Opus 1.5 Audio Codec Able To Make Extensive Use Of Machine Learning

[Chewi]

D'oh. Assuming this is any good, it's slightly annoying when you already re-ripped most of your CD collection into Opus a few month back.

[jmvalin]

What about RISC-V vector extension support?

That would be nice. Patch welcome :-)

[Anux]

D'oh. Assuming this is any good, it's slightly annoying when you already re-ripped most of your CD collection into Opus a few month back.

Most of the gains are for voice and error correction, probably only single digit gains for your offline music collection. If your chasing every single improvement you will never be done.

[jmvalin]

D'oh. Assuming this is any good, it's slightly annoying when you already re-ripped most of your CD collection into Opus a few month back.

No need to re-rip your CD collection. Most of the improvements in Opus 1.5 are related to real-time communications. For offline music encoding, you're not going to see a difference.

[varikonniemi]

I thought this was an interesting development until I read this line and then some trepidation creeped in:



The processing part is fine with me, but having a ML-backed ability to generate sounds seems creepy, amost like some of the "AI trash" that some companies are trying to foist on us all ... and only getting government inquiries for their trouble and public mea culpas by their CEO.

Any sort of AI technology seems to be reaching the point where only fools rush in while the cautious remain just that.

AI is like saying: instead of using standard commands i will use SIMD. Look how much more performant these special algorithms are?!=

Legacy learning: 1TOPS
AI learning: 10TOPS

"AI" must be next stage in human evolution, see how it can do things 10 times faster! I wonder if it is conscious`?

[Quackdoc]

time too re-do the testing all over again

[skeevy420]

Look up https://www.theverge.com/2024/2/21/2...ate-historical if you want to know what I tried to say. Since I don't know what words exactly are the reason for the censorship, I had no other choice than to guess.

You seriously think I can't tell block and black from that context

To be pedantic, the question specifically said a 1943 German soldier and not a Nazzy. Technically speaking, there were Germain soldiers of color during WW2, mainly African, Arabic, and Indian.

Ironically, this was from 24 years ago:

Woke AI is about as spot on as South Park

[Anux]

Technically speaking, there were Germain soldiers of color during WW2, mainly African, Arabic, and Indian.

What?? You probably have a non AI picture to prove that? If you would know anything about german raze theory you would also not think that is possible. There were partisan groups in occupied countries like ukraine and probably africa too, but those were not germans, did not speak german and had different uniforms (but they were probably as bad as their german counterparts).

Edit: yes southpark as always was right on spot to predict social abnormalities. Sad how this was a funny fantasy and now it's happening.

[dragorth]

The female LACE sample is impressive at the lowest bitrate. The male not so much.

The big one is the 50% packet loss, that one is amazing and you think it can't get better, but 80% is where the surprise really comes in.

Great work guys!

[skeevy420]

What?? You probably have a non AI picture to prove that? If you would know anything about german raze theory you would also not think that is possible. There were partisan groups in occupied countries like ukraine and probably africa too, but those were not germans, did not speak german and had different uniforms (but they were probably as bad as their german counterparts).

There are videos about it...

"My enemy's enemy is my friend." Not everyone was happy under British colonial rule.