hello all....
especially sebi ......
i follow your posts ....keep it up . you are doin a great job ..

coming to my question ...
what is normalization in audio ...
i mean normalization of floating point values in audio .....
when you have some samples divide it with certain constant ....this is one form of normlization right ?
if any one is familiar with ARM assembly can you tell me how its performed in ARM assembly.

thanks .......

Basic normalization is amplifying all the audio linearly to bring the highest peaks to some set value, that is multiply the value of each sample by some constant factor. It is generally expressed in either % of dB: one can normalize to 100% or 99% or 95% or whatever, or alternately to 0dB or 99.5dB, etc. The operation and results are the same whether the file is fixed point or floating point.

RMS normalization is a bit more complicated and something I've never played with. It is done across a number of files/tracks to equalize their average RMS level rather than the peak values. I'm pretty sure it involves calculating the RMS average of each, then amplifying by either a positive or negative factor as needed to make each file's value the same.