cogl-gst-video-sink: Premultiply the colors coming from the video

Commit 99a53c82e9ab0a1e5 removed the internal format argument when
uploading a video frame to a texture so that the format will just be
determined automatically from the image format. However this also
leaves the premultiplied state at the default and the default is TRUE.
That means that when we upload RGBA data Cogl will do a premultiplied
conversion on the CPU. We probably don't want to be putting a CPU
conversion in the way of video frames so this patch changes it to set
the premultiplied state to FALSE on the textures and then do the
premultiplied conversion in the shader.

This is particularly important for AYUV which uses the alpha channel
for the V component so doing a premultiplied conversion on the CPU
just creates garbage and messes up the image.

The RGB and RGBA renderers have each been split into two; one that
uses GLSL and one that uses a regular pipeline. The RGBA pipeline
without GLSL is then changed to use 2 layers so we that we can do the
premultiplied conversion in the second layer with a special layer
combine string.

Reviewed-by: Robert Bragg <robert@linux.intel.com>
(cherry picked from commit 07a57f26596c72507035369c90ed6d62568330b5)
This commit is contained in:
Neil Roberts 2014-01-22 15:11:55 +00:00
parent 0af13af653
commit 2543a86ba1
2 changed files with 125 additions and 38 deletions

View File

@ -421,40 +421,6 @@ clear_frame_textures (CoglGstVideoSink *sink)
priv->frame_dirty = TRUE; priv->frame_dirty = TRUE;
} }
static void
cogl_gst_rgb_setup_pipeline (CoglGstVideoSink *sink,
CoglPipeline *pipeline)
{
CoglGstVideoSinkPrivate *priv = sink->priv;
if (cogl_has_feature (priv->ctx, COGL_FEATURE_ID_GLSL))
{
static SnippetCache snippet_cache;
SnippetCacheEntry *entry = get_cache_entry (sink, &snippet_cache);
if (entry == NULL)
{
char *source;
source =
g_strdup_printf ("vec4\n"
"cogl_gst_sample_video%i (vec2 UV)\n"
"{\n"
" return texture2D (cogl_sampler%i, UV);\n"
"}\n",
priv->custom_start,
priv->custom_start);
entry = add_cache_entry (sink, &snippet_cache, source);
g_free (source);
}
setup_pipeline_from_cache_entry (sink, pipeline, entry, 1);
}
else
setup_pipeline_from_cache_entry (sink, pipeline, NULL, 1);
}
static inline CoglBool static inline CoglBool
is_pot (unsigned int number) is_pot (unsigned int number)
{ {
@ -508,9 +474,46 @@ video_texture_new_from_data (CoglContext *ctx,
cogl_object_unref (bitmap); cogl_object_unref (bitmap);
cogl_texture_set_premultiplied (tex, FALSE);
return tex; return tex;
} }
static void
cogl_gst_rgb24_glsl_setup_pipeline (CoglGstVideoSink *sink,
CoglPipeline *pipeline)
{
CoglGstVideoSinkPrivate *priv = sink->priv;
static SnippetCache snippet_cache;
SnippetCacheEntry *entry = get_cache_entry (sink, &snippet_cache);
if (entry == NULL)
{
char *source;
source =
g_strdup_printf ("vec4\n"
"cogl_gst_sample_video%i (vec2 UV)\n"
"{\n"
" return texture2D (cogl_sampler%i, UV);\n"
"}\n",
priv->custom_start,
priv->custom_start);
entry = add_cache_entry (sink, &snippet_cache, source);
g_free (source);
}
setup_pipeline_from_cache_entry (sink, pipeline, entry, 1);
}
static void
cogl_gst_rgb24_setup_pipeline (CoglGstVideoSink *sink,
CoglPipeline *pipeline)
{
setup_pipeline_from_cache_entry (sink, pipeline, NULL, 1);
}
static CoglBool static CoglBool
cogl_gst_rgb24_upload (CoglGstVideoSink *sink, cogl_gst_rgb24_upload (CoglGstVideoSink *sink,
GstBuffer *buffer) GstBuffer *buffer)
@ -546,6 +549,17 @@ map_fail:
} }
} }
static CoglGstRenderer rgb24_glsl_renderer =
{
"RGB 24",
COGL_GST_RGB24,
COGL_GST_RENDERER_NEEDS_GLSL,
GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("{ RGB, BGR }")),
1, /* n_layers */
cogl_gst_rgb24_glsl_setup_pipeline,
cogl_gst_rgb24_upload,
};
static CoglGstRenderer rgb24_renderer = static CoglGstRenderer rgb24_renderer =
{ {
"RGB 24", "RGB 24",
@ -553,10 +567,61 @@ static CoglGstRenderer rgb24_renderer =
0, 0,
GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("{ RGB, BGR }")), GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("{ RGB, BGR }")),
1, /* n_layers */ 1, /* n_layers */
cogl_gst_rgb_setup_pipeline, cogl_gst_rgb24_setup_pipeline,
cogl_gst_rgb24_upload, cogl_gst_rgb24_upload,
}; };
static void
cogl_gst_rgb32_glsl_setup_pipeline (CoglGstVideoSink *sink,
CoglPipeline *pipeline)
{
CoglGstVideoSinkPrivate *priv = sink->priv;
static SnippetCache snippet_cache;
SnippetCacheEntry *entry = get_cache_entry (sink, &snippet_cache);
if (entry == NULL)
{
char *source;
source =
g_strdup_printf ("vec4\n"
"cogl_gst_sample_video%i (vec2 UV)\n"
"{\n"
" vec4 color = texture2D (cogl_sampler%i, UV);\n"
/* Premultiply the color */
" color.rgb *= color.a;\n"
" return color;\n"
"}\n",
priv->custom_start,
priv->custom_start);
entry = add_cache_entry (sink, &snippet_cache, source);
g_free (source);
}
setup_pipeline_from_cache_entry (sink, pipeline, entry, 1);
}
static void
cogl_gst_rgb32_setup_pipeline (CoglGstVideoSink *sink,
CoglPipeline *pipeline)
{
CoglGstVideoSinkPrivate *priv = sink->priv;
char *layer_combine;
setup_pipeline_from_cache_entry (sink, pipeline, NULL, 1);
/* Premultiply the texture using the a special layer combine */
layer_combine = g_strdup_printf ("RGB=MODULATE(PREVIOUS, TEXTURE_%i[A])\n"
"A=REPLACE(PREVIOUS[A])",
priv->custom_start);
cogl_pipeline_set_layer_combine (pipeline,
priv->custom_start + 1,
layer_combine,
NULL);
g_free(layer_combine);
}
static CoglBool static CoglBool
cogl_gst_rgb32_upload (CoglGstVideoSink *sink, cogl_gst_rgb32_upload (CoglGstVideoSink *sink,
GstBuffer *buffer) GstBuffer *buffer)
@ -592,14 +657,25 @@ map_fail:
} }
} }
static CoglGstRenderer rgb32_glsl_renderer =
{
"RGB 32",
COGL_GST_RGB32,
COGL_GST_RENDERER_NEEDS_GLSL,
GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("{ RGBA, BGRA }")),
1, /* n_layers */
cogl_gst_rgb32_glsl_setup_pipeline,
cogl_gst_rgb32_upload,
};
static CoglGstRenderer rgb32_renderer = static CoglGstRenderer rgb32_renderer =
{ {
"RGB 32", "RGB 32",
COGL_GST_RGB32, COGL_GST_RGB32,
0, 0,
GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("{ RGBA, BGRA }")), GST_STATIC_CAPS (GST_VIDEO_CAPS_MAKE ("{ RGBA, BGRA }")),
1, /* n_layers */ 2, /* n_layers */
cogl_gst_rgb_setup_pipeline, cogl_gst_rgb32_setup_pipeline,
cogl_gst_rgb32_upload, cogl_gst_rgb32_upload,
}; };
@ -782,6 +858,8 @@ cogl_gst_ayuv_glsl_setup_pipeline (CoglGstVideoSink *sink,
" color.r = y + 1.59765625 * v;\n" " color.r = y + 1.59765625 * v;\n"
" color.g = y - 0.390625 * u - 0.8125 * v;\n" " color.g = y - 0.390625 * u - 0.8125 * v;\n"
" color.b = y + 2.015625 * u;\n" " color.b = y + 2.015625 * u;\n"
/* Premultiply the color */
" color.rgb *= color.a;\n"
" return color;\n" " return color;\n"
"}\n", priv->custom_start, "}\n", priv->custom_start,
priv->custom_start); priv->custom_start);
@ -842,8 +920,13 @@ cogl_gst_build_renderers_list (CoglContext *ctx)
int i; int i;
static CoglGstRenderer *const renderers[] = static CoglGstRenderer *const renderers[] =
{ {
/* These are in increasing order of priority so that the
* priv->renderers will be in decreasing order. That way the GLSL
* renderers will be preferred if they are available */
&rgb24_renderer, &rgb24_renderer,
&rgb32_renderer, &rgb32_renderer,
&rgb24_glsl_renderer,
&rgb32_glsl_renderer,
&yv12_glsl_renderer, &yv12_glsl_renderer,
&i420_glsl_renderer, &i420_glsl_renderer,
&ayuv_glsl_renderer, &ayuv_glsl_renderer,
@ -920,6 +1003,8 @@ cogl_gst_find_renderer_by_format (CoglGstVideoSink *sink,
CoglGstRenderer *renderer = NULL; CoglGstRenderer *renderer = NULL;
GSList *element; GSList *element;
/* The renderers list is in decreasing order of priority so we'll
* pick the first one that matches */
for (element = priv->renderers; element; element = g_slist_next (element)) for (element = priv->renderers; element; element = g_slist_next (element))
{ {
CoglGstRenderer *candidate = (CoglGstRenderer *) element->data; CoglGstRenderer *candidate = (CoglGstRenderer *) element->data;

View File

@ -82,7 +82,9 @@
* called cogl_gst_sample_video0 which takes a single vec2 argument. * called cogl_gst_sample_video0 which takes a single vec2 argument.
* This can be used by custom snippets set the by the application to * This can be used by custom snippets set the by the application to
* sample from the video. The vec2 argument represents the normalised * sample from the video. The vec2 argument represents the normalised
* coordinates within the video. * coordinates within the video. The function returns a vec4
* containing a pre-multiplied RGBA color of the pixel within the
* video.
* *
* Since: 1.16 * Since: 1.16
*/ */