Blackfin Audio Line In-Out

[noahgrant]

Hello, I am having problems manipulating an audio signal, even using the Line In-Out example that comes with LabView Help.  When I go straight through the audio signal with no processing, it works, but when I try to manipulate a signal--as soon as I just connect a wire to the Separate Audio Channels VI output, I get a white noise looping sound, and nothing works.  Can anyone give me any advice?  I've got the same thing running on my actual program, so if I can get the Line In-Out example to work, I can get my code to work.  Thank you so much!!

Noah

[LaRisa_M]

Hi Noah,

Which Line In-Out example are you using (i.e. which Blackfin board do you have)? Can you post a screen shot of your block diagram with the signal manipulation modifications you've made? When you say "nothing works" do you mean the program doesn't run, or does it run but not give the results you're expecting? What are you're expected results?

[noahgrant]

Hi LaRisa!

Here is an updated version of our code.  The goal of the code is to take a note from an electric guitar that is out of tune, correct the pitch (adjust every frequency in the FFT accordingly) and send it back out.  The code has four modes: (0) bypass, where what goes in comes right back out, (1) a mode where the open E string is played, and some data is stored, including the fundamental frequency (which may be out of tune), (2) the same as mode 1 for the A string, (3) the 'live' mode, where a note on the E or A string is played, the code recognizes the string and adjust the frequency according the fundamental frequency recorded in either modes 1 or 2.  the blackfin is connected to a pedal with a switch that governs the modes.  our biggest problem right now is this:

1) in mode 3, we seem to be getting the code to recognize the string (as tracked by the LEDs on I/O pins), but are getting no data out of the inverse FFT after the frequency transformation takes place.  is the an aliasing effect because we are entering more than 22050 data points in the new array (we are at a resolution of 1 Hz)?  We are really clueless as to what is going on here.

2) we are not able to see the graph of the FFT for the E string and A string input notes of modes 1 and 2.  How can we fix this?

Number one is a big deal, and we spent 10 hours today working on it.  If you can get back to us ASAP we would SO SO SO SO SO appreciate it!  Thanks so much!!

Noah

[LaRisa_M]

Hi Noah,

I'm sorry, but I've got to admit that I don't know a lot about signal manipulation, FFT's, etc. I work more closely with NI's embedded products, like our Blackfin kits.

I put together the attached VI just to observe how the signal manipulation VI's you're using work. I noticed a couple of things that I hope are helpful. First, I think that the FFT (Peak) measurement that you're doing is ignoring any negative frequencies, and so it may be cutting out half of your points. Take a look at the difference between Method 1 (the way you're doing it now) and Method 2 in my test VI. I recommend that you try the 2nd method.

Message Edited by LaRisa_S on 05-01-2009 06:26 PM