LAME 3.100

>The current release version of LAME is 3.100
lame.sourceforge.net/download.php

Where were you when MP3 has just gotten better? LAME 3.100 has been released. How can logg, oupss and ahahacc even compete anymore?

Other urls found in this thread:

svn.code.sf.net/p/lame/svn/trunk/lame/doc/html/history.html
twitter.com/SFWRedditVideos

lame thread namefag

no updates that affect quality. opus 4 lyfe doe.

>can't encode 44.1khz
>not universally supported like mp3
>resource hog

nah, i'll stick with what works

lol, it's shit

Who the fuck cares?

daily reminder that mp3 is now a free format

I really hate that we are still using lossy quality compression algorithms for sound and images. Sure, it maybe works for the great masses who are just gonna listen to it, but as soon as you want to do something with it, like change frequency, filter or increase contrast, it shows.

It's soon 2018, why the fuck don't we just send over the pixels/bitsamples? We could easily do it for at least images and sound. I'm allowing people to cheat with video until we develop faster bandwitdh connections.

Anything but lossy images for photography is difficult. I believe all camera manufacturers have their own proprietary RAW formats. We need a standard for RAW photographs if possible

>2017
>mp3

Why cant they just send the pixels over? Three matrices for each color. Sometimes I think people make things more complicated than they should be.

Sure, add XML information if you really, really want metadata.

>mp3 is now a free format
Ikr

why didnt you post the changelog?

For MP3 encoding LAME rules above all, there's no denying that. But nowadays there are better lossy encoders for other formats, the best one (per bitrate vs quality) is Opus and I moved to using that years ago.

LAME, you were always awesome, but your time has passed.

Opus, welcome to the party, pal.

opus sounds weird even at the highest bitrate

TITS OR GTFO.

>we are still using lossy quality compression algorithms for sound and images
You realize there's a limit to how far you can compress data right? How are you going to store it on portable media, or stream it with no buffering?

More than likely your ears or headphones are shit.

>now you'll respond with some bullshit about how you have brand headphones and listen to bitrate lossless files
>yadda yadda yadda
>get the fuck over it, kid

>can't encode 44.1khz
Explain this shiet. This cannot be true.

i use speakers and listen to cds and lossless files. ogg vorbis at high bitrates is good enough but opus sounds too weird

I swear I read that opus automatically discards anything over 20khz, yet it can only use 48khz minimum, not 44.1

I know it's lossy, but that sounds stupid if I'm right.

Cause she's stupid cunt interested only in numbers and not changes. Also, it's week old news and it's mostly bug fixes. svn.code.sf.net/p/lame/svn/trunk/lame/doc/html/history.html

Weird.
Thanks.

Stop being a streamfag then. MUH CLOUDD

>she
didnt even notice the name because i read the post contents first. i should always read the name first so i can hide namefag threads..

Yep. From the OpusFAQ on XiphWiki

>However, files at these rates are internally converted to 48 kHz and then only frequencies up to 20 kHz are encoded.

If the point of a lossy codec is to sound okay and be as small as possible, and they're brickwalling/cutting off any sound of 20khz why not use 44.1? I think you can only use 44.1khz with some legacy/compatability plugin too.

I like to listen to music on my phone when I go
>outside
Also, it's not a cloud if it's your own home server.

>only frequencies up to 20 kHz are encoded
Isn't this above the limit of human hearing anyway? If I remember, MP3 was also cut at this frequency for its compression. I don't have golden ears but any OPUS file I have listened to sounds amazing.

it's okay, but not really great. vorbis is the shit.

Will my re-encoded mp3s to aac to opus suffer from rotational velocidensity effects, gee?

You mean just store things as raw bitmaps?

I'd like to see existing compression schemes used properly. My png encoder will generally reduce the size of a given file by at least 30%, within 30 seconds. More time might remove another 5-10%, depending on deflate block count.

Simple heuristics to select apodization method during flac encoding can greatly reduce file size as well. It's been over a decade and arithmetic coded jpegs still aren't widely supported, despite long since being in the public domain. LZMA (eg 7zip) has been around for years, and can reduce certain types of data by 99%. Instead you'll usually see a sloppy DEFLATE run on this data.

It's a pathetic mess, like the rest of the "tech" world. Existing standards are entrenched, and adoption of new techniques is poor to nonexistent. FLIF is coming along for images, not sure about audio. Video is often constrained by realtime / hardware decoding and encoding considerations, and need for non-sequential playback.

I'm working on a general purpose compression algorithm, but even if it's as effectivce as I think it can be, I don't expect much mainstream response.

That's true. absolutely. But why limit it to 48khz? Why not just keep it 44.1, which is probably what a lot of source content will be?

Didn't 48khz become the standard with DVD audio? I don't know about BluRay audio but I think 48khz is now just used as the optimal default and their compression is optimized around this frequency band.

A lot of DACs work at 48khz.

you mean opus?

>literally 2 codecs glued together
they even admit it on their site
because:
-uncompressed audio/video needs fuckloadloads of bandwidth like 200MB/s for 444 and around 100MB/s for 420 for 1080@30
-even if you were to compress it losslessly it would give you roughly 3.2-3.6:1 "compression ratio". it's still roughly 30MB/s
>but you may "it's fucking retarded why they just save the difference of consective frame"
-so you save the "raw" difference between frames a then compress it. so you "save" another 50%. now you at 15MB/s.
>"what the fuck it's still to much"
>"hmm some regions of frame change and others do not"
-split to blocks. flag them if they change or not. another 20% shaved. now 10MB/s
>"you know fuck this people won't notice small changes in color" so it's fucking pointless to encode them
-apply DCT/DST to every block then just remove higher frequencies, by diving the matrix after applying transform. -40-55%
>"holy fuck that worked pretty good. but.. some blocks look similar to each other, bt are in slightly different locations"
-bam, motion vectors -60%.
so now you're at 4-1.5MB/s, while still not using half tricks from your sleeve

i understand your frustration, but it's just not really feasible to send nearly 800GB through network just so few people can watch a movie in studio-quality

however BPG/HEIF looks pretty good desu

Not the guy you're responding to but:

Isn't this problem found everywhere else though? The whole "technically better" vs "first to market/widespread adoption/compatability/support" is everywhere.

idiots, that's why.

>CTRL+F
>only two mentions of vorbis

sad!
vorbis has native gapless playback and a huge array of quality options during encoding. a big step up from MP3.

Because 48kHz is the standard on DVDs and Blu-rays.

it's because the devs are idiots and couldn't figure out how to code 44100 hz support

Who the fuck doesn't use lossless audio these days.

can lame finally resample while decoding?

...

opus-custom supports 44.1, but it's not advisable to use it because it degrades quality.

it's dogshit. only great for low bitrates.
Fucking useful when we have 10mbps upload on every package around the world now with 0 packet loss. Very fucking useful indeed.

>lossy audio
top kek

>we have 10mbps upload on every package around the world now
ha, i wish

...it doesn't work this way. the 'pixels' would be simple bmp. or png if you want a bit of universal compression. but raw includes more

No, Vorbis. The Silk portion of Opus makes it great for speech/voice uses to the point I wouldn't use anything else, but the Celt part is not a Vorbis killer.

CDs are 44khz, and even most digital stores like iTunes sell music with lossy codecs at 44khz.

DVD-A should have replaced CD-A but didn't because replacing two decades worth of consumer hardware wasn't going to happen when the average listener wouldn't notice a difference.

From the FAQ
>Note that it's generally preferable for a decoder to output at 48kHz, even when you know the original input was 44.1kHz. This is not only because you can skip resampling, but also because many cheaper audio interfaces have poor quality output for 44.1kHz.