Bläddra i källkod

Make a number of key optimizations

 - Changed glMapBuffer to glMapBufferRange to allow invalidation.  Using
   just glMapBuffer alone was causing some unacceptable stalls.

 - Changed dynamic buffers from GL_DYNAMIC_WRITE to GL_STREAM_WRITE
   because I had misunderstood the OpenGL specification

 - Added _OPENGL and _D3D11 builtin preprocessor macros to effects to
   allow special processing if needed

 - Added fmod support to shaders (NOTE: D3D and GL do not function
   identically with negative numbers when using this.  Positive numbers
   however function identically)

 - Created a planar conversion shader that converts from packed YUV to
   planar 420 right on the GPU without any CPU processing.  Reduces
   required GPU download size to approximately 37.5% of its normal rate
   as well.  GPU usage down by 10 entire percentage points despite the
   extra required pass.
jp9000 11 år sedan
förälder
incheckning
2dbbffe4a2

+ 2 - 0
.gitignore

@@ -53,6 +53,8 @@ ltmain.sh
 .depend
 tags
 *.trace
+*.vsp
+*.psess
 *.swp
 *.dat
 *.clbin

+ 136 - 0
build/data/libobs/format_conversion.effect

@@ -0,0 +1,136 @@
+/******************************************************************************
+    Copyright (C) 2014 by Hugh Bailey <[email protected]>
+
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation, either version 2 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with this program.  If not, see <http://www.gnu.org/licenses/>.
+******************************************************************************/
+
+//#define DEBUGGING
+
+uniform float4x4  ViewProj;
+
+uniform float     u_plane_offset;
+uniform float     v_plane_offset;
+
+uniform float     width;
+uniform float     height;
+uniform float     width_i;
+uniform float     height_i;
+uniform float     width_d2;
+uniform float     height_d2;
+uniform float     width_d2_i;
+uniform float     height_d2_i;
+uniform float     input_width;
+uniform float     input_height;
+
+uniform texture2d image;
+
+sampler_state def_sampler {
+	Filter   = Linear;
+	AddressU = Clamp;
+	AddressV = Clamp;
+};
+
+struct VertInOut {
+	float4 pos : POSITION;
+	float2 uv  : TEXCOORD0;
+};
+
+VertInOut VSDefault(VertInOut vert_in)
+{
+	VertInOut vert_out;
+	vert_out.pos = mul(float4(vert_in.pos.xyz, 1.0), ViewProj);
+	vert_out.uv  = vert_in.uv;
+	return vert_out;
+}
+
+float4 PSPlanar420(VertInOut vert_in) : TARGET
+{
+#ifdef _OPENGL
+	float v_mul = floor((1.0 - vert_in.uv.y) * input_height);
+#else
+	float v_mul = floor(vert_in.uv.y * input_height);
+#endif
+
+	float byte_offset = floor((v_mul + vert_in.uv.x) * input_width) * 4.0;
+
+	float2 sample_pos[4];
+
+	if (byte_offset < u_plane_offset) {
+#ifdef DEBUGGING
+		return float4(1.0f, 1.0f, 1.0f, 1.0f);
+#endif
+
+		float lum_u = fmod(byte_offset, width)     * width_i;
+		float lum_v = floor(byte_offset * width_i) * height_i;
+
+		/* move to texel centers to sample the 4 pixels properly */
+		lum_u += width_i  * 0.5;
+		lum_v += height_i * 0.5;
+
+		sample_pos[0] = float2(lum_u,            lum_v);
+		sample_pos[1] = float2(lum_u += width_i, lum_v);
+		sample_pos[2] = float2(lum_u += width_i, lum_v);
+		sample_pos[3] = float2(lum_u +  width_i, lum_v);
+
+	} else {
+#ifdef DEBUGGING
+		return ((byte_offset < v_plane_offset) ?
+				float4(0.5f, 0.5f, 0.5f, 0.5f) :
+				float4(0.2f, 0.2f, 0.2f, 0.2f));
+#endif
+
+		float new_offset = byte_offset -
+				((byte_offset < v_plane_offset) ?
+				u_plane_offset : v_plane_offset);
+
+		float ch_u = fmod(new_offset, width_d2)     * width_d2_i;
+		float ch_v = floor(new_offset * width_d2_i) * height_d2_i;
+		float width_i2 = width_i*2.0;
+
+		/* move to the borders of each set of 4 pixels to force it
+		 * to do bilinear averaging */
+		ch_u += width_i;
+		ch_v += height_i;
+
+		sample_pos[0] = float2(ch_u,             ch_v);
+		sample_pos[1] = float2(ch_u += width_i2, ch_v);
+		sample_pos[2] = float2(ch_u += width_i2, ch_v);
+		sample_pos[3] = float2(ch_u +  width_i2, ch_v);
+	}
+
+	float4x4 out_val = {
+		image.Sample(def_sampler, sample_pos[0]),
+		image.Sample(def_sampler, sample_pos[1]),
+		image.Sample(def_sampler, sample_pos[2]),
+		image.Sample(def_sampler, sample_pos[3])
+	};
+
+	out_val = transpose(out_val);
+
+	if (byte_offset < u_plane_offset)
+		return out_val[1];
+	else if (byte_offset < v_plane_offset)
+		return out_val[0];
+	else
+		return out_val[2];
+}
+
+technique Planar420
+{
+	pass
+	{
+		vertex_shader = VSDefault(vert_in);
+		pixel_shader  = PSPlanar420(vert_in);
+	}
+}

+ 1 - 0
libobs-d3d11/d3d11-exports.h

@@ -21,6 +21,7 @@
 
 extern "C" {
 
+EXPORT const char *device_preprocessor_name(void);
 EXPORT device_t device_create(struct gs_init_data *data);
 EXPORT void device_destroy(device_t device);
 EXPORT void device_entercontext(device_t device);

+ 5 - 0
libobs-d3d11/d3d11-subsystem.cpp

@@ -421,6 +421,11 @@ gs_device::gs_device(gs_init_data *data)
 	device_setrendertarget(this, NULL, NULL);
 }
 
+const char *device_preprocessor_name(void)
+{
+	return "_D3D11";
+}
+
 gs_device *device_create(gs_init_data *data)
 {
 	gs_device *device = NULL;

+ 1 - 0
libobs-opengl/gl-exports.h

@@ -19,6 +19,7 @@
 
 #include <util/c99defs.h>
 
+EXPORT const char *device_preprocessor_name(void);
 EXPORT device_t device_create(struct gs_init_data *data);
 EXPORT void device_destroy(device_t device);
 EXPORT void device_entercontext(device_t device);

+ 6 - 2
libobs-opengl/gl-helpers.c

@@ -155,8 +155,12 @@ bool update_buffer(GLenum target, GLuint buffer, void *data, size_t size)
 	if (!gl_bind_buffer(target, buffer))
 		return false;
 
-	ptr = glMapBuffer(target, GL_WRITE_ONLY);
-	success = gl_success("glMapBuffer");
+	/* glMapBufferRange with these flags will actually give far better
+	 * performance than a plain glMapBuffer call */
+	ptr = glMapBufferRange(target, 0, size,
+			GL_MAP_WRITE_BIT |
+			GL_MAP_INVALIDATE_BUFFER_BIT);
+	success = gl_success("glMapBufferRange");
 	if (success && ptr) {
 		memcpy(ptr, data, size);
 		glUnmapBuffer(target);

+ 3 - 1
libobs-opengl/gl-shaderparser.c

@@ -239,7 +239,7 @@ static inline void gl_write_structs(struct gl_shader_parser *glsp)
  *   clip     -> (unsupported)
  *   ddx      -> dFdx
  *   ddy      -> dFdy
- *   fmod     -> (unsupported)
+ *   fmod     -> mod (XXX: these are different if sign is negative)
  *   frac     -> fract
  *   lerp     -> mix
  *   lit      -> (unsupported)
@@ -367,6 +367,8 @@ static bool gl_write_intrinsic(struct gl_shader_parser *glsp,
 		dstr_cat(&glsp->gl_string, "fract");
 	} else if (strref_cmp(&token->str, "lerp") == 0) {
 		dstr_cat(&glsp->gl_string, "mix");
+	} else if (strref_cmp(&token->str, "fmod") == 0) {
+		dstr_cat(&glsp->gl_string, "mod");
 	} else if (strref_cmp(&token->str, "rsqrt") == 0) {
 		dstr_cat(&glsp->gl_string, "inversesqrt");
 	} else if (strref_cmp(&token->str, "saturate") == 0) {

+ 5 - 0
libobs-opengl/gl-subsystem.c

@@ -163,6 +163,11 @@ void convert_sampler_info(struct gs_sampler_state *sampler,
 	               info->max_anisotropy, sampler->max_anisotropy);
 }
 
+const char *device_preprocessor_name(void)
+{
+	return "_OPENGL";
+}
+
 device_t device_create(struct gs_init_data *info)
 {
 	struct gs_device *device = bzalloc(sizeof(struct gs_device));

+ 1 - 1
libobs-opengl/gl-vertexbuffer.c

@@ -20,7 +20,7 @@
 
 static bool create_buffers(struct gs_vertex_buffer *vb)
 {
-	GLenum usage = vb->dynamic ? GL_DYNAMIC_DRAW : GL_STATIC_DRAW;
+	GLenum usage = vb->dynamic ? GL_STREAM_DRAW : GL_STATIC_DRAW;
 	size_t i;
 
 	if (!gl_create_buffer(GL_ARRAY_BUFFER, &vb->vertex_buffer,

+ 15 - 0
libobs/graphics/effect-parser.c

@@ -914,11 +914,26 @@ error:
 
 static bool ep_compile(struct effect_parser *ep);
 
+extern const char *gs_preprocessor_name(void);
+
 bool ep_parse(struct effect_parser *ep, effect_t effect,
               const char *effect_string, const char *file)
 {
 	bool success;
 
+	const char *graphics_preprocessor = gs_preprocessor_name();
+
+	if (graphics_preprocessor) {
+		struct cf_def def;
+
+		cf_def_init(&def);
+		def.name.str.array = graphics_preprocessor;
+		def.name.str.len   = strlen(graphics_preprocessor);
+
+		strref_copy(&def.name.unmerged_str, &def.name.str);
+		cf_preprocessor_add_def(&ep->cfp.pp, &def);
+	}
+
 	ep->effect = effect;
 	if (!cf_parser_parse(&ep->cfp, effect_string, file))
 		return false;

+ 1 - 0
libobs/graphics/graphics-imports.c

@@ -40,6 +40,7 @@ bool load_graphics_imports(struct gs_exports *exports, void *module,
 {
 	bool success = true;
 
+	GRAPHICS_IMPORT(device_preprocessor_name);
 	GRAPHICS_IMPORT(device_create);
 	GRAPHICS_IMPORT(device_destroy);
 	GRAPHICS_IMPORT(device_entercontext);

+ 1 - 0
libobs/graphics/graphics-internal.h

@@ -24,6 +24,7 @@
 #include "matrix4.h"
 
 struct gs_exports {
+	const char *(*device_preprocessor_name)(void);
 	device_t (*device_create)(struct gs_init_data *data);
 	void (*device_destroy)(device_t device);
 	void (*device_entercontext)(device_t device);

+ 6 - 0
libobs/graphics/graphics.c

@@ -865,6 +865,12 @@ void gs_perspective(float angle, float aspect, float near, float far)
 
 /* ------------------------------------------------------------------------- */
 
+const char *gs_preprocessor_name(void)
+{
+	graphics_t graphics = thread_graphics;
+	return graphics->exports.device_preprocessor_name();
+}
+
 swapchain_t gs_create_swapchain(struct gs_init_data *data)
 {
 	graphics_t graphics = thread_graphics;

+ 12 - 0
libobs/obs-internal.h

@@ -87,11 +87,14 @@ struct obs_core_video {
 	stagesurf_t                     copy_surfaces[NUM_TEXTURES];
 	texture_t                       render_textures[NUM_TEXTURES];
 	texture_t                       output_textures[NUM_TEXTURES];
+	texture_t                       convert_textures[NUM_TEXTURES];
 	bool                            textures_rendered[NUM_TEXTURES];
 	bool                            textures_output[NUM_TEXTURES];
 	bool                            textures_copied[NUM_TEXTURES];
+	bool                            textures_converted[NUM_TEXTURES];
 	struct source_frame             convert_frames[NUM_TEXTURES];
 	effect_t                        default_effect;
+	effect_t                        conversion_effect;
 	stagesurf_t                     mapped_surface;
 	int                             cur_texture;
 
@@ -99,6 +102,15 @@ struct obs_core_video {
 	pthread_t                       video_thread;
 	bool                            thread_initialized;
 
+	bool                            gpu_conversion;
+	const char                      *conversion_tech;
+	uint32_t                        conversion_height;
+	uint32_t                        plane_offsets[3];
+	uint32_t                        plane_sizes[3];
+	uint32_t                        plane_linewidth[3];
+
+	uint32_t                        output_width;
+	uint32_t                        output_height;
 	uint32_t                        base_width;
 	uint32_t                        base_height;
 

+ 160 - 4
libobs/obs-video.c

@@ -64,6 +64,7 @@ static inline void render_displays(void)
 static inline void set_render_size(uint32_t width, uint32_t height)
 {
 	gs_enable_depthtest(false);
+	gs_enable_blending(false);
 	gs_setcullmode(GS_NEITHER);
 
 	gs_ortho(0.0f, (float)width, 0.0f, (float)height, -100.0f, 100.0f);
@@ -137,17 +138,79 @@ static inline void render_output_texture(struct obs_core_video *video,
 	video->textures_output[cur_texture] = true;
 }
 
-static inline void stage_output_texture(struct obs_core_video *video,
+static inline void set_eparam(effect_t effect, const char *name, float val)
+{
+	eparam_t param = effect_getparambyname(effect, name);
+	effect_setfloat(effect, param, val);
+}
+
+static void render_convert_texture(struct obs_core_video *video,
 		int cur_texture, int prev_texture)
 {
 	texture_t   texture = video->output_textures[prev_texture];
-	stagesurf_t copy    = video->copy_surfaces[cur_texture];
+	texture_t   target  = video->convert_textures[cur_texture];
+	float       fwidth  = (float)video->output_width;
+	float       fheight = (float)video->output_height;
+	size_t      passes, i;
 
-	unmap_last_surface(video);
+	effect_t    effect  = video->conversion_effect;
+	eparam_t    image   = effect_getparambyname(effect, "image");
+	technique_t tech    = effect_gettechnique(effect,
+			video->conversion_tech);
 
 	if (!video->textures_output[prev_texture])
 		return;
 
+	set_eparam(effect, "u_plane_offset", (float)video->plane_offsets[1]);
+	set_eparam(effect, "v_plane_offset", (float)video->plane_offsets[2]);
+	set_eparam(effect, "width",  fwidth);
+	set_eparam(effect, "height", fheight);
+	set_eparam(effect, "width_i",  1.0f / fwidth);
+	set_eparam(effect, "height_i", 1.0f / fheight);
+	set_eparam(effect, "width_d2",  fwidth  * 0.5f);
+	set_eparam(effect, "height_d2", fheight * 0.5f);
+	set_eparam(effect, "width_d2_i",  1.0f / (fwidth  * 0.5f));
+	set_eparam(effect, "height_d2_i", 1.0f / (fheight * 0.5f));
+	set_eparam(effect, "input_width", fwidth);
+	set_eparam(effect, "input_height", (float)video->conversion_height);
+
+	effect_settexture(effect, image, texture);
+
+	gs_setrendertarget(target, NULL);
+	set_render_size(video->output_width, video->conversion_height);
+
+	passes = technique_begin(tech);
+	for (i = 0; i < passes; i++) {
+		technique_beginpass(tech, i);
+		gs_draw_sprite(texture, 0, video->output_width,
+				video->conversion_height);
+		technique_endpass(tech);
+	}
+	technique_end(tech);
+
+	video->textures_converted[cur_texture] = true;
+}
+
+static inline void stage_output_texture(struct obs_core_video *video,
+		int cur_texture, int prev_texture)
+{
+	texture_t   texture;
+	bool        texture_ready;
+	stagesurf_t copy = video->copy_surfaces[cur_texture];
+
+	if (video->gpu_conversion) {
+		texture = video->convert_textures[prev_texture];
+		texture_ready = video->textures_converted[prev_texture];
+	} else {
+		texture = video->output_textures[prev_texture];
+		texture_ready = video->output_textures[prev_texture];
+	}
+
+	unmap_last_surface(video);
+
+	if (!texture_ready)
+		return;
+
 	gs_stage_texture(copy, texture);
 
 	video->textures_copied[cur_texture] = true;
@@ -163,9 +226,13 @@ static inline void render_video(struct obs_core_video *video, int cur_texture,
 
 	render_main_texture(video, cur_texture);
 	render_output_texture(video, cur_texture, prev_texture);
+	if (video->gpu_conversion)
+		render_convert_texture(video, cur_texture, prev_texture);
+
 	stage_output_texture(video, cur_texture, prev_texture);
 
 	gs_setrendertarget(NULL, NULL);
+	gs_enable_blending(true);
 
 	gs_endscene();
 }
@@ -186,6 +253,90 @@ static inline bool download_frame(struct obs_core_video *video,
 	return true;
 }
 
+static inline uint32_t calc_linesize(uint32_t pos, uint32_t linesize)
+{
+	uint32_t size = pos % linesize;
+	return size ? size : linesize;
+}
+
+static void copy_dealign(
+		uint8_t *dst, uint32_t dst_pos, uint32_t dst_linesize,
+		const uint8_t *src, uint32_t src_pos, uint32_t src_linesize,
+		uint32_t remaining)
+{
+	while (remaining) {
+		uint32_t src_remainder = src_pos % src_linesize;
+		uint32_t dst_offset = dst_linesize - src_remainder;
+		uint32_t src_offset = src_linesize - src_remainder;
+
+		if (remaining < dst_offset) {
+			memcpy(dst + dst_pos, src + src_pos, remaining);
+			src_pos += remaining;
+			dst_pos += remaining;
+			remaining = 0;
+		} else {
+			memcpy(dst + dst_pos, src + src_pos, dst_offset);
+			src_pos += src_offset;
+			dst_pos += dst_offset;
+			remaining -= dst_offset;
+		}
+	}
+}
+
+static inline uint32_t make_aligned_linesize_offset(uint32_t offset,
+		uint32_t dst_linesize, uint32_t src_linesize)
+{
+	uint32_t remainder = offset % dst_linesize;
+	return (offset / dst_linesize) * src_linesize + remainder;
+}
+
+static void fix_gpu_converted_alignment(struct obs_core_video *video,
+		struct video_frame *frame, int cur_texture)
+{
+	struct source_frame *new_frame = &video->convert_frames[cur_texture];
+	uint32_t src_linesize = frame->linesize[0];
+	uint32_t dst_linesize = video->output_width * 4;
+	uint32_t src_pos      = 0;
+
+	for (size_t i = 0; i < 3; i++) {
+		if (video->plane_linewidth[i] == 0)
+			break;
+
+		src_pos = make_aligned_linesize_offset(video->plane_offsets[i],
+				dst_linesize, src_linesize);
+
+		copy_dealign(new_frame->data[i], 0, dst_linesize,
+				frame->data[0], src_pos, src_linesize,
+				video->plane_sizes[i]);
+	}
+
+	/* replace with cached frames */
+	for (size_t i = 0; i < MAX_AV_PLANES; i++) {
+		frame->data[i]     = new_frame->data[i];
+		frame->linesize[i] = new_frame->linesize[i];
+	}
+}
+
+static bool set_gpu_converted_data(struct obs_core_video *video,
+		struct video_frame *frame, int cur_texture)
+{
+	if (frame->linesize[0] == video->output_width*4) {
+		for (size_t i = 0; i < 3; i++) {
+			if (video->plane_linewidth[i] == 0)
+				break;
+
+			frame->linesize[i] = video->plane_linewidth[i];
+			frame->data[i] =
+				frame->data[0] + video->plane_offsets[i];
+		}
+
+	} else {
+		fix_gpu_converted_alignment(video, frame, cur_texture);
+	}
+
+	return true;
+}
+
 static bool convert_frame(struct obs_core_video *video,
 		struct video_frame *frame,
 		const struct video_output_info *info, int cur_texture)
@@ -223,9 +374,14 @@ static inline void output_video_data(struct obs_core_video *video,
 	const struct video_output_info *info;
 	info = video_output_getinfo(video->video);
 
-	if (format_is_yuv(info->format))
+	if (video->gpu_conversion) {
+		if (!set_gpu_converted_data(video, frame, cur_texture))
+			return;
+
+	} else if (format_is_yuv(info->format)) {
 		if (!convert_frame(video, frame, info, cur_texture))
 			return;
+	}
 
 	video_output_frame(video->video, frame);
 }

+ 101 - 8
libobs/obs.c

@@ -47,16 +47,92 @@ static inline void make_video_info(struct video_output_info *vi,
 	vi->height  = ovi->output_height;
 }
 
+#define PIXEL_SIZE 4
+
+#define GET_ALIGN(val, align) \
+	(((val) + (align-1)) & ~(align-1))
+
+static inline void set_420p_sizes(const struct obs_video_info *ovi)
+{
+	struct obs_core_video *video = &obs->video;
+	uint32_t chroma_pixels;
+	uint32_t total_bytes;
+
+	chroma_pixels = (ovi->output_width * ovi->output_height / 4);
+	chroma_pixels = GET_ALIGN(chroma_pixels, PIXEL_SIZE);
+
+	video->plane_offsets[0] = 0;
+	video->plane_offsets[1] = ovi->output_width * ovi->output_height;
+	video->plane_offsets[2] = video->plane_offsets[1] + chroma_pixels;
+
+	video->plane_linewidth[0] = ovi->output_width;
+	video->plane_linewidth[1] = ovi->output_width/2;
+	video->plane_linewidth[2] = ovi->output_width/2;
+
+	video->plane_sizes[0] = video->plane_offsets[1];
+	video->plane_sizes[1] = video->plane_sizes[0]/4;
+	video->plane_sizes[2] = video->plane_sizes[1];
+
+	total_bytes = video->plane_offsets[2] + chroma_pixels;
+
+	video->conversion_height =
+		(total_bytes/PIXEL_SIZE + ovi->output_width-1) /
+		ovi->output_width;
+
+	video->conversion_height = GET_ALIGN(video->conversion_height, 2);
+	video->conversion_tech = "Planar420";
+}
+
+static inline void calc_gpu_conversion_sizes(const struct obs_video_info *ovi)
+{
+	obs->video.conversion_height = 0;
+	memset(obs->video.plane_offsets, 0, sizeof(obs->video.plane_offsets));
+	memset(obs->video.plane_sizes, 0, sizeof(obs->video.plane_sizes));
+	memset(obs->video.plane_linewidth, 0,
+		sizeof(obs->video.plane_linewidth));
+
+	switch ((uint32_t)ovi->output_format) {
+	case VIDEO_FORMAT_I420:
+		set_420p_sizes(ovi);
+	}
+}
+
+static bool obs_init_gpu_conversion(struct obs_video_info *ovi)
+{
+	struct obs_core_video *video = &obs->video;
+
+	calc_gpu_conversion_sizes(ovi);
+
+	if (!video->conversion_height) {
+		blog(LOG_INFO, "GPU conversion not available for format: %u",
+				(unsigned int)ovi->output_format);
+		video->gpu_conversion = false;
+		return true;
+	}
+
+	for (size_t i = 0; i < NUM_TEXTURES; i++) {
+		video->convert_textures[i] = gs_create_texture(
+				ovi->output_width, video->conversion_height,
+				GS_RGBA, 1, NULL, GS_RENDERTARGET);
+
+		if (!video->convert_textures[i])
+			return false;
+	}
+
+	return true;
+}
+
 static bool obs_init_textures(struct obs_video_info *ovi)
 {
 	struct obs_core_video *video = &obs->video;
 	bool yuv = format_is_yuv(ovi->output_format);
+	uint32_t output_height = video->gpu_conversion ?
+		video->conversion_height : ovi->output_height;
 	size_t i;
 
 	for (i = 0; i < NUM_TEXTURES; i++) {
 		video->copy_surfaces[i] = gs_create_stagesurface(
-				ovi->output_width, ovi->output_height,
-				GS_RGBA);
+				ovi->output_width, output_height, GS_RGBA);
 
 		if (!video->copy_surfaces[i])
 			return false;
@@ -92,8 +168,12 @@ static bool obs_init_graphics(struct obs_video_info *ovi)
 	int errorcode;
 
 	make_gs_init_data(&graphics_data, ovi);
-	video->base_width  = ovi->base_width;
-	video->base_height = ovi->base_height;
+	video->base_width    = ovi->base_width;
+	video->base_height   = ovi->base_height;
+	video->output_width  = ovi->output_width;
+	video->output_height = ovi->output_height;
+
+	video->gpu_conversion = ovi->gpu_conversion;
 
 	errorcode = gs_create(&video->graphics, ovi->graphics_module,
 			&graphics_data);
@@ -106,7 +186,9 @@ static bool obs_init_graphics(struct obs_video_info *ovi)
 
 	gs_entercontext(video->graphics);
 
-	if (!obs_init_textures(ovi))
+	if (ovi->gpu_conversion && !obs_init_gpu_conversion(ovi))
+		success = false;
+	if (success && !obs_init_textures(ovi))
 		success = false;
 
 	if (success) {
@@ -114,8 +196,16 @@ static bool obs_init_graphics(struct obs_video_info *ovi)
 		video->default_effect = gs_create_effect_from_file(filename,
 				NULL);
 		bfree(filename);
+
+		filename = find_libobs_data_file("format_conversion.effect");
+		video->conversion_effect = gs_create_effect_from_file(filename,
+				NULL);
+		bfree(filename);
+
 		if (!video->default_effect)
 			success = false;
+		if (!video->conversion_effect)
+			success = false;
 	}
 
 	gs_leavecontext();
@@ -195,15 +285,18 @@ static void obs_free_graphics(void)
 		for (i = 0; i < NUM_TEXTURES; i++) {
 			stagesurface_destroy(video->copy_surfaces[i]);
 			texture_destroy(video->render_textures[i]);
+			texture_destroy(video->convert_textures[i]);
 			texture_destroy(video->output_textures[i]);
 			source_frame_free(&video->convert_frames[i]);
 
-			video->copy_surfaces[i]   = NULL;
-			video->render_textures[i] = NULL;
-			video->output_textures[i] = NULL;
+			video->copy_surfaces[i]    = NULL;
+			video->render_textures[i]  = NULL;
+			video->convert_textures[i] = NULL;
+			video->output_textures[i]  = NULL;
 		}
 
 		effect_destroy(video->default_effect);
+		effect_destroy(video->conversion_effect);
 		video->default_effect = NULL;
 
 		gs_leavecontext();

+ 3 - 0
libobs/obs.h

@@ -118,6 +118,9 @@ struct obs_video_info {
 	uint32_t            adapter;
 
 	struct gs_window    window;        /**< Window to render to */
+
+	/** Use shaders to convert to different color formats */
+	bool                gpu_conversion;
 };
 
 /**

+ 1 - 1
libobs/util/cf-lexer.c

@@ -1295,7 +1295,7 @@ void cf_preprocessor_add_def(struct cf_preprocessor *pp, struct cf_def *def)
 		cf_def_free(existing);
 		memcpy(existing, def, sizeof(struct cf_def));
 	} else {
-		da_push_back(pp->defines, &def);
+		da_push_back(pp->defines, def);
 	}
 }
 

+ 7 - 6
obs/window-basic-main.cpp

@@ -282,16 +282,17 @@ bool OBSBasic::InitGraphics()
 	App()->GetConfigFPS(ovi.fps_num, ovi.fps_den);
 
 	ovi.graphics_module = App()->GetRenderModule();
-	ovi.base_width    = (uint32_t)config_get_uint(GetGlobalConfig(),
+	ovi.base_width     = (uint32_t)config_get_uint(GetGlobalConfig(),
 			"Video", "BaseCX");
-	ovi.base_height   = (uint32_t)config_get_uint(GetGlobalConfig(),
+	ovi.base_height    = (uint32_t)config_get_uint(GetGlobalConfig(),
 			"Video", "BaseCY");
-	ovi.output_width  = (uint32_t)config_get_uint(GetGlobalConfig(),
+	ovi.output_width   = (uint32_t)config_get_uint(GetGlobalConfig(),
 			"Video", "OutputCX");
-	ovi.output_height = (uint32_t)config_get_uint(GetGlobalConfig(),
+	ovi.output_height  = (uint32_t)config_get_uint(GetGlobalConfig(),
 			"Video", "OutputCY");
-	ovi.output_format = VIDEO_FORMAT_I420;
-	ovi.adapter       = 0;
+	ovi.output_format  = VIDEO_FORMAT_I420;
+	ovi.adapter        = 0;
+	ovi.gpu_conversion = true;
 
 	QTToGSWindow(ui->preview, ovi.window);
 

+ 0 - 1
plugins/obs-ffmpeg/obs-ffmpeg-output.c

@@ -20,7 +20,6 @@
 
 #include <libavformat/avformat.h>
 #include <libswscale/swscale.h>
-#include <libavutil/frame.h>
 
 struct ffmpeg_data {
 	AVStream           *video;

+ 3 - 0
vs/2013/OBS.sln

@@ -39,6 +39,8 @@ Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "obs-ffmpeg", "obs-ffmpeg\ob
 EndProject
 Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "obs-studio", "obs-studio\obs-studio.vcxproj", "{B12702AD-ABFB-343A-A199-8E24837244A3}"
 EndProject
+Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Solution Items", "Solution Items", "{6AE25DF7-A903-48AD-8DA4-9902010AEB11}"
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Win32 = Debug|Win32
@@ -102,6 +104,7 @@ Global
 		{76226D20-1972-4789-A595-EDACC7A76DC3}.Release|Win32.ActiveCfg = Release|Win32
 		{76226D20-1972-4789-A595-EDACC7A76DC3}.Release|Win32.Build.0 = Release|Win32
 		{76226D20-1972-4789-A595-EDACC7A76DC3}.Release|x64.ActiveCfg = Release|Win32
+		{76226D20-1972-4789-A595-EDACC7A76DC3}.Release|x64.Build.0 = Release|Win32
 		{36970254-B1E5-4BE6-A561-301B6E7CDCE3}.Debug|Win32.ActiveCfg = Debug|Win32
 		{36970254-B1E5-4BE6-A561-301B6E7CDCE3}.Debug|Win32.Build.0 = Debug|Win32
 		{36970254-B1E5-4BE6-A561-301B6E7CDCE3}.Debug|x64.ActiveCfg = Debug|x64