listen to audio samples from different mobile devices and usb microphones
First, for dharma talks, use mono instead of stereo. Otherwise your
file size doubles. Also use the 16 bit quantization instead of the
default 32 bits.
Telephone speech quality is set to 8K, which is based on the
assumption that most human speech are understandable within 0-4000 Hz
frequency range, and for male voice it works Ok. But if you have
music, etc., then I'll experiment with a higher sampling rate. You can
do the math your self: file size = sampling rate * length in seconds *
bits per sample.
You may find that the quality is not satisfying with a lower sampling
rate. To improve quality, the best strategy is to find a good
microphone, perhaps a headset that prevents amplitude variation due to
head movement. Logitech sells one for $40. Having a quiet room also
helps. Some audio recorder also has built-in digital signal processing
that cancels noise based on their microphone receptive properties.
I didn't see your question about mp3. MP3 compression is based on a
quantization scheme that devotes more bits to the lower frequency
range. Mp3 file size is determined by bits per second * length in
seconds. I'll try either 64, 96 or 128kbps. But your microphone and
the quality of the raw audio (e.g. WAV) are going to be the deciding
factors for the quality of your mp3.