#include "formats.h"
#include "internal.h"
#include "video.h"
+#include "w3fdif.h"
typedef struct W3FDIFContext {
const AVClass *class;
AVFrame *prev, *cur, *next; ///< previous, current, next frames
int32_t **work_line; ///< lines we are calculating
int nb_threads;
+
+ W3FDIFDSPContext dsp;
} W3FDIFContext;
#define OFFSET(x) offsetof(W3FDIFContext, x)
return ff_set_common_formats(ctx, fmts_list);
}
+static void filter_simple_low(int32_t *work_line,
+ uint8_t *in_lines_cur[2],
+ const int16_t *coef, int linesize)
+{
+ int i;
+
+ for (i = 0; i < linesize; i++) {
+ *work_line = *in_lines_cur[0]++ * coef[0];
+ *work_line++ += *in_lines_cur[1]++ * coef[1];
+ }
+}
+
+static void filter_complex_low(int32_t *work_line,
+ uint8_t *in_lines_cur[4],
+ const int16_t *coef, int linesize)
+{
+ int i;
+
+ for (i = 0; i < linesize; i++) {
+ *work_line = *in_lines_cur[0]++ * coef[0];
+ *work_line += *in_lines_cur[1]++ * coef[1];
+ *work_line += *in_lines_cur[2]++ * coef[2];
+ *work_line++ += *in_lines_cur[3]++ * coef[3];
+ }
+}
+
+static void filter_simple_high(int32_t *work_line,
+ uint8_t *in_lines_cur[3],
+ uint8_t *in_lines_adj[3],
+ const int16_t *coef, int linesize)
+{
+ int i;
+
+ for (i = 0; i < linesize; i++) {
+ *work_line += *in_lines_cur[0]++ * coef[0];
+ *work_line += *in_lines_adj[0]++ * coef[0];
+ *work_line += *in_lines_cur[1]++ * coef[1];
+ *work_line += *in_lines_adj[1]++ * coef[1];
+ *work_line += *in_lines_cur[2]++ * coef[2];
+ *work_line++ += *in_lines_adj[2]++ * coef[2];
+ }
+}
+
+static void filter_complex_high(int32_t *work_line,
+ uint8_t *in_lines_cur[5],
+ uint8_t *in_lines_adj[5],
+ const int16_t *coef, int linesize)
+{
+ int i;
+
+ for (i = 0; i < linesize; i++) {
+ *work_line += *in_lines_cur[0]++ * coef[0];
+ *work_line += *in_lines_adj[0]++ * coef[0];
+ *work_line += *in_lines_cur[1]++ * coef[1];
+ *work_line += *in_lines_adj[1]++ * coef[1];
+ *work_line += *in_lines_cur[2]++ * coef[2];
+ *work_line += *in_lines_adj[2]++ * coef[2];
+ *work_line += *in_lines_cur[3]++ * coef[3];
+ *work_line += *in_lines_adj[3]++ * coef[3];
+ *work_line += *in_lines_cur[4]++ * coef[4];
+ *work_line++ += *in_lines_adj[4]++ * coef[4];
+ }
+}
+
+static void filter_scale(uint8_t *out_pixel, const int32_t *work_pixel, int linesize)
+{
+ int j;
+
+ for (j = 0; j < linesize; j++, out_pixel++, work_pixel++)
+ *out_pixel = av_clip(*work_pixel, 0, 255 * 256 * 128) >> 15;
+}
+
static int config_input(AVFilterLink *inlink)
{
AVFilterContext *ctx = inlink->dst;
return AVERROR(ENOMEM);
for (i = 0; i < s->nb_threads; i++) {
- s->work_line[i] = av_calloc(s->linesize[0], sizeof(*s->work_line[0]));
+ s->work_line[i] = av_calloc(FFALIGN(s->linesize[0], 32), sizeof(*s->work_line[0]));
if (!s->work_line[i])
return AVERROR(ENOMEM);
}
+ s->dsp.filter_simple_low = filter_simple_low;
+ s->dsp.filter_complex_low = filter_complex_low;
+ s->dsp.filter_simple_high = filter_simple_high;
+ s->dsp.filter_complex_high = filter_complex_high;
+ s->dsp.filter_scale = filter_scale;
+
+ if (ARCH_X86)
+ ff_w3fdif_init_x86(&s->dsp);
+
return 0;
}
const int dst_line_stride = out->linesize[plane];
const int start = (height * jobnr) / nb_jobs;
const int end = (height * (jobnr+1)) / nb_jobs;
- int i, j, y_in, y_out;
+ int j, y_in, y_out;
/* copy unchanged the lines of the field */
y_out = start + (s->field == cur->top_field_first) - (start & 1);
out_line = dst_data + (y_out * dst_line_stride);
while (y_out < end) {
- /* clear workspace */
- memset(s->work_line[jobnr], 0, sizeof(*s->work_line[jobnr]) * linesize);
-
/* get low vertical frequencies from current field */
for (j = 0; j < n_coef_lf[filter]; j++) {
y_in = (y_out + 1) + (j * 2) - n_coef_lf[filter];
work_line = s->work_line[jobnr];
switch (n_coef_lf[filter]) {
case 2:
- for (i = 0; i < linesize; i++) {
- *work_line += *in_lines_cur[0]++ * coef_lf[filter][0];
- *work_line++ += *in_lines_cur[1]++ * coef_lf[filter][1];
- }
+ s->dsp.filter_simple_low(work_line, in_lines_cur,
+ coef_lf[filter], linesize);
break;
case 4:
- for (i = 0; i < linesize; i++) {
- *work_line += *in_lines_cur[0]++ * coef_lf[filter][0];
- *work_line += *in_lines_cur[1]++ * coef_lf[filter][1];
- *work_line += *in_lines_cur[2]++ * coef_lf[filter][2];
- *work_line++ += *in_lines_cur[3]++ * coef_lf[filter][3];
- }
+ s->dsp.filter_complex_low(work_line, in_lines_cur,
+ coef_lf[filter], linesize);
}
/* get high vertical frequencies from adjacent fields */
work_line = s->work_line[jobnr];
switch (n_coef_hf[filter]) {
case 3:
- for (i = 0; i < linesize; i++) {
- *work_line += *in_lines_cur[0]++ * coef_hf[filter][0];
- *work_line += *in_lines_adj[0]++ * coef_hf[filter][0];
- *work_line += *in_lines_cur[1]++ * coef_hf[filter][1];
- *work_line += *in_lines_adj[1]++ * coef_hf[filter][1];
- *work_line += *in_lines_cur[2]++ * coef_hf[filter][2];
- *work_line++ += *in_lines_adj[2]++ * coef_hf[filter][2];
- }
+ s->dsp.filter_simple_high(work_line, in_lines_cur, in_lines_adj,
+ coef_hf[filter], linesize);
break;
case 5:
- for (i = 0; i < linesize; i++) {
- *work_line += *in_lines_cur[0]++ * coef_hf[filter][0];
- *work_line += *in_lines_adj[0]++ * coef_hf[filter][0];
- *work_line += *in_lines_cur[1]++ * coef_hf[filter][1];
- *work_line += *in_lines_adj[1]++ * coef_hf[filter][1];
- *work_line += *in_lines_cur[2]++ * coef_hf[filter][2];
- *work_line += *in_lines_adj[2]++ * coef_hf[filter][2];
- *work_line += *in_lines_cur[3]++ * coef_hf[filter][3];
- *work_line += *in_lines_adj[3]++ * coef_hf[filter][3];
- *work_line += *in_lines_cur[4]++ * coef_hf[filter][4];
- *work_line++ += *in_lines_adj[4]++ * coef_hf[filter][4];
- }
+ s->dsp.filter_complex_high(work_line, in_lines_cur, in_lines_adj,
+ coef_hf[filter], linesize);
}
/* save scaled result to the output frame, scaling down by 256 * 128 */
work_pixel = s->work_line[jobnr];
out_pixel = out_line;
- for (j = 0; j < linesize; j++, out_pixel++, work_pixel++)
- *out_pixel = av_clip(*work_pixel, 0, 255 * 256 * 128) >> 15;
+ s->dsp.filter_scale(out_pixel, work_pixel, linesize);
/* move on to next line */
y_out += 2;
projects / ffmpeg / blobdiff