[music-dsp] Floating point dynamic range
S.W.Harris at ecs.soton.ac.uk
Fri Oct 3 01:14:00 EDT 2003
On Thu, Oct 02, 2003 at 06:42:15 +0000, David Ross wrote:
> what about a fixed point 32 bit representation of a signed integer (just
> for comparison)...because a float gives more dynamic range, right?
Depends on the "dot" position. fixed point numbers are expressed as, eg
4.24, ie. 4 integer bits and 24 fractional bits.
I dont know wether they are generally 1's or 2's complement.
More information about the music-dsp