Alpha premul pipeline cleanup

This assumptions are now made:
- Internally float buffers are always linear alpha-premul colors
- Readers should worry about delivering float buffers with that
  assumptions.
- There's an input image setting to say whether it's stored with
  straight/premul alpha on the disk.
- Byte buffers are now assumed have straight alpha, readers should
  deliver straight alpha.

Some implementation details:

- Removed scene's color unpremultiply setting, which was very
  much confusing and was wrong for default settings.
  Now all renderers assumes to deliver premultiplied alpha.

- IMB_buffer_byte_from_float will now linearize alpha when
  converting from buffer.

- Sequencer's effects were changed to assume bytes have got
  straight alpha. Most of effects will work with bytes still,
  however for glow it was more tricky to avoid data loss, so
  there's a commented out glow implementation which converts
  byte buffer to floats first, operates on floats and returns
  bytes back. It's slower and not sure if it should actually
  be used -- who're using glow on alpha anyway?

- Sequencer modifiers should also be working nice with straight
  bytes now.

- GLSL preview will predivide float textures to make nice shading,
  shading with byte textures worked nice (GLSL was assuming straight
  alpha).

- Blender Internal will set alpha=1 to the whole sky. The same
  happens in Cycles and there's no way to avoid this -- sky is
  neither straight nor premul and doesn't fit color pipeline well.

- Straight alpha mode for render result was also eliminated.

- Conversion to correct alpha need to be done before linearizing
  float buffer.

- TIFF will now load and save files with proper alpha mode setting
  in file meta data header.

- Remove Use Alpha from texture mapping and replaced with image
  datablock setting.

  Behaves much more predictable and clear from code point of view
  and solves possible regressions when non-premultiplied images were
  used as textures with ignoring alpha channel.

1 files changed, 54 insertions, 27 deletions

diff --git a/source/blender/imbuf/intern/divers.c b/source/blender/imbuf/intern/divers.c
index f049c404e2d..f0d8b7cac72 100644
--- a/source/blender/imbuf/intern/divers.c
+++ b/source/blender/imbuf/intern/divers.c
@@ -39,6 +39,7 @@
 #include "IMB_imbuf_types.h"
 #include "IMB_imbuf.h"
 #include "IMB_allocimbuf.h"
+#include "IMB_filter.h"
 
 #include "IMB_colormanagement.h"
 #include "IMB_colormanagement_intern.h"
@@ -249,11 +250,25 @@ void IMB_buffer_byte_from_float(uchar *rect_to, const float *rect_from,
 			uchar *to = rect_to + stride_to * y * 4;
 
 			if (profile_to == profile_from) {
+				float straight[4];
+
 				/* no color space conversion */
-				if (dither) {
+				if (dither && predivide) {
+					for (x = 0; x < width; x++, from += 4, to += 4) {
+						premul_to_straight_v4(straight, from);
+						float_to_byte_dither_v4(to, straight, di);
+					}
+				}
+				else if (dither) {
 					for (x = 0; x < width; x++, from += 4, to += 4)
 						float_to_byte_dither_v4(to, from, di);
 				}
+				else if (predivide) {
+					for (x = 0; x < width; x++, from += 4, to += 4) {
+						premul_to_straight_v4(straight, from);
+						rgba_float_to_uchar(to, straight);
+					}
+				}
 				else {
 					for (x = 0; x < width; x++, from += 4, to += 4)
 						rgba_float_to_uchar(to, from);
@@ -262,10 +277,12 @@ void IMB_buffer_byte_from_float(uchar *rect_to, const float *rect_from,
 			else if (profile_to == IB_PROFILE_SRGB) {
 				/* convert from linear to sRGB */
 				unsigned short us[4];
+				float straight[4];
 
 				if (dither && predivide) {
 					for (x = 0; x < width; x++, from += 4, to += 4) {
-						linearrgb_to_srgb_ushort4_predivide(us, from);
+						premul_to_straight_v4(straight, from);
+						linearrgb_to_srgb_ushort4(us, from);
 						ushort_to_byte_dither_v4(to, us, di);
 					}
 				}
@@ -277,7 +294,8 @@ void IMB_buffer_byte_from_float(uchar *rect_to, const float *rect_from,
 				}
 				else if (predivide) {
 					for (x = 0; x < width; x++, from += 4, to += 4) {
-						linearrgb_to_srgb_ushort4_predivide(us, from);
+						premul_to_straight_v4(straight, from);
+						linearrgb_to_srgb_ushort4(us, from);
 						ushort_to_byte_v4(to, us);
 					}
 				}
@@ -526,7 +544,6 @@ void IMB_buffer_byte_from_byte(uchar *rect_to, const uchar *rect_from,
 
 void IMB_rect_from_float(ImBuf *ibuf)
 {
-	int predivide = (ibuf->flags & IB_cm_predivide);
 	float *buffer;
 	const char *from_colorspace;
 
@@ -548,7 +565,10 @@ void IMB_rect_from_float(ImBuf *ibuf)
 	buffer = MEM_dupallocN(ibuf->rect_float);
 
 	/* first make float buffer in byte space */
-	IMB_colormanagement_transform(buffer, ibuf->x, ibuf->y, ibuf->channels, from_colorspace, ibuf->rect_colorspace->name, predivide);
+	IMB_colormanagement_transform(buffer, ibuf->x, ibuf->y, ibuf->channels, from_colorspace, ibuf->rect_colorspace->name, TRUE);
+
+	/* convert from float's premul alpha to byte's straight alpha */
+	IMB_unpremultiply_rect_float(buffer, ibuf->planes, ibuf->x, ibuf->y);
 
 	/* convert float to byte */
 	IMB_buffer_byte_from_float((unsigned char *) ibuf->rect, buffer, ibuf->channels, ibuf->dither, IB_PROFILE_SRGB, IB_PROFILE_SRGB,
@@ -565,7 +585,6 @@ void IMB_partial_rect_from_float(ImBuf *ibuf, float *buffer, int x, int y, int w
 {
 	float *rect_float;
 	uchar *rect_byte;
-	int predivide = (ibuf->flags & IB_cm_predivide);
 	int profile_from = IB_PROFILE_LINEAR_RGB;
 
 	/* verify we have a float buffer */
@@ -588,12 +607,12 @@ void IMB_partial_rect_from_float(ImBuf *ibuf, float *buffer, int x, int y, int w
 
 		/* and do color space conversion to byte */
 		IMB_buffer_byte_from_float(rect_byte, rect_float,
-		                           4, ibuf->dither, IB_PROFILE_SRGB, profile_from, predivide,
+		                           4, ibuf->dither, IB_PROFILE_SRGB, profile_from, TRUE,
 		                           w, h, ibuf->x, w);
 	}
 	else {
 		IMB_buffer_float_from_float(buffer, rect_float,
-		                            ibuf->channels, IB_PROFILE_SRGB, profile_from, predivide,
+		                            ibuf->channels, IB_PROFILE_SRGB, profile_from, TRUE,
 		                            w, h, w, ibuf->x);
 
 		/* XXX: need to convert to image buffer's rect space */
@@ -608,8 +627,6 @@ void IMB_partial_rect_from_float(ImBuf *ibuf, float *buffer, int x, int y, int w
 
 void IMB_float_from_rect(ImBuf *ibuf)
 {
-	int predivide = (ibuf->flags & IB_cm_predivide);
-
 	/* verify if we byte and float buffers */
 	if (ibuf->rect == NULL)
 		return;
@@ -634,22 +651,12 @@ void IMB_float_from_rect(ImBuf *ibuf)
 
 	/* then make float be in linear space */
 	IMB_colormanagement_colorspace_to_scene_linear(ibuf->rect_float, ibuf->x, ibuf->y, ibuf->channels,
-	                                               ibuf->rect_colorspace, predivide);
-
-	BLI_unlock_thread(LOCK_COLORMANAGE);
-}
-
-/* no profile conversion */
-void IMB_float_from_rect_simple(ImBuf *ibuf)
-{
-	int predivide = (ibuf->flags & IB_cm_predivide);
+	                                               ibuf->rect_colorspace, FALSE);
 
-	if (ibuf->rect_float == NULL)
-		imb_addrectfloatImBuf(ibuf);
+	/* byte buffer is straight alpha, float should always be premul */
+	IMB_premultiply_rect_float(ibuf->rect_float, ibuf->planes, ibuf->x, ibuf->y);
 
-	IMB_buffer_float_from_byte(ibuf->rect_float, (uchar *)ibuf->rect,
-	                           IB_PROFILE_SRGB, IB_PROFILE_SRGB, predivide,
-	                           ibuf->x, ibuf->y, ibuf->x, ibuf->x);
+	BLI_unlock_thread(LOCK_COLORMANAGE);
 }
 
 /* use when you need to get a buffer with a certain profile
@@ -660,7 +667,6 @@ void IMB_float_from_rect_simple(ImBuf *ibuf)
  */
 float *IMB_float_profile_ensure(ImBuf *ibuf, int profile, int *alloc)
 {
-	int predivide = (ibuf->flags & IB_cm_predivide);
 	int profile_from = IB_PROFILE_LINEAR_RGB;
 	int profile_to;
 
@@ -686,12 +692,13 @@ float *IMB_float_profile_ensure(ImBuf *ibuf, int profile, int *alloc)
 
 		if (ibuf->rect_float == NULL) {
 			IMB_buffer_float_from_byte(fbuf, (uchar *)ibuf->rect,
-			                           profile_to, profile_from, predivide,
+			                           profile_to, profile_from, FALSE,
 			                           ibuf->x, ibuf->y, ibuf->x, ibuf->x);
+			IMB_premultiply_rect_float(ibuf->rect_float, ibuf->planes, ibuf->x, ibuf->y);
 		}
 		else {
 			IMB_buffer_float_from_float(fbuf, ibuf->rect_float,
-			                            4, profile_to, profile_from, predivide,
+			                            4, profile_to, profile_from, TRUE,
 			                            ibuf->x, ibuf->y, ibuf->x, ibuf->x);
 		}
 
@@ -727,6 +734,26 @@ void IMB_buffer_float_clamp(float *buf, int width, int height)
 	}
 }
 
+void IMB_buffer_float_unpremultiply(float *buf, int width, int height)
+{
+	int total = width * height;
+	float *cp = buf;
+	while (total--) {
+		premul_to_straight_v4(cp, cp);
+		cp += 4;
+	}
+}
+
+void IMB_buffer_float_premultiply(float *buf, int width, int height)
+{
+	int total = width * height;
+	float *cp = buf;
+	while (total--) {
+		straight_to_premul_v4(cp, cp);
+		cp += 4;
+	}
+}
+
 /**************************** alter saturation *****************************/
 
 void IMB_saturation(ImBuf *ibuf, float sat)