lavaball has quit [Remote host closed the connection]
Mister_Magister has quit [Quit: bye]
Mister_Magister has joined #ffmpeg
irc_user has joined #ffmpeg
qaph has joined #ffmpeg
kron has quit [Ping timeout: 252 seconds]
kron has joined #ffmpeg
qaph has quit [Ping timeout: 264 seconds]
Jhonny2x4 has quit [Ping timeout: 240 seconds]
waleee has joined #ffmpeg
psykose has quit [Remote host closed the connection]
psykose has joined #ffmpeg
minimal has quit [Quit: Leaving]
thilo has quit [Ping timeout: 240 seconds]
thilo has joined #ffmpeg
jarthur has quit [Quit: jarthur]
navi has quit [Quit: WeeChat 4.0.4]
nigetilly has joined #ffmpeg
lucasta has quit [Quit: Leaving]
Keshl has quit [Read error: Connection reset by peer]
Keshl_ has joined #ffmpeg
Keshl_ is now known as Keshl
qaph has joined #ffmpeg
kron has quit [Ping timeout: 256 seconds]
kron has joined #ffmpeg
qaph has quit [Ping timeout: 260 seconds]
nigetilly has quit [Ping timeout: 252 seconds]
HarshK23 has joined #ffmpeg
lemourin has quit [Read error: Connection reset by peer]
lemourin has joined #ffmpeg
JanC_ has joined #ffmpeg
JanC is now known as Guest7764
Guest7764 has quit [Killed (lead.libera.chat (Nickname regained by services))]
JanC_ is now known as JanC
irc_user has quit [Quit: Connection closed for inactivity]
emmanuelux has quit [Read error: Connection reset by peer]
emmanuelux has joined #ffmpeg
Hackerpcs has quit [Quit: Hackerpcs]
ossifrage has quit [Remote host closed the connection]
ossifrage has joined #ffmpeg
Hackerpcs has joined #ffmpeg
Hackerpcs has quit [Max SendQ exceeded]
ossifrage has quit [Remote host closed the connection]
ossifrage has joined #ffmpeg
Hackerpcs has joined #ffmpeg
bitbinge has quit [Ping timeout: 240 seconds]
qaph has joined #ffmpeg
Ogobaga has quit [Quit: Konversation terminated!]
Ogobaga has joined #ffmpeg
kron has quit [Ping timeout: 260 seconds]
qaph is now known as kron
Muimi has joined #ffmpeg
bitoff_ has quit [Ping timeout: 240 seconds]
LionEagle has quit [Quit: Leaving]
noonien85 has quit [Ping timeout: 245 seconds]
epony has quit [Ping timeout: 264 seconds]
nigetilly has joined #ffmpeg
bitoff_ has joined #ffmpeg
epony has joined #ffmpeg
Coinflipper has quit [Quit: ]
Coinflipper has joined #ffmpeg
praise_the_sun has quit [Quit: Client closed]
FH_thecat has quit [Quit: Leaving]
cryptic has quit [Ping timeout: 260 seconds]
nigetilly has quit [Ping timeout: 252 seconds]
praise_the_sun has joined #ffmpeg
vampirefrog has quit [Ping timeout: 255 seconds]
epony has quit [Remote host closed the connection]
epony has joined #ffmpeg
hseg has joined #ffmpeg
<hseg>
Hi. Is there a way to match _either_ an audio stream by language, or all audio streams if they aren't tagged by language?
<hseg>
because -map m:language:jpn? seems to not map anything if it doesn't find a japanese-language stream
mven9 has quit [Ping timeout: 276 seconds]
nigetilly has joined #ffmpeg
AbleBacon has quit [Read error: Connection reset by peer]
mven9 has joined #ffmpeg
FH_thecat has joined #ffmpeg
Jhonny2x4 has joined #ffmpeg
<hseg>
guess i could always ffprobe and determine appropriate cmdline from there
Jhonny2x4 has quit [Read error: Connection reset by peer]
Jhonny2x4 has joined #ffmpeg
Jhonny2x4 has quit [Read error: Connection reset by peer]
Jhonny2x4 has joined #ffmpeg
Jhonny2x4 has quit [Read error: Connection reset by peer]
nigetilly has quit [Ping timeout: 252 seconds]
Jhonny2x4 has joined #ffmpeg
hseg has quit [Quit: WeeChat 4.1.2]
hseg has joined #ffmpeg
Jhonny2x4 has quit [Read error: Connection reset by peer]
Jhonny2x4 has joined #ffmpeg
nigetilly has joined #ffmpeg
hseg has quit [Ping timeout: 256 seconds]
Jhonny2x4 has quit [Read error: Connection reset by peer]
Jhonny2x4 has joined #ffmpeg
Jhonny2x4 has quit [Read error: Connection reset by peer]
Jhonny2x4 has joined #ffmpeg
Jhonny2x4 has quit [Read error: Connection reset by peer]
Jhonny2x4 has joined #ffmpeg
MootPoot has joined #ffmpeg
Jhonny2x4 has quit [Read error: Connection reset by peer]
waleee has quit [Ping timeout: 252 seconds]
Jhonny2x4 has joined #ffmpeg
nigetilly has quit [Remote host closed the connection]
Jhonny2x4 has quit [Read error: Connection reset by peer]
Muimi has quit [Quit: Going offline, see ya! (www.adiirc.com)]
emanuele6 has quit [Quit: WeeChat 4.1.2]
emanuele6 has joined #ffmpeg
fling has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
fling has joined #ffmpeg
apteryx has quit [Ping timeout: 260 seconds]
mrelcee has quit [Ping timeout: 245 seconds]
mrelcee has joined #ffmpeg
vampirefrog has joined #ffmpeg
apteryx has joined #ffmpeg
<tm512>
seems like I need to add a resampling filter to the command I'm using to make AAC encodes for my iPod, since I just got a couple random 96KHz files mixed in with this album that's otherwise entirely 44.1KHz
<tm512>
resampling 44.1KHz -> 44.1KHz will be a no-op right? like it won't somehow change the audio that then gets passed to the encoder?
<tm512>
I'll probably use soxr since I compiled that into my ffmpeg build
<tm512>
probably don't need to be too fussy with the resampler when the output is going into FDK at like 100-180kbps VBR, tbh, but whatever
fling has quit [Remote host closed the connection]
raccct has joined #ffmpeg
fling has joined #ffmpeg
<raccct>
`ffmpeg -i A.mkv -codec mpeg4 A.mp4` made the file shrink by a third but no text is readable anymore. Any way I could preserve the quality and still go to a format I can use with messengers?
elastic_dog has quit [Ping timeout: 268 seconds]
<CounterPillow>
Are you sure you even want to be using mpeg4? I assume H.264 will be more widely supported as a video codec
<raccct>
CounterPillow: Thanks for the hint. I'm open to try that.
<CounterPillow>
Your choice of -codec mpeg4 had nothing to do with the output container format also named mp4, it referred to the "MPEG-4 Part 2" video codec which is related to divx and H.263
<raccct>
Unfortunately the conversion breaks with: Error while opening encoder - maybe incorrect parameters such as bit_rate, rate, width or height.
<CounterPillow>
Can you show me the full output in a paste?
<raccct>
I had the same before simpy using `convert`
<CounterPillow>
(I recommend that over using the high444 profile for compatibility reasons)
<raccct>
CounterPillow: I added high444 and -pix. Same error. Does -pix need a particular position in flag sequence?
<CounterPillow>
You should not add both
<CounterPillow>
you should only either add -pix_fmt yuv420p, or replace -profile:v high with -profile:v high444
<raccct>
ok, sry. I tried only -pix. same error. I try only high444 now.
<CounterPillow>
it's -pix_fmt, not -pix.
<CounterPillow>
this will work: ffmpeg -i A.mkv -c:v libx264 -pix_fmt yuv420p -crf 21 -preset veryslow -profile:v high -level 4.2 -movflags +faststart output.mp4
<CounterPillow>
this will also work, but have less compatibility across devices: ffmpeg -i A.mkv -c:v libx264 -crf 21 -preset veryslow -profile:v high444 -level 4.2 -movflags +faststart output.mp4
<psykose>
if signal is limited by the same stuff as chromium (since electron) then it wouldn't be able to play high444 mp4 for sure
<psykose>
of course it's easy to test so just do that
<raccct>
high444 is working and looking good. The second last proposal still fails with: [libx264 @ 0x559b23593b40] width not divisible by 2 (383x717)
<raccct>
I guess I'm fine with high444. Thanks for the help!
<CounterPillow>
right, weird resolution, you'll probably have to insert a pad filter
MootPoot has quit [Quit: Connection closed for inactivity]
<raccct>
CounterPillow: that might be it. thanks for the hint.
BetweenUs has joined #ffmpeg
BillTorvalds has joined #ffmpeg
Ogobaga has quit [Quit: Konversation terminated!]
Ogobaga has joined #ffmpeg
Tinos has joined #ffmpeg
<raccct>
tried to even my odd resolution by adding -filter:v pad="((2*iw)/2):((2*ih)/2)". Not very successful. Ideas?
Blacker47 has joined #ffmpeg
jab416171 has quit [Ping timeout: 276 seconds]
jab416171 has joined #ffmpeg
<vincejv>
i'm having issues with ffmpeg, it uses 1-2 cpu core on my quad core cpu (with HT on) and got this warning "x265 [warning]: No thread pool allocated, --wpp disabled, x265 [warning]: No thread pool allocated, --lookahead-slices disabled"
<vincejv>
anyway to fix? I encoded the same video on another machine, same CPU and it didn't have the error and uses multiple cores
stevenliu_ has joined #ffmpeg
<CounterPillow>
This sounds like multithreading APIs are broken on your OS install.
<vincejv>
not really an expert but does FFMPEG need CONFIG_NUMA=y in kernel for that threading API to work?
<CounterPillow>
It's not a problem in FFmpeg, it's in x265
<CounterPillow>
and yes it does.
<vincejv>
okay thank you CounterPillow! seems like the issue is with my distro not enabling that flag. I'll see what the devs think after i submit the PR, it's OpenWRT btw. Idk why they disabled it on x86 but enabled it for arm.
LionEagle has joined #ffmpeg
bitbinge has joined #ffmpeg
lavaball has joined #ffmpeg
<psykose>
it probably isn't really needed on an average 2-core x86 router i'd guess
<CounterPillow>
Yeah but when you start disabling random kernel features everyone else enables, you'll run into badly designed fallback paths like this one, where x265 presumably assumes "I can't tell how many NUMA nodes there are" means "I shouldn't use a thread pool at all"
bitbinge has quit [Remote host closed the connection]
bitbinge has joined #ffmpeg
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
rv1sr has joined #ffmpeg
Tinos has quit [Quit: Client closed]
ZedHedTed has quit [Remote host closed the connection]
ZedHedTed has joined #ffmpeg
Jhonny2x4 has joined #ffmpeg
navi has joined #ffmpeg
fling has quit [Remote host closed the connection]
Starz0r has quit [Ping timeout: 240 seconds]
fling has joined #ffmpeg
Arsen has quit [Quit: Quit.]
Arsen has joined #ffmpeg
Starz0r has joined #ffmpeg
Ogobaga has quit [Quit: Konversation terminated!]
Ogobaga has joined #ffmpeg
praise_the_sun has quit [Quit: Client closed]
bitbinge has quit [Ping timeout: 240 seconds]
praise_the_sun has joined #ffmpeg
elastic_dog has quit [Ping timeout: 240 seconds]
Kroko has joined #ffmpeg
praise_the_sun has quit [Quit: Client closed]
Kroko has quit [Remote host closed the connection]
praise_the_sun has joined #ffmpeg
elastic_dog has joined #ffmpeg
bitbinge has joined #ffmpeg
bitbinge has quit [Remote host closed the connection]
bitbinge has joined #ffmpeg
AbleBacon has joined #ffmpeg
JanC has quit [Ping timeout: 255 seconds]
\\Mr_C\\ has joined #ffmpeg
JanC has joined #ffmpeg
praise_the_sun has quit [Quit: Client closed]
user23 has joined #ffmpeg
<aleeku_>
where is avradio branch? cant find it anywhere...
rvalue has joined #ffmpeg
omegatron has joined #ffmpeg
praise_the_sun has joined #ffmpeg
praise_the_sun has quit [Client Quit]
praise_the_sun has joined #ffmpeg
praise_the_sun63 has joined #ffmpeg
praise_the_sun63 has quit [Client Quit]
praise_the_sun28 has joined #ffmpeg
praise_the_sun28 has quit [Write error: Broken pipe]
praise_the_sun63 has joined #ffmpeg
praise_the_sun63 has quit [Client Quit]
BetweenUs has quit [Ping timeout: 264 seconds]
praise_the_sun has quit [Ping timeout: 250 seconds]
kus has quit [Read error: Connection reset by peer]
<jokoon>
just some python module
<CounterPillow>
SuperCollider is an environment and programming language originally released in 1996 by James McCartney for real-time audio synthesis and algorithmic composition.[4][5]
<CounterPillow>
It would be pretty impressive if supercollider was "just some python module", on account of it being C++.
praise_the_sun has quit [Quit: Client closed]
praise_the_sun has joined #ffmpeg
LionEagle has quit [Ping timeout: 246 seconds]
LionEagle has joined #ffmpeg
Sl4yer has joined #ffmpeg
l4yer has quit [Ping timeout: 276 seconds]
LionEagle has quit [Client Quit]
BetweenUs has quit [Quit: Leaving]
sentriz has joined #ffmpeg
<sentriz>
hi all, when converting some flac -> mp3. is it possible for ffmpeg to write the original track duration/length to the mp3 file such that ffprobe wont say "Estimating duration from bitrate"?
<CounterPillow>
I don't think so, as MP3 as a format lacks that capability
<CounterPillow>
MP3 has no real container whatsoever, all the metadata you can add is basically a hack. If there is an ID3 tag for length then I'm not sure everything will support it
<furq>
yeah there's no actual header for duration in mp3
<furq>
there are id3 tags for it but it's probably best not to trust them
rsx has quit [Quit: rsx]
<sentriz>
ah i see
<sentriz>
though what is the tag that ffprobe is trying to read, that it can't find? when it falls back to bitrate estimation
jpb has joined #ffmpeg
<minimal>
furq: TLEN in ID3v2?
<furq>
yeah
<furq>
but it's just a metadata tag so it could be incorrect
<furq>
vbr mp3 has the frame count stored in the xing header
<furq>
and for cbr the estimation should be correct anyway
<minimal>
I'm assuming it's the encoder that sets the TLEN (correctly)
<jpb>
general video question - i'm streaming and saving H.265 from an axis camera. the video playback (ffplay and vlc) exhibits a glitchy jump forward of about one quarter second every now and then? any ideas?
<furq>
sure but it's probably better if ffprobe doesn't assume that
<jpb>
gaaak - streaming/saving H.264, not H.265
j45 has joined #ffmpeg
<Marth64>
gaaak: might be a setting on the camera side to adjust GOP (I-frame interval)
<Marth64>
you might be able to tune it to your liking to spread the jump out but cameras are a mixed bag; only assuming this is the cause since I had a camera like that
jokoon has quit [Ping timeout: 255 seconds]
<jpb>
Marth64: thanks. will check the GOP (I-frame interval)
<Marth64>
np good luck
<Marth64>
wouldn't make it too high or you'll get a delay starting the stream again later
<CounterPillow>
I assume ffprobe simply ignores it.
<Marth64>
just have to tune and test
<CounterPillow>
err, not ffprobe, but ffplay
<CounterPillow>
or ffprobe, nvm
jokoon has joined #ffmpeg
vincent has joined #ffmpeg
<furq>
i don't get that warning on a recent ffmpeg/ffprobe
<furq>
so maybe that changed recently
<sentriz>
CounterPillow: ah you think ffprobe is just ignoring it. that makes sense
<furq>
no i think something changed
<cristian_c>
hi
<CounterPillow>
or what furq said, the behaviour changed recently and you're using an older ffmpeg version
<furq>
if i encode with -write_xing 0 then i get that warning
<furq>
i think previously the xing header was only written for vbr
<cristian_c>
I've found that minimal playback is working, so, I've to try recording and so I use avconv/ffmpeg to achieve that. There is a script to do that, I've just changed from avconv to ffpmeg command, and set pal in place of ntsc
<vincent>
hey sentriz, i'm coming from the github issue. please ping me if i can help somehow
<furq>
of course the real answer is why are you using mp3 in 2024
<vincent>
haha
<vincent>
i feel threatened
<CounterPillow>
Yeah, unless you're targeting an awful MP3 player from the 00s you should not use MP3
<furq>
there's like three valid reasons to do it
<furq>
if you're not bound by those then please use something newer
<sentriz>
haha i already suggested opus >:)
<Marth64>
only one I can think of is browser support for basically every browser. Mozilla docs basically says MP3 has best compatbiility
<sentriz>
CounterPillow: i'm on n6.1.1. at least with this version, writing flac -> mp3, ffprobe estimates the length to be 5:44, puddletag says 4:30 from (probably difference is artwork or something) its CBR
<cristian_c>
when I run the script, output reaches line 80 and nothing happens.. If I press Ctrl+C, the missing output part is printed in the terminal
<cristian_c>
any ideas?
<Marth64>
cristian_c: Are your input files sockets or actual files? Hard to tell. Can't debug your script really but if you can break down what you're sending/doing with ffmpeg itself might be able to help
<Marth64>
err, pipes not sockets. just making an ssumption here, don't really know what the script does
<cristian_c>
Marth64, they are 4 pipes
<tm512>
regarding Opus on iOS, according to wikipedia: "On iOS 17: Support for playing .opus from Files; supports opus in mp4 (but not branded m4a) container"
<Marth64>
cristian_c: okay, that makes sense. so ffmpeg is picking up an EOF signal from those pipes (end of file)
<Marth64>
so it's actually terminating gracefully
<tm512>
I'm gathering that .opus files won't play nice with the regular music app or whatever it's called
<cristian_c>
Marth64, I remember I tried to get rid (in order to debug ffmpeg) of mkfifo and buffer commands in past (this means not using pipes) but things made even worse
<CounterPillow>
try -follow 1 on the inputs
<Marth64>
You should be able to use pipes, nothing wrong with pipe. Would try what CounterPillow suggested
<tm512>
AAC isn't too bad technically. as an end-user, there's like so little to be concerned about regarding patents, unless you're just wanting to avoid it for ideological reasons
qaph has joined #ffmpeg
<cristian_c>
Marth64, when you suggest to break down the script for debugging purpose, as you suggest to edit the script/ffmpeg cmmand?
<Marth64>
cristian_c: I meant just to explain it here in a summary, but I get the idea now kinda. thx
<cristian_c>
CounterPillow, I try to take a look at url
Haxxa has quit [Quit: Haxxa flies away.]
kron has quit [Ping timeout: 260 seconds]
<cristian_c>
CounterPillow, so, instead of -i .video_buffer and -i .audio_buffer
<cristian_c>
ah, ok, sorry, i've made a mistake
<CounterPillow>
try putting -follow 1 before each -i
<cristian_c>
follow
<cristian_c>
If set to 1, the protocol will retry reading at the end of the file, allowing reading files that still are being written. In order for this to terminate, you either need to use the rw_timeout option, or use the interrupt callback (for API users).
<cristian_c>
ok
<Marth64>
tm512: I admit to ideological reasons for avoiding it. biased and only personal opinion, but I have just had bad experience with AAC overall and hate how it sounds
<CounterPillow>
oh god we now have vinyl people for old bad codecs?
<bencoh>
:D
<tm512>
Marth64: I prefer using free codecs, though I've been using AAC a lot the last few weeks (got myself an iPod)
<tm512>
I haven't really been bothered by the sound quality, though I haven't really dipped into the bitrates where opus is clearly superior
Haxxa has joined #ffmpeg
foul_owl has joined #ffmpeg
qaph has quit [Ping timeout: 260 seconds]
<Marth64>
ya nothing wrong with that, gotta enjoy the music
kron has joined #ffmpeg
<cristian_c>
it seems nothing has changed
<CounterPillow>
I'm suspicious of any statement about a codec's "sound" that ignores bitrate as a factor because the xiphmont used to be able to distinguish different MP3 encoders in a blind AB test which, if not a cool party trick, should tell you a lot about what the important variable to control for is.
<CounterPillow>
-the
<Marth64>
Of course, there's a ton of optimization/tuning/alternative encoders can be used to get it right. My opinion is purely based on personal anecdote/bias and not scientifc
<tm512>
might go with rockbox (like my original reason for grabbing an old DAP) though I'm less sure of that. at least as long as I'm on the stock firmware, AAC is it unfortunately. not a fan of MP3 and the iPod doesn't seem to be a fan either. keep having stuttering issues with 320kbit MP3 or higher-bitrate VBR
kron has quit [Remote host closed the connection]
qaph has joined #ffmpeg
<CounterPillow>
cristian_c: out of ideas then, sorry :(
<cristian_c>
Marth64, CounterPillow btw, if I try video capture (not recording), it works: sudo somagic-capture | mplayer -vf yadif,screenshot -demuxer rawvideo -rawvideo "pal:format=uyvy:fps=25" -aspect 4:3 -
<cristian_c>
if i try recording using ffmpeg, it doesn't work
<tm512>
192kbit ABR from LAME plays back fine from my limited experience but FDK AAC sounds just as good to me with its VBR 4 mode which ends up closer to 140-160kbit on average and it's true VBR
<Marth64>
cristian_c: My suggestion is to track down where those EOF signals are coming from or at least why ffmpeg is thinking they are there
jokoon has quit [Quit: Leaving]
qaph has quit [Read error: Connection reset by peer]
kron has joined #ffmpeg
<tm512>
so the files are smaller and I don't have to worry as much about quality dropping because LAME decided it couldn't bump the bitrate on a difficult section without bringing the average bitrate too high
<znf>
BtbN: well, I came over, installed Ubuntu 22.04 fresh, enabled lingering, still won't limit ram :(
<tm512>
would prefer ~128kbit opus though which would get the files even smaller
<BtbN>
Probably time to ask in the systemd channel about it
<cristian_c>
Marth64, do you mean after "File 'video.avi' already exists. Overwrite? [y/N] y
<cristian_c>
"?
<Marth64>
cristian_c: `[vist#0:0/rawvideo @ 0x5561eb75d600] Decoder thread received EOF packet`
<Marth64>
Find thesource of this EOF packet\
<cristian_c>
[vist#0:0/rawvideo @ 0x5561eb75d600] Decoder thread received EOF packet
<cristian_c>
[vist#0:0/rawvideo @ 0x5561eb75d600] Decoder returned EOF, finishing
<cristian_c>
Marth64, could it depends on the fact I press Ctrl+c?
<cristian_c>
*depend on
<tm512>
didn't get an answer to this last night but I'm still curious. for the command I use to batch encode from lossless, do I have to worry about putting everything through a resampler? like if the output sample rate matches the input, it should just be a no-op, and won't affect the sound?
qaph has joined #ffmpeg
<cristian_c>
ctrl+c -> EOF, I mean
kron has quit [Ping timeout: 260 seconds]
<Marth64>
cristian_c: Possibly but there is a script involved around it so hard to tell how things are flowing. Sorry
<tm512>
ran into a couple 96KHz flacs in this album I downloaded and of course AAC will happily encode 96KHz audio but it produces files that won't play back on my iPod
<Marth64>
What the ffmpeg logs show is that it's starting up, seeing nothing to process but an EOF, and writing an empty file
<cristian_c>
Marth64, ffmpeg gets .video_buffer and .audio_buffer as inpts
<cristian_c>
so, I should look into these two pipes
<Marth64>
yes I would try that
<cristian_c>
but since they are pipes I can't look into them. Am I wrong?
<tm512>
I guess maybe I'll have to decode this 44.1KHz flac and mux it to raw PCM, then output the same file to raw PCM that's been passed through the resampler set to 44.1KHz
<tm512>
I think I had a similar concern in the past with the video scaler, whether it'd needlessly burn CPU cycles if I universally applied the scaler even when the input and output sizes were the same
kron has joined #ffmpeg
waleee has joined #ffmpeg
<cristian_c>
Marth64, unfortunately, .video_buffer and .audio_buffer print nothing if I try to cat them
qaph has quit [Ping timeout: 260 seconds]
<Marth64>
You sure it's getting data? if ffmpeg and cat both get nothing from it, something might be wrong with whats being fed to the pipe
<CounterPillow>
while I was on the toilet I in fact just thought that maybe your application feeding the fifos is exiting immediately, closing the fifos, which may result in an eof
<cristian_c>
Marth64, maybe I should run every command line-by-line but I remember I could do that in past
kron has quit [Ping timeout: 252 seconds]
jtgd has quit [Ping timeout: 245 seconds]
<cristian_c>
I've to try to convert somagic-both --pal 1>.video 2>.audio & pid=$! from bash to a real command
<cristian_c>
1> should be stdout and 2> should be stderr so this should stay unchanged
<cristian_c>
I'm not sure about pid=$!
kron has joined #ffmpeg
<cristian_c>
"$! is the PID of the last program your shell ran in the background (e.g. myprog &)"
<CounterPillow>
that just sets a variable within the script, you can ignore it
<cristian_c>
so, just: sudo somagic-both --pal 1>.video 2>.audio
<cristian_c>
it hangs
<cristian_c>
so, just: sudo somagic-both --pal 1>.video 2>.audio &
<cristian_c>
[1] 43790 (in background)
<cristian_c>
buffer < .video > .video_buffer &
<cristian_c>
[2] 43795
<cristian_c>
cat .video_buffer is still empty
kron has quit [Ping timeout: 255 seconds]
<cristian_c>
buffer < .audio > .audio_buffer &
<cristian_c>
[3] 43815
<cristian_c>
I suspect there is an issue with sudo
LionEagle has joined #ffmpeg
Ogobaga has quit [Quit: Konversation terminated!]
praise_the_sun has quit [Quit: Client closed]
<cristian_c>
CounterPillow, Marth64 ok, I've managed to reproduce the issue
<cristian_c>
CounterPillow, Marth64 , i've tested with a root shell (through sudo su -)
<cristian_c>
then I've run the rm, mkfifo and somagic-both commands (without sudo)
<cristian_c>
(meanwhile in a second terminal tab, i've run cat .video_buffer)
<cristian_c>
I've run buffer < .video > .video_buffer & and nothing happens in cat .video_buffer output
Ogobaga has joined #ffmpeg
Ogobaga has quit [Client Quit]
<cristian_c>
but after I run buffer < .audio > .audio_buffer &, an impressive amount of garbage is printed in cat .video_buffer output
<cristian_c>
:O
Ogobaga has joined #ffmpeg
<cristian_c>
CounterPillow, Marth64 I'm not an expert but my tests seem to reveal it should be a sudo related issue
Sl4yer has quit [Ping timeout: 255 seconds]
<cristian_c>
so if I run the full script from a root shell, I hope I don't catch such EOFs?
<cristian_c>
*.
fling_ has joined #ffmpeg
fling has quit [Ping timeout: 240 seconds]
kus has joined #ffmpeg
<cristian_c>
Marth64, CounterPillow it works!!!!
<cristian_c>
Marth64, CounterPillow you're clever
<cristian_c>
video and audio seeem to be working
kron has joined #ffmpeg
fling_ is now known as fling
<Marth64>
awesome
<Marth64>
!
acovrig60 has quit [Quit: Ping timeout (120 seconds)]
acovrig60 has joined #ffmpeg
<cristian_c>
Marth64, unfortunately, i'm not able to do that
<cristian_c>
I don't know what triggered
<cristian_c>
when I have done that, I had run cat .video_buffer in another tab
<cristian_c>
but it seems unrelated, in my opinion
<cristian_c>
I should look for the culprit
Sl4yer has joined #ffmpeg
kron has quit [Remote host closed the connection]
kron has joined #ffmpeg
Sl4yer has quit [Client Quit]
ZedHedTed has quit [Quit: leaving]
LionEagle has quit [Ping timeout: 260 seconds]
l4yer has joined #ffmpeg
irrgit_ has quit [Remote host closed the connection]
irrgit_ has joined #ffmpeg
hpkn has quit [Remote host closed the connection]
hpkn has joined #ffmpeg
Jhonny2x4 has quit [Read error: Connection reset by peer]
jtgd has joined #ffmpeg
ZedHedTed has joined #ffmpeg
Blacker47 has quit [Quit: Life is short. Get a V.90 modem fast!]
ACI-XCIX_0001 has quit [Quit: Client closed]
<cristian_c>
Marth64, it turns out recording starts only when I run sudo cat .video_buffer in another tab at the same time
<cristian_c>
it's very odd
<cristian_c>
audio result is good but video is damaged (picture is not stable)
BillTorvalds has quit [Quit: Leaving]
omegatron has quit [Quit: Power is a curious thing. It can be contained, hidden, locked away, and yet it always breaks free.]
Jhonny2x4 has joined #ffmpeg
sixecho has joined #ffmpeg
<sixecho>
is there a way to select frames using entropy? something like -vf `select=gte(entropy\,0.4)`
irrgit_ has quit [Remote host closed the connection]
irrgit_ has joined #ffmpeg
user23 has joined #ffmpeg
<BtbN>
What do you mean by "using entropy"?
<BtbN>
As in... at random?
<sixecho>
BtbN: i am already extracing various scene frames from a video as images, and was post-processing those images using imagemagick identify to calculate/filter-on the entropy of each image -- and noticed ffmpeg has entropy in filters; new to this syntax/etc
<BtbN>
I still don't understand what you mean by entropy
<BtbN>
specifying a dri device for cuda also makes no sense
<BtbN>
And you will also specify more parameters, the defaults are not great quality
<zumba_addict>
Got an error after retrying again `[h264_nvenc @ 0x55ecafc884c0] EncodePicture failed!: unsupported param (12): ` `Error submitting video frame to the encoder`
<zumba_addict>
this command works - `ffmpeg -h encoder=h264_nvenc`
krabador has quit [Quit: Leaving]
<zumba_addict>
it works now after removing high444
<zumba_addict>
wpw. ot
<zumba_addict>
wow, it's fast!
<zumba_addict>
This is the command I copied `ffmpeg -i input -c:v h264_nvenc -profile high444p -pixel_format yuv444p -preset default output.mp4`
<BtbN>
You are still not providing any quality parameters, so it will pick rather poor defaults
<zumba_addict>
it's done, wow, it's like less than 1.5 mins versus 45 mins ;D
<BtbN>
With likely substantially worse quality
<BtbN>
There is not really any reason to use a hardware encoder for non-real-time stuff
<BtbN>
It'll always be worse than a software encoder
<zumba_addict>
oh wow, yep, i just played it. It's like crf 35, hahaha
<zumba_addict>
really? we can't improve it?
<BtbN>
It's a hardware encoder.
<BtbN>
It is what it is, to change it, you change the hardware
<zumba_addict>
we can't specify crf?
<BtbN>
crf is an x264 concept
<BtbN>
some encoders have adopted it, but generally, only x264 has crf
<zumba_addict>
card is geforce gtx 1080. This is a good card right?
<Marth64>
libx264 is pretty fast on its own right, did you test it with -preset veryfast before switching to hw?
<BtbN>
Doesn't matter how good the card is. All cards from the same generation have the same encoder.
<BtbN>
A 1050 and a 1080 will perform near identically
<BtbN>
the lower end cards often even a bit better, due to higher clock
<zumba_addict>
and quality will always be similar to like crf 35 output?
<BtbN>
quality will be what you configure it to be
<BtbN>
but for the same bitrate, x264 will always be better
<zumba_addict>
oh
<Marth64>
every card generation is differnt. only constant is hardware will almost always be less efficient than software like BtbN is explaining
GingerGeek has joined #ffmpeg
<zumba_addict>
I want to try so I can compare the speed. Please help me improve the quality. Right now, it looks like crf 35 output `ffmpeg -i fpv-freestyle-juicy-sbang-LITH.mp4 -c:v h264_nvenc -pixel_format yuv420p -preset default output.mp4`
<Marth64>
all the knobs and switches are explained here `ffmpeg -h encoder=h264_nvenc
<zumba_addict>
what do I add?
<BtbN>
see the hwaccel wiki page, and the help of the encoder
<zumba_addict>
k
<BtbN>
There's way too many knobs and different kind of needs to just give a cookie cutter solution
<zumba_addict>
I tried -preset lossless and quality is the same as original. The bitrate is 649180.0kbits/s. And it is like 70x faster than my x264. It's because my i7 cpu is from 2009
<Marth64>
IIRC nvenc lossless is not actually lossless
<Marth64>
but if you are happy with the file size and outcome, it's good enough for most people
<zumba_addict>
but it's also not good in terms of filesize
<zumba_addict>
so big :D
<zumba_addict>
I guess, I really just need to upgrade to a better cpu
<Marth64>
zumba_Addict: Try this and tune from there. I don't know what your encoding goals are and don't expect best results but it might be fine and at least a start for VBR. -c:v h264_nvenc -preset p7 -tune hq -profile main -b_ref_mode middle -rc vbr -cq 22
<Marth64>
if your card doesn't support it remove -b_ref_mode middle
<zumba_addict>
ok
<zumba_addict>
I'm just playing and just so curious since I just got this card
<zumba_addict>
trying it now
<Marth64>
i use p7 preset a lot, it's not actually slow for average use
<zumba_addict>
the param worked
<zumba_addict>
it's fast
<Marth64>
also you are doing 4k in h264 which itself will need some extra consideration for space
<zumba_addict>
55000kbit/s
<Marth64>
bump up the -cq value till you are happy with the outcome
<zumba_addict>
will do
<zumba_addict>
wow, the first one you gave is already very nice
<sonicrules1234>
These days I just encode in av1
<Marth64>
:)
<Marth64>
nvenc is ok for what it is
<zumba_addict>
in -cq, lower better or higher better?
<Marth64>
lower better
<zumba_addict>
cool
<Marth64>
quality wise (which means ++size)
<zumba_addict>
yep
<Marth64>
av1 best
<zumba_addict>
-cq av1
<Marth64>
nonono hahaha
<sonicrules1234>
Lol
<zumba_addict>
:D
<Marth64>
i meant av1 is a good, beloved codec
<zumba_addict>
ah
<Traneptora>
av1 is great if you have a cpu that can support encoding it. H.264's design goals included low enough complexity that hardware at the time could handle it (in 2003) but that sort went out the window after the fact
<zumba_addict>
it's weird, my output doesn't have audio even after adding -acodec aac
bitbinge has quit [Ping timeout: 240 seconds]
<zumba_addict>
cool info
<Traneptora>
does your input have audio? what's the full, uncut command and output
<zumba_addict>
yes, it has audio
<Traneptora>
paste the full (uncut) command, and output, to a paste site and drop a link here