Index: source/libvpx/vp9/encoder/vp9_denoiser.c |
=================================================================== |
--- source/libvpx/vp9/encoder/vp9_denoiser.c (revision 281795) |
+++ source/libvpx/vp9/encoder/vp9_denoiser.c (working copy) |
@@ -9,38 +9,90 @@ |
*/ |
#include <assert.h> |
+#include <limits.h> |
#include "vpx_scale/yv12config.h" |
#include "vpx/vpx_integer.h" |
+#include "vp9/common/vp9_reconinter.h" |
#include "vp9/encoder/vp9_denoiser.h" |
+/* The VP9 denoiser is a work-in-progress. It currently is only designed to work |
+ * with speed 6, though it (inexplicably) seems to also work with speed 5 (one |
+ * would need to modify the source code in vp9_pickmode.c and vp9_encoder.c to |
+ * make the calls to the vp9_denoiser_* functions when in speed 5). |
+ * |
+ * The implementation is very similar to that of the VP8 denoiser. While |
+ * choosing the motion vectors / reference frames, the denoiser is run, and if |
+ * it did not modify the signal to much, the denoised block is copied to the |
+ * signal. |
+ */ |
+ |
+#ifdef OUTPUT_YUV_DENOISED |
+static void make_grayscale(YV12_BUFFER_CONFIG *yuv); |
+#endif |
+ |
static const int widths[] = {4, 4, 8, 8, 8, 16, 16, 16, 32, 32, 32, 64, 64}; |
static const int heights[] = {4, 8, 4, 8, 16, 8, 16, 32, 16, 32, 64, 32, 64}; |
-int vp9_denoiser_filter() { |
- return 0; |
+static int absdiff_thresh(BLOCK_SIZE bs, int increase_denoising) { |
+ (void)bs; |
+ return 3 + (increase_denoising ? 1 : 0); |
} |
-static int update_running_avg(const uint8_t *mc_avg, int mc_avg_stride, |
- uint8_t *avg, int avg_stride, |
- const uint8_t *sig, int sig_stride, |
- int increase_denoising, BLOCK_SIZE bs) { |
+static int delta_thresh(BLOCK_SIZE bs, int increase_denoising) { |
+ (void)bs; |
+ (void)increase_denoising; |
+ return 4; |
+} |
+ |
+static int noise_motion_thresh(BLOCK_SIZE bs, int increase_denoising) { |
+ (void)bs; |
+ (void)increase_denoising; |
+ return 25 * 25; |
+} |
+ |
+static unsigned int sse_thresh(BLOCK_SIZE bs, int increase_denoising) { |
+ return widths[bs] * heights[bs] * (increase_denoising ? 60 : 40); |
+} |
+ |
+static int sse_diff_thresh(BLOCK_SIZE bs, int increase_denoising, |
+ int mv_row, int mv_col) { |
+ if (mv_row * mv_row + mv_col * mv_col > |
+ noise_motion_thresh(bs, increase_denoising)) { |
+ return 0; |
+ } else { |
+ return widths[bs] * heights[bs] * 20; |
+ } |
+} |
+ |
+static int total_adj_strong_thresh(BLOCK_SIZE bs, int increase_denoising) { |
+ return widths[bs] * heights[bs] * (increase_denoising ? 3 : 2); |
+} |
+ |
+static int total_adj_weak_thresh(BLOCK_SIZE bs, int increase_denoising) { |
+ return widths[bs] * heights[bs] * (increase_denoising ? 3 : 2); |
+} |
+ |
+static VP9_DENOISER_DECISION denoiser_filter(const uint8_t *sig, int sig_stride, |
+ const uint8_t *mc_avg, |
+ int mc_avg_stride, |
+ uint8_t *avg, int avg_stride, |
+ int increase_denoising, |
+ BLOCK_SIZE bs) { |
int r, c; |
- int diff, adj, absdiff; |
- int shift_inc1 = 0, shift_inc2 = 1; |
+ const uint8_t *sig_start = sig; |
+ const uint8_t *mc_avg_start = mc_avg; |
+ uint8_t *avg_start = avg; |
+ int diff, adj, absdiff, delta; |
int adj_val[] = {3, 4, 6}; |
int total_adj = 0; |
- if (increase_denoising) { |
- shift_inc1 = 1; |
- shift_inc2 = 2; |
- } |
- |
+ // First attempt to apply a strong temporal denoising filter. |
for (r = 0; r < heights[bs]; ++r) { |
for (c = 0; c < widths[bs]; ++c) { |
diff = mc_avg[c] - sig[c]; |
absdiff = abs(diff); |
- if (absdiff <= 3 + shift_inc1) { |
+ if (absdiff <= absdiff_thresh(bs, increase_denoising)) { |
avg[c] = mc_avg[c]; |
total_adj += diff; |
} else { |
@@ -68,7 +120,47 @@ |
avg += avg_stride; |
mc_avg += mc_avg_stride; |
} |
- return total_adj; |
+ |
+ // If the strong filter did not modify the signal too much, we're all set. |
+ if (abs(total_adj) <= total_adj_strong_thresh(bs, increase_denoising)) { |
+ return FILTER_BLOCK; |
+ } |
+ |
+ // Otherwise, we try to dampen the filter if the delta is not too high. |
+ delta = ((abs(total_adj) - total_adj_strong_thresh(bs, increase_denoising)) |
+ >> 8) + 1; |
+ if (delta > delta_thresh(bs, increase_denoising)) { |
+ return COPY_BLOCK; |
+ } |
+ |
+ mc_avg = mc_avg_start; |
+ avg = avg_start; |
+ sig = sig_start; |
+ for (r = 0; r < heights[bs]; ++r) { |
+ for (c = 0; c < widths[bs]; ++c) { |
+ diff = mc_avg[c] - sig[c]; |
+ adj = abs(diff); |
+ if (adj > delta) { |
+ adj = delta; |
+ } |
+ if (diff > 0) { |
+ avg[c] = MAX(0, avg[c] - adj); |
+ total_adj += adj; |
+ } else { |
+ avg[c] = MIN(UINT8_MAX, avg[c] + adj); |
+ total_adj -= adj; |
+ } |
+ } |
+ sig += sig_stride; |
+ avg += avg_stride; |
+ mc_avg += mc_avg_stride; |
+ } |
+ |
+ // We can use the filter if it has been sufficiently dampened |
+ if (abs(total_adj) <= total_adj_weak_thresh(bs, increase_denoising)) { |
+ return FILTER_BLOCK; |
+ } |
+ return COPY_BLOCK; |
} |
static uint8_t *block_start(uint8_t *framebuf, int stride, |
@@ -76,51 +168,174 @@ |
return framebuf + (stride * mi_row * 8) + (mi_col * 8); |
} |
-void copy_block(uint8_t *dest, int dest_stride, |
- uint8_t *src, int src_stride, BLOCK_SIZE bs) { |
- int r, c; |
+static void copy_block(uint8_t *dest, int dest_stride, |
+ const uint8_t *src, int src_stride, BLOCK_SIZE bs) { |
+ int r; |
for (r = 0; r < heights[bs]; ++r) { |
- for (c = 0; c < widths[bs]; ++c) { |
- dest[c] = src[c]; |
- } |
+ vpx_memcpy(dest, src, widths[bs]); |
dest += dest_stride; |
src += src_stride; |
} |
} |
+static VP9_DENOISER_DECISION perform_motion_compensation(VP9_DENOISER *denoiser, |
+ MACROBLOCK *mb, |
+ BLOCK_SIZE bs, |
+ int increase_denoising, |
+ int mi_row, |
+ int mi_col) { |
+ int mv_col, mv_row; |
+ int sse_diff = denoiser->zero_mv_sse - denoiser->best_sse; |
+ MV_REFERENCE_FRAME frame; |
+ MACROBLOCKD *filter_mbd = &mb->e_mbd; |
+ MB_MODE_INFO *mbmi = &filter_mbd->mi[0]->mbmi; |
+ |
+ // We will restore these after motion compensation. |
+ MB_MODE_INFO saved_mbmi = *mbmi; |
+ struct buf_2d saved_dst = filter_mbd->plane[0].dst; |
+ struct buf_2d saved_pre[2]; |
+ saved_pre[0] = filter_mbd->plane[0].pre[0]; |
+ saved_pre[1] = filter_mbd->plane[0].pre[1]; |
+ |
+ mv_col = denoiser->best_sse_mv.as_mv.col; |
+ mv_row = denoiser->best_sse_mv.as_mv.row; |
+ |
+ frame = denoiser->best_reference_frame; |
+ |
+ // If the best reference frame uses inter-prediction and there is enough of a |
+ // difference in sum-squared-error, use it. |
+ if (frame != INTRA_FRAME && |
+ sse_diff > sse_diff_thresh(bs, increase_denoising, mv_row, mv_col)) { |
+ mbmi->ref_frame[0] = denoiser->best_reference_frame; |
+ mbmi->mode = denoiser->best_sse_inter_mode; |
+ mbmi->mv[0] = denoiser->best_sse_mv; |
+ } else { |
+ // Otherwise, use the zero reference frame. |
+ frame = denoiser->best_zeromv_reference_frame; |
+ |
+ mbmi->ref_frame[0] = denoiser->best_zeromv_reference_frame; |
+ mbmi->mode = ZEROMV; |
+ mbmi->mv[0].as_int = 0; |
+ |
+ denoiser->best_sse_inter_mode = ZEROMV; |
+ denoiser->best_sse_mv.as_int = 0; |
+ denoiser->best_sse = denoiser->zero_mv_sse; |
+ } |
+ |
+ // Set the pointers in the MACROBLOCKD to point to the buffers in the denoiser |
+ // struct. |
+ filter_mbd->plane[0].pre[0].buf = |
+ block_start(denoiser->running_avg_y[frame].y_buffer, |
+ denoiser->running_avg_y[frame].y_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[0].pre[0].stride = denoiser->running_avg_y[frame].y_stride; |
+ |
+ filter_mbd->plane[1].pre[0].buf = |
+ block_start(denoiser->running_avg_y[frame].u_buffer, |
+ denoiser->running_avg_y[frame].uv_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[1].pre[0].stride = denoiser->running_avg_y[frame].uv_stride; |
+ |
+ filter_mbd->plane[2].pre[0].buf = |
+ block_start(denoiser->running_avg_y[frame].v_buffer, |
+ denoiser->running_avg_y[frame].uv_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[2].pre[0].stride = denoiser->running_avg_y[frame].uv_stride; |
+ |
+ filter_mbd->plane[0].pre[1].buf = |
+ block_start(denoiser->running_avg_y[frame].y_buffer, |
+ denoiser->running_avg_y[frame].y_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[0].pre[1].stride = denoiser->running_avg_y[frame].y_stride; |
+ |
+ filter_mbd->plane[1].pre[1].buf = |
+ block_start(denoiser->running_avg_y[frame].u_buffer, |
+ denoiser->running_avg_y[frame].uv_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[1].pre[1].stride = denoiser->running_avg_y[frame].uv_stride; |
+ |
+ filter_mbd->plane[2].pre[1].buf = |
+ block_start(denoiser->running_avg_y[frame].v_buffer, |
+ denoiser->running_avg_y[frame].uv_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[2].pre[1].stride = denoiser->running_avg_y[frame].uv_stride; |
+ |
+ filter_mbd->plane[0].dst.buf = |
+ block_start(denoiser->mc_running_avg_y.y_buffer, |
+ denoiser->mc_running_avg_y.y_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[0].dst.stride = denoiser->mc_running_avg_y.y_stride; |
+ |
+ filter_mbd->plane[1].dst.buf = |
+ block_start(denoiser->mc_running_avg_y.u_buffer, |
+ denoiser->mc_running_avg_y.uv_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[1].dst.stride = denoiser->mc_running_avg_y.y_stride; |
+ |
+ filter_mbd->plane[2].dst.buf = |
+ block_start(denoiser->mc_running_avg_y.v_buffer, |
+ denoiser->mc_running_avg_y.uv_stride, |
+ mi_row, mi_col); |
+ filter_mbd->plane[2].dst.stride = denoiser->mc_running_avg_y.y_stride; |
+ |
+ vp9_build_inter_predictors_sby(filter_mbd, mv_row, mv_col, bs); |
+ |
+ // Restore everything to its original state |
+ filter_mbd->plane[0].pre[0] = saved_pre[0]; |
+ filter_mbd->plane[0].pre[1] = saved_pre[1]; |
+ filter_mbd->plane[0].dst = saved_dst; |
+ *mbmi = saved_mbmi; |
+ |
+ mv_row = denoiser->best_sse_mv.as_mv.row; |
+ mv_col = denoiser->best_sse_mv.as_mv.col; |
+ |
+ if (denoiser->best_sse > sse_thresh(bs, increase_denoising)) { |
+ return COPY_BLOCK; |
+ } |
+ if (mv_row * mv_row + mv_col * mv_col > |
+ 8 * noise_motion_thresh(bs, increase_denoising)) { |
+ return COPY_BLOCK; |
+ } |
+ return FILTER_BLOCK; |
+} |
+ |
void vp9_denoiser_denoise(VP9_DENOISER *denoiser, MACROBLOCK *mb, |
int mi_row, int mi_col, BLOCK_SIZE bs) { |
- int decision = COPY_BLOCK; |
- |
+ VP9_DENOISER_DECISION decision = FILTER_BLOCK; |
YV12_BUFFER_CONFIG avg = denoiser->running_avg_y[INTRA_FRAME]; |
+ YV12_BUFFER_CONFIG mc_avg = denoiser->mc_running_avg_y; |
+ uint8_t *avg_start = block_start(avg.y_buffer, avg.y_stride, mi_row, mi_col); |
+ uint8_t *mc_avg_start = block_start(mc_avg.y_buffer, mc_avg.y_stride, |
+ mi_row, mi_col); |
struct buf_2d src = mb->plane[0].src; |
- update_running_avg(denoiser->mc_running_avg_y.y_buffer, |
- denoiser->mc_running_avg_y.y_stride, |
- denoiser->running_avg_y[INTRA_FRAME].y_buffer, |
- denoiser->running_avg_y[INTRA_FRAME].y_stride, |
- mb->plane[0].src.buf, mb->plane[0].src.stride, 0, bs); |
+ decision = perform_motion_compensation(denoiser, mb, bs, |
+ denoiser->increase_denoising, |
+ mi_row, mi_col); |
if (decision == FILTER_BLOCK) { |
- // TODO(tkopp) |
+ decision = denoiser_filter(src.buf, src.stride, |
+ mc_avg_start, mc_avg.y_stride, |
+ avg_start, avg.y_stride, |
+ 0, bs); |
} |
- if (decision == COPY_BLOCK) { |
- copy_block(block_start(avg.y_buffer, avg.y_stride, mi_row, mi_col), |
- avg.y_stride, src.buf, src.stride, bs); |
+ |
+ if (decision == FILTER_BLOCK) { |
+ copy_block(src.buf, src.stride, avg_start, avg.y_stride, bs); |
+ } else { // COPY_BLOCK |
+ copy_block(avg_start, avg.y_stride, src.buf, src.stride, bs); |
} |
} |
static void copy_frame(YV12_BUFFER_CONFIG dest, const YV12_BUFFER_CONFIG src) { |
- int r, c; |
+ int r; |
const uint8_t *srcbuf = src.y_buffer; |
uint8_t *destbuf = dest.y_buffer; |
assert(dest.y_width == src.y_width); |
assert(dest.y_height == src.y_height); |
for (r = 0; r < dest.y_height; ++r) { |
- for (c = 0; c < dest.y_width; ++c) { |
- destbuf[c] = srcbuf[c]; |
- } |
+ vpx_memcpy(destbuf, srcbuf, dest.y_width); |
destbuf += dest.y_stride; |
srcbuf += src.y_stride; |
} |
@@ -155,9 +370,27 @@ |
} |
} |
-void vp9_denoiser_update_frame_stats() { |
+void vp9_denoiser_reset_frame_stats(VP9_DENOISER *denoiser) { |
+ denoiser->zero_mv_sse = UINT_MAX; |
+ denoiser->best_sse = UINT_MAX; |
} |
+void vp9_denoiser_update_frame_stats(VP9_DENOISER *denoiser, MB_MODE_INFO *mbmi, |
+ unsigned int sse, PREDICTION_MODE mode) { |
+ // TODO(tkopp): Use both MVs if possible |
+ if (mbmi->mv[0].as_int == 0 && sse < denoiser->zero_mv_sse) { |
+ denoiser->zero_mv_sse = sse; |
+ denoiser->best_zeromv_reference_frame = mbmi->ref_frame[0]; |
+ } |
+ |
+ if (mbmi->mv[0].as_int != 0 && sse < denoiser->best_sse) { |
+ denoiser->best_sse = sse; |
+ denoiser->best_sse_inter_mode = mode; |
+ denoiser->best_sse_mv = mbmi->mv[0]; |
+ denoiser->best_reference_frame = mbmi->ref_frame[0]; |
+ } |
+} |
+ |
int vp9_denoiser_alloc(VP9_DENOISER *denoiser, int width, int height, |
int ssx, int ssy, int border) { |
int i, fail; |
@@ -170,6 +403,9 @@ |
vp9_denoiser_free(denoiser); |
return 1; |
} |
+#ifdef OUTPUT_YUV_DENOISED |
+ make_grayscale(&denoiser->running_avg_y[i]); |
+#endif |
} |
fail = vp9_alloc_frame_buffer(&denoiser->mc_running_avg_y, width, height, |
@@ -178,6 +414,10 @@ |
vp9_denoiser_free(denoiser); |
return 1; |
} |
+#ifdef OUTPUT_YUV_DENOISED |
+ make_grayscale(&denoiser->running_avg_y[i]); |
+#endif |
+ denoiser->increase_denoising = 0; |
return 0; |
} |
@@ -196,3 +436,22 @@ |
vp9_free_frame_buffer(&denoiser->mc_running_avg_y); |
} |
} |
+ |
+#ifdef OUTPUT_YUV_DENOISED |
+static void make_grayscale(YV12_BUFFER_CONFIG *yuv) { |
+ int r, c; |
+ uint8_t *u = yuv->u_buffer; |
+ uint8_t *v = yuv->v_buffer; |
+ |
+ // The '/2's are there because we have a 440 buffer, but we want to output |
+ // 420. |
+ for (r = 0; r < yuv->uv_height / 2; ++r) { |
+ for (c = 0; c < yuv->uv_width / 2; ++c) { |
+ u[c] = UINT8_MAX / 2; |
+ v[c] = UINT8_MAX / 2; |
+ } |
+ u += yuv->uv_stride + yuv->uv_width / 2; |
+ v += yuv->uv_stride + yuv->uv_width / 2; |
+ } |
+} |
+#endif |