I'm confused about the use of the biquad functions from the RTL for the ADSP-21479.
In the RTL Manual, on page 2-45, it says the biquad function takes an input array, output array, filter coefficient array, state array, samples, and sections. Looking at biquad_vec.asm, though, the word "sections" only appears in the prototype at the top of the file, but it looks like it's been replaced everywhere else by "taps." Is this a typo? I thought taps should be the number of delay elements, not the number of sections.
From the example code in the RTL manual, it looks like you would use the scalar code if you planned to pass in one piece of data per function call, and you would use the vector code if you want to perform batch processing. Is that correct, or is the vector form useful if you have 2 or more data streams that you're processing in parallel?
Is sample code available showing the use of either (or both) the scalar and vector code? The documentation's examples are not sufficient for me to understand how a real implementation would use them.