Hello All,
i'm using the bf 518 with eval. board.
recently i used the fft (i used the library function - rfft_fr16) on an audio
input stream of different sizes. I then took the samples and the results of the
fft and compared it to the results i get on matlab.
I noticed a few odd things some related to the the Codec. and I'm wondering
if anyone has an explanation..
The Codec:
a) the sampling rate of the codec. I found a pdf SSM2602. and found out how to
control the SR - SAMPLING RATE, ADDRESS 0x08 (page 26).
i noticed that for each SR i enter, the actual SR i half (96KHz--> 48 KHz, for
example). for a 1KHz sine wave and a SR of 48K - i would expect to get 48
samples per cycle but instead i got 24 samples.
Why is it so? could the reason be that we sample both the left and right
channels?
b)when I sampled a zero value ( no audio input) i got a straight line (as
expected) but the line had a slite slope. as if the x-axis itself has some kind
of a slope. ideas?
c)when entering a non symetrical audio signal, I saw that the signal was
'mirrored down' (towards the X-axis) again compared to the matlab
results. for example:
____|_|*__ and ___________
| |*
if i added a minus before one of the sets of samples, i would ge the same
waveform. i hope i explaind myself.
any help would be appreciated.
Chagai