[Dspforum] IIR audio output question

[Wenstrand, Doug S.]

Same thing as with the FIR filter -- use the pragma must_iterate, data_align, restricts, or intrinsics where appropriate.  The DSPlib has float to q15, and q15 to float, which do this as well..

Doug

________________________________________
From: dspforum-bounces at echelonembedded.com [dspforum-bounces at echelonembedded.com] On Behalf Of Alan Ding [alan33d at gmail.com]
Sent: Monday, November 09, 2009 9:56 PM
To: dspforum at echelonembedded.com
Subject: Re: [Dspforum] IIR audio output question

1)      Conversion between floating point and integer formats shall be done using DSPlib functions, or optimized C routines

I'm guessing we are suppose to use the DSPF functions....  How do you optimize C routines to do casting between types?

~Alan

On Sun, Nov 8, 2009 at 4:43 PM, O'Connor, Sean R. <Sean.OConnor at jhuapl.edu<mailto:Sean.OConnor at jhuapl.edu>> wrote:

Even with no audio input, I hear a faint eerie resonance on both channels.  The oscillations are of the correct frequency, namely ~697Hz on the left channel and ~770Hz on the right channel.  The design seems to work when I input test tones, but the resonance is always present (it is even quieter with only one biquad).  It seems like there is some instability in the filter, although the output does not 'run away' or saturate the output.  Any thoughts on this?

Also, is the float to integer conversion supposed to be done with DSPF_fltoq15()?  Currently I am just casting the float filter output to an integer, this is the approach taken in the book.

Thanks,
Sean
_______________________________________________
Dspforum mailing list
Dspforum at echelonembedded.com<mailto:Dspforum at echelonembedded.com>
http://echelonembedded.com/mailman/listinfo/dspforum_echelonembedded.com



--
-------------------------------------------------------
Alan Ding
Master of Sci, Electrical Engineering
Johns Hopkins University

Email: alan33d at gmail.com<mailto:alan33d at gmail.com>
Contact #: 4844593290