rizino has quit [Read error: Connection reset by peer]
rizino_ is now known as rizino
lucasta has quit [Quit: Leaving]
YuGiOhJCJ has quit [Remote host closed the connection]
YuGiOhJCJ has joined #ffmpeg
elastic_dog has joined #ffmpeg
ecapi_ has joined #ffmpeg
ecapi has joined #ffmpeg
graphitemaster has quit [Server closed connection]
graphitemaster has joined #ffmpeg
lexano has quit [Ping timeout: 255 seconds]
Muimi has joined #ffmpeg
twb has joined #ffmpeg
ecapi has quit [Ping timeout: 252 seconds]
<twb>
Is there a reason I can say "udp://239.255.3.17:1234?localaddr=10.0.0.1" but not "rtp://239.255.3.17:1234?localaddr=10.0.0.1" ? Per https://ffmpeg.org/ffmpeg-protocols.html#rtp
<twb>
wait that IS there...
<twb>
It's just not in the older version I'm running.... :/
GenTooMan has quit [Ping timeout: 272 seconds]
GenTooMan has joined #ffmpeg
dangmoo has joined #ffmpeg
finsternis has quit [Ping timeout: 255 seconds]
ParkerR has quit [Ping timeout: 260 seconds]
tofran has joined #ffmpeg
mobivme has joined #ffmpeg
elastic_dog has quit [Ping timeout: 252 seconds]
Kruppt has quit [Remote host closed the connection]
kurosu has quit [Quit: Connection closed for inactivity]
elastic_dog has joined #ffmpeg
Shuriko has joined #ffmpeg
MrZeus_ has quit [Ping timeout: 264 seconds]
Shuriko has quit [Read error: Connection reset by peer]
Shuriko has joined #ffmpeg
thilo has quit [Ping timeout: 240 seconds]
thilo has joined #ffmpeg
navi has quit [Quit: WeeChat 4.0.4]
luva6 has quit [Quit: Ping timeout (120 seconds)]
pikapika has quit [Ping timeout: 240 seconds]
luva6 has joined #ffmpeg
kmikita has quit [Ping timeout: 255 seconds]
moviuro has quit [Ping timeout: 272 seconds]
pikapika has joined #ffmpeg
carpenter has quit [Ping timeout: 240 seconds]
carpenter has joined #ffmpeg
wangbin has joined #ffmpeg
moviuro has joined #ffmpeg
Tano has quit [Quit: WeeChat 4.0.4]
Vonter has joined #ffmpeg
YuGiOhJCJ has quit [Ping timeout: 256 seconds]
AbleBacon has quit [Read error: Connection reset by peer]
kmikita has joined #ffmpeg
wangbin has quit [Ping timeout: 240 seconds]
lemourin has quit [Read error: Connection reset by peer]
lemourin has joined #ffmpeg
Ogobaga has quit [Quit: Konversation terminated!]
Ogobaga has joined #ffmpeg
YuGiOhJCJ has joined #ffmpeg
dangmoo has quit [Quit: Connection closed for inactivity]
Muimi has quit [Quit: Going offline, see ya! (www.adiirc.com)]
LuKaRo has quit [Ping timeout: 240 seconds]
LuKaRo has joined #ffmpeg
mobivme has quit [Ping timeout: 255 seconds]
maxim_d33 has quit [Ping timeout: 240 seconds]
DeHackEd has quit [Server closed connection]
maxim_d33 has joined #ffmpeg
DeHackEd has joined #ffmpeg
moviuro has quit [Quit: Reboot? Or did my jail(8) just die?]
pikapika has quit [Ping timeout: 258 seconds]
kmikita has quit [Ping timeout: 252 seconds]
pikapika has joined #ffmpeg
moviuro has joined #ffmpeg
Reventlov has quit [Server closed connection]
Reventlov has joined #ffmpeg
ecapi has joined #ffmpeg
pzich has quit [Server closed connection]
pzich has joined #ffmpeg
hijacker has quit [Server closed connection]
hjckr has joined #ffmpeg
ecapi has quit [Ping timeout: 264 seconds]
ecapi_ has quit [Ping timeout: 258 seconds]
durandal_1707 has quit [Ping timeout: 264 seconds]
durandal_1707 has joined #ffmpeg
moviuro has quit [Ping timeout: 252 seconds]
pikapika has quit [Ping timeout: 240 seconds]
waleee has quit [Ping timeout: 255 seconds]
ecapi has joined #ffmpeg
pikapika has joined #ffmpeg
moviuro has joined #ffmpeg
ecapi_ has joined #ffmpeg
ecapi has quit [Ping timeout: 264 seconds]
mobivme has joined #ffmpeg
ecapi__ has joined #ffmpeg
Shine has joined #ffmpeg
kmikita has joined #ffmpeg
ecapi_ has quit [Ping timeout: 264 seconds]
mobivme has quit [Ping timeout: 252 seconds]
<aaabbb>
i'm converting an audio codec at 22050 hz sample rate (and wideband) to opus. i see in the spectrum that there are some artifacts above 11khz, is there any benefit to having a low pass filter to remove them since they obviously weren't in the source audio?
<aaabbb>
*MB not WB
<twb>
Maybe a dumb answer, but do two test runs, one with and one without, and just listen to them?
<aaabbb>
my ears are bad and i personally cannot tell the difference
<twb>
ask your grandkids to listne to it for you? ;P
<aaabbb>
well i would but the audio isn't necessarily appropriate for kid's ears xD
<aaabbb>
i do 'showspectrum=mode=separate:color=intensity:scale=cbrt:s=640x360:fps=60' on the input audio, which is 22050 hz aac, so of course at about 12 khz the spectrum goes black, but there are still little spikes going above 12 khz which means surely they are artifacts
<microchip_>
oi
<aaabbb>
and i'm not sure if i want opus to be trying to encode those artifacts
<aaabbb>
9
ParkerR has joined #ffmpeg
<aaabbb>
also for low resolution (480p) videos is there any reason not to use mpdecimate to get rid of similar frames when there is very little motion happening?
<twb>
That doesn't happen automatically?
<aaabbb>
twb: no, it still encodes frames, it just takes less bits because there is very little motion
<twb>
"The main use of this filter is for very-low-bitrate encoding (e.g. streaming over dialup modem)"
<aaabbb>
i know but it's useful to remove duplicate frames when someone reencodes vfr as cfr, or records a video with screen capture
<twb>
Maybe I'm a slob but I generally don't mess too much with defaults, except maybe for one-off cases
<aaabbb>
a video captured at 60fps of a webcam recording captured at an average of 8fps sent through several bad transcodes so even identical frames differ slightly is literally dozens of times larger if i transcode without mpdecimate
<aaabbb>
shrinking a 200mb video to 60mb (hevc, no mpdecimate) or shrinking it to 5mb (hevc, with mpdecimate), with minimal quality loss
adendrag33209987 has joined #ffmpeg
<aaabbb>
i think because the transcodes it went through introduced far more quality loss than i ever could lol
stolen has joined #ffmpeg
<twb>
interesting
cis_ has quit [Remote host closed the connection]
<twb>
I thought libvpx (or whatever the hevc equivalent is) was smart enough to detect that automatically
<twb>
at least in "typical" cases and not, like, a slow pan across waving tall grass while on a bumpy train
<aaabbb>
it won't, if you tell it to output 60 fps then it'll output a frame every 60th of a second
<aaabbb>
if two frames are identical, then it'll only take a small number of bits but it won't deduplicate the frame
<twb>
oh hang on I see
<twb>
your 8fps of input became 60fps but those frames *weren't* "0 changes since the last keyframe"
<twb>
even though they should logically be
<aaabbb>
it became 60fps because some idiot was screen recording a video from a browser
<twb>
Incidentally I was fucking annoyed that right after I settled on vp9 I found out iOS can't do anything but H.26x
<furq>
if you want to be even more annoyed then ios has supported vp8/vp9 for ages but only through a private api that only safari is allowed to use and only for webrtc
<furq>
so that's nice of them isn't it
Mohammad has joined #ffmpeg
<aaabbb>
wow wtf
<furq>
i think at some point there was a whitelist that safari could use so that netflix would work
<furq>
idk if that's still the case
<furq>
but either way it wasn't something anyone else could use
Shine has quit [Read error: Connection reset by peer]
<furq>
with that said i would generally expect x265 to be faster and better than libvpx
Vonter has quit [Ping timeout: 272 seconds]
<aaabbb>
x265 is much better in almost all circumstances
Tano has joined #ffmpeg
<twb>
furq: ugh
goldsoft has joined #ffmpeg
mobivme has joined #ffmpeg
FilipSz has joined #ffmpeg
FilipSz has quit [Client Quit]
Freejack79 has joined #ffmpeg
Freejack79 has quit [Client Quit]
vlm has joined #ffmpeg
YuGiOhJCJ has quit [*.net *.split]
dgcampea has quit [*.net *.split]
chiselfuse has quit [*.net *.split]
beaver has quit [*.net *.split]
bitbinge has quit [*.net *.split]
hussein1 has quit [*.net *.split]
hussein1 has joined #ffmpeg
lusciouslover has quit [Ping timeout: 240 seconds]
vlm has quit [Remote host closed the connection]
chiselfuse has joined #ffmpeg
Keshl has quit [Read error: Connection reset by peer]
Keshl_ has joined #ffmpeg
NaviTheFairy has quit [Quit: Oh no. A thing broke.]
NaviTheFairy has joined #ffmpeg
Shine has joined #ffmpeg
paddymahoney has quit [Remote host closed the connection]
ivanich has joined #ffmpeg
bitbinge has joined #ffmpeg
durko03 has joined #ffmpeg
durko03 has quit [Remote host closed the connection]
lavaball has joined #ffmpeg
ivanich has quit [Remote host closed the connection]
Muimi has joined #ffmpeg
mobivme has quit [Ping timeout: 255 seconds]
kraftwerk28 has quit [Quit: *disconnects*]
kraftwerk28 has joined #ffmpeg
paddymahoney has joined #ffmpeg
Muimi has quit [Quit: Going offline, see ya! (www.adiirc.com)]
mobivme has joined #ffmpeg
Shuriko has quit [Ping timeout: 252 seconds]
kurosu has joined #ffmpeg
Lazenca has joined #ffmpeg
mobivme has quit [Remote host closed the connection]
mobivme has joined #ffmpeg
mobivme has quit [Remote host closed the connection]
mobivme has joined #ffmpeg
Capstan has joined #ffmpeg
stolen has quit [Quit: Connection closed for inactivity]
bnchs is now known as Cindy
Shine has quit [Read error: Connection reset by peer]
<twb>
When I do this: "ffmpeg -i tmp.ts -f rtp_mpegts -codec:a copy rtp://239.255.77.88:1234"
<twb>
...it seems to play back the video too fast (speed=26.3x)
<furq>
-re before -i
<twb>
Is there an option like "in real time"?
<twb>
Thanks
<twb>
I was looking for it under the mpegts or rtp settings
<twb>
Nice.
<twb>
Hopefully this means I can throw away ingests and multicat
<JEEB>
multicat is really nice and simple tho
<JEEB>
while this parses the whole input, and then re-generates the output
<JEEB>
so a *lot* can go wrong
<twb>
I guess
<JEEB>
multicat will not touch the data as much as possible
<twb>
My two reasons for doing so are 1) having to run ingests afterwards is annoying; and 2) I want to write to both rtp://239.255.77.88:1234?localaddr=10.128.0.1 *and* rtp://239.255.77.88:1234?localaddr=10.0.0.1 at the same time
<twb>
The current time-shifting process seems to be: multicat -d <duration × 27_000_000> @239.255.x.y:1234 foo.raw.ts; avconv -y -i foo.raw.ts -async 500 -vf yadif -q 4 foo.ts; ingests -p 8192 foo.ts; duration = $(lasts foo.aux)
<twb>
and then when playing it back, "multicat -t 2 -k foo.ts 239.255.x.y:1234"
<aaabbb>
is there a guide somewhere to understand how to correct metadata in an untagged video? something that tells how to detect if color range really is full despite being tagged as limited, whether something not tagged as progressive actually is, etc? for example i use signalstats=out=brng:color=red to determine if pixels fall out of the 16-235 range
Capstan has quit [Ping timeout: 250 seconds]
<aaabbb>
only the ones that can be detected unambiguously of course, so not things like dar
<furq>
there's the idet filter for progressive/interlaced but i usually check by eye
<furq>
and for things like primaries you just have to make an educated guess or compare to another source
<furq>
e.g. for an hdtv source it's either bt.709 or someone has made a mistake
<aaabbb>
mostly not hdtv, stuff recorded on cameras, phones, etc
<aaabbb>
lots of videos i have marked as yuv420p(tv) actually have pixels outside 16-235 and something to automatically correct that (and other things like chroma location, i had one that said it was center but it was actually left)
<aaabbb>
or at least just to erase all the metadata, because a*lot* of the time when it says tv, it's actually full pc range
<galad>
it's not strange to have values outside 16-235, even if it's not full range
<aaabbb>
i thought limited meant constrained to 16-235?
<JEEB>
16-235|240 but due to lossy video coding you might have some slightly out of range values
<galad>
cameras fill those too, so for example in post-production you could save some highlights or similar things if you got the exposure wrong
<furq>
it just means values outside that range will be clamped to 16 or 235
<JEEB>
yea
<furq>
encoders are free to use that space however they want
<furq>
and cameras etc
<aaabbb>
so ignalstats=out=brng:color=red isn't an accurate way to determine the range?
<aaabbb>
*signalstats=out=brng:color=red
<furq>
there'd be no need to flag the range otherwise
<furq>
do you ever say something and then immediately realise it's wrong
<aaabbb>
?
<aaabbb>
i'm still learning so a lot of what i say will be wrong without me realizing
<furq>
i meant what i said
Shine has joined #ffmpeg
<furq>
i meant that the thing i said was wrong
<aaabbb>
oh
<twb>
practically every day
<twb>
"It's not the DRESS that makes you look fat, honey, it's your huge arse"
<furq>
that's terrible
<furq>
massive arse has a much better rhythm to it
beaver has joined #ffmpeg
<twb>
Well if you want to go there, there's always "callipigean"
Lazenca has quit [Remote host closed the connection]
a0z has quit [Ping timeout: 255 seconds]
Shine has quit [Read error: Connection reset by peer]
deltab has quit [Ping timeout: 258 seconds]
ecapi__ has quit [Read error: Connection reset by peer]
deltab has joined #ffmpeg
Blacker47 has joined #ffmpeg
mobivme has quit [Remote host closed the connection]
mobivme has joined #ffmpeg
LeoNice1981 has joined #ffmpeg
mobivme has quit [Remote host closed the connection]
mobivme has joined #ffmpeg
kraftwerk28 has quit [Quit: *disconnects*]
kraftwerk28 has joined #ffmpeg
mobivme has quit [Remote host closed the connection]
mobivme has joined #ffmpeg
Ogobaga has quit [Quit: Konversation terminated!]
Ogobaga has joined #ffmpeg
pookie has quit [Quit: All for nothing]
olspookishmagus has joined #ffmpeg
drv has quit [Server closed connection]
drv has joined #ffmpeg
vlm has joined #ffmpeg
navi has joined #ffmpeg
ZedHedTed has quit [Remote host closed the connection]
ZedHedTed has joined #ffmpeg
mobivme has quit [Read error: Connection reset by peer]
lexano has joined #ffmpeg
Magissia has joined #ffmpeg
MrZeus_ has joined #ffmpeg
Ogobaga has quit [Quit: Konversation terminated!]
Ogobaga has joined #ffmpeg
rv1sr has joined #ffmpeg
user23 has joined #ffmpeg
ivanich has joined #ffmpeg
user23 has quit [Remote host closed the connection]
user23 has joined #ffmpeg
user23 has quit [Remote host closed the connection]
user23 has joined #ffmpeg
Arokh has quit [Server closed connection]
Arokh has joined #ffmpeg
Coinflipper has quit [Quit: ]
Coinflipper has joined #ffmpeg
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg
AbleBacon has joined #ffmpeg
minimal has joined #ffmpeg
user23 has quit [Remote host closed the connection]
user23 has joined #ffmpeg
Capstan has joined #ffmpeg
lexano has quit [Ping timeout: 240 seconds]
carpenter has quit [Ping timeout: 255 seconds]
lexano has joined #ffmpeg
Shine has joined #ffmpeg
stolen has joined #ffmpeg
galad has quit [Ping timeout: 258 seconds]
galad has joined #ffmpeg
DarkG has quit [Excess Flood]
DarkG has joined #ffmpeg
LuKaRo has quit [Ping timeout: 252 seconds]
waleee has joined #ffmpeg
fossdd_ has joined #ffmpeg
vlm_ has joined #ffmpeg
fossdd has quit [Ping timeout: 240 seconds]
fossdd_ is now known as fossdd
vlm_ has quit [Client Quit]
vlm has quit [Quit: Leaving]
chainik has quit [Server closed connection]
chainik has joined #ffmpeg
lexano has quit [Ping timeout: 258 seconds]
iderik has joined #ffmpeg
Nixkernal_ has joined #ffmpeg
<Trel>
But callipygian doesn't mean big, it means shapely.
Nixkernal has quit [Ping timeout: 258 seconds]
lexano has joined #ffmpeg
carpenter has joined #ffmpeg
beaver has quit [Remote host closed the connection]
beaver has joined #ffmpeg
LuKaRo has joined #ffmpeg
beaver has quit [Remote host closed the connection]
beaver has joined #ffmpeg
lexano has quit [Ping timeout: 252 seconds]
beaver has quit [Remote host closed the connection]
beaver has joined #ffmpeg
jtgd has quit [Ping timeout: 260 seconds]
minimal has quit [Quit: Leaving]
Shine has quit [Ping timeout: 240 seconds]
JEEB has quit [Quit: ブーン～]
lexano has joined #ffmpeg
JEEB has joined #ffmpeg
waleee has quit [Ping timeout: 252 seconds]
Capstan has quit [Quit: Client closed]
kron has joined #ffmpeg
flom84 has joined #ffmpeg
flom84 has quit [Remote host closed the connection]
flom84 has joined #ffmpeg
justThanks is now known as justTesting
emanuele6 has quit [Ping timeout: 264 seconds]
stolen has quit [Quit: Connection closed for inactivity]
microchip__ has joined #ffmpeg
microchip_ has quit [Ping timeout: 252 seconds]
microchip__ is now known as microchip_
Capstan has joined #ffmpeg
esc--ape-- has joined #ffmpeg
esc_ape has quit [Ping timeout: 245 seconds]
frankplow has quit [Server closed connection]
frankplow has joined #ffmpeg
waleee has joined #ffmpeg
Haxxa has quit [Quit: Haxxa flies away.]
HarshK23 has quit [Quit: Connection closed for inactivity]
Ogobaga has quit [Quit: Konversation terminated!]
Ogobaga has joined #ffmpeg
Haxxa has joined #ffmpeg
Ogobaga has quit [Remote host closed the connection]
Ogobaga has joined #ffmpeg
dangmoo has joined #ffmpeg
user23 has quit [Remote host closed the connection]
user23 has joined #ffmpeg
ivanich has quit [Ping timeout: 255 seconds]
MrZeus__ has joined #ffmpeg
<znf>
Still no support for AV1 in MPEGTS :(
yazooq has joined #ffmpeg
MrZeus_ has quit [Ping timeout: 252 seconds]
<JEEB>
znf: yea nobody from the developers has so far had interest in implementing it, although one of the developers is a co-author of the spec
<znf>
but -- once implemented on the "encoder" part, does it also need support on the decoder/demuxer (aka: the video player)?
<JEEB>
definitely
flom84 has quit [Ping timeout: 245 seconds]
<znf>
so, even if it lands in tomorrow, it's like still 1-2 years away
<JEEB>
?!
<JEEB>
if someone implements it, they generally will want to implement both demux and mux
<znf>
I mean, in the wild
<JEEB>
in the wild a lot of people utilize master
<znf>
that is if they actually use ffmpeg/libav and they release a new version of their $thing, and... so on
<znf>
and probably tv-manufacturers won't care to implement it for older models either
<JEEB>
and now you've gotten outside of FFmpeg :D
<znf>
yup
<JEEB>
personally while I do like the idea of AV1 in MPEG-TS, the custom startcode format and other such things just make me go "meh"
<JEEB>
and I have plenty of other stuff on my plate like finishing up avctx side data for encoder configuration
<znf>
what I'm trying to say is that, even if it lands tomorrow in ffmpeg, I probably won't see support for it for consumer devices any time soon, even if, for example, almost all new mobile devices ship with an AV1 decoder
<znf>
(hardware decoder)
<BtbN>
existing consumer devices won't receive updates for AV1 anyway
<BtbN>
their CPU is _WAY_ too weak, and you can't patch in a hardware decoder
<znf>
yup, obviously
<znf>
but, still, I think a lot of SoC for Smart TVs released in the last 2-3 years support AV1 hardware decoding, considering both Netflix and YouTube are pushing it
<JEEB>
anyways if you care about it, feel free to make a stab at implementing it yourself?
<JEEB>
you may get guidance about it and all :P
<JEEB>
after all, if you care about something you then actually end up happy if you hack it
<znf>
I'm very not a programmer :)
user23 has quit [Remote host closed the connection]
<znf>
my C knowledge is close to 0
user23 has joined #ffmpeg
<JEEB>
everyone goes through that point :D
<znf>
I'm too old for this, I just spent 1 hour in fixing a custom patch :D
<znf>
(which is under 100 lines)
rv1sr has quit []
kraftwerk28 has quit [Quit: *disconnects*]
<znf>
BtbN, I don't assume there's a way to *hasten* makeimage.sh? Apart from the actual compilation, everything else seems to underutilize all my 100+ threads
<znf>
oh wait, there's max-parallelism = 4
<JEEB>
I think with MPEG-TS you need to 1) write the stream type and descriptor (there are plenty examples of those in both the mpegts writer and reader, just searching for AV_CODEC_ID such as H264 or HEVC will bring you that stuff) and 2) handle the OBU to start code format conversion
<znf>
is that for local builds, too?
<JEEB>
of which I expect 2 to be more complex
lavaball has quit [Remote host closed the connection]
sm1999__ has quit [Quit: WeeChat 4.2.0-dev]
yazooq has quit [Ping timeout: 252 seconds]
sm1999__ has joined #ffmpeg
iive has joined #ffmpeg
<BtbN>
znf: only increase that if you got south of 64GB of RAM.
<BtbN>
Since it will happily spawn max-parallelism * nproc compilers
<znf>
I'm good, got 512
<BtbN>
might be able to safely set it to 8 with that
<znf>
Uhm
<znf>
Ok I might not be good
<znf>
I set it to 32 and went away
<BtbN>
You said you got 100 threads or something. So the ram consumption of every individual build can already be over 100GB or something
<znf>
If it crashes, it crashes! I'm not going back upstairs now