Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Is there a way to filter quantization noise and get better audio? (Read 10701 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Re: Is there a way to filter quantization noise and get better audio?

Reply #25 –
Quote from: mzso on
I just remembered this old thread.
I wonder... Is there something like an "AI upscale" sort of thing nowadays that can improve the sound of low-fi audio talked about here?
(I would expect AI to work well on something like this with acoustic music, since there aren't an infinite variety of instruments. The question is whether anyone worked on something like this.)

Another thing which would be interesting in this topic is AI assisted audio compression. Instead of coding actual high bandwidth audio data, represent it as parametric data as much as possible, and let the AI "dream" the audio back.
Voice codecs seems to already benefit from this approach, eg. Google's WaveNET, Lyra codecs.

Re: Is there a way to filter quantization noise and get better audio?

Reply #26 –
Quote from: darkbyte on
Quote from: mzso on
I just remembered this old thread.
I wonder... Is there something like an "AI upscale" sort of thing nowadays that can improve the sound of low-fi audio talked about here?
(I would expect AI to work well on something like this with acoustic music, since there aren't an infinite variety of instruments. The question is whether anyone worked on something like this.)

Another thing which would be interesting in this topic is AI assisted audio compression. Instead of coding actual high bandwidth audio data, represent it as parametric data as much as possible, and let the AI "dream" the audio back.
Voice codecs seems to already benefit from this approach, eg. Google's WaveNET, Lyra codecs.
Is bandwidth still that costly to be worth dealing with such? For 100-150 kbps you can get near perfect stereo audio.