<aaabbb>
if you cloned the one i linked before when i said idk if it's a fork, then probably haha
<cedd>
damnit
<aaabbb>
just clean it and clone the right one
<aaabbb>
mkdir build; cd build; cmake ../; make; sudo make install
<cedd>
omg i think it worked
<aaabbb>
ffmpeg's configure is happy now?
<aaabbb>
the configure script doesn't actually do anything with libgme it only checks if it's there. in theory it could compile even if libgme wasn't there, but then when you try to run the new ffmpeg binary it'll fail to start because the linker can't find libgme on your system
<cedd>
well i installed libgme right in the ffmpeg dir
<cedd>
it seemed to recognize it now
<cedd>
i cant find the ffmpeg executable anywhere though
<aaabbb>
if you did sudo make install then it'll be in /usr/local/lib/ probably
<cedd>
no sudo on this one
<aaabbb>
are you trying to compile ligbme into a static ffmpeg?
<cedd>
i do see libgme there though
<aaabbb>
in where?
<cedd>
/usr/local/lib/
<aaabbb>
then you probably compiled and installed as root. either way that's a place ffmpeg's configure script will look
<cedd>
yea i just dont know where ffmpeg is supposed to be lol
<aaabbb>
did you "make" with ffmpeg?
<cedd>
yea
<aaabbb>
it's probably in bin
<aaabbb>
./bin
<cedd>
wait maybe i didnt
<cedd>
im doing it again regardless
<cedd>
hmm i must have messed something up
<cedd>
if i run ffmpeg it shows it was configured with '--enable-libgme'
<cedd>
but it doesnt look like libgme was included
<aaabbb>
run "ldd" on the new ffmpeg binary
<aaabbb>
and see if libgme is an entry in there
<cedd>
will dcfldd work?
<cedd>
im on a mac btw
<psykose>
that's not related to ldd
<aaabbb>
ldd, not dd
<psykose>
otool -L on macos i think
<cedd>
hmm i grepped gme on the output and it shows it
<cedd>
so maybe it did include it
<aaabbb>
then you've got it
<cedd>
not sure why spc isnt listed in the codecs though
<aaabbb>
fmpeg -decoders | grep gme
flotwig has quit [Ping timeout: 260 seconds]
<cedd>
doesnt show it
<aaabbb>
idk if it'd be called gme... but grep for whatever
<cedd>
i guess i could just try an spc file now
<cedd>
1 sec
<psykose>
gme is a demuxer i think
<psykose>
ffmpeg -demuxers
<furq>
yes it is
<cedd>
omg it worked
flotwig has joined #ffmpeg
<cedd>
thank you guys!!
<aaabbb>
yw :)
<cedd>
and special thanks to aaabbb you went above and beyond on this one lol
<cedd>
ah this is so cool
<aaabbb>
now you can convert these all to opus. make sure not to delete the original spc files!
<cedd>
yes!
<cedd>
smallest archival files ever
<aaabbb>
i bet if you gzip them they'll be even smaller lol
<cedd>
at 64kb its not worth it lol
<aaabbb>
true, but if you put them all in an archive, 64kb * 1000 can shrink from... oh 64mb isn't very big either lol
<cedd>
every time i scan from my laptop i have to load a 200mb program into ram
<cedd>
just to make a comparison with how bad things have gotten
<aaabbb>
firefox's image decoders take up more memory now for a single medium sized image than an entire operating system took in 2000
<cedd>
yea we cant go on like this
<aaabbb>
then a single web page loads friking 10mb of compact javascript and templates just to create a fade in-fade out menu
<aaabbb>
which could be done natively in css
<cedd>
lol yea
<galad>
eh, we are shipping a full web browser with almost each app nowadays
<aaabbb>
galad: electron can go die in a pit
<galad>
aaabbb: unfortunately reality has got other plans
<cedd>
i remember seeing it happen like 2 decades ago, where people were making the argument "we have so many cpu cycles now it doesnt matter if our code is efficient"
<aaabbb>
reality can go suck it. damn kids get off my lawn!
<cedd>
i didnt realize at the time it was a memetic virus that would spread out of control
<aaabbb>
there were days when each cycle was counted and every instruction was special. your spc files are like that!
<aaabbb>
then programs got big and complicated, but not to fear, optimizing compilers to the rescue!
<aaabbb>
now we have highly optimized and easy to write high level code, yet people choose to use gigantic monster frameworks running scripting languages on top of scripting languages
<cedd>
i know i want to go back to those days, way more fun imo
<cedd>
there was real craftsmanship then
<cedd>
theres still good stuff in the linux world though thankfully
<aaabbb>
it's still doable today. it's not like everyone has to go back to hand-optimized assembly
<aaabbb>
people just have to quit using jquery for literally everything and to write code in the lowest level language that you can comfortably write it in
<cedd>
it just feels like people are shipping prototypes now, the high level stuff is good as a start but when you ship to consumers there should be more concern for quality and efficiency
<cedd>
some of the music on these spc's btw are masterpieces
<cedd>
so much can be done with very little
<cedd>
but not many are trying these days
<aaabbb>
chiptunes is amazing
<aaabbb>
doesn't matter if it's spc or nsf or some fm card
<aaabbb>
when all you have is a few triangle waves, a few square waves, and an lfsr for noise, you can do amazing things
<cedd>
the creativity of some of these composers is amazing
<cedd>
4 channels including drums and your ears arent wanting for anything
rv1sr has joined #ffmpeg
<aaabbb>
never has nearest neighbor interpolation sounded so good lol
noobaroo has quit [Quit: Konversation terminated!]
hbbs has joined #ffmpeg
hbbs has quit [Ping timeout: 252 seconds]
hbbs has joined #ffmpeg
ivanich has joined #ffmpeg
Tinos has joined #ffmpeg
MisterMinister has quit [Ping timeout: 256 seconds]
l4yer has joined #ffmpeg
dv_ has joined #ffmpeg
<dv_>
bitrate control theory question. let's say there is a live camera feed where the contents are pretty much unchanging for extended periods of time. then, suddenly, the camera moves, and thus, the entire frame suddenly changes a lot. or, something on screen moves, and that movement covers the entire screen. I suppose that in such a case, it is really hard to avoid noticeable bitrate spikes, because
<dv_>
these changes are so sudden. is this correct?
noobaroo has joined #ffmpeg
<dv_>
I mean, the P frame content is almost totally made of zeros most of the time, and then, suddenly, bam, lots of motion
Raz- has joined #ffmpeg
<dv_>
without good motion estimation I can imagine that all of a sudden P frames are much bigger
cedd has quit [Ping timeout: 240 seconds]
xx has joined #ffmpeg
<JEEB>
dv_: more new content. so more bits, and you either have a really small buffer and make the quality really bad, or you have proper buffer and things are OK
xx has quit [Remote host closed the connection]
xx has joined #ffmpeg
<aaabbb>
dv_: before that, most macroblocks will be "skip" macroblocks which only takes a few bytes (or at most, macroblocks with very little residual). so yes, there will be a spike
blaze has quit [Quit: WeeChat 3.5]
<aaabbb>
it also depends on your rate control algorithm. it will be different for cbr vs abr vs cqp vs crf
<aaabbb>
and also depends on if you have a vbv buffer in use (that can constrain it to not have a spike get TOO big)
blaze has joined #ffmpeg
<aaabbb>
if the motion that covers the entire frame is brief then having a large number of references will be beneficial because the first frame after the movement can just reference the last frame before the movement (avc has a max of 16 reference frames)
noobaroo has quit [Quit: Konversation terminated!]
noobaroo has joined #ffmpeg
Sciencentistguy has quit [Quit: o/]
Sciencentistguy has joined #ffmpeg
<JEEB>
for a live feed you should always have HRD/VBV so I did not even consider the case where that is not configured :P
<JEEB>
also please do not utilize the word "CBR"
<JEEB>
it means so many different things to different people and systems
<aaabbb>
isn't that the term in libx264 (i assume is what he uses)?
<JEEB>
no
<aaabbb>
i know it is for libx265 at lesat
<aaabbb>
least*
<aaabbb>
then "abr with super duper tight constraints" is better or?
<JEEB>
CBR (whatever the definition of it is picked by the speaker) is defined as (average) bit rate being defined as rate control mode
<JEEB>
there is no separate CBR rate control mode
<JEEB>
I think the only option where "cbr" is mentioned is nal-hrd modes
<JEEB>
basically what I'm saying is that CBR is such an overloaded term it should not be utilized
<JEEB>
for some people it is that each slice takes exactly X amount of bytes
<aaabbb>
in x265 it's also mentioned with some "hard-cbr" mode that isn't actually hard cbr, it's just 5% tolerance for bitrate variance. anyway thanks is hould get out of that habit to call it cbr
<JEEB>
then you have people who are ok with bitrate+maxrate+bufsize which is ~1 second of buffer over an average (if you are always hitting that bit rate)
<dv_>
JEEB: yeah, HRD is active, but this is a hardware encoder that seems to do a very basic rate control algorithm
<JEEB>
s/+/=/
<dv_>
so I'm trying if I can tune something, perhaps more aggressive motion estimation or such
<JEEB>
aaabbb: btw sorry if I feel harsh but I already get depressed enough when people here ask "how to do CBR"
lavaball has joined #ffmpeg
<aaabbb>
JEEB: no worries i know it's not truly cbr but i'm used to audio where mp3 can be called cbr even with its bit resivor
<aaabbb>
so i know it's a bad habit to call it cbr :)
Tinos has quit [Remote host closed the connection]
Tinos has joined #ffmpeg
<JEEB>
aaabbb:
<JEEB>
asdf
<JEEB>
aaabbb: it wouldn't be a problem if CBR was actually defined
<JEEB>
as in, "a bufsize of X seconds can be considered CBR"
<JEEB>
it never is, an all those gosh darn streaming services that people then utilize just tell them to "do CBR", and then they ask here how to "do CBR"
<JEEB>
it's not an end user problem of course, it's the problem that this CBR thing is on the other hand somehow expected to be defined, yet it is not explicitly defined anywhere
<aaabbb>
it would be a fun learning experience to show someone "true" cbr where each frame takes the exact same number of bits, so they can see all the nasty problems of things like reverse keyframe pumping
JanC has quit [Ping timeout: 256 seconds]
MightyBOB has quit [Remote host closed the connection]
j45 has quit [Ping timeout: 260 seconds]
j45 has joined #ffmpeg
j45 has quit [Ping timeout: 268 seconds]
j45 has joined #ffmpeg
pvn has left #ffmpeg [#ffmpeg]
HarshK23 has joined #ffmpeg
j45 has quit [Ping timeout: 260 seconds]
EmleyMoor has quit [Ping timeout: 256 seconds]
j45 has joined #ffmpeg
ttys000 has quit [Read error: Connection reset by peer]
ttys000 has joined #ffmpeg
theobjectivedad has quit [Remote host closed the connection]
theobjectivedad has joined #ffmpeg
lavaball has quit [Remote host closed the connection]
noobaroo has quit [Quit: Konversation terminated!]
noobaroo has joined #ffmpeg
noobaroo has quit [Ping timeout: 260 seconds]
noobaroo has joined #ffmpeg
rv1sr has quit []
vampirefrog has joined #ffmpeg
alexherbo2 has joined #ffmpeg
Blacker47 has joined #ffmpeg
ppw has left #ffmpeg [#ffmpeg]
JanC has joined #ffmpeg
<bcn_>
I want to get a 360 camerathat I could set in the middle of a conf room table, and capture (basicaly) a wide wide panorama of the round table. Can anyone recommend a 360 cam that comes with software compatible with linux?
<bcn_>
I'd like to "render" the video as three stacked 120-degree videos, or two 180 degree videos so it "fits" in a normal 16x9 window
noobaroo has quit [Quit: Konversation terminated!]
noobaroo has joined #ffmpeg
Livio has joined #ffmpeg
e^pi-1 has quit [Ping timeout: 256 seconds]
kus has joined #ffmpeg
lavaball has joined #ffmpeg
vlm has joined #ffmpeg
Livio has quit [Ping timeout: 260 seconds]
j45 has quit [Ping timeout: 268 seconds]
j45 has joined #ffmpeg
ivanich has quit [Ping timeout: 240 seconds]
Tinos has quit [Remote host closed the connection]
AlexandrKirilov has quit [Ping timeout: 268 seconds]
AlexandrKirilov has joined #ffmpeg
minimal has joined #ffmpeg
JanC has quit [Ping timeout: 256 seconds]
noobaroo has quit [Quit: Konversation terminated!]
noobaroo has joined #ffmpeg
Marth64 has joined #ffmpeg
noobaroo has quit [Quit: Konversation terminated!]
ivanich has joined #ffmpeg
kus has quit [Ping timeout: 260 seconds]
cedd has joined #ffmpeg
noobaroo has joined #ffmpeg
namazso has quit [Remote host closed the connection]
junaid_ has joined #ffmpeg
namazso has joined #ffmpeg
Traneptora has quit [Quit: Quit]
rv1sr has joined #ffmpeg
lucasta has joined #ffmpeg
Muimi has joined #ffmpeg
JanC has joined #ffmpeg
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
five61848033 has quit [Remote host closed the connection]
five61848033 has joined #ffmpeg
minimal has quit [Quit: Leaving]
lucasta has quit [Quit: Leaving]
lucasta has joined #ffmpeg
cedd has quit [Ping timeout: 272 seconds]
lucasta has quit [Remote host closed the connection]