lavfi/nlmeans: improve the performance

Remove the pdiff_lut_scale in nlmeans and increase weight_lut table size
from 2^9 to 500000, this change will avoid using pdiff_lut_scale in
nlmeans_slice() for weight_lut table search, improving the performance
by about 12%. (in 1080P size picture case).

Use the profiling command like:

perf stat -a -d -r 5 ./ffmpeg -i input -an -vf nlmeans=s=30 -vframes 10 \
-f null /dev/null

without this change:
when s=1.0(default value) 63s
     s=30.0               72s

after this change:
     s=1.0(default value) 56s
     s=30.0               63s

Reviewed-by: Carl Eugen Hoyos <ceffmpeg@gmail.com>
Signed-off-by: Jun Zhao <mypopydev@gmail.com>
Signed-off-by: Clément Bœsch <u@pkh.me>
This commit is contained in:
Jun Zhao 2019-02-01 10:45:24 +08:00 committed by Clément Bœsch
parent ed20fbcd48
commit 6c586303a1
1 changed files with 9 additions and 8 deletions

View File

@ -43,8 +43,12 @@ struct weighted_avg {
float sum;
};
#define WEIGHT_LUT_NBITS 9
#define WEIGHT_LUT_SIZE (1<<WEIGHT_LUT_NBITS)
/*
* Note: WEIGHT_LUT_SIZE must be larger than max_meaningful_diff
* (log(255)*max(h)^2, which is approximately 500000 with the current
* maximum sigma of 30).
*/
#define WEIGHT_LUT_SIZE 500000
typedef struct NLMeansContext {
const AVClass *class;
@ -63,7 +67,6 @@ typedef struct NLMeansContext {
struct weighted_avg *wa; // weighted average of every pixel
ptrdiff_t wa_linesize; // linesize for wa in struct size unit
float weight_lut[WEIGHT_LUT_SIZE]; // lookup table mapping (scaled) patch differences to their associated weights
float pdiff_lut_scale; // scale factor for patch differences before looking into the LUT
uint32_t max_meaningful_diff; // maximum difference considered (if the patch difference is too high we ignore the pixel)
NLMeansDSPContext dsp;
} NLMeansContext;
@ -401,8 +404,7 @@ static int nlmeans_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs
const uint32_t patch_diff_sq = e - d - b + a;
if (patch_diff_sq < s->max_meaningful_diff) {
const unsigned weight_lut_idx = patch_diff_sq * s->pdiff_lut_scale;
const float weight = s->weight_lut[weight_lut_idx]; // exp(-patch_diff_sq * s->pdiff_scale)
const float weight = s->weight_lut[patch_diff_sq]; // exp(-patch_diff_sq * s->pdiff_scale)
wa[x].total_weight += weight;
wa[x].sum += weight * src[x];
}
@ -527,10 +529,9 @@ static av_cold int init(AVFilterContext *ctx)
s->pdiff_scale = 1. / (h * h);
s->max_meaningful_diff = -log(1/255.) / s->pdiff_scale;
s->pdiff_lut_scale = 1./s->max_meaningful_diff * WEIGHT_LUT_SIZE;
av_assert0((s->max_meaningful_diff - 1) * s->pdiff_lut_scale < FF_ARRAY_ELEMS(s->weight_lut));
av_assert0((s->max_meaningful_diff - 1) < FF_ARRAY_ELEMS(s->weight_lut));
for (i = 0; i < WEIGHT_LUT_SIZE; i++)
s->weight_lut[i] = exp(-i / s->pdiff_lut_scale * s->pdiff_scale);
s->weight_lut[i] = exp(-i * s->pdiff_scale);
CHECK_ODD_FIELD(research_size, "Luma research window");
CHECK_ODD_FIELD(patch_size, "Luma patch");