4 ms·

I took a quick look and couldn't find anything obvious, but I'd be surprised if someone wasn't using machine learning to improve psychoacoustic models. A fundam

• jandrese 9y ago
  You could calculate the difference between the original and uncompressed and then weight the differences by sensitivity levels of the human ear at various frequencies... It wouldn't be able to measure "warmth" or "heart" or "anger" or anything like that. It would be able to tell you which model most closely matches the original at the frequencies you can hear. Once you have a metric, you can use it as the basis of a parameter space search. Not exactly machine learning, but maybe some preliminaries.
  • jacobush 9y ago
    What's also in vogue is to gather huge datasets, in a fashion only really available to Google and their ilk. I'd imagine a "click-on-all-vehicles"-affair, only with sound instead of images.