For microphone voice-comms you will need to use some form of DLL, check in the userlibs forum section because I'm pretty much certain there has been a fair amount on this in the past.
When it comes to analysing sound, there's a HUGE difference between Siri and a music analyser.
Music analysers at their heart essentially just analyse the use and repetition of various frequencies of sound. These values can be read from the particular music file, provided you know how to ascertain the various information from the file Headers or (in the case of MP3's) 'Frames'. Things like Sampling frequency, Bitrate, number of channels, and byte-size are all important factors in being able to correctly read the sound data, and thereby retrieve the frequency of the soundwave at a given point.
However, voice-recognition requires a completely different outlook, since this concerns overall "groups of sounds, "phonemes" rather than the finer specifics of individual frequencies - to discern these phonemes would require:
a) "Training" to be particular to a voice/dialect
b) A database of sorts to correlate the phonemes with words/phrases
c) Some further logical coding in order to process the results
Both features are definitely not anything simple or straightforward, and would involve a lot of work indeed.
|
