ASIC/FPGA Design and Verification Out Source Services
FFT system C project - 8 point FFT using DIT algorithm.
- The first stage of the FFT, requires addition and subtractions only (twiddle factors of 1 or -1.
- The second stage twiddle factors are 1, -1, j or -j. Again this stage, as the former one, does not have fractional twiddle factors.
- Last stage has twiddle factors of the combination of 1+j, -1+j, etc ...
- The fractional .070710678 is multiplied by 8 and rounded to 6, which also reduces the accuracy. In another example a multiply factor of 16 is used. This improves the accuracy as the number 0.070710678 is now represented by two digits of precision (11 instead of 6). FFT accuracy issues
- In the code therefor you either see a multiplication of 8 (twiddle factor of 1, -1, j or -j) or 6.
-
A complex multiplier (two-dimensional Cartesian coordinate is used at
both of its inputs and outputs)
using only three
scalar multipliers, to save chip area, is also available
on this site:
from:
- Return to the FFT main page. FFT main page .
Please let me know what you think on this FFT work.
Contact me now at: |