I found a new paper about this FFT based Reed-Solomon Codes. They present erasure-and-error decoding of an (n, k) RS code. The decoding complexity is with only O(n log n + (n − k) log2 (n − k)). As reference, complexity of current erasure decoding is O(n log n).
"On fast Fourier transform-based decoding of Reed-Solomon codes"
by Yunghsiang S. Han, Chao Chen, Sian-Jheng Lin, Baoming Bai
http://ct.ee.ntust.edu.tw/IJAHUC2021.pdf
I post the link here as a notice. If you know it already, just ignore this issue. I cannot understand the theory in the paper. If you implement the error correction also in your leopard-RS library, it may be useful rarely. While erasure correction is enough mostly, possibility of error correction is welcome.
I found a new paper about this FFT based Reed-Solomon Codes. They present erasure-and-error decoding of an (n, k) RS code. The decoding complexity is with only O(n log n + (n − k) log2 (n − k)). As reference, complexity of current erasure decoding is O(n log n).
"On fast Fourier transform-based decoding of Reed-Solomon codes"
by Yunghsiang S. Han, Chao Chen, Sian-Jheng Lin, Baoming Bai
http://ct.ee.ntust.edu.tw/IJAHUC2021.pdf
I post the link here as a notice. If you know it already, just ignore this issue. I cannot understand the theory in the paper. If you implement the error correction also in your leopard-RS library, it may be useful rarely. While erasure correction is enough mostly, possibility of error correction is welcome.