libavcodec: Implementation of 32 bit fixed point FFT

Multimedia / FFmpeg - Nedeljko Babic [mips.com] - 4 August 2013 07:01 UTC

Iterative implementation of 32 bit fixed point split-radix FFT. Max FFT that can be calculated currently is 2^12.

18d7074 libavcodec: Implementation of 32 bit fixed point FFT
doc/mips.txt | 5 +
libavcodec/Makefile | 4 +-
libavcodec/fft-fixed32-test.c | 21 ++++
libavcodec/fft-internal.h | 21 +++-
libavcodec/fft-test.c | 4 +
libavcodec/fft.c | 181 +++++++++++++++++++++++++++++++++-
libavcodec/fft.h | 19 +++-
libavcodec/fft_fixed.c | 1 +
libavcodec/fft_fixed_32.c | 52 ++++++++++
libavcodec/fft_float.c | 1 +
libavcodec/fft_init_table.c | 200 ++++++++++++++++++++++++++++++++++++++
libavcodec/fft_table.h | 66 +++++++++++++
libavcodec/mdct.c | 4 +
libavcodec/mdct_fixed.c | 1 +
libavcodec/mdct_fixed_32.c | 52 ++++++++++
libavcodec/mdct_float.c | 1 +
libavcodec/mips/Makefile | 1 -
libavcodec/mips/fft_init_table.c | 67 -------------
libavcodec/mips/fft_mips.c | 2 +-
libavcodec/mips/fft_table.h | 63 ------------
tests/fate/fft.mak | 21 +++-
21 files changed, 647 insertions(+), 140 deletions(-)

  • Share