Session 1: hardware research, content measurement, codec decision, MAME harness

Verified GVRAM is one word-access per pixel in ALL color modes; chose 256-color
256x192 with movem.l bursts (page 1 sacrificed as double-buffer).

Measured 8 scenes from the Blu-ray source: blit costs under 8% of the 12fps
cycle budget, so I/O is the bottleneck, not CPU. Naive delta+RLE reaches only
3.2:1 (365 KB/s, 470MB) -> decision to use 4x4 vector quantization (~30 KB/s).

"Shot on twos" assumption failed: the transfer has zero duplicate frames, so
12fps requires explicit decimation.

Documents three false measurement results and their root causes (per-frame
Floyd-Steinberg dithering, temporal denoise, exact-match dedupe on noisy source).

MAME Lua injection harness works and is reusable for cycle-cost measurement;
the IOCS _B_READ disk benchmark is blocked returning -1.

Claude-Session: https://claude.ai/code/session_01194oWYW8DQXK1SZ2DnChW6

tools/bench/sweep.s

@@ -0,0 +1,34 @@
+FLAG    = $18000
+RESULT  = $18100
+BUF     = $20000
+        org     $10000
+start:
+        move.l  #1,FLAG.l
+        lea     RESULT,a2
+        moveq   #0,d4
+outer:
+        move.l  d4,d1              ; encoding A: PDA in bits 31-24
+        add.l   #$80,d1
+        swap    d1
+        lsl.l   #8,d1
+        bsr.s   doread
+        move.l  d0,(a2)+
+        move.l  d4,d1              ; encoding B: PDA in bits 15-8
+        add.l   #$80,d1
+        lsl.l   #8,d1
+        bsr.s   doread
+        move.l  d0,(a2)+
+        addq.l  #1,d4
+        cmpi.l  #16,d4
+        bne.s   outer
+        move.l  #$FF,FLAG.l
+stop:   bra.s   stop
+doread:
+        movem.l d2-d7/a2-a6,-(sp)
+        moveq   #$46,d0
+        moveq   #0,d2
+        move.l  #256,d3
+        lea     BUF,a1
+        trap    #15
+        movem.l (sp)+,d2-d7/a2-a6
+        rts