Hello Bob,

For uniform quantization, the S/N is 6n+4.77+(Smean/Speak)

Smean/Speak is the mean to peak ratio for the signal. For a sine wave it is
just -3 dB; however for speech it will be much higher. It will be more
like -13dB. And of course music will probably be closer to the speech value
than the sine wave value.

Look at Rabiner & Schafer "Digital Processing of Speech Signals" for
details. "Speech Processing" edited by Chris Rowden covers this as well.

Clay




"BobGardner" <bobgardner@aol.com> wrote in message
news:20031121215550.28390.00000721@mb-m14.aol.com...
> >
> >If I am given a 16-bit fixed-point number of uniform quantization, I
> >can surely say that the maximum theoritic dynamic range I can get is
> >96.329 dB
>
> Minor quibble... 96 dB is 6 dB per bit for 16 bits...the max signal is
really
> +-32767 or +-15 bits, right? so its 90dB....