2025-03-03 01:04
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 7.1.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
00:01
<
kasper93 >
I left it as is, sent other janitorial patches
00:12
minimal has quit [Quit: Leaving]
00:17
kasper93 has quit [Quit: kasper93]
00:22
<
BtbN >
The whip developer, Jack Lau, is starting to annoy me... He keeps arguing "we need convention breaking thing XYZ in udp.c/tls_openssl.c/... cause WHIP/WebRTC"
00:22
vjaquez has quit [Ping timeout: 252 seconds]
00:22
<
BtbN >
He is seemingly unable to think outside of that box, only developing straight towards the goal he has in mind, taking no prisoners on the way there
00:23
<
BtbN >
"(Yeah, using the dtls:// protocol normally is impossible this way,) but WebRTC profits from it". Like what
00:25
kasper93 has joined #ffmpeg-devel
00:37
<
cone-473 >
ffmpeg Jack Lau master:7afe1167e5d5: avformat/tls_openssl: set tlsext host name after init ssl
00:37
<
cone-473 >
ffmpeg Timo Rothenpieler master:95eae5dd507f: avformat/tls_openssl: unset nonblock flag on correct URLContext during dtls handshake
00:37
<
cone-473 >
ffmpeg Jack Lau master:7bd3bdbd6224: avformat/whip: free udp socket after dtls free
00:37
<
cone-473 >
ffmpeg Jack Lau master:dda91b87e86b: avformat/tls_openssl: replace 1 with TLS_ST_OK to be more clear
00:37
<
cone-473 >
ffmpeg Jack Lau master:e6af82c49895: avformat/tls_openssl: remove all redundant "TLS: " in log with AVClass
01:16
carinobra has joined #ffmpeg-devel
01:21
<
kasper93 >
oh, so patchwork can send an email when build fails
01:21
<
kasper93 >
but cannot send me reset password? lol
01:26
<
carinobra >
Hello. I'm trying to pipe down a MPEG Dash (.mpd manifest) that is encrypted with Widevine. I have KID and KEY information. How do I pass those keys to ffmpeg to decrypt the stream?
01:43
acryo has quit [Remote host closed the connection]
01:44
acryo has joined #ffmpeg-devel
01:44
acryo has joined #ffmpeg-devel
01:44
acryo has quit [Changing host]
01:45
secondcreek has quit [Ping timeout: 248 seconds]
01:45
acryo has quit [Remote host closed the connection]
01:45
System_Error has quit [Remote host closed the connection]
01:46
acryo has joined #ffmpeg-devel
01:46
acryo has joined #ffmpeg-devel
01:46
acryo has quit [Remote host closed the connection]
01:48
acryo has joined #ffmpeg-devel
01:48
acryo has joined #ffmpeg-devel
01:48
acryo has quit [Remote host closed the connection]
01:49
carinobra has quit [Ping timeout: 272 seconds]
01:51
acryo has joined #ffmpeg-devel
01:51
acryo has joined #ffmpeg-devel
01:51
acryo has quit [Changing host]
01:52
System_Error has joined #ffmpeg-devel
02:13
beastd has quit [Ping timeout: 252 seconds]
02:14
<
compn >
hmm if that carinobra person comes back send him my way
02:26
beastd has joined #ffmpeg-devel
02:27
galqiwi has joined #ffmpeg-devel
02:28
galqiwi has quit [Client Quit]
03:19
jamrial has quit []
03:31
<
Yalda >
i will keep an eye as well
03:37
cone-473 has quit [Quit: transmission timeout]
04:04
mkver has joined #ffmpeg-devel
04:30
secondcreek has joined #ffmpeg-devel
04:32
mkver has quit [Ping timeout: 252 seconds]
05:01
_whitelogger has joined #ffmpeg-devel
06:48
<
compn >
a lot of mails from ffmpeg-lists going into my spam filter. lets just 'report as not spam' and hope things work out
06:49
vjaquez has joined #ffmpeg-devel
06:49
<
compn >
roundcube mail sucks as a webmail.
06:52
MisterMinister has quit [Ping timeout: 252 seconds]
07:05
<
compn >
oh now i'm getting kierank's duplicate mails :D
07:07
<
compn >
except no dupes on michaels. weird stuff.
07:24
Teukka has quit [Read error: Connection reset by peer]
07:25
bsFFFFFF has joined #ffmpeg-devel
07:27
Teukka has joined #ffmpeg-devel
07:27
Teukka has quit [Changing host]
07:27
Teukka has joined #ffmpeg-devel
07:32
blb has quit [Ping timeout: 272 seconds]
07:41
ngaullier has joined #ffmpeg-devel
07:41
ngaullier has quit [Remote host closed the connection]
07:45
ngaullier has joined #ffmpeg-devel
08:00
indecisiveturtle has quit [Quit: indecisiveturtle]
08:37
blb has joined #ffmpeg-devel
10:13
indecisiveturtle has joined #ffmpeg-devel
10:28
cone-566 has joined #ffmpeg-devel
10:28
<
cone-566 >
ffmpeg Niklas Haas master:9251af058a04: tests/checkasm: add scene_sad checkasm test
10:28
<
cone-566 >
ffmpeg Niklas Haas master:e44a1aaeecc1: avfilter/x86/scene_sad: add high bit depth AVX2/AVX512 version
10:28
<
cone-566 >
ffmpeg Niklas Haas master:dc61b74c1d30: avfilter/scene_sad: pass true depth to ff_scene_sad_get_fn()
10:28
<
cone-566 >
ffmpeg Niklas Haas master:91f2d146d418: avfilter/x86/scene_sad: add AVX512 implementation
10:32
<
haasn >
It seems `make fate` no longer implies `make ffprobe`? If I run `make clean && make fate` I get an error
10:33
<
haasn >
But after `make ffprobe` it passes
10:35
<
cone-566 >
ffmpeg Niklas Haas master:bc8d06d54188: avfilter/vf_thumbnail: unroll and use multiple histograms
10:55
compn has quit [Read error: Connection reset by peer]
10:56
compn has joined #ffmpeg-devel
10:56
bsFFFFFF has quit [Ping timeout: 252 seconds]
10:58
bsFFFFFF has joined #ffmpeg-devel
11:14
_whitelogger has joined #ffmpeg-devel
11:32
hbbs has quit [Quit: bye]
11:46
hbbs has joined #ffmpeg-devel
11:46
hbbs has quit [Changing host]
11:46
hbbs has joined #ffmpeg-devel
12:12
<
cone-566 >
ffmpeg Jack Lau master:abb274b154a1: avformat/tls_openssl: make tls and dtls use one close function
12:12
<
cone-566 >
ffmpeg Timo Rothenpieler master:e6d2c67cd096: avformat/tls_openssl: avoid unusual inline-if style
12:12
Xe_ has joined #ffmpeg-devel
12:12
Xe has quit [Ping timeout: 244 seconds]
12:17
Xe has joined #ffmpeg-devel
12:19
Xe_ has quit [Ping timeout: 276 seconds]
12:31
hbbs has quit [Ping timeout: 248 seconds]
12:35
<
cone-566 >
ffmpeg Jack Lau master:25e710c61ebf: avformat/whip: force NONBLOCK for rtp
12:40
<
haasn >
Lynne: what is going on with GSoC? I got emails asking about midterm evaluations
12:41
<
Lynne >
oh, yes, I need to do them
12:41
<
haasn >
As far as I can tell I have had no contact at all with the relevant students
12:41
hbbs has joined #ffmpeg-devel
12:41
hbbs has quit [Changing host]
12:41
hbbs has joined #ffmpeg-devel
12:42
jamrial has joined #ffmpeg-devel
12:56
<
Lynne >
students show up here once in a while, averne and indecisiveturtle
13:02
hbbs has quit [Quit: bye]
13:14
hbbs has joined #ffmpeg-devel
13:14
hbbs has joined #ffmpeg-devel
13:14
hbbs has quit [Changing host]
13:29
ccawley2011 has joined #ffmpeg-devel
13:35
minimal has joined #ffmpeg-devel
13:42
Guest4 has joined #ffmpeg-devel
13:44
<
indecisiveturtle >
Hi xD
13:45
Guest4 has quit [Client Quit]
14:20
MisterMinister has joined #ffmpeg-devel
14:25
kurosu has joined #ffmpeg-devel
14:38
ccawley2011 has quit [Read error: Connection reset by peer]
14:40
ccawley2011 has joined #ffmpeg-devel
15:05
ccawley2011 has quit [Read error: Connection reset by peer]
15:06
ccawley2011 has joined #ffmpeg-devel
15:29
jarthur has joined #ffmpeg-devel
15:33
ramiro has joined #ffmpeg-devel
15:35
cone-566 has quit [Quit: transmission timeout]
15:59
<
jamrial >
Lynne: that heif sample you linked is 6000x4000 with no tiles
15:59
<
jamrial >
crazy that a phone can handle that. i assumed their hardware en/decoders were limited to at most 4k
16:00
<
Lynne >
its from a camera
16:01
<
Lynne >
the hardware can do 8k 3:2 420p at 30ish fps, still pretty crazy
16:02
<
jamrial >
8k non tiled? nice
16:02
<
Lynne >
240fps at 1080p is much more impressive though
16:02
<
Lynne >
its 800Mbps, 2x the bitrate of 8k open gate
16:02
<
jamrial >
not really, samsung phones can do 960fps for their super slowmo mode
16:03
<
Lynne >
wow, at 1080p?
16:05
realies9 has quit [Quit: ~]
16:07
<
nevcairiel >
It was first introduced at 720p, but later phone models upped it to 1080p
16:10
<
Lynne >
phones sure move fast these days
16:10
<
Lynne >
wonder if they overheat
16:11
<
jamrial >
they limit recording super slow mo to only a few seconds
16:12
<
Lynne >
oof, about as much as I thought they would
16:13
<
Lynne >
shorter actually, I expected them to last a minute
16:25
ccawley2011 has quit [Read error: Connection reset by peer]
16:27
ccawley2011 has joined #ffmpeg-devel
16:34
kurosu has quit [Quit: Connection closed for inactivity]
17:12
realies9 has joined #ffmpeg-devel
17:38
ngaullier has quit [Remote host closed the connection]
17:38
hpkn has joined #ffmpeg-devel
17:41
rvalue has quit [Read error: Connection reset by peer]
17:42
rvalue has joined #ffmpeg-devel
17:54
galad has quit [Ping timeout: 276 seconds]
18:02
ccawley2011 has quit [Ping timeout: 244 seconds]
18:08
secondcreek has quit [Quit: secondcreek]
18:24
ccawley2011 has joined #ffmpeg-devel
18:26
<
haasn >
I am actually dying of cringe
18:26
<
haasn >
I don't know why this throwaway line is getting so much attention
18:26
<
haasn >
it's literally just a random simd routine for a bespoke filter that probably nobody else cares about
18:26
<
haasn >
and the C version was just shit
18:27
<
haasn >
"The obscurity of the function means it hadn’t been prioritized by the devs until now."
18:27
<
haasn >
dude it didn't exist until now
18:28
<
haasn >
phoronix quality research as always
18:28
<
haasn >
oh, it's not phoronix
18:28
<
haasn >
my apologies to phoronix
18:28
<
haasn >
"They would later go on to elaborate that the functionality, which might enjoy a 100% speed boost" 100x = 100% apparently now
18:29
<
kierank >
might be AI
18:30
<
kierank >
phoronix is not that bad
18:35
<
BtbN >
Phoronix picked it up as well though
18:35
<
BtbN >
cause of course they did
18:57
BradleyS has quit [Quit: quit]
19:00
BradleyS has joined #ffmpeg-devel
19:05
ccawley2011 has quit [Read error: Connection reset by peer]
19:06
ccawley2011 has joined #ffmpeg-devel
19:12
<
kasper93 >
sorry, but this is just crazy
19:13
averne has joined #ffmpeg-devel
19:14
<
kasper93 >
not to discard haasn achivement here, but this is not that important of filter, lol
19:15
<
kasper93 >
> Is this for encode or decode? Because for encoding, that's a big deal.
19:15
<
kasper93 >
yeah, about that
19:23
iive has joined #ffmpeg-devel
19:23
secondcreek has joined #ffmpeg-devel
19:25
mkver has joined #ffmpeg-devel
19:33
veritgo235 has joined #ffmpeg-devel
19:54
kurosu has joined #ffmpeg-devel
20:18
mkver has quit [Ping timeout: 248 seconds]
20:40
ccawley2011 has quit [Ping timeout: 276 seconds]
20:46
ccawley2011 has joined #ffmpeg-devel
20:54
bsFFFFFF has quit [Quit: bsFFFFFF]
21:00
Kwiboo has quit [Quit: .]
21:02
Kwiboo has joined #ffmpeg-devel
21:24
<
haasn >
there was exactly one intelligent reply on X
21:24
<
haasn >
the guy that pointed out that "speedup" is an artificial metric because there is no limit to how inefficient you can make a C version
22:04
<
BtbN >
"C version relies on bogosort"
22:04
kurosu has quit [Quit: Connection closed for inactivity]
22:04
witchymary has quit [Remote host closed the connection]
22:07
witchymary has joined #ffmpeg-devel
22:32
<
CounterPillow >
Lmao at the comments having some intrinsics enthusiast herping a derp
22:35
aaabbb has quit [Server closed connection]
22:36
aaabbb has joined #ffmpeg-devel
22:41
ccawley2011 has quit [Read error: Connection reset by peer]
23:17
kasper93 has quit [Ping timeout: 260 seconds]
23:21
kasper93 has joined #ffmpeg-devel
23:30
Yalda has quit [Quit: updates]
23:41
<
kasper93 >
save the click for lazy people: scene_sad8_avx512: 3686.6 ( 4.48x)
23:45
Marth64 has joined #ffmpeg-devel
23:52
Marth64 is now known as Yalda