DEMOs

Hero (513)

03-10-2022, 00:07

PingPong wrote:

msx came out on 1983 they should have choosen a better alternative, even a 6845 can do better

We'll never know the reasons why they made the choice. I suspect cost/time were the main reasons given that the old IBM CGA cards had 50+ TTL ICs on them.
But I don't think the 6845, at least the Amstrad implementation, is better: IMHO I find the Amstrad games to have both the defects of the C64 (low res/blocky graphics in multicolor mode) and Spectrum (no hardware sprites, CPU has to brute-force all).

PingPong wrote:

It was, however a failed attempt to go more far than space invaders itself.

Nope, in fact they were quite successful, as the device could do far more than simply mimic Space Invaders.
Just take a look at the Colecovision: It had the best games until the Famicom came out.

Van Micha

Expert (98)

03-10-2022, 08:52

If I had to name one thing that I regret about the 9918 then it is that is has no dark colors at all. One of the yellows (color 10 or 11; they are almost equal) should have been a dark grey (or maybe a dark brown). This would have made a big difference in the look of most MSX1 games, and could have been implemented without any trade-offs.
For the rest it is a great chip considering the cost and the year it was on the market.

Van PingPong

Prophet (4093)

03-10-2022, 09:25

Quote:

The V9938 and the VIC-II do scroll horizontal exactly the same way: it's just a screen offset. But while on the V9938 you have 16 pixels of offset, the VIC-II only have 8.

I disagree, completely: the purpose of adjust register IS TO CENTER THE IMAGE, not to scroll.
To be clear things are as follow: the v9938 lacks horizontal scroll and the se use a hack (the horz adjust register) to approximate the scroll suppport. But the purpose of the adjust registers is different!
Following your reasoning i can say that even a msx1 has smooth scrolling support, by tricking the tiles definitions i can achieve a similar effect!.

on the VIC-II the scrolling registers DO NOT AFFECT the x position of the image on the screen because those registers are meant to support scrolling, not image centering on screen!

this is also proved by the fact that the v9958 had scrolling support and it works exactly like v9938 horizontal scrolling

Saying the V9938 do the scroll in the same way the VIC-II do is like saying that i can use a small car to carry 1 ton of bricks instead of a truck. Of course you can carry bricks with a car, but if you use a car for this purpose you get a poor and limited truck

Van PingPong

Prophet (4093)

03-10-2022, 10:01

Pentarou wrote:

We'll never know the reasons why they made the choice. I suspect cost/time were the main reasons given that the old IBM CGA cards had 50+ TTL ICs on them.

the msx was derived from spectravideo prototype, they used pratically the entire existing architecture.

Quote:

But I don't think the 6845, at least the Amstrad implementation, is better: IMHO I find the Amstrad games to have both the defects of the C64 (low res/blocky graphics in multicolor mode) and Spectrum (no hardware sprites, CPU has to brute-force all).

that is the point : most cpc games (with no hw support for games) look more colorful than msx ones (which should have hw support, at least sprites). In theory should be the opposite, but look for example to this: https://www.youtube.com/watch?v=2hfXxIExK_I , can we at least reach the same on msx1? nope, can do this on msx2, (maybe, not sure).
this is an example of what i mean when i say "failed attempt". the TMS should be better, but it's worse

Van mars2000you

Enlighted (6430)

03-10-2022, 11:22

As mentioned several times by Kay Nishi, the machine that he wanted to release is the MSX2, but in 1983, the VDP chip with all the features was not ready, that's why TMS9918 was choosed as temporary solution and MSX1 is not the machine that corresponds to the wishes of the main MSX creator.

Van santiontanon

Paragon (1770)

03-10-2022, 16:00

Oh, is it that time of the year again where there is the annual VDP thrashing thread? hahaha

Van AxelStone

Prophet (3189)

03-10-2022, 16:19

Very nice CGA demo! I really think that CRTC has a lot of potential hidden.

Van Pentarou

Hero (513)

03-10-2022, 17:48

PingPong wrote:

the msx was derived from spectravideo prototype, they used pratically the entire existing architecture.

We already know that.
So at first you said they could have chosen another VDP, and when I replied that they had cost/time constraints you deflected the question...
.
So now you will start blaming spectravideo?

PingPong wrote:

that is the point : most cpc games (with no hw support for games) look more colorful than msx ones (which should have hw support, at least sprites)

ROFL! The link you posted proves my points: 160x200 blocky graphics (with a gigantic hud) and a game that scuttles away at 10fps! (*)

mars2000you wrote:

As mentioned several times by Kay Nishi, the machine that he wanted to release is the MSX2,

He's not the most trustable source, given his revisionist habit.

(*) The game is a decent conversion given the Amstrad capabilities. But I'm more impressed by the Spectrum port.

Van Parn

Paladin (833)

03-10-2022, 18:52

I, for one, don't envy other systems' CRTC. The MSX VDPs have their own character and particular strengths. I would never wish my MSX were the same as a Spectrum or a CPC, much less (God forbid!) a C64 with its weird-ass colors and garbage BASIC and disk drives. Of course, I like those machines too. But MSX is much more fun!

Van PingPong

Prophet (4093)

03-10-2022, 20:37

Quote:

Pentarou wrote:

PingPong wrote:

the msx was derived from spectravideo prototype, they used pratically the entire existing architecture.

Cost/time constraints? MSX is not an incredibly cheap machine so cost limits is not a real reason.
Let's say the most probable reason. "OK, we need to create a new standard for home computing.... ok, let's copy the spectravideo without taking in account the weak parts and the strengths of it".
Unfortunately the weak part on spectravideo was the TMS VDP...

Quote:

PingPong wrote:

that is the point : most cpc games (with no hw support for games) look more colorful than msx ones (which should have hw support, at least sprites)

ROFL! The link you posted proves my points: 160x200 blocky graphics (with a gigantic hud) and a game that scuttles away at 10fps! (*)

the only thing that it proves is the absence of colour clash. And to be clear. A similar game is *IMPOSSIBLE* on TMS VDP (and hard on V9938) where thanks to the hw sprite support (max monochrome sprites/scanline) you will be forced to switch to sw sprites, and with the I/O driver you could also forget the 10fps (be lucky if you get 7-8fps), and be delighted with colour clash or squared-like sprites moving at steps of 8 px. What's better?

The main limit on msx1 is exactly the hw sprite support. It's here and theoretically a good thing. but the stupid limit of 4 sprites / scanline make a lot of games too much difficult to create, forcing the developers to sw sprite modes and the usual spectrum/cpc approach. With a big difference: you get a very slow I/O to vram.