mån 2025-06-09 klockan 20:06 +1000 skrev Peter Ross:
> +static void convolve(float *tgt, const float *src, int len, int n)
> +{
> + for (; n >= 0; n--)
> + tgt[n] = ff_scalarproduct_float_c(src, src - n, len);
> +
> +}
This should probably use an FFT since this implementation is O(n²).
T
intended for use by RealAudio 2.0 (28.8k) and G.728 decoders.
---
libavcodec/g728_template.c | 65 ++
libavcodec/ra288.c | 50 ++---
2 files changed, 68 insertions(+), 47 deletions(-)
create mode 100644 libavcodec/g728_template.c
