On Thu, 14 Feb 2002, Oliver Wittchow wrote:
> data are uncompressed so it's simple to calculate the
> possible sound lenght:
> the usual format is 16 bit, 8 khz, mono which means that
> one second would take 16,000 bytes, plus some 100 bytes
> for the MFi header.
This would be quite amenable to compression, however. The first
and simplest thing to do would be to use eight-bit samples, since
a 96 dB dynamic range is hardly necessary for playback from a phone;
48 dB would be plenty. So store or down-load your eight-bit data,
and convert it to 16-bit on the phone by putting zero in the LSB
of the 16-bit sample.
After that, there are various other compression techniques that
you could use to get a further reduction in data size. If you're
willing to use some of the more advanced encoding techniques, the
reduction in size could be quite spectacular. (Making sampled voice
data as small as possible has been a topic of research at Bell Labs
for decades now, so there are some pretty good techniques out there.)
Regardless, you're definitely going to want to download the data
from the server; it would take a miracle to get more than a few
seconds of sound into 10 KB.
Curt Sampson <cjs_at_cynic.net> +81 90 7737 2974 http://www.netbsd.org
Don't you know, in this new Dark Age, we're all light. --XTC
Received on Mon Feb 4 06:15:23 2002