It is currently Mon Sep 16, 2019 5:07 am

All times are UTC - 7 hours





Post new topic Reply to topic  [ 56 posts ]  Go to page Previous  1, 2, 3, 4
Author Message
PostPosted: Fri Jan 18, 2019 1:42 am 
Offline

Joined: Mon Jan 23, 2012 11:27 pm
Posts: 169
i have a mmc3c socketed cart for testing. the text on the opening screens is jumpy. here's a test run:

https://twitter.com/FrankWDoom/status/1 ... 9824610309

(sorry, don't know a better way to post a video here)

recorded on a crt on an original nes (has a copynes installed but i don't think that matters)


Top
 Profile  
 
PostPosted: Fri Jan 18, 2019 3:05 am 
Offline

Joined: Thu May 19, 2005 11:30 am
Posts: 975
Thank-you very much for that video. That means that the Everdrive N8's MMC3 scanline counter is not completely accurate.

Now the challenge for me will be to implement the delayed rendering change in a way that gets its desired effect of preventing glitches in Micro Machines while keeping the jumpy text in Crash Dummies when emulated with MMC3C instead of MC-ACC.


Top
 Profile  
 
PostPosted: Fri Jan 25, 2019 7:21 pm 
Offline

Joined: Tue Nov 23, 2004 9:35 pm
Posts: 742
This video https://www.youtube.com/watch?v=8x5IHXsJ2js was almost certainly taken with a real cartridge and shows "some jumpiness", but not as much as in FrankWDoom's video.

_________________
Nerdly Pleasures - My Vintage Video Game & Computing Blog


Top
 Profile  
 
PostPosted: Fri Jan 25, 2019 10:32 pm 
Offline
User avatar

Joined: Sun Sep 19, 2004 9:28 pm
Posts: 4208
Location: A world gone mad
I'm sorry that I couldn't help more on this issue (considering it's 5 years old) -- I've since sold all of my NES carts. Also, the YouTube videos in questions were lost during some clean-up I did. Again, sorry that I couldn't help more or re-do the tests (especially since I now have a 1080p 60fps digital camera).


Top
 Profile  
 
PostPosted: Fri Aug 16, 2019 5:05 am 
Offline
User avatar

Joined: Sun Jun 18, 2017 3:33 am
Posts: 5
NewRisingSun wrote:
Thank-you very much for that video. That means that the Everdrive N8's MMC3 scanline counter is not completely accurate.
when emulated with MMC3C instead of MC-ACC.


This is because you can't force N8 to use regular MMC by changing byte in the end of rom. Full rom CRC calculation is too slow for handling on system CPU, instead N8 used few bytes checks in certain rom area for specific rom detection. With regular MMC3 screen shakes will back.


Top
 Profile  
 
PostPosted: Fri Aug 16, 2019 6:01 am 
Offline

Joined: Thu May 19, 2005 11:30 am
Posts: 975
Ah, interesting. I had no way of knowing.

Would it be possible to add NES 2.0 Submappers to properly differentiate between mapper variants, and limit such checks to ROMs with no NES 2.0 header?


Top
 Profile  
 
PostPosted: Fri Aug 16, 2019 6:10 am 
Offline
User avatar

Joined: Sun Jun 18, 2017 3:33 am
Posts: 5
NewRisingSun wrote:
Ah, interesting. I had no way of knowing.

Would it be possible to add NES 2.0 Submappers to properly differentiate between mapper variants, and limit such checks to ROMs with no NES 2.0 header?


Actually i thinking about including nes 2.0 support


Top
 Profile  
 
PostPosted: Wed Sep 11, 2019 3:50 am 
Offline
User avatar

Joined: Sun Jun 18, 2017 3:33 am
Posts: 5
I did a small research around acclaim mapper. Seems like besides of neg edge triggering it ignores a12 if rendering turning off, also it expects a12 to be in low state more than two cpu cycles. As we know problems with Crash Dummies on regular MMC3 related to edge of a12 triggering, but with Mickey's Safari in Letterland is another story. It also shakes menu bar on regular mmc3, but not due the triggering edge of a12 , it happens if a12 clocking hits to the mement when PPU turn off the rendering. On attached screenshots shown such moment. Red color shown when ppu turned off and yelow dots shown false triggering which counted by MMC3 logic and ignored by MC-ACC.
However, i not sure how exactly MC-ACC ignores those cycles during turned off PPU, because it does not have enough connected address lines to decode PPU registers addresses and intercept writes to $2001, may be it require specific timings between A12 clocking or something

EDIT:
Ok, seems like i figure out how it working. Acclaim mapper counts every A12 pulse instead of waiting for long phase in low state. Counter clocking happens once per 8 A12 cycles at first cycle. Writing to $C001 resets pulse counter. On attached screenshot shown how real acclaim chip triggers IRQ not at first A12 pulse. Also i looked at few emulators and they all emulate this behavior incorrectly. In Mickey emulators fires second IRQ too early. It easy to check: With proper timings in bottom bar should be two lines between "in" and "letterland", but emulators have only one line. Also those 3 black boxes in right side of bar should have size 8x8 pixels, but emulators likely will render them 8x7


Attachments:
File comment: Two lines with proper timings
proper-timings.jpg
proper-timings.jpg [ 103.04 KiB | Viewed 149 times ]
File comment: Shifted counter phase
11.png
11.png [ 44.63 KiB | Viewed 149 times ]
File comment: Phase shift happens due the short rendering disabling
DS1Z_QuickPrin99.png
DS1Z_QuickPrin99.png [ 41.35 KiB | Viewed 182 times ]
Top
 Profile  
 
PostPosted: Thu Sep 12, 2019 12:13 pm 
Offline

Joined: Thu May 19, 2005 11:30 am
Posts: 975
This recording, which I believe to be from real hardware based on the line on the left and the buzzing sound in the audio, although none too clear, seems to show only one line between "in" and "letterland".

Quote:
Acclaim mapper counts every A12 pulse instead of waiting for long phase in low state. Counter clocking happens once per 8 A12 cycles at first cycle.
"Unfiltered PA12 rise (or fall) with prescaler" would mean that the AC-MCC's scanline counter acts more like the J.Y. Company ASIC's.


Top
 Profile  
 
PostPosted: Thu Sep 12, 2019 3:15 pm 
Offline
User avatar

Joined: Sun Jun 18, 2017 3:33 am
Posts: 5
NewRisingSun wrote:
This recording, which I believe to be from real hardware based on the line on the left and the buzzing sound in the audio, although none too clear, seems to show only one line between "in" and "letterland".


This video for sure recorded not from MC-ACC cart. Besides of lost line, bottom bar shakes time to time. I have real Mickey cartridge and it shows two lines and no shakes


Top
 Profile  
 
PostPosted: Thu Sep 12, 2019 3:28 pm 
Offline

Joined: Thu May 19, 2005 11:30 am
Posts: 975
What did the author of that video record it from, in 2007? The console seems to be real, but maybe the cartridge is not. Was the PowerPak available back then?

Anyway, this video shows two scanlines, also ostensibly from real hardware.


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 56 posts ]  Go to page Previous  1, 2, 3, 4

All times are UTC - 7 hours


Who is online

Users browsing this forum: No registered users and 3 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group