I'm wondering how I may obtain musical information such as amplitude from a audio file?
Suppose we have a raw audio file, what I want to extract data from the file which allows me to generate a curve such as http://www1.icsi.berkeley.edu/Speech/mr/nearfar.html. Once I have obtained this curve, I will perform FFT, etc.
I have been trying to search for solutions in Java Sound, but what I only know so far is that I can pour the data into byte[]
using AudioInputStream
.
But how do I translate that byte[]
into a double[]
which contains actual information about the sound? In this case, the values in the double[]
represent amplitude.
Could anyone please provide a solution to solve this problem in Java?
Typical wav file is 16 bit little endian, so you can take each two consecutive bytes, b1 and b2, and get the amplitude as follows:
if the input is stereo, first two bytes would be left channel, then two bytes for right channel.