[PATCH] Make the crop filter accept parametric expressions.
Stefano Sabatini
stefano.sabatini-lala
Thu Sep 9 01:05:04 CEST 2010
Also change the syntax of the crop filter, from:
x:y:w:h
to
w:h:x:y
which looks more useful.
---
Changelog | 1 +
doc/ffmpeg-doc.texi | 8 +-
doc/filters.texi | 103 ++++++++++++++++++++++-----
ffmpeg.c | 4 +-
libavfilter/vf_crop.c | 176 ++++++++++++++++++++++++++++++++++++++++-----
tests/lavfi-regression.sh | 10 ++--
6 files changed, 254 insertions(+), 48 deletions(-)
diff --git a/Changelog b/Changelog
index 2a8dcaa..f082c1b 100644
--- a/Changelog
+++ b/Changelog
@@ -37,6 +37,7 @@ version <next>:
- R10k video decoder
- ocv_smooth filter
- frei0r wrapper filter
+- make crop filter parametric and change its syntax to w:h:x:y
version 0.6:
diff --git a/doc/ffmpeg-doc.texi b/doc/ffmpeg-doc.texi
index d9c5c14..e43e87e 100644
--- a/doc/ffmpeg-doc.texi
+++ b/doc/ffmpeg-doc.texi
@@ -226,13 +226,13 @@ The following abbreviations are recognized:
@item -aspect @var{aspect}
Set aspect ratio (4:3, 16:9 or 1.3333, 1.7777).
- at item -croptop @var{size} (deprecated - use -vf crop=x:y:width:height instead)
+ at item -croptop @var{size} (deprecated - use the crop filter instead)
Set top crop band size (in pixels).
- at item -cropbottom @var{size} (deprecated - use -vf crop=x:y:width:height instead)
+ at item -cropbottom @var{size} (deprecated - use the crop filter instead)
Set bottom crop band size (in pixels).
- at item -cropleft @var{size} (deprecated - use -vf crop=x:y:width:height instead)
+ at item -cropleft @var{size} (deprecated - use the crop filter instead)
Set left crop band size (in pixels).
- at item -cropright @var{size} (deprecated - use -vf crop=x:y:width:height instead)
+ at item -cropright @var{size} (deprecated - use the crop filter instead)
Set right crop band size (in pixels).
@item -padtop @var{size}
@item -padbottom @var{size}
diff --git a/doc/filters.texi b/doc/filters.texi
index 2309c56..1dc222a 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -26,34 +26,101 @@ Below is a description of the currently available video filters.
@section crop
-Crop the input video to @var{x}:@var{y}:@var{width}:@var{height}.
+Crop the input video to @var{out_w}:@var{out_h}:@var{x}:@var{y}.
- at example
-./ffmpeg -i in.avi -vf "crop=0:0:0:240" out.avi
- at end example
+The parameters are expressions containing the following constants:
- at var{x} and @var{y} specify the position of the top-left corner of the
-output (non-cropped) area.
+ at table @option
+ at item E, PI, PHI
+the corresponding mathematical approximated values for e
+(euler number), pi (greek PI), PHI (golden ratio)
-The default value of @var{x} and @var{y} is 0.
+ at item x, y
+the computed values for @var{x} and @var{y}. They are evaluated for
+each new frame.
+
+ at item in_w, in_h
+the input width and heigth
+
+ at item iw, ih
+same as @var{in_w} and @var{in_h}
+
+ at item out_w, out_h
+the output (cropped) width and heigth
+
+ at item ow, oh
+same as @var{out_w} and @var{out_h}
+
+ at item n
+the number of input frame, starting from 0
+
+ at item pos
+the position in the file of the input frame, NAN if unknown
+
+ at item t
+timestamp expressed in seconds, NAN if the input timestamp is unknown
+
+ at end table
-The @var{width} and @var{height} parameters specify the width and height
-of the output (non-cropped) area.
+The @var{out_w} and @var{out_h} parameters specify the expressions for
+the width and height of the output (cropped) video. They are
+evaluated just at the configuration of the filter.
-A value of 0 is interpreted as the maximum possible size contained in
-the area delimited by the top-left corner at position x:y.
+The default value of @var{out_w} is "in_w", and the default value of
+ at var{out_h} is "in_h".
-For example the parameters:
+The expression for @var{out_w} may depend on the value of @var{out_h},
+and the expression for @var{out_h} may depend on @var{out_w}, but they
+cannot depend on @var{x} and @var{y}, as @var{x} and @var{y} are
+evaluated after @var{out_w} and @var{out_h}.
+The @var{x} and @var{y} parameters specify the expressions for the
+position of the top-left corner of the output (non-cropped) area. They
+are evaluated for each frame. If the evaluated value is not valid, it
+is approximated to the nearest valid value.
+
+The default value of @var{x} is "(in_w-out_w)/2", and the default
+value for @var{y} is "(in_h-out_h)/2", which set the cropped area at
+the center of the input image.
+
+The expression for @var{x} may depend on @var{y}, and the expression
+for @var{y} may depend on @var{x}.
+
+Follow some examples:
@example
-"crop=100:100:0:0"
- at end example
+# crop the central input area with size 100x100
+crop=100:100
-will delimit the rectangle with the top-left corner placed at position
-100:100 and the right-bottom corner corresponding to the right-bottom
-corner of the input image.
+# crop the central input area with size 2/3 of the input video
+"crop=2/3*in_w:2/3*in_h"
-The default value of @var{width} and @var{height} is 0.
+# crop the input video central square
+crop=in_h
+
+# delimit the rectangle with the top-left corner placed at position
+# 100:100 and the right-bottom corner corresponding to the right-bottom
+# corner of the input image.
+crop=in_w-100:in_h-100:100:100
+
+# crop 10 pixels from the lefth and right borders, and 20 pixels from
+# the top and bottom borders
+"crop=in_w-2*10:in_h-2*20"
+
+# keep only the bottom right quarter of the input image
+"crop=in_w/2:in_h/2:in_w/2:in_h/2"
+
+# crop height for getting Greek harmony
+"crop=in_w:1/PHI*in_w"
+
+# trembling effect
+"crop=in_w/2:in_h/2:(in_w-out_w)/2+((in_w-out_w)/2)*sin(n/10):(in_h-out_h)/2 +((in_h-out_h)/2)*sin(n/7)"
+
+# erratic camera effect depending on timestamp and position
+"crop=in_w/2:in_h/2:(in_w-out_w)/2+((in_w-out_w)/2)*sin(t*10):(in_h-out_h)/2 +((in_h-out_h)/2)*sin(t*13)"
+
+# set x depending on the value of y
+"crop=in_w/2:in_h/2:y:10+10*sin(n/10)"
+ at end example
@section fifo
diff --git a/ffmpeg.c b/ffmpeg.c
index 9d7b2e5..4b5dc1c 100644
--- a/ffmpeg.c
+++ b/ffmpeg.c
@@ -430,9 +430,7 @@ static int configure_filters(AVInputStream *ist, AVOutputStream *ost)
last_filter = ist->input_video_filter;
if (ost->video_crop) {
- snprintf(args, 255, "%d:%d:%d:%d", ost->leftBand, ost->topBand,
- codec->width,
- codec->height);
+ snprintf(args, 255, "%d:%d:%d:%d", codec->width, codec->height, ost->leftBand, ost->topBand);
if ((ret = avfilter_open(&filter, avfilter_get_by_name("crop"), NULL)) < 0)
return ret;
if ((ret = avfilter_init_filter(filter, args, NULL)) < 0)
diff --git a/libavfilter/vf_crop.c b/libavfilter/vf_crop.c
index 4f8a382..d190d72 100644
--- a/libavfilter/vf_crop.c
+++ b/libavfilter/vf_crop.c
@@ -24,8 +24,43 @@
*/
#include "avfilter.h"
+#include "libavutil/eval.h"
+#include "libavutil/avstring.h"
+#include "libavutil/libm.h"
#include "libavcore/imgutils.h"
+static const char *var_names[] = {
+ "E",
+ "PHI",
+ "PI",
+ "in_w", "iw", ///< width of the input video
+ "in_h", "ih", ///< height of the input video
+ "out_w", "ow", ///< width of the cropped video
+ "out_h", "oh", ///< height of the cropped video
+ "x",
+ "y",
+ "n", ///< number of frame
+ "pos", ///< position in the file
+ "t", ///< timestamp expressed in seconds
+ NULL
+};
+
+enum var_name {
+ E,
+ PHI,
+ PI,
+ IN_W, IW,
+ IN_H, IH,
+ OUT_W, OW,
+ OUT_H, OH,
+ X,
+ Y,
+ N,
+ POS,
+ T,
+ VARS_NB
+};
+
typedef struct {
int x; ///< x offset of the non-cropped area with respect to the input area
int y; ///< y offset of the non-cropped area with respect to the input area
@@ -34,6 +69,9 @@ typedef struct {
int max_step[4]; ///< max pixel step for each plane, expressed as a number of bytes
int hsub, vsub; ///< chroma subsampling
+ char x_expr[256], y_expr[256], ow_expr[256], oh_expr[256];
+ AVExpr *x_pexpr, *y_pexpr; /* parsed expressions for x and y */
+ double var_values[VARS_NB];
} CropContext;
static int query_formats(AVFilterContext *ctx)
@@ -72,44 +110,114 @@ static av_cold int init(AVFilterContext *ctx, const char *args, void *opaque)
{
CropContext *crop = ctx->priv;
+ av_strlcpy(crop->ow_expr, "iw", sizeof(crop->ow_expr));
+ av_strlcpy(crop->oh_expr, "ih", sizeof(crop->oh_expr));
+ av_strlcpy(crop->x_expr, "(in_w-out_w)/2", sizeof(crop->x_expr));
+ av_strlcpy(crop->y_expr, "(in_h-out_h)/2", sizeof(crop->y_expr));
+
if (args)
- sscanf(args, "%d:%d:%d:%d", &crop->x, &crop->y, &crop->w, &crop->h);
+ sscanf(args, "%255[^:]:%255[^:]:%255[^:]:%255[^:]", crop->ow_expr, crop->oh_expr, crop->x_expr, crop->y_expr);
return 0;
}
+static av_cold void uninit(AVFilterContext *ctx)
+{
+ CropContext *crop = ctx->priv;
+
+ av_free_expr(crop->x_pexpr); crop->x_pexpr = NULL;
+ av_free_expr(crop->y_pexpr); crop->y_pexpr = NULL;
+}
+
+static inline int normalize_double(int *n, double d)
+{
+ int ret = 0;
+
+ if (isnan(d)) {
+ ret = AVERROR(EINVAL);
+ } else if (d > INT_MAX || d < INT_MIN) {
+ *n = d > INT_MAX ? INT_MAX : INT_MIN;
+ ret = AVERROR(EINVAL);
+ } else
+ *n = round(d);
+
+ return ret;
+}
+
static int config_input(AVFilterLink *link)
{
AVFilterContext *ctx = link->dst;
CropContext *crop = ctx->priv;
const AVPixFmtDescriptor *pix_desc = &av_pix_fmt_descriptors[link->format];
+ int ret;
+ const char *expr;
+ double res;
+
+ crop->var_values[E] = M_E;
+ crop->var_values[PHI] = M_PHI;
+ crop->var_values[PI] = M_PI;
+ crop->var_values[IN_W] = crop->var_values[IW] = ctx->inputs[0]->w;
+ crop->var_values[IN_H] = crop->var_values[IH] = ctx->inputs[0]->h;
+ crop->var_values[X] = NAN;
+ crop->var_values[Y] = NAN;
+ crop->var_values[OUT_W] = crop->var_values[OW] = NAN;
+ crop->var_values[OUT_H] = crop->var_values[OH] = NAN;
+ crop->var_values[N] = 0;
+ crop->var_values[T] = NAN;
+ crop->var_values[POS] = NAN;
av_image_fill_max_pixsteps(crop->max_step, NULL, pix_desc);
crop->hsub = av_pix_fmt_descriptors[link->format].log2_chroma_w;
crop->vsub = av_pix_fmt_descriptors[link->format].log2_chroma_h;
- if (crop->w == 0)
- crop->w = link->w - crop->x;
- if (crop->h == 0)
- crop->h = link->h - crop->y;
-
- crop->x &= ~((1 << crop->hsub) - 1);
- crop->y &= ~((1 << crop->vsub) - 1);
+ if ((ret = av_parse_and_eval_expr(&res, (expr = crop->ow_expr),
+ var_names, crop->var_values,
+ NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0) goto fail_expr;
+ crop->var_values[OUT_W] = crop->var_values[OW] = res;
+ if ((ret = av_parse_and_eval_expr(&res, (expr = crop->oh_expr),
+ var_names, crop->var_values,
+ NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0) goto fail_expr;
+ crop->var_values[OUT_H] = crop->var_values[OH] = res;
+ /* evaluate again ow as it may depend on oh */
+ if ((ret = av_parse_and_eval_expr(&res, (expr = crop->ow_expr),
+ var_names, crop->var_values,
+ NULL, NULL, NULL, NULL, NULL, 0, ctx)) < 0) goto fail_expr;
+ crop->var_values[OUT_W] = crop->var_values[OW] = res;
+ if (normalize_double(&crop->w, crop->var_values[OUT_W]) < 0 ||
+ normalize_double(&crop->h, crop->var_values[OUT_H]) < 0) {
+ av_log(ctx, AV_LOG_ERROR,
+ "Too big value or invalid expression for out_w/ow or out_h/oh. "
+ "Maybe the expression for out_w:'%s' or for out_h:'%s' is self-referencing.\n",
+ crop->ow_expr, crop->oh_expr);
+ return AVERROR(EINVAL);
+ }
+ crop->w &= ~((1 << crop->hsub) - 1);
+ crop->h &= ~((1 << crop->vsub) - 1);
- av_log(link->dst, AV_LOG_INFO, "x:%d y:%d w:%d h:%d\n",
- crop->x, crop->y, crop->w, crop->h);
+ if ((ret = av_parse_expr(&crop->x_pexpr, crop->x_expr, var_names,
+ NULL, NULL, NULL, NULL, 0, ctx)) < 0 ||
+ (ret = av_parse_expr(&crop->y_pexpr, crop->y_expr, var_names,
+ NULL, NULL, NULL, NULL, 0, ctx)) < 0)
+ return AVERROR(EINVAL);
- if (crop->x < 0 || crop->y < 0 ||
- crop->w <= 0 || crop->h <= 0 ||
- (unsigned)crop->x + (unsigned)crop->w > link->w ||
- (unsigned)crop->y + (unsigned)crop->h > link->h) {
+ if (crop->w <= 0 || crop->h <= 0 ||
+ crop->w > link->w || crop->h > link->h) {
av_log(ctx, AV_LOG_ERROR,
- "Output area %d:%d:%d:%d not within the input area 0:0:%d:%d or zero-sized\n",
- crop->x, crop->y, crop->w, crop->h, link->w, link->h);
+ "Invalid too big or non positive size for width '%d' or height '%d'\n",
+ crop->w, crop->h);
return AVERROR(EINVAL);
}
+ /* set default, required in the case the first computed value for x/y is NAN */
+ crop->x = (link->w - crop->w) / 2;
+ crop->y = (link->h - crop->h) / 2;
+ crop->x &= ~((1 << crop->hsub) - 1);
+ crop->y &= ~((1 << crop->vsub) - 1);
return 0;
+
+fail_expr:
+ av_log(NULL, AV_LOG_ERROR, "Error when evaluating the expression '%s'\n", expr);
+ return ret;
}
static int config_output(AVFilterLink *link)
@@ -124,13 +232,35 @@ static int config_output(AVFilterLink *link)
static void start_frame(AVFilterLink *link, AVFilterBufferRef *picref)
{
- CropContext *crop = link->dst->priv;
+ AVFilterContext *ctx = link->dst;
+ CropContext *crop = ctx->priv;
AVFilterBufferRef *ref2 = avfilter_ref_buffer(picref, ~0);
int i;
picref->video->w = crop->w;
picref->video->h = crop->h;
+ /* FIXME: when the TB will be settable */
+ crop->var_values[T] = picref->pts == AV_NOPTS_VALUE ? NAN : (double)picref->pts / AV_TIME_BASE;
+ crop->var_values[POS] = picref->pos == -1 ? NAN : picref->pos;
+ crop->var_values[X] = av_eval_expr(crop->x_pexpr, crop->var_values, NULL);
+ crop->var_values[Y] = av_eval_expr(crop->y_pexpr, crop->var_values, NULL);
+ crop->var_values[X] = av_eval_expr(crop->x_pexpr, crop->var_values, NULL);
+
+ normalize_double(&crop->x, crop->var_values[X]);
+ normalize_double(&crop->y, crop->var_values[Y]);
+
+ if (crop->x < 0) crop->x = 0;
+ if (crop->y < 0) crop->y = 0;
+ if ((unsigned)crop->x + (unsigned)crop->w > link->w) crop->x = link->w - crop->w;
+ if ((unsigned)crop->y + (unsigned)crop->h > link->h) crop->y = link->h - crop->h;
+ crop->x &= ~((1 << crop->hsub) - 1);
+ crop->y &= ~((1 << crop->vsub) - 1);
+
+ av_log(ctx, AV_LOG_DEBUG,
+ "n:%d t:%f x:%d y:%d x+w:%d y+h:%d\n",
+ (int)crop->var_values[N], crop->var_values[T], crop->x, crop->y, crop->x+crop->w, crop->y+crop->h);
+
ref2->data[0] += crop->y * ref2->linesize[0];
ref2->data[0] += crop->x * crop->max_step[0];
@@ -170,19 +300,29 @@ static void draw_slice(AVFilterLink *link, int y, int h, int slice_dir)
avfilter_draw_slice(ctx->outputs[0], y - crop->y, h, slice_dir);
}
+static void end_frame(AVFilterLink *link)
+{
+ CropContext *crop = link->dst->priv;
+
+ crop->var_values[N] += 1.0;
+ avfilter_end_frame(link->dst->outputs[0]);
+}
+
AVFilter avfilter_vf_crop = {
.name = "crop",
- .description = NULL_IF_CONFIG_SMALL("Crop the input video to x:y:width:height."),
+ .description = NULL_IF_CONFIG_SMALL("Crop the input video to width:height:x:y."),
.priv_size = sizeof(CropContext),
.query_formats = query_formats,
.init = init,
+ .uninit = uninit,
.inputs = (AVFilterPad[]) {{ .name = "default",
.type = AVMEDIA_TYPE_VIDEO,
.start_frame = start_frame,
.draw_slice = draw_slice,
+ .end_frame = end_frame,
.get_video_buffer = avfilter_null_get_video_buffer,
.config_props = config_input, },
{ .name = NULL}},
diff --git a/tests/lavfi-regression.sh b/tests/lavfi-regression.sh
index 511e0ae..a3e75dd 100755
--- a/tests/lavfi-regression.sh
+++ b/tests/lavfi-regression.sh
@@ -22,15 +22,15 @@ do_lavfi() {
fi
}
-do_lavfi "crop" "crop=100:100"
-do_lavfi "crop_scale" "crop=100:100,scale=400:-1"
-do_lavfi "crop_scale_vflip" "null,null,crop=200:200,crop=20:20,scale=200:200,scale=250:250,vflip,vflip,null,scale=200:200,crop=100:100,vflip,scale=200:200,null,vflip,crop=100:100,null"
-do_lavfi "crop_vflip" "crop=100:100,vflip"
+do_lavfi "crop" "crop=iw-100:ih-100:100:100"
+do_lavfi "crop_scale" "crop=iw-100:ih-100:100:100,scale=400:-1"
+do_lavfi "crop_scale_vflip" "null,null,crop=iw-200:ih-200:200:200,crop=iw-20:ih-20:20:20,scale=200:200,scale=250:250,vflip,vflip,null,scale=200:200,crop=iw-100:ih-100:100:100,vflip,scale=200:200,null,vflip,crop=iw-100:ih-100:100:100,null"
+do_lavfi "crop_vflip" "crop=iw-100:ih-100:100:100,vflip"
do_lavfi "null" "null"
do_lavfi "scale200" "scale=200:200"
do_lavfi "scale500" "scale=500:500"
do_lavfi "vflip" "vflip"
-do_lavfi "vflip_crop" "vflip,crop=100:100"
+do_lavfi "vflip_crop" "vflip,crop=iw-100:ih-100:100:100"
do_lavfi "vflip_vflip" "vflip,vflip"
do_lavfi_pixfmts(){
--
1.7.1
--r5Pyd7+fXNt84Ff3--
More information about the ffmpeg-devel
mailing list