Re: [stella] asymmetrical playfields

Subject: Re: [stella] asymmetrical playfields
From: Glenn Saunders <cybpunks2@xxxxxxxxxxxxx>
Date: Fri, 21 Sep 2001 00:11:00 -0700
At 06:56 PM 9/20/2001 -0700, you wrote:
I just tried this, and all I had to do was write 1 instead of 0 to
CTRLPF.  I even tried some different bitmaps to make sure
the transition was right -- it's on EXACTLY the right cycle
with the code unmodified. Of course this is like the first 6502

The existing code didn't work for me on a real VCS by just doing that. I had to add more NOPs to find the spot where the PF2 rewrite should be, then move the PF0 and PF1 rewrites earlier than the PF2 rewrite.

According to my math, in an asymmetrical reflected playfield:

The STA PF2b had to start exactly on cycle 45, done by cycle 47.

STA PF0b can be rewritten as early as cycle 23.  In my mod it's at 31.
STA PF1b can be rewritten as early as cycle 29.  In my mod it's at 38.

One thing to add to the Stella v2 document is a chart of cycle counts on exactly when each playfield register becomes visible and finishes and when it's safe to rewrite the registers (cycle ranges) in both modes. It's been discussed here and there on the list but never formally documented. It's also a little tricky because the terminology for suggested target cycles has to take into account the store operation cycle count (2 cycles) so you need to be clear whether the cycle range refers to when either the STA can commence or when it needs to have finished, or both.

Archives (includes files) at
Unsub & more at

Current Thread