BtbN changed the topic of #ffmpeg to: Welcome to the FFmpeg USER support channel | Development channel: #ffmpeg-devel | Bug reports: https://ffmpeg.org/bugreports.html | Wiki: https://trac.ffmpeg.org/ | This channel is publically logged | FFmpeg 7.0 is released
alexherbo2 has joined #ffmpeg
zamba has quit [Ping timeout: 245 seconds]
Some_Person has quit [Ping timeout: 248 seconds]
System_Error has quit [Ping timeout: 260 seconds]
kasper93 has joined #ffmpeg
<ThePendulum>
I get a warning "-vsync/fps_mode drop is deprecated" but the pipe to imagemagick breaks without it, is there anything that replaces it?
Peppi has quit [Read error: Connection reset by peer]
tokyovigilante has quit [Read error: Connection reset by peer]
Rena has quit [Read error: Connection reset by peer]
noobaroo has joined #ffmpeg
tokyovigilante has joined #ffmpeg
Rena has joined #ffmpeg
Dagger has quit [Ping timeout: 260 seconds]
lucasta has quit [Quit: Leaving]
<ThePendulum>
fps_mode vfr seems to 'work' but I don't know if that's a suitable replacement or I'm just lucky it worked this time
Tano has joined #ffmpeg
<noobaroo>
Weird question, i remember about ~10 years ago when the xbox360 vs ps3 debate wasnt ancient, that xbox360 stores non-HD scenery (not sure if just pictures or just video or both, or some dynamic game rendering engine that is neither of the the above, or all of the above) and then upscales to HD in realtime. Every once in a while I find a movie that is 480p or less but then when fullscreen I see the detail of their pores, acne, etc, in impressive
<noobaroo>
quality.
<noobaroo>
Is there any way to do this on modern PC technology without paying thousands/millions of dollars for some dedicated software?
<ThePendulum>
topaz seems to be popular, it's not thousands but still a couple hundred
<ThePendulum>
there's video2x which is free, don't know how good it is
Some_Person has joined #ffmpeg
<ThePendulum>
/good/ 480p is pretty watchable imo, it's just that a lot of 480p comes either from degraded VHS tapes/players or poorly transcoded DVDs
<noobaroo>
Im not talking about re-encoding the video
<ThePendulum>
real-time in the player?
<noobaroo>
Yeah exactly, i know lanczos can be used and mpv --profile=gpu-hq, but Xbox360 did it for everything and often outscored PS3's native 1080p in side-by-side comparisons
<noobaroo>
Personally i think thats fascinating. They reduced 25GB bluray games to 4GB that were higher quality than the bluray 1080p ones
<noobaroo>
by video2x do you mean waifu2x ?
<ThePendulum>
apparently it's based on waifu2x, seems a lot of upscaling happens with anime lol
<ThePendulum>
a modern TV will probably be the most accessible affordable real-time upscaler you can get atm tbh
<noobaroo>
Personally i cant tell the difference between --profike=gpu-hq or not, or lanczos or bicubic, etc. These things pale in comparison to whatever sort of thing xbox360 did starting in 2005. Its probably somewhat different because its CGI not actual photos, so I think there can be texture equations to multiply across the screen from very little data, but AFAIK that can be done regardless of resolution
lucasta has joined #ffmpeg
<ThePendulum>
yeah there are quite a few gaming upscalers out there, real-world video in real time is more obscure
<noobaroo>
Yeah, I can definitely see how it can be 100x easier to work with, honestly I feel like it should be more like 10,000x easier to work with or even more zeros x easier. I honestly don't understand whats going on when I have live action movies that re-encode to 1/4th the bitrate of some anime. It seriously perplexes me.
<ThePendulum>
with the same settings or going by subjective quality?
<noobaroo>
But even with CGI having and advantage when it comes to this sort of thing (less truly random data) Then it should be possible to use some xbox360-inspired technology on animated content/CGI/cartoons
<noobaroo>
Neither, objective quality
<ThePendulum>
the sharp lines can be a problem for some codecs
<noobaroo>
Which ones?
<noobaroo>
Probably AV1, i notice it usually makes video more soft and blurry, whereas x265 has more blocky artifacts
<ThePendulum>
this was a rumor from a discussion from 2002 tbf lol, but it adds up with that you don't generally use jpg for icons and logos because they start to look like shit almost immediately, while photos without hard edges can be compressed a ton before it stands out
<ThePendulum>
managed to get by video slideshow extractor working without intermediate files, replacing -c:v png with -c:v ppm was the key in the end https://www.pastery.net/jkzqgk/
<ThePendulum>
*my
<ThePendulum>
I understand the png option dumps all the frames as one giant stream of png data that imagemagick can't separate, while ppm has the file separation data it needs
Sakura`Kinomoto has quit [Ping timeout: 246 seconds]
kus has joined #ffmpeg
minimal has quit [Quit: Leaving]
System_Error has joined #ffmpeg
Dagger has joined #ffmpeg
ndufresne has joined #ffmpeg
xx has quit [Ping timeout: 260 seconds]
cmc has quit [Remote host closed the connection]
cmc has joined #ffmpeg
realies has quit [Read error: Connection reset by peer]
realies has joined #ffmpeg
realies has quit [Quit: ~]
hightower3 has joined #ffmpeg
mrelcee has quit [Quit: I want Waffles!]
mrelcee has joined #ffmpeg
jarthur_ has joined #ffmpeg
jarthur_ has quit [Client Quit]
jarthur has quit [Ping timeout: 260 seconds]
hightower2 has quit [Ping timeout: 255 seconds]
jarthur has joined #ffmpeg
realies has joined #ffmpeg
alexherbo2 has quit [Remote host closed the connection]
StephenLynx has quit [Remote host closed the connection]
realies has quit [Read error: Connection reset by peer]
realies has joined #ffmpeg
Sakura`Kinomoto has joined #ffmpeg
ewomer has quit [Quit: WeeChat 4.4.3]
realies has quit [Ping timeout: 246 seconds]
realies has joined #ffmpeg
coldfeet has joined #ffmpeg
Suchiman has quit [Quit: Connection closed for inactivity]
Sakura`Kinomoto has quit [Remote host closed the connection]
Sakura`Kinomoto has joined #ffmpeg
Juesto has joined #ffmpeg
Juest has quit [Ping timeout: 252 seconds]
Juesto is now known as Juest
stolen has joined #ffmpeg
jarthur has quit [Ping timeout: 252 seconds]
Some_Person has quit [Ping timeout: 260 seconds]
coldfeet has quit [Remote host closed the connection]
Some_Person has joined #ffmpeg
makidoll has quit [Remote host closed the connection]
lavaball has joined #ffmpeg
Some_Person has quit [Ping timeout: 252 seconds]
makidoll has joined #ffmpeg
Some_Person has joined #ffmpeg
j45 has quit [Ping timeout: 276 seconds]
j45 has joined #ffmpeg
Krusher has joined #ffmpeg
jafa has quit [Ping timeout: 260 seconds]
j45 has quit [Read error: Connection reset by peer]
j45 has joined #ffmpeg
lavaball has quit [Remote host closed the connection]
System_Error has quit [Ping timeout: 260 seconds]
realies has quit [Ping timeout: 272 seconds]
cmc_ has joined #ffmpeg
cmc has quit [Remote host closed the connection]
ciphrCat has left #ffmpeg [#ffmpeg]
militantorc has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net]
Magissia has joined #ffmpeg
xx has joined #ffmpeg
xx has quit [Max SendQ exceeded]
xx has joined #ffmpeg
System_Error has joined #ffmpeg
stolen has quit [Quit: Connection closed for inactivity]
vincejv has quit [Remote host closed the connection]
rvalue has quit [Read error: Connection reset by peer]
rvalue has joined #ffmpeg
MisterMinister has quit [Ping timeout: 244 seconds]
MisterMinister has joined #ffmpeg
System_Error has quit [Remote host closed the connection]
hightower3 has quit [Remote host closed the connection]
System_Error has joined #ffmpeg
Sketch has quit [Remote host closed the connection]
Sketch has joined #ffmpeg
<noobaroo>
This sounds like a stupid question, but is there any chance the ffmpeg cmd syntax for aresample lowpass as described on the ffmpeg wiki is outdated ? As described here: https://ffmpeg.org/ffmpeg-resampler.html
<JEEB>
it could be, unfortunately documentation is not auto-generated from actual code and how it works :)
<noobaroo>
Ive tried various combinations of `aresample=resampler=soxr:osf=s16:osr=48000:cutoff=0.2:precision=28` , with sox and with swr, and also tried removing it and just specifying as its own separate arg like `-cutoff 0.2` which I thought for sure would work, but it didnt
<noobaroo>
It says its supposed to be a float between 0 and 1, with the default 0.9 something. But no matter what i use, the result seems to be entire spectrum
<JEEB>
if you are trying to control a *filter* then separate arguments won't work as those are passed only to avformat and avcodec components (or if it's a cli custom then to whatever that is mapped to - like the display matrix stuff)
<noobaroo>
Oh okay... Well I have seen on stackexchange answers it being used either way, as part of the aresample colon separated list and as individual args, and also the ffmpeg webpage seems to say its okay either way too. But you definitely know more. Im just saying that it *should* work from the info available publicly
<noobaroo>
But I cant get it to work at all, no matter which way i use
<JEEB>
`ffmpeg -h filter=aresample`
<noobaroo>
There are no errors, but in audacity its identical and shows same dB all the way up to 20+ kHz, even the file sizes are identical
<JEEB>
that outputs the options actually defined by the module itself in whatever version you have
<JEEB>
and yes, avfilter options I think were =: based, so that I think should be OK
<noobaroo>
It says the same thing as the webpage, "set cutoff frequency ratio (from 0 to 1) (default 0)" except i do noticed it doesnt say float, it says <double> idk what a double is?
<JEEB>
double is a 64bit float
<JEEB>
or well, whatever the size on the architecture you are using :)
<JEEB>
generally float is 32bit, double is 64bit
<JEEB>
does not change the fact it's floating point
xx has quit [Ping timeout: 260 seconds]
<noobaroo>
So... any idea why its having no effect?
xx has joined #ffmpeg
<JEEB>
nope
<JEEB>
but the options are just going directly to swresample it seems
<noobaroo>
Maybe i need to specify the the input sample rate, i havent tried that so far
<noobaroo>
i thought that was only needed for raw data though, im doing from a flac file
<JEEB>
no you don't need to
<noobaroo>
This is cool though thank you, i wish i could find songs in flac that would actually benefit from being flac
<noobaroo>
Then what could it be? its worth a try
<JEEB>
I quickly checked and there's no logging of whether the option got applied to swresample or not (logging there), so I'd recommend adding an av_log call there to verify whether whatever you're passing in is getting there
<JEEB>
if it is, then it's code time
<JEEB>
the header does have further definition of the value tho
<JEEB>
but it's the internal header so not sure if it's relevant to the external value passed in :P
<JEEB>
thus in any case the primary thing you should do is add a log line in the init of swresample that would then log the configured value for that parameter
<noobaroo>
Well i specified isr=96000 and doesnt seem to make a difference
<JEEB>
that way you see whether it's getting passed to begin with
<JEEB>
well d'uh
<JEEB>
things would *fail* if the input info was not there
<noobaroo>
Thanks for checking, its giving up time for me, I am in too deep, was just trying to play around
Blacker47 has joined #ffmpeg
MoC has joined #ffmpeg
Tano has quit [Quit: WeeChat 4.4.2]
Suchiman has joined #ffmpeg
nobiz has quit [Ping timeout: 252 seconds]
<noobaroo>
JEEB in audacious, i can choose bit depth for the output plugin, for alsa plugin it only supports 16 or 32. For pipewire/pulseaudio it supports anything. Weirdly I feel like the best combination is 16bit output through either pulse or pipe. I like 16bit a lot. But I think pipewire internally mixes everything into float32 so wouldnt the "float" output theoretically reduce having to re-convert to f32 in the pipewire server?
<noobaroo>
Hopefully i explained that well
nobiz has joined #ffmpeg
Sakura`Kinomoto has quit [Remote host closed the connection]
<noobaroo>
I also kind of feel like the pipewire/alsa configuration (and first and foremost before this, the actual speaker hardware brand/model/size, but this probably doesnt need to be said) is way more important than choosing opus/aac/vorbis/mp3/flac ...
<noobaroo>
I also feel like the sample rate is more important than whether its lossless or lossy... Probably because from my understanding, the latency is measured by periodsize/samplerate so higher samplerate = less latency, of course this is only the case if its not resampled by pipewire settings, i have it so the hardware always plays at native sample rates, i cant play 88200hz files because of this
deus0ww has quit [Ping timeout: 252 seconds]
deus0ww has joined #ffmpeg
StephenLynx has joined #ffmpeg
MoC has quit [Quit: Konversation terminated!]
Lynne has quit [Read error: Connection reset by peer]
Sakura`Kinomoto has joined #ffmpeg
Sakura`Kinomoto has quit [Ping timeout: 252 seconds]
Some_Person has quit [Ping timeout: 276 seconds]
stolen has joined #ffmpeg
Some_Person has joined #ffmpeg
realies has joined #ffmpeg
HerbY_NL has joined #ffmpeg
Sakura`Kinomoto has joined #ffmpeg
HerbY_NL has quit [Read error: Connection reset by peer]
HerbY_NL_ has joined #ffmpeg
Lynne has joined #ffmpeg
Lynne has quit [Remote host closed the connection]