Howdy Ross, > Hmmm, I guess I wasn't clear that I was asking in terms of my > original stated > project. I'm not about to believe that the effects of MP3 > and Ogg Vorbis on > human speech sounds have been studied since the 20's ::-)
I'm not sure why not? While MP3 and Vorbis are products of the 90's, they _work_ (or don't) because of theory developed in the 20's. > And the goal isn't to examine the theory, but the actuality. Which is all fine and good. But how do you propose to understand or even usefully examine the actuality without any knowledge of the theory? The subjectivity of your analysis is going to require that you examine the theory, unless you just want to be another one of the guys with colorful frequency transform plots that look nice but tell you just about nothing about how things sound. > If I use oggenc > on a speech sample today, does it loose any critical (from an > acoustic phonetic > perspective) information? How are you going to judge what is critical and what isn't without using psychoacoustics? > While I'm sure the general theory > of lossy audio > compression is very interesting, I have neither time, knowledge, nor > inclination to approach the question this generally. I suspect you may want to approach a different question. And no, I'm not really trying to deter you from doing this sort of study - I just want you to be aware of what you're getting into if you want to do the study in a useful way. In fact, I'd love to know why MP3 (using the ATT psych model) does a particularly poor job of handling fricative sounds in speech, since I'm trying to patch that right now. But I don't know how you're going to come to any reasonable conclusions without a pretty heavy dose of psychoacoustic and audio signal processing theory. And I say that as someone who has done speech & speaker recognition, MPEG audio, and audio quality analysis work for the last eight or so years in a variety of capacities. > My choice of 'lossy audio compression' was an atempt at > inclusive language, > not generality ::-) If only it were so simple. 'Lossy audio compression' (well, and speech recognition) is basically the pinnacle of all audio related work in the past century. It's way, waaay complicated. Or, as Billy Crystal said, "Have fun stormin' the castle!" Hope that helps, Alex _______________________________________________ mp3encoder mailing list [EMAIL PROTECTED] http://minnie.tuhs.org/mailman/listinfo/mp3encoder
