From bb7e454ce6f4d1723a2e925744682b82a1c656e9 Mon Sep 17 00:00:00 2001
From: Aras Pranckevicius <aras@nesnausk.org>
Date: Fri, 12 Jan 2024 12:33:20 +0200
Subject: [PATCH] VSE: speed up solid color effect

It's pretty simple, but threading it, and making it write out whole
pixel at a time (instead of one byte at a time) still makes it faster.

4K resolution, five Color strips blended over each other, playback
on Windows/VS2022, Ryzen 5950X:
- Playback 9.2FPS -> 11.5FPS
- do_solid_color for one effect, median time 7.7ms -> 3.8ms

Additionally, the solid color on byte output was not doing float->byte
color rounding & clamping properly, and on float output it was writing
255.0 into alpha instead of 1.0. So fix that too.
---
 source/blender/sequencer/intern/effects.cc | 60 +++++++++-------------
 1 file changed, 25 insertions(+), 35 deletions(-)

diff --git a/source/blender/sequencer/intern/effects.cc b/source/blender/sequencer/intern/effects.cc
index f8ef6a668fc..be57c036372 100644
--- a/source/blender/sequencer/intern/effects.cc
+++ b/source/blender/sequencer/intern/effects.cc
@@ -1925,51 +1925,41 @@ static ImBuf *do_solid_color(const SeqRenderData *context,
                              ImBuf *ibuf2,
                              ImBuf *ibuf3)
 {
+  using namespace blender;
   ImBuf *out = prepare_effect_imbufs(context, ibuf1, ibuf2, ibuf3);
 
   SolidColorVars *cv = (SolidColorVars *)seq->effectdata;
 
-  int x = out->x;
-  int y = out->y;
+  threading::parallel_for(IndexRange(out->y), 64, [&](const IndexRange y_range) {
+    if (out->byte_buffer.data) {
+      /* Byte image. */
+      uchar color[4];
+      rgb_float_to_uchar(color, cv->col);
+      color[3] = 255;
 
-  if (out->byte_buffer.data) {
-    uchar color[4];
-    color[0] = cv->col[0] * 255;
-    color[1] = cv->col[1] * 255;
-    color[2] = cv->col[2] * 255;
-    color[3] = 255;
-
-    uchar *rect = out->byte_buffer.data;
-
-    for (int i = 0; i < y; i++) {
-      for (int j = 0; j < x; j++) {
-        rect[0] = color[0];
-        rect[1] = color[1];
-        rect[2] = color[2];
-        rect[3] = color[3];
-        rect += 4;
+      uchar *dst = out->byte_buffer.data + y_range.first() * out->x * 4;
+      uchar *dst_end = dst + y_range.size() * out->x * 4;
+      while (dst < dst_end) {
+        memcpy(dst, color, sizeof(color));
+        dst += 4;
       }
     }
-  }
-  else if (out->float_buffer.data) {
-    float color[4];
-    color[0] = cv->col[0];
-    color[1] = cv->col[1];
-    color[2] = cv->col[2];
-    color[3] = 255;
+    else {
+      /* Float image. */
+      float color[4];
+      color[0] = cv->col[0];
+      color[1] = cv->col[1];
+      color[2] = cv->col[2];
+      color[3] = 1.0f;
 
-    float *rect_float = out->float_buffer.data;
-
-    for (int i = 0; i < y; i++) {
-      for (int j = 0; j < x; j++) {
-        rect_float[0] = color[0];
-        rect_float[1] = color[1];
-        rect_float[2] = color[2];
-        rect_float[3] = color[3];
-        rect_float += 4;
+      float *dst = out->float_buffer.data + y_range.first() * out->x * 4;
+      float *dst_end = dst + y_range.size() * out->x * 4;
+      while (dst < dst_end) {
+        memcpy(dst, color, sizeof(color));
+        dst += 4;
       }
     }
-  }
+  });
 
   out->planes = R_IMF_PLANES_RGB;
 
-- 
2.30.2