Hi,
I am having some issues while running demo_ntsc.nes. The "middle line" as per rom documentation doesn't appear in middle when I run this in my emulator.
One thing that I was doubting was the NMI timing, I ran all the NMI-VBL-PPU sync/timing tests (that I could find) and I am passing all of them, but note that I am not passing cpu_interrupt_v2 tests.
I have made sure that NMI is triggered at the correct PPU cycle w.r.t to PPU Frame Start (scanline = 1 and ppu cycle = 0). As expected, its taking <82522, 82523> cycles from the PPU Frame Start to trigger the NMI after accounting for even-odd frames.
Other aspect that I am suspecting is OAM DMA timing (Is there any rom to test this?)
Any insight in debugging this would be helpful. Have attached the gif below for reference.
================================================================================================================================
Edit1: Major part of the issue mentioned above is resolved now ("middle line" appears in the middle as expected), Issue was syncing of PPU Cycles to OAM DMA cycles (was wrong by factor of 3). After this modification, the only difference I see w.r.t expected pattern is that the middle line is still shifted more towards the left of the screen. Still some minute timing issues I guess...
I am having some issues while running demo_ntsc.nes. The "middle line" as per rom documentation doesn't appear in middle when I run this in my emulator.
One thing that I was doubting was the NMI timing, I ran all the NMI-VBL-PPU sync/timing tests (that I could find) and I am passing all of them, but note that I am not passing cpu_interrupt_v2 tests.
I have made sure that NMI is triggered at the correct PPU cycle w.r.t to PPU Frame Start (scanline = 1 and ppu cycle = 0). As expected, its taking <82522, 82523> cycles from the PPU Frame Start to trigger the NMI after accounting for even-odd frames.
Other aspect that I am suspecting is OAM DMA timing (Is there any rom to test this?)
Any insight in debugging this would be helpful. Have attached the gif below for reference.
Edit1: Major part of the issue mentioned above is resolved now ("middle line" appears in the middle as expected), Issue was syncing of PPU Cycles to OAM DMA cycles (was wrong by factor of 3). After this modification, the only difference I see w.r.t expected pattern is that the middle line is still shifted more towards the left of the screen. Still some minute timing issues I guess...
Statistics: Posted by kotambail_hegde — Tue Dec 31, 2024 1:30 am — Replies 4 — Views 104