it really looks too tight :(

dither4.s

@@ -156,13 +156,15 @@ each_scanline:
         pha
     .endmacro
 
-    fake:
-    lda #18
-    pha
-    pha
+    .macro audio_play
+        ldy #0
+        lda (sample_ptr),y
+        sta AUDC1
+    .endmacro
 
     .macro inner_scanline
         ldy scanline
+        inc scanline
 
         ; Leisurely memory fetches
         lda palette1,y
@@ -181,21 +183,21 @@ each_scanline:
         sty COLPF2
         
         ; Audio sample
-        pla
-        sta AUDC1
-
-        inc scanline
+        ;pla
+        ;lda temp1
+        ;sta AUDC1
     .endmacro
 
+    audio_play
     inner_scanline
 
     .macro audio_inc
+    ;; toooo slllooowwww
         ; Increment sample ptr
-        jmp audio_cont
         inc sample_ptrl
         bne audio_cont
-        inc sample_ptrh
 
+        inc sample_ptrh
         lda sample_ptrh
         cmp #.hibyte(audio_samples_end)
         bne audio_cont
@@ -208,10 +210,12 @@ each_scanline:
         sta sample_ptrl
         lda #.hibyte(audio_samples)
         sta sample_ptrh
-    .endmacro
 
         audio_cont:
+    .endmacro
 
+
+    ;audio_inc
     inner_scanline
 
     lda VCOUNT

pack-wav.js

@@ -15,7 +15,7 @@ function byte2byte(arr) {
 }
 
 function to4bit(val8) {
-    let val = val8 + 7 >> 4;
+    let val = (val8 + 7) >> 4;
     if (val > 15) {
         return 15;
     }
@@ -24,12 +24,19 @@ function to4bit(val8) {
 
 function pack(audio) {
     let packed = [];
+    /*
     for (let i = 0; i < audio.length; i += 2) {
         let low = to4bit(audio[i]);
         let high = to4bit(audio[i + 1]);
         let byte = low | (high << 4);
         packed.push(byte);
     }
+    */
+    for (let i = 0; i < audio.length; i += 2) {
+        let val = to4bit(audio[i]);
+        let byte = val | 0x10;
+        packed.push(byte);
+    }
     return packed;
 }