Cool Tech
Digital audio listening tests
A lot of people claim that their newfangled codec is the same/better at 64kbps than MP3 is as 128kbps. A new listening test shows that no single codec really lives up to that claim. The results from the listening test gives you all the gory details.
This is probably the most objective and realistic comparison between codecs I've seen in a while. These codecs use psychoacoustic models that mimick the human ear and remove portions of the audio spectrum that the human ear cannot perceive in order to compress the audio. For instance, if you have a really loud sound followed immediately by a quiet sound, the human ear will no be able to preceive that quiet sound. So why bother keeping it around if the human ear can't hear it?
This is the basis for modem audio codecs and thus it's imperative that humans listen to the output of the codecs in order to qualitatively judge the codec. A machine can't do it, and that is why I've always had issues with people referring to Rob Leslie's MPEG audio compliance page as an actual measure of the quality of the codec.
Posted by Mayhem at September 23, 2003 04:36 PM