]
<Lynne> duke?
<Lynne> all the time
MisterMinister has joined #ffmpeg-devel
darkapex has quit [Ping timeout: 246 seconds]
darkapex has joined #ffmpeg-devel
<another|> Some "Kai" wants to sign me up to his AI newsletter, praising my ffmpeg contributions.
dykai has joined #ffmpeg-devel
dykai has quit [Client Quit]
AbleBacon has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
<balling> Wrong, Kai is not AI generated, its ampersand is Apple style, not present in ASCII. ChatGPT will generate normal ' sign, just checked. Probably not spam either, but I disabled Gmail app spam filter.
balling has quit [Quit: Connection closed]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
wangbin has quit [K-Lined]
balling has quit [Ping timeout: 268 seconds]
balling has joined #ffmpeg-devel
xxpor has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
gust82 has joined #ffmpeg-devel
llyyr has quit [Server closed connection]
llyyr has joined #ffmpeg-devel
Krowl has joined #ffmpeg-devel
<balling> Can somebody apply patch in (Carl?) https://trac.ffmpeg.org/ticket/9468
xxpor has quit [Quit: WeeChat 4.1.1]
<courmisch> another|: Duke Abbadon AI yes, I don't see the point that is trying to achieve
<courmisch> I procmailed it to oblivion after a while
<elenril> duke abaddon is weird
<aaabbb> what happens with the 2nd vote since several private vote urls were posted in public?
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
<elenril> ignore it and it will go away
<elenril> speaking of which
<elenril> michaelni: I'm still waiting for a list of extra GA members from you
<aaabbb> someone not authorized to vote voted on both public posted private urls
<elenril> both out of 4?
<aaabbb> i only improperly voted on^W^Wsaw 2
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
<balling> Lynne: BTW, when we talked on email you still did not tell me why you cannot apply legacy mp3 (not even aac) apple style iTunSMPB. First, they are used by author's project, he just sent a patch when I asked. 2nd, the FATE files for that patch are there in FATE. So files are in FATE that are unused.
<Lynne> I don't remember anything about mp3 skip
<Lynne> it should be a constant 1144 samples iirc, though, shouldn't it?
balling has quit [Ping timeout: 256 seconds]
balling has joined #ffmpeg-devel
<balling> @Lynne Apple encoder does not signal it like the main mp3 encoder called LAME does. There are other encoders of mp3, you know. FFmpeg patch is used in Cog. https://github.com/losnoco/Cog/blob/main/ThirdParty/ffmpeg/patches/0005-avformat-parse-iTunes-gapless-information.patch
<Lynne> the issues we have are for aac, not mp3
<Lynne> I'd like to leave mp3 for someone else to do after we've designed the new API properly
<balling> @Lynne AAC also uses that iTunesSMPB. Mp3 file is here, and the patch adds test for that file. http://fate-suite.ffmpeg.org/gapless/
<balling> Ffprobe already parses and prints iTunesSMPB, just not uses it.
<balling> QAAC compied with Visual Studio 2022 supports writing iTunesSMPB
<balling> And there are no any strange issues with SBR "additional" delay
<Lynne> there are, since our decoder also has a delay which isn't accounted for
<balling> FFmpeg decoder is not even close to itunes from Microsoft store a.k.a QAAC a.k.a. Apple Core Audio
<balling> *FFmpeg encoder
<balling> Apple encoder is better than even FDK: https://wiki.hydrogenaud.io/index.php?title=AAC_encoders
<aaabbb> balling: that's not really true anymore, fdk and qaac are extremely close these days
<balling> aaabbb: they were close even before, we were talking twoloop *native* encoder of ffmpeg
<Lynne> I have new analysis code written which is probably enough to outdo qaac
<aaabbb> ah, yeah that one is not good lmao
<Lynne> but with this whole preskip brokenness, and me doing xhe-aac now, I'm not sure I'd like to rewrite our aac encoder
<balling> @Lynne I hope you still merge 960 patch
<balling> 960 patch is for HE/v2 decoder
<Lynne> yup, I'll finish that, it's even related to xhe-aac in a way
Krowl has joined #ffmpeg-devel
<aaabbb> Lynne: that would cause the native encoder to outdo qaac?
<balling> I though you finished 960 patch already? The second patch from another person was complete, but you pointed out some bugs
<balling> aaabbb: no, 960 patch is for digital radio mondial
<balling> *Mondiale
<aaabbb> oh
<balling> It is for SDR with gnuradio. You decode digital stream from ether. Did you know that even decodibg FM radio is very complicated, because it has very different spectrum is different countries and to decode stereo you need a RTX 2080 Ti at least? Ahaha
<balling> *Mondiale is called DRM too, very funny
<Lynne> balling: it wasn't complete, it wasn't compliant
<balling> You pointed out there was 1 bug or 2. And then sent the patch in reply.
<Lynne> it was more serious than a bug or two, as I found out while merging the code into my patch
<balling> But FDK decoder is compliant?
<balling> Anyway, mp3 patch is here, looks very simple, I desribed format of iTunesSMPB, will copy https://patchwork.ffmpeg.org/project/ffmpeg/patch/20220128052107.1678032-1-kode54@gmail.com/
<balling> format is here, comment 18. "the byte offset from the first audio frame to the 8th-from-last frame" also means something, though https://trac.ffmpeg.org/ticket/2325#comment:18
balling has quit [Remote host closed the connection]
balling has joined #ffmpeg-devel
balling has quit [Client Quit]
balling has joined #ffmpeg-devel
balling has quit [Client Quit]
balling has joined #ffmpeg-devel
ldts has quit []
haxar has quit [Server closed connection]
haxar has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
pross has quit [Server closed connection]
pross has joined #ffmpeg-devel
philipl has quit [Remote host closed the connection]
philipl has joined #ffmpeg-devel
gust82 has quit [Remote host closed the connection]
gust82 has joined #ffmpeg-devel
marcj has quit [Quit: ZNC 1.7.5+deb4 - https://znc.in]
marcj has joined #ffmpeg-devel
dellas has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
Swedaniel has joined #ffmpeg-devel
balling has joined #ffmpeg-devel
kylophone has quit [Server closed connection]
kylophone has joined #ffmpeg-devel
Swedaniel has quit [Quit: Connection closed]
<durandal_1707> Quick favor from an FFmpeg user
<durandal_1707> This message seems dangerous
<durandal_1707> Similar messages were used to steal people's personal information. Avoid clicking links, downloading attachments, or replying with personal information.
Krowl has quit [Read error: Connection reset by peer]
<aaabbb> durandal_1707: which message?
<JEEB> I think one of my emails also got that message
balling has quit [Remote host closed the connection]
balling has joined #ffmpeg-devel
balling has quit [Remote host closed the connection]
balling has joined #ffmpeg-devel
ringo_ has quit [Server closed connection]
ringo_ has joined #ffmpeg-devel
balling has quit [Ping timeout: 268 seconds]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
<courmisch> more x86 weirdness http://lock.cmpxchg8b.com/reptar.html
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
michaelni has quit [Server closed connection]
michaelni has joined #ffmpeg-devel
balling has joined #ffmpeg-devel
navi has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
Krowl has joined #ffmpeg-devel
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
dellas has quit [Remote host closed the connection]
<Lynne> xhe-aac reads more like a D&D ruleset than specifications with all heuristics
pal has quit [Server closed connection]
pal has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
<Lynne> only they don't mention that it's based on another old ruleset, they just expect you to know everything about the old ruleset too, without telling you what's clearly different, just what is the new way
rvalue has quit [Ping timeout: 256 seconds]
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
balling has quit [Ping timeout: 256 seconds]
balling has joined #ffmpeg-devel
Daemon404 has joined #ffmpeg-devel
rvalue has joined #ffmpeg-devel
gust82 has quit [Remote host closed the connection]
gust82 has joined #ffmpeg-devel
gust82 has quit [Remote host closed the connection]
gust82 has joined #ffmpeg-devel
<balling> Android already has xHE decoder. Part of fdk aac. In fact decoder with ffmpeg works, just seek is broken, because probe is still from ffmpeg https://trac.ffmpeg.org/ticket/8690
jamrial has joined #ffmpeg-devel
mkver has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
dellas has joined #ffmpeg-devel
<JEEB> except fdk-aac's license is not GPL compatible so you cannot distribute an FFmpeg built with any GPL components (x264, x265 etc) with it. so adding support in the avcodec decoder makes sense in other ways than just "the person is interested in it"
balling has quit [Read error: Connection reset by peer]
cosminaught has quit [Server closed connection]
cosminaught has joined #ffmpeg-devel
balling has joined #ffmpeg-devel
<elenril> michaelni: re your private email, i believe it is a VERY bad idea to put people on the ballot when they did not actively agree to being on it
<elenril> I do not insist for that active agreement to be public (though then again I see no good reason for it not to be)
<elenril> but there must be consent from their side
<elenril> otherwise it's spam and borderline harrassment
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
<BBB> durandal_1707: I got that also
<BBB> I have to admit, it was a well-crafted email
<BBB> the "ffmpeg github" was kind of weird though, and using ffmpeg just for downscaling seems ... uh ... mis-informed?
<BBB> maybe we each got a different variant of the same message, too
<elenril> people do all kinds of weird things with ffmpeg
<balling> @JEEB it is more simple to open a server in Iran where GPL cannot be enforced, because the country is not part of Berne convention. I will ppint out that we all know non-free source of fdk ffmpeg binary on github, right?
<JEEB> I know people break the license restrictions. And as usual, any license related stuff is just extra if the person has an interest in implementing something
lexano has joined #ffmpeg-devel
<balling> @JEEB the point is it destributes configure non-free binaries with fdk and x264. In USA jurisdiction. 🤗
<JEEB> yes? doesn't make it any less correct. it is annoying that fdk has a weird custom license. but for some reason instead of people complaining to GOOG just do not follow the license?
<JEEB> but as I said, even if you consider the reason of licensing irrelevant
<JEEB> if someone *wants* to implement something due to interest, then none of your reasoning has any game to play there
Krowl has joined #ffmpeg-devel
<balling> At least openssl since 3.0.0 is Apache now. Compatible, even on windows. FIXES horrible bug with schannel about failed to send close TLS message, and wireshark looks very nice too compared to Windows crypto code. I had spend some type finding out how to leak the encryption key from schannel. Oogh.
dellas has quit [Remote host closed the connection]
<balling> @BBB if you have account on github it links to Sign-off-by as author (can be 3 and more authors, each with own Co-authored-by) and also if you are the committer it also links.
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
LaserEyess has quit [Server closed connection]
LaserEyess has joined #ffmpeg-devel
<BBB> I understand how he got my email address :) it's not hard to find
user23 has joined #ffmpeg-devel
<BBB> it's more about "why would you refer to ffmpeg's github when talking about how great ffmpeg is"
<BBB> github is not ffmpeg's official source code location
dellas has joined #ffmpeg-devel
<elenril> so many people talking to themselves today
ccawley2011 has joined #ffmpeg-devel
user23 has quit [Remote host closed the connection]
user23 has joined #ffmpeg-devel
j45 has quit [Server closed connection]
j45 has joined #ffmpeg-devel
j45 has quit [Changing host]
j45 has joined #ffmpeg-devel
gust82 has quit [Remote host closed the connection]
gust82 has joined #ffmpeg-devel
balling has quit [Remote host closed the connection]
balling has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 260 seconds]
ccawley2011 has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
balling has joined #ffmpeg-devel
<balling> @BBB it is not about source code. He looks at social network profiles. Github is a social network. It has public profiles with links to X, Facebook, etc. Patchwork does not allow users to insert their social.
<BBB> elenril: I was talking to balling :)
<BBB> to/with etc.
cone-878 has joined #ffmpeg-devel
<cone-878> ffmpeg Rémi Denis-Courmont master:45d0eb3f70fe: lavc/llauddsp: R-V V scalarproduct_and_madd_int16
<cone-878> ffmpeg Rémi Denis-Courmont master:6720a509a77f: checkasm: add lossless audio DSP
<cone-878> ffmpeg Rémi Denis-Courmont master:d076517056ca: lavc/llauddsp: R-V V scalarproduct_and_madd_int32
<elenril> at?
Lypheo has quit [Server closed connection]
Lypheo has joined #ffmpeg-devel
Krowl has quit [Read error: Connection reset by peer]
gust82 has quit [Remote host closed the connection]
gust82 has joined #ffmpeg-devel
dellas has quit [Remote host closed the connection]
Krowl has joined #ffmpeg-devel
wellsakus has quit [Server closed connection]
<michaelni> elenril, its a little late you bring this up, i can ask them but we need to give them time to reply.
<elenril> I am not bringing this up now
<elenril> I told you a week ago
<elenril> everybody told you
<elenril> various people have different exact criteria, but I see nobody except you saying it's ok to put people on the ballot without their agreement
<michaelni> then i misunderstood.
<michaelni> to me it seemed that peopel who just didnt make the threshold would be ok and others require to post in public
<michaelni> i mean it seemed to me that was what people asked for
<elenril> >there should be some confirmation that these people are actually interested in having voting rights.
<elenril> >Thing is, if they are subscribed and still don't ask to be added to the GA, I don't think they should be "volunteered" by someone else. They are of course welcome to apply.
<michaelni> "Adding the people who were part of the GA before; or are just under the threshold, but were active before; or are part of the infra teams and packaging; seem sane to add to the ballot."
<michaelni> thats what jb said, and i refered to it multiple times noone said "no" IIRC
<elenril> in my understanding that was with assumption that they agree to it
<elenril> or at least know about it
<elenril> it seem very rude to me to put someone's name on a ballot without even telling them
<michaelni> sure but noone said this a week ago. and i think i never said that i intended to contact them, just the people who fell outside of jbs set of rules
<elenril> you did say you would contac them
<michaelni> yep i did not say i will contact them
<michaelni> "otherwise I will contact them" not i will contact all
<michaelni> i only said i will contact the ones outside jbs definitions
<elenril> I suggest you start using punctuation then
<elenril> because my reading of that email was that you would ask all the people whether they want to be on the list
<elenril> I am VERY strongly opposed to putting people's names on the ballot with no indication that they know about it
<michaelni> thats ok but what do we do now ? should i contact them? and how long do they have time ? also i have no working email for at least ting but maybe more
wellsakus has joined #ffmpeg-devel
<elenril> if you really want them to be on the ballot, then you should contact them, yes
<elenril> if we have no working email for someone, then IMO that is a good indication that they should NOT be on the ballot
<michaelni> true
<JEEB> and yea, my understanding was as well that such people are open to be included (within some common sense limits) but I'd expect some sort of request/will for it. being pushed into a system in some cases after years - without any sort of previous contact - is something I'd consider quite odd.
<BBB> how would they vote without a working email?
<BBB> so we must know their email :)
<BBB> otherwise it's logically impossible to vote, right?
jdek has quit [Server closed connection]
jdek has joined #ffmpeg-devel
ePirat has left #ffmpeg-devel [Textual IRC Client: www.textualapp.com]
ePirat has joined #ffmpeg-devel
<michaelni> not impossible but close (they could be added to the GA and later provide a working address) but i agree we should drop them from this vote now as we cannot tell them about it and doing that seems what everyone wants
balling has quit [Read error: Connection reset by peer]
lexano has quit [Ping timeout: 256 seconds]
<michaelni> JEEB, i thought we vote on them first and if they pass we tell them they where selected to be extra members of the GA and if they are ok with that. I mean FFmpeg has also won all kinds of contests we never have been asked ahead of time
<michaelni> but ill mail all of them and ill CC anton or something
<JEEB> yea but contest is a bit different from actually being put into a position of decision making within a project
<JEEB> contest is where you get told how great you are and possibly receive something. end of story. being put onto a decision making body is kind of in a different ballpark :D
<elenril> FFmpeg is not a person
<elenril> being put on a ballot can be stressful for some people
<elenril> especially since everyone can see the results
Krowl has quit [Read error: Connection reset by peer]
<BBB> michaelni: I agree they may not all have time to reply before the vote commences. but we can always add the people that confirmed interest, and then afterwards if anyone responds "late", we can have a new vote for them in a little while, right?
<BBB> I also understand we don't want to have 100s of votes. but one vote now (which was already planned) and another one in a couple of weeks (?) seems not too bad
<michaelni> BBB, yes
balling has joined #ffmpeg-devel
balling has quit [Read error: Connection reset by peer]
dellas has joined #ffmpeg-devel
Traneptora has joined #ffmpeg-devel
balling has joined #ffmpeg-devel
Venemo has quit [Server closed connection]
Venemo has joined #ffmpeg-devel
balling has quit [Remote host closed the connection]
balling has joined #ffmpeg-devel
<cosminaught> Another problem with the fdk-aac decoder as implemented is that you can't do `ffmpeg -i someinput` and have it work automatically if that input happens to be xhe. Would need to ffprobe first to know (and ffprobe doesn't really pickup xhe properly anyway) and then override the decoder
AbleBacon has joined #ffmpeg-devel
<michaelni> elenril, others is that mail ok with you: https://pastebin.com/bgspU6Ty ? or should i change something before sending ?
<kierank> quite an demeaning quote
<michaelni> kierank, :)
<kierank> not funny
lexano has joined #ffmpeg-devel
<michaelni> heres the quote in question for the record: "Those who are too smart to engage in politics are punished by being governed by those who are dumber. -- Plato"
<elenril> michaelni: in the first paragraph, mention what it means - i.e. that GA members can vote
<elenril> "if you have more the 50%" is vague and not really accurate
<elenril> a better phrasing is 'if more people are in favor than against'
<courmisch> well that's actually a real problem. In VideoLAN, if you abstain on a membership request, it's as if you vote no
<elenril> why is that a problem?
<courmisch> because it means you can't be neutral, you have to vote yes or no
<elenril> does videolan also use CIVS?
<courmisch> and if a lot of members don't vote, then nobody can ever be voted in
<courmisch> no
balling has quit [Ping timeout: 260 seconds]
<elenril> well, the way I'm planning to map this vote onto CIVS is that there'll be a list of candidates and a special 'NOT CHOSEN' option
<courmisch> I mean it's a problem more like it needs to be clarified, not it's a fatal issue
<elenril> people who rank higher than NOT CHOSEN are chosen
<elenril> you can totally express not caring in that system
elastic_dog has quit [Ping timeout: 268 seconds]
<michaelni> version 2: https://pastebin.com/WCPYBUdN
<elenril> fun fact: plato was a fascist
<elenril> in today's terms
<elenril> s/until the next such vote/for 2 years/
<elenril> s/Yes please reply to/please reply to/
<Daemon404> isnt the point of the GA to be current
<Daemon404> not people missing for 10 years
<elenril> you are free to express that opinion by voting against adding them to GA
<Daemon404> assuming we get to vote again
<elenril> we do
<Daemon404> rather than endless FUD and french philisophical winging
<elenril> next vote is scheduled to start on monday
<Daemon404> on the ML
elastic_dog has joined #ffmpeg-devel
<Daemon404> which vote it this
<elenril> on adding extra people to GA
<courmisch> yeah, I agree, lets have Greek and Chinese philosophy instead
<elenril> i.e. the think we're discussing right now
<Daemon404> to my knowledge, the problematic parties dont even count the last vote as valid, so any further vote based on the GA will meet the same fate.
<michaelni> elenril, changes made, ill send the mail after eating a bit if no further big changes are requested otherwise i post a version 3
<durandal_1707> FFmpeg - where voting takes your rights
<Daemon404> durandal_1707, bexitmpeg?
<Daemon404> brexit, even
<durandal_1707> FFmpeg - where voting gives fake sense of made choices
<bencoh> isn't that every democracy?
<Daemon404> belarusmpeg then
<elenril> ffexit?
<Daemon404> ffexit(1);
funman has quit [Server closed connection]
funman has joined #ffmpeg-devel
<courmisch> ffraise(SIGKILL)
<Daemon404> ffraise(AVSIGNAL(SIGKILL))
<courmisch> you mean AVSIGNAL_KILL
<Daemon404> no that would be too clean
<Daemon404> see: AVERROR()
<courmisch> this ffproject sure fflikes to ffreinvent the ffwheel
<courmisch> ffreminds me of Apache HTTPd
<Daemon404> ffmpeg has an httpd i think
dellas has quit [Remote host closed the connection]
<courmisch> TIL Intel APX
<courmisch> Totally Not Copying AArch64 Features Into x86
ccawley2011_ has joined #ffmpeg-devel
ccawley2011 has quit [Ping timeout: 260 seconds]
<courmisch> - 32 GPRs
<courmisch> - nondestructive ops
<courmisch> - push/pop 2 GPRs at a time
<courmisch> where have I seen this before
dellas has joined #ffmpeg-devel
<durandal_1707> o great thou Leader you are, thoe strongest of Leaders all
<kurosu> checkasm is very unreliable for benchmarking purposes. The solution in eg dav1d was to increase the number of iterations if benchmarking is requested, but I don't see a way to do it in ffmpeg ?
<kierank> kurosu: are you sure it's not frequency scaling
<kierank> or powersaving on your pc
<kierank> the avx512 nucs have that disabled
<courmisch> kierank: the benchmarks are ridiculously short
<courmisch> it's okay if you have an accurate cycle counter, but not so much if you just have a clock
<kurosu> it could be, albeit I have set to the highest power/performance setting my laptop - which indeed is still a large source of unreliability
kekePower has quit [Read error: Connection reset by peer]
user23 has quit [Remote host closed the connection]
user23 has joined #ffmpeg-devel
kekePower has joined #ffmpeg-devel
<kurosu> (and I have other benchmarks that, being longer, do not see the same variance)
kekePower6 has joined #ffmpeg-devel
user23 has quit [Ping timeout: 250 seconds]
kekePower has quit [Ping timeout: 260 seconds]
kekePower6 has quit [Ping timeout: 260 seconds]
kekePower has joined #ffmpeg-devel
<courmisch> where did you guys hide all the half-precision DSP codez?
<courmisch> the RVV supports vector half-precision
<courmisch> +board
kekePower4 has joined #ffmpeg-devel
kekePower has quit [Read error: Connection reset by peer]
kekePower4 is now known as kekePower
kekePower0 has joined #ffmpeg-devel
kekePower has quit [Read error: Connection reset by peer]
kekePower0 is now known as kekePower
<durandal_1707> courmisch: there is only conversion from/to half
<courmisch> durandal_1707: seems scalar only though?
kekePower has quit [Client Quit]
kekePower has joined #ffmpeg-devel
<durandal_1707> courmisch: uses luts
<durandal_1707> on CPU
cone-878 has quit [Quit: transmission timeout]
kekePower has quit [Read error: Connection reset by peer]
kekePower has joined #ffmpeg-devel
kekePower has quit [Read error: Connection reset by peer]
kekePower has joined #ffmpeg-devel
kekePower has quit [Read error: Connection reset by peer]
MrZeus_ has joined #ffmpeg-devel
kekePower has joined #ffmpeg-devel
kekePower has quit [Read error: Connection reset by peer]
kekePower has joined #ffmpeg-devel
cone-673 has joined #ffmpeg-devel
<cone-673> ffmpeg sunyuechi master:afb967b81e41: af_afir: RISC-V V fcmul_add
<courmisch> gotta love how gitee.com search matches *any* keyword you git rather than *all*
<JEEB> :D
<JEEB> magical ORs
dellas has quit [Remote host closed the connection]
tufei has quit [Remote host closed the connection]
tufei has joined #ffmpeg-devel
<ePirat> anyone has a HAP codec sample file for me?
<ePirat> the official ones on their website are 404…
MisterMinister has quit [Ping timeout: 256 seconds]
MetaNova has quit [Server closed connection]
MetaNova has joined #ffmpeg-devel
MisterMinister has joined #ffmpeg-devel
zsoltiv_ has quit [Server closed connection]
another| has quit [Server closed connection]
another| has joined #ffmpeg-devel
zsoltiv_ has joined #ffmpeg-devel
<kurosu> courmisch: anyway, that add_pref_left or whatever is on x86 4.2x for SSSE3 and 8x for AVX2. And I hadn't relied, and wouldn't rely on checkasm with its L1d-hot data. Usually, much more cumbersome, but I surround a call in a codec using a DSP function, and run it on a file that is several hundreds frames
<kurosu> Captain obvious, but otherwise, memory loads appear a bit cheaper...
beastd has quit [Server closed connection]
beastd has joined #ffmpeg-devel
<cosminaught> there's a variety of HAP codec samples linked from Github at https://github.com/Vidvox/hap and I clicked on at least a few links and they seem to work
<cosminaught> I also spot checked a few on the links on https://docs.vidvox.net/vdmx_sample_media.html#media-files and they seem to resolve for me as well
cheesylard has joined #ffmpeg-devel
<cheesylard> hi, i found a bug in 4.2.9
<cheesylard> here is the code to repro: https://pastebin.com/raw/yQr3WWvX
<cheesylard> macOS arm64
<cheesylard> what's crazy about this is, when i inspect it with lldb, index is equal to -1. so that block shouldnt even be running.
<cheesylard> the input file is any FLAC file
<cheesylard> and this only repros if you try to seek forward. if you seek backwards, it doesnt repro. see line 114 of my repro code
<cheesylard> i realize there is a bug tracker, but i am more used to IRC, let me know if i should just create an issue in the tracker for this
<cheesylard> ALSO, this does not repro on ffmpeg 6.1
<cheesylard> so i guess it got fixed at some point, but the changes weren't cherrypicked
<cosminaught> I suspect this was the fix
<cheesylard> wow incredible. ok so im not crazy
<cheesylard> this was driving me nuts lol
<cheesylard> let me try
cone-673 has quit [Quit: transmission timeout]
<cheesylard> yep that fixed
<cheesylard> it
<wbs> that commit was a rather interesting case to learn about yet another way how UB can silently bite you
<cheesylard> yea those bugs are the worst, at first i was thinking that maybe my code was doing UB
<BtbN> it seems like compiler hostility more than anything
<BtbN> same with that no-strict-aliasing stuff
gust82 has quit [Remote host closed the connection]
<cheesylard> aight i sent a patch to the ffmpeg-devel@
<kierank> 4.2 lol
<cheesylard> yeah.... i learned my lesson, sticking with latest now
<durandal_1707> 4.2 is best
gust82 has joined #ffmpeg-devel
cone-957 has joined #ffmpeg-devel
<cone-957> ffmpeg Paul B Mahol master:82be1e5c0d82: avfilter/af_adynamicequalizer: do gain calculations in log domain
<cone-957> ffmpeg Paul B Mahol master:08e97dae205d: avfilter/af_adynamicequalizer: add adaptive detection mode
ccawley2011_ has quit [Read error: Connection reset by peer]
<durandal_1707> Lynne: see ^ i can give you mpv cmd to use to listen noise reduction/denoised audio
gust82 has quit [Remote host closed the connection]
gust82 has joined #ffmpeg-devel
navi has quit [Quit: WeeChat 4.0.4]
elenril has joined #ffmpeg-devel
<Lynne> durandal_1707: sure, could you paste it?
gust82 has quit [Remote host closed the connection]
gust82 has joined #ffmpeg-devel
Daemon404 has left #ffmpeg-devel [#ffmpeg-devel]