diff --git a/wled00/FX.cpp b/wled00/FX.cpp
index 34cd661e..c763177c 100644
--- a/wled00/FX.cpp
+++ b/wled00/FX.cpp
@@ -10532,8 +10532,6 @@ uint16_t mode_particlegalaxy(void) {
   // Particle System settings
   PartSys->updateSystem(); // update system properties (dimensions and data pointers)
   uint8_t particlesize = SEGMENT.custom1;
-  if(SEGMENT.check3)
-    particlesize =  SEGMENT.custom1 ? 1 : 0; // set size to 0 (single pixel) or 1 (quad pixel) so motion blur works and adds streaks
   PartSys->setParticleSize(particlesize); // set size globally
   PartSys->setMotionBlur(250 * SEGMENT.check3); // adds trails to single/quad pixel particles, no effect if size > 1
 
@@ -10601,7 +10599,7 @@ uint16_t mode_particlegalaxy(void) {
   PartSys->update(); // update and render
   return FRAMETIME;
 }
-static const char _data_FX_MODE_PARTICLEGALAXY[] PROGMEM = "PS Galaxy@!,!,Size,,Color,,Starfield,Trace;;!;2;pal=59,sx=80,c1=2,c3=4";
+static const char _data_FX_MODE_PARTICLEGALAXY[] PROGMEM = "PS Galaxy@!,!,Size,,Color,,Starfield,Trace;;!;2;pal=59,sx=80,c1=1,c3=4";
 
 #endif //WLED_DISABLE_PARTICLESYSTEM2D
 #endif // WLED_DISABLE_2D
diff --git a/wled00/FX_2Dfcn.cpp b/wled00/FX_2Dfcn.cpp
index 1e7505cb..d12153d5 100644
--- a/wled00/FX_2Dfcn.cpp
+++ b/wled00/FX_2Dfcn.cpp
@@ -368,7 +368,7 @@ void IRAM_ATTR_YN Segment::setPixelColorXY(int x, int y, uint32_t col) //WLEDMM:
 
 // WLEDMM setPixelColorXY(float x, float y, uint32_t col, ..) is depricated. use wu_pixel(x,y,col) instead.
 // anti-aliased version of setPixelColorXY()
-void Segment::setPixelColorXY(float x, float y, uint32_t col, bool aa, bool fast) // WLEDMM some speedups due to fast int and faster sqrt16
+void Segment::setPixelColorXY(float x, float y, uint32_t col, bool aa, bool fast) // WLEDMM some speedups due to fast int and faster sqrt32_bw
 {
   if (Segment::maxHeight==1) return; // not a matrix set-up
   if (x<0.0f || x>1.0f || y<0.0f || y>1.0f) return; // not normalized
@@ -400,10 +400,10 @@ void Segment::setPixelColorXY(float x, float y, uint32_t col, bool aa, bool fast
         setPixelColorXY(xL, yB, color_blend(col, cXLYB, uint8_t(sqrtf(dL*dB)*255.0f))); // blend BL pixel
         setPixelColorXY(xR, yB, color_blend(col, cXRYB, uint8_t(sqrtf(dR*dB)*255.0f))); // blend BR pixel
       } else {
-        setPixelColorXY(xL, yT, color_blend(col, cXLYT, uint8_t(sqrt16(dL*dT*65025.0f)))); // blend TL pixel     // WLEDMM: use faster sqrt16 for integer; perform multiplication by 255^2 before sqrt
-        setPixelColorXY(xR, yT, color_blend(col, cXRYT, uint8_t(sqrt16(dR*dT*65025.0f)))); // blend TR pixel     //         this is possible because sqrt(a) * sqrt(b)  =  sqrt(a * b)
-        setPixelColorXY(xL, yB, color_blend(col, cXLYB, uint8_t(sqrt16(dL*dB*65025.0f)))); // blend BL pixel
-        setPixelColorXY(xR, yB, color_blend(col, cXRYB, uint8_t(sqrt16(dR*dB*65025.0f)))); // blend BR pixel
+        setPixelColorXY(xL, yT, color_blend(col, cXLYT, uint8_t(sqrt32_bw(dL*dT*65025.0f)))); // blend TL pixel     // WLEDMM: use faster sqrt32_bw for integer; perform multiplication by 255^2 before sqrt
+        setPixelColorXY(xR, yT, color_blend(col, cXRYT, uint8_t(sqrt32_bw(dR*dT*65025.0f)))); // blend TR pixel     //         this is possible because sqrt(a) * sqrt(b)  =  sqrt(a * b)
+        setPixelColorXY(xL, yB, color_blend(col, cXLYB, uint8_t(sqrt32_bw(dL*dB*65025.0f)))); // blend BL pixel
+        setPixelColorXY(xR, yB, color_blend(col, cXRYB, uint8_t(sqrt32_bw(dR*dB*65025.0f)))); // blend BR pixel
       }
     } else if (xR!=xL && yT==yB) {
       setPixelColorXY(xR, yT, color_blend(col, cXLYT, uint8_t(dL*255.0f))); // blend L pixel