| |
MagerValp
Registered: Dec 2001
Posts: 1074 |
C1/DTV PAL compatibility
Here's a little something for all you people who are curious about C1 and DTV C64 compatibility:
http://www.ling.gu.se/~pelle/deusex-palbeta-xvid.zip
It's actually running on the DTV development board, but the same core will be compiled for the C1 when it's done. As it's still in beta, a few things aren't working right:
* PAL phase alternation isn't working, so the colours are all wrong.
* Video output is "dirty" causing moving/flickering bands.
* 2-pixel $D016 delay is causing vertical bars to the left and right of the screen.
* An init bug makes the top left sprite black when the basic screen is moved out to the sides. It works if you power cycle before running, so it could be a bug in the demo and not the core.
* My video capture card has issues with $D016 interlace - this is not the core's fault :)
* $3fff should be covering the interference pattern, but doesn't.
* A few minor timing issues still remain, causing flicker in certain
I think you'll agree that C64 compatibility is decent though :)
PS: I didn't have a suitable cable, so I'm afraid there's no sound. I'll try to record some music later. |
|
... 11 posts hidden. Click here to view all posts....
|
| |
MagerValp
Registered: Dec 2001
Posts: 1074 |
Yes, there is no emulation anywhere in the DTV, it's all cloned hw. A real 1541 is used for testing.
|
| |
JackAsser
Registered: Jun 2002
Posts: 2014 |
Exactly how is it cloned? Does Jeri physically look at the VIC-chip and reverse engineer it or is it "simply" based on the knowledges such as the vic-article and VICE? What I'm really asking is if the approach differs from VICE and other emus? Will the DTB be more compatible and in the end 100% accurate? Can I trust to find new VIC-effects by using the DTV or do I still have to use a real C= for exploring the VIC-chip? |
| |
MagerValp
Registered: Dec 2001
Posts: 1074 |
It's cloned with trial and error :) Jeri's good at figuring out how hardware works, and yes the VIC Article has helped (though she says it has several inaccuracies). If she doesn't know how something works, she asks us coders and we write test programs that are run on both the DTV and a real 64 (well, VICE most of the time). The DTV won't be 100% compatible, since it's a commercial product and there's just isn't any justification for making it better - it already runs all the games in the joystick. The C1 should see continued development, and who knows, maybe it'll be perfect at some point.
|
| |
JackAsser
Registered: Jun 2002
Posts: 2014 |
So basically we're talking emulation but emulated by using HW instead of pure SW. To me, cloning is like when they open the casing of a chip and reverse engineer it using a microscope or something. At least a little bit deeper that trial and error. Anyway, cool stuff! I'll buy when the PAL version is out! =) |
| |
taper
Registered: Dec 2001
Posts: 119 |
I got the two NTSC c64dtv's I ordered from Magervalp yesterday (thanks!), and I'm very pleased with it. I'll be doing some modifications on one of them when I have the time for it, and probably buy one or two PAL versions aswell when they are released in the shops here.
It is a little box of wonders... :)
You can clearly hear that it's not a real SID in it though, even if the sound is okay. And I am pretty certain Cybernoid and Cybernoid II weren't NTSC fixed, as the songs ran too fast compaed to how I remembered them... :) But that's not the hardware's fault. |
| |
MagerValp
Registered: Dec 2001
Posts: 1074 |
Quote: So basically we're talking emulation but emulated by using HW instead of pure SW. To me, cloning is like when they open the casing of a chip and reverse engineer it using a microscope or something. At least a little bit deeper that trial and error. Anyway, cool stuff! I'll buy when the PAL version is out! =)
Unfortunately you can't clone chips like that. The C= chips are done with NMOS, and chips today are done with a different process - even if you were to clone the logic gates like a carbon copy, you wouldn't get the same functionality. And fwiw Jeri has studied the VIC and SID die scans that are out there, and apparently it helped her figure out a thing or two. |
| |
JackAsser
Registered: Jun 2002
Posts: 2014 |
Ahh, ok. Interesting! :D |
| |
MagerValp
Registered: Dec 2001
Posts: 1074 |
<Six> I would love to see a rundown of the errors you found in the Vic Article while you were duplicating the VIC. I bet those differences are the key to some of the as-yet undiscovered stuff.
<Jeri> I think it's pretty accurate. They didn't quite get all the hardware the way it really is.
<Jeri> They have more counters in the VIC artical than exist in the real thing.
<Six> Really?
<Jeri> also the bad line ram is a fifo, not an addressable memory.
<Jeri> those are the big differences I can find
|
| |
JackAsser
Registered: Jun 2002
Posts: 2014 |
A bit more in depth would be really interesting. Also, ask Jeri if she has some logical explanation on how sprite shrinking works, i.e. that the sprite stop render before it's 21 lines. |
| |
hollowman
Registered: Dec 2001
Posts: 474 |
Quote: A bit more in depth would be really interesting. Also, ask Jeri if she has some logical explanation on how sprite shrinking works, i.e. that the sprite stop render before it's 21 lines.
perhaps radman can ask in the interview for arts =)
http://www.pouet.net/topic.php?which=1859&page=1&x=26&y=7 |
| Previous - 1 | 2 | 3 - Next |
