Hello, I have two questions for using fir accelerator:

1. I downloaded a example code from FIR Accelerator for 21489 ..... and there are some piece of code like this,

Float_Out_L = **circptr**(Float_Out_L,1,input_left, TAPS + AUDIO_BLOCK_SIZE - 1);

Float_Out_R = **circptr**(Float_Out_R,1,input_right,TAPS + AUDIO_BLOCK_SIZE - 1);

Why I need to use circptr function, and I looked the datasheet which instroduced fir input data stored in the memory is:

x[n-(N-1)], x[n-(N-2)]

....

x[n-1], x[n], x[n+1]

....

x[n+W-1]

As this, the WINDOWS_SIZE sample could be always put in the end of input buffer(x[n], x[n+1]....x[n+W-1]).. However, I have ever try it, the output is a mess. But I use circptr it works.

2. the second question is about the FIR delay time, I try use library function fir and measure the delay time is (N)/2 * sample interval .. It's right . But I use FIR accelerator ,the delay time is unstable, is different within my every measure. I feel it is unthinkable.

Sorry for the delay in getting back. Let me look into this and get back to you.