shave some cycles (skip if copymask=0, fake bit 3)

src/fx/fx.dhgr.precomputed.1bit.a

@@ -15,19 +15,15 @@
          inx
          bne   -
 
-         lda   #%10000001
-         sta   .copymasks+$60
-         sta   .mirror_copymasks+$60
-
-         lda   #%10000110
+         lda   #%00000111
          sta   .copymasks+$80
          sta   .mirror_copymasks+$40
 
-         lda   #%10011000
+         lda   #%00011000
          sta   .copymasks+$A0
          sta   .mirror_copymasks+$20
 
-         lda   #%11100000
+         lda   #%01100000
          sta   .copymasks+$C0
          sta   .mirror_copymasks
 
@@ -81,50 +77,58 @@ input=*+1
          sty   <y
          clc
 bankloop
+         lda   copymasks, x
+         beq   +
+         sta   <copymask
+
          ; main 1x2 block in top-left quadrant
 src1=*+1
          lda   $FDFD, y
          eor   (<dest1), y
-         and   copymasks, x
+copymask=*+1
+         and   #$FD                  ; SMC
          eor   (<dest1), y
 dest1=*+1
          sta   $FDFD, y
 src2=*+1
          lda   $FDFD, y
          eor   (<dest2), y
-         and   copymasks, x
+         and   <copymask
          eor   (<dest2), y
 dest2=*+1
          sta   $FDFD, y
 
-         ; corresponding 1x2 block in top-right quadrant (same row, opposite column)
+         ; corresponding 1x2 block in bottom-left quadrant (opposite row, original column)
+         +COPY_BIT_ZP mirror_src1, mirror_dest1, copymask
+         +COPY_BIT_ZP mirror_src2, mirror_dest2, copymask
+
++
+         lda   mirror_copymasks, x
+         beq   +
+         sta   <mirror_copymask
+         ; corresponding 1x2 block in bottom-right quadrant (opposite row, opposite column)
          lda   mirror_cols, y
          tay
-         +COPY_BIT src1, dest1, mirror_copymasks
-         +COPY_BIT src2, dest2, mirror_copymasks
-
-         ; corresponding 1x2 block in bottom-right quadrant (opposite row, opposite column)
 mirror_src1=*+1
          lda   $FDFD, y
          eor   (<mirror_dest1), y
-         and   mirror_copymasks, x
+mirror_copymask=*+1
+         and   #$FD                             ; SMC
          eor   (<mirror_dest1), y
 mirror_dest1=*+1
          sta   $FDFD, y
 mirror_src2=*+1
          lda   $FDFD, y
          eor   (<mirror_dest2), y
-         and   mirror_copymasks, x
+         and   <mirror_copymask
          eor   (<mirror_dest2), y
 mirror_dest2=*+1
          sta   $FDFD, y
 
-         ; corresponding 1x2 block in bottom-left quadrant (opposite row, original column)
-         lda   mirror_cols, y
-         tay
-         +COPY_BIT mirror_src1, mirror_dest1, copymasks
-         +COPY_BIT mirror_src2, mirror_dest2, copymasks
-
+         ; corresponding 1x2 block in top-right quadrant (same row, opposite column)
+         +COPY_BIT_ZP src1, dest1, mirror_copymask
+         +COPY_BIT_ZP src2, dest2, mirror_copymask
++
          bcs   +
          sta   $C003
          sta   $C005

src/fx/macros.a

@@ -124,6 +124,14 @@
          sta   (.dest1),y            ; write the result
 }
 
+!macro COPY_BIT_ZP .src1, .dest1, .zpcopymask {
+         lda   (.src1),y
+         eor   (.dest1),y            ; merge source and destination bits
+         and   <.zpcopymask          ; isolate the bits to replace, zero the rest
+         eor   (.dest1),y            ; unmerge source and destination bits, leaves 'to keep' destination bits intact
+         sta   (.dest1),y            ; write the result
+}
+
 !macro SWITCH_TO_MASKS .copy {
          lda   #<.copy
          sta   CopyMaskAddr