generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Add parameter search to global motion computation

Browse Source 
Change-Id: I66ea5a819ab54ecb5327eee20f798d7d7f0833d3
This commit is contained in:
Sarah Parker
2016-09-06 19:09:13 -07:00
parent 1d22837fbd
commit ecb0afc838
1 changed files with 83 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

84
av1/encoder/encodeframe.c
View File

@@ -4403,7 +4403,80 @@ static int input_fpmb_stats(FIRSTPASS_MB_STATS *firstpass_mb_stats,
#if CONFIG_GLOBAL_MOTION
#define MIN_TRANS_THRESH 8
#define GLOBAL_MOTION_ADVANTAGE_THRESH 0.60
#define GLOBAL_MOTION_MODEL TRANSLATION
#define GLOBAL_MOTION_MODEL ROTZOOM
static void refine_integerized_param(WarpedMotionParams *wm,
#if CONFIG_AOM_HIGHBITDEPTH
int use_hbd, int bd,
#endif // CONFIG_AOM_HIGHBITDEPTH
uint8_t *ref, int r_width, int r_height,
int r_stride, uint8_t *dst, int d_width,
int d_height, int d_stride,
int n_refinements) {
int i = 0, p;
int n_params = n_trans_model_params[wm->wmtype];
int16_t *param_mat = (int16_t *)wm->wmmat;
double step_error;
int step;
int16_t *param;
int16_t curr_param;
int16_t best_param;
double best_error =
av1_warp_erroradv(wm,
#if CONFIG_AOM_HIGHBITDEPTH
use_hbd, bd,
#endif // CONFIG_AOM_HIGHBITDEPTH
ref, r_width, r_height, r_stride, dst, 0, 0, d_width,
d_height, d_stride, 0, 0, 16, 16);
for (p = 0; p < n_params; ++p) {
param = param_mat + p;
step = 1 << (n_refinements + 1);
curr_param = *param;
best_param = curr_param;
for (i = 0; i < n_refinements; i++) {
// look to the left
*param = curr_param - step;
step_error =
av1_warp_erroradv(wm,
#if CONFIG_AOM_HIGHBITDEPTH
use_hbd, bd,
#endif // CONFIG_AOM_HIGHBITDEPTH
ref, r_width, r_height, r_stride, dst, 0, 0,
d_width, d_height, d_stride, 0, 0, 16, 16);
if (step_error < best_error) {
step >>= 1;
best_error = step_error;
best_param = *param;
curr_param = best_param;
continue;
}
// look to the right
*param = curr_param + step;
step_error =
av1_warp_erroradv(wm,
#if CONFIG_AOM_HIGHBITDEPTH
use_hbd, bd,
#endif // CONFIG_AOM_HIGHBITDEPTH
ref, r_width, r_height, r_stride, dst, 0, 0,
d_width, d_height, d_stride, 0, 0, 16, 16);
if (step_error < best_error) {
step >>= 1;
best_error = step_error;
best_param = *param;
curr_param = best_param;
continue;
}
// no improvement found-> means we're either already at a minimum or
// step is too wide
step >>= 1;
}
*param = best_param;
}
}
static void convert_to_params(double *H, TransformationType type,
int16_t *model) {
int i;
@@ -4416,6 +4489,7 @@ static void convert_to_params(double *H, TransformationType type,
model[1] = (int16_t)clamp(model[1], GM_TRANS_MIN, GM_TRANS_MAX) *
GM_TRANS_DECODE_FACTOR;
// TODO(sarahparker) 1 should be subtracted here
for (i = 2; i < n_params; ++i) {
model[i] = (int16_t)floor(H[i] * (1 << GM_ALPHA_PREC_BITS) + 0.5);
model[i] = (int16_t)clamp(model[i], GM_ALPHA_MIN, GM_ALPHA_MAX) *
@@ -4476,6 +4550,14 @@ static void encode_frame_internal(AV1_COMP *cpi) {
convert_model_to_params(H, GLOBAL_MOTION_MODEL,
&cm->global_motion[frame]);
if (get_gmtype(&cm->global_motion[frame]) > GLOBAL_ZERO) {
refine_integerized_param(
&cm->global_motion[frame].motion_params,
#if CONFIG_AOM_HIGHBITDEPTH
xd->cur_buf->flags & YV12_FLAG_HIGHBITDEPTH, xd->bd,
#endif // CONFIG_AOM_HIGHBITDEPTH
ref_buf->y_buffer, ref_buf->y_width, ref_buf->y_height,
ref_buf->y_stride, cpi->Source->y_buffer, cpi->Source->y_width,
cpi->Source->y_height, cpi->Source->y_stride, 3);
// compute the advantage of using gm parameters over 0 motion
double erroradvantage = av1_warp_erroradv(
&cm->global_motion[frame].motion_params,