FFmpeg: AVDCT Struct Reference

Several optimized IDCTs need a permutated input (relative to the normal order of the reference IDCT). This permutation must be performed before the idct_put/add. Note, normally this can be merged with the zigzag/alternate scan
An example to avoid confusion:

(->decode coeffs -> zigzag reorder -> dequant -> reference IDCT -> ...)
(x -> reference DCT -> reference IDCT -> x)
(x -> reference DCT -> simple_mmx_perm = idct_permutation -> simple_idct_mmx -> x)
(-> decode coeffs -> zigzag reorder -> simple_mmx_perm -> dequant -> simple_idct_mmx -> ...)

Definition at line 48 of file avdct.h.

Referenced by filter().

void(* AVDCT::fdct)(int16_t *block)

Definition at line 50 of file avdct.h.

Referenced by filter().

int AVDCT::dct_algo

DCT algorithm.

must use AVOptions to set this field.

Definition at line 57 of file avdct.h.

Referenced by avcodec_dct_init().

int AVDCT::idct_algo

IDCT algorithm.

must use AVOptions to set this field.

Definition at line 63 of file avdct.h.

Referenced by avcodec_dct_init().

void(* AVDCT::get_pixels)(int16_t *block, const uint8_t *pixels, ptrdiff_t line_size)

Definition at line 65 of file avdct.h.

Referenced by filter().

int AVDCT::bits_per_sample

Definition at line 69 of file avdct.h.

Referenced by avcodec_dct_init().

The documentation for this struct was generated from the following file:

libavcodec/avdct.h

Generated on Wed Jun 10 2015 01:57:19 for FFmpeg by doxygen 1.8.6