Syswip
Advanced Member level 4
Hello,
I'm looking a way to calculate my fixed point FFT absolute error.
I don't use any scaling (increasing the word length after each stage) in my implementation. But there is still error due to a fixed point multiplication with
twiddle factor.
Is there a way to calculate absolute error depends on twiddle factor bit width?
Thanks in advance,
Tiksan
I'm looking a way to calculate my fixed point FFT absolute error.
I don't use any scaling (increasing the word length after each stage) in my implementation. But there is still error due to a fixed point multiplication with
twiddle factor.
Is there a way to calculate absolute error depends on twiddle factor bit width?
Thanks in advance,
Tiksan