Change Sample Rate

Chris Barrett wrote:

> How do I use the sinc function to change the sample rate of a signal?

Julius O. Smith has a description of one algorithm:

   http://ccrma.stanford.edu/~jos/resample/

> I've been told that to change the sample rate of a signal I'm suppose to
> convolve it with the sinc function.

Sample rate conversion requires filtering. One way to get an approriate
FIR filter is to apply a window to a scaled sinc function.

Erik
-- 
-----------------------------------------------------------------
Erik de Castro Lopo
-----------------------------------------------------------------
"Mutable state is actually another form of manual memory management: every
time you over-write a value you are making a decision that the old value is
now garbage, regardless of what other part of the program might have been
using it." -- Paul Johnson

Erik de Castro Lopo wrote:

> Julius O. Smith has a description of one algorithm:
> 
>    http://ccrma.stanford.edu/~jos/resample/

I should also mention that there is working source code here:

    http://www.mega-nerd.com/SRC/

The code is licensed under the GNU General Public License or
alternatively a very affordable commercial use license is also
available.

Erik
-- 
-----------------------------------------------------------------
Erik de Castro Lopo
-----------------------------------------------------------------
"Patents are bad for the software industry.."
-- Bill Gates in 1991

"Ron N." <rhnlogic@yahoo.com> wrote in news:1184191178.484927.283240
@j4g2000prf.googlegroups.com:

>   The full ft of
> your rect chopped interp result will actually have
> more high frequency alias content then the "ringy"
> looking sincresample result.

???

>> fftsinc=fft(outsinc);
>> fftinterp=fft(outinterp);
>> plot(20*log10(abs(fftsinc)));
>> hold
Current plot held
>> plot(20*log10(abs(fftinterp)),'r');

The interp result looks a solid 40 dB better at high frequency for the rect 
window.  When I apply a 64 point hamming to the original signal, much of 
the "ringing" goes obviously goes away, but I still get maybe 30 dB better 
high freq performance from interp.

>> y=sin(2*pi*3*x).*window(@hamming,64)';
>> outsinc=sincresample(y',4*length(y));
>> outinterp=interp(y,4);
>> fftinterphamming=fft(outinterp);
>> fftsinchamming=fft(outsinc);
>> plot(20*log10(abs(fftinterphamming)),'r');
>> hold
Current plot held
>> plot(20*log10(abs(fftsinchamming)));

Dead on, though, about the bandlimited gauss windowed square wave.  




-- 
Scott
Reverse name to reply

As Jerry's post mentioned, any sample rate change is associated with
filtering.
Typically,
- for upsampling, we insert zeros and filter (for removing the
images).
- for downsampling, we filter (for removing aliases) and skip
samples.
The type of filter used depends on the rejection requirements on the
'undesired'
spectrum.
Maybe the post on multirate signal processing @ dspguru will be
helpful
http://www.dspguru.com/info/faqs/mrfaq.htm
Also, you can find some small code snippets @
http://dsplog.blogspot.com/2007/05/polyphase-filters-for-interpolation.html

Krishna
http://dsplog.blogspot.com