VSU news

Can sound strange given the whole incredible specs; but for me the most impressive fact about the VSU is that it doesn't need a separate video output, all that chips sharing the same output is really great!

I'll guess that thats the job of the FPGA

and what about adding a M68000 or a small Pentium ?
just to support the z80....

You guys missing the point...

You load everything into VRAM once, then it's a matter of sending commands to copy data (which doesn't use much CPU - and won't have those horrid waits for the VDP to finish them, as you can change to another VDP and set it to copy more data) - or update the tiles in the tilemap. Stuff like copying data from RAM to VRAM will be a thing from the past... You will actually get MORE Z80 to work with!

Take a look... that mickey demo was made in BASIC, and it's several SCREEN 8 planes! IN B-A-S-I-C. Now think what can be done in ASM!

And btw: it's 80 sprites PER LINE

Check the Mickey Mouse thing: http://www.youtube.com/watch?v=GfMbWi3TBPM

How much is actually happening on screen? While I dunno the source, or what chips have been used here, I'd say: the action/frontlayer is probably a big G9k VRAM map which only requires you to change the scroll registers to the appropriate coordinate. So, this means no custom plotting tiles when you scroll out of the screen.

So far, it can indeed be done in BASIC. Question is: *what* can be done in assembly if you want a great number of actions and activities in your graphics? This level may look big because the scroll is slow, but what if your level is supposed to be super-big? (either because of an adventure argument or simply because the scroll is fast (think Jazz Jackrabbit)), then you can't just store the whole map in VRAM and you actually have to work with tilemaps (like we do now already on any MSX). If you also want the same tilemap thing to work for other layers, then I do wonder whether the Z80 can cope with it..

Can sound strange given the whole incredible specs; but for me the most impressive fact about the VSU is that it doesn't need a separate video output, all that chips sharing the same output is really great!Not to mention the VGA output. That's a very interesting point. It would be a nice hardware even for classic software.

uhm... gonna pas on this thing. It is fun no doubt but, nope not for me. I stick to just one v9990, nice and simple. But it is compatible with gfx9000 right? So no worries every one can still play my games!

ps: any potentional market growt for buying my verry expensive games wich gonna make me very but then I do mean very rich is great news.

Its an amazing piece of hardware, but to be honest I don't think a Z80 is good enough to actually make good use of the possibilities with the VSU.

With the very nice graphics potential of the VSU you probably also want to have a good physics engine and AI. For this you really need something more than a Z80. I like the idea of adding a PPC or an ARM7 or something, to help out with physics and AI. Then the Z80 can be the hub in the system...You see, that's the thing. The bottleneck always lies somewhere; the current V9938 VDP is too slow for the Z80. Then again, when the VDP is upgraded, the Z80 is too slow. Reasoning, we should all move to R800 equipped turboR computers but then it's still too slow. Let's upgrade the CPU again and find that the 6 VDPs are then insufficient. This whole story can go on forever. I see the VSU as a suitable and worthy addition to the current repository of MSX hardware, especially because it is compatible with both Moonsound and GFX9000.
This makes for comparisons with the MSX turboR standard, it is compatible with MSX, MSX2 and MSX2+, so, there's more content for those machines. Albeit not exclusively.

that Mickey sure takes a lot 'o loading time . . .

I would settle for a VSU Light just a combination of GFX9000+Moonsound combined into 1 cart.

Price is a major factor. A very low price would mean hundreds of users with that card and dozens of MSX developers creating nice stuff.

Lot's of V9990 and OPL4 still available. Oazem has purchased over a 100 of them

I think that a board with a v9958,v9990 and an opl4 would already be very intteresseting, but maybe more is better

Very nice, but I still want a "FM Sound Stereo" for use on regular MSX personal computers ...

After all, stereo output should be good enough!

You are missing the point guys... really... In the last few years all sw/hw projects that have succeeded had a low price: Matra "shockwave" cartridges, MSX dev cartridges, other games on cartridge like Manbow2 (with SCC bulit-in!), IDE interfaces (few people own a Mega SCSI due to its high price), MP3 cartridges, etc, etc.

Make a cheap -yet powerful- videocard for MSX and the GFX9000 will be history.

I've just one remark: the TMS9918 programming is so easy when compared to the V9938. V9938 is not as complex as V9958 new screen modes. V9990 is far more complex than V9958. Now imagine how complex could it be to code for a system with a bunch of mixed VDP, each with its own separate VRAM. Problem here is the increased development complexity, not the higher specs. Of course, lack of implementation within the community would be an unbearable handicap.

If you have an MSX, then code for an MSX and expect MSX games. If it is another thing, then it is not MSX. I could connect my MSX to a PC and use the MSX as a "terminal" for playing PC games. Is then the NVidia GeForce a new MSX graphic card?

I'll say it again for the last time: VSU will be cheaper than buying a GFX9000 and an Moonsound, it's all in one cartridge instead of two, and fully compatible with both.

And actually, developing for gfx9000 (or many of them) is easier than on MSX1, since you won't have to worry about sprites limits (80 sprites per line! And all of the GFX9000 sprites are multicolored ones!), you can easily do several layers without screensplits, and so on.

Again, that Mickey demo is in BASIC. Cant get easier than that.

The two 9958 are kept for two reasons: for people with MSX1 to be able to turn it into an MSX2+ with 9990 and OPL4, just by replacing the ROMs and connecting the VSU - and to be able not only to use BASIC/DOS/whatever on VSU, but also to give the option to "improve" some old MSX games, giving them more colorfull sprites, more layers, or whatever comes into the mind of the developer

That's the more important thing... if you don't want to use all the VSU horsepower and just develope something for gfx9000 + moonsound, nothing stops you to do so! If you want to use only the two 9958 to make a pseudo "multi layer MSX1 game" - you can also make it.

It's just a new hardware - fully compatible with everything prior to it - that gives new possibilities to us, the developers. If you think it's "too much", then use only what you think it's "enough".

GFX9000 / VSU also uses tiled modes, but multicolored ones, instead of those 2 colors on every 8x1 block on MSX1... and the sprites themselves are multicolored, no "single color sprite"!

I can't see how that is more complicated than on MSX1.

When a new hardware is released, the software will most likely NOT explore it entirely.
I would take some time (maybe a lot of time... maybe the eternity). I believe we are used
to VDP limitations... with VSU the video will be not the limitation for sure, but the Z80
(or R800). The limits of what can be done with the VSU will be dictated by the main
processor.

On the other hand, some programmers are not interested in what impress someone,
but in developing beautiful looking, nice and *easily programmed* games and software...
something that has nothing to do with limitations.

Personally, I like to program MSX because its concept is simple, I like to develop in ASM
(and MSX ASM is very straightforward) and I curse everyday the VDP limitations I have to
cope with. I do not program MSX because it is "limited". I am not concerned in squeezing
MSX1 (that's why I seldom develop MSX1 exclusive software). And don't tell me to
develop PC software; I am used to PC and I do not think it's a cool machine to develop
games. PC has much power? Yes. PC is nice to program? I don't think so... (my opinion).

Of course many MSX developers do not agree with me, but I keep my focus on the result,
a good-looking and enjoyable game... and not "wow, how could he do this using only a
PSG and a dumb VDP?". Of course the "wtf! factor" is a nice thing to explore, but I do not
think in it as the only one.

I will be very GLAD if impressive games could be developed using simple BASIC
commands... I will be very glad bringing other plataform's games using VSU, and
will be very glad even if no one ever reach the limits of VSU. I do not even expect it
to happen.

Agree with me or not, this is your choice... this a personal POV after all; but it is
always COOL to have options to choose. No one will ever be obligated to program
for VSU in the same way no one was ever obligated to develop turbo R software.

Peace!

Well .. of course .. if VSU is cheaper than GFX9000+Moonsound and have 4V9990+2V9958+FGA+OPL4 ok!! for me is great .. the only problem in my opinion is the size of the PCB .. i think that will be tremendous. Or am i wrong? In any case somebody in relation with VSU could answer my questions?

**I would like to know how the VSU priorize the VDPs output signal. I suppose that there is a VDP that will cover all the screen if i put a white box (for example)
And i would like to know how the resolution and screen modes work with all the VDPs because each VDP could have a different screen mode, right?**

Thank you very much.

I see.... So, why is so difficult to get fine MSX1 games if it's easy? In my opinion 80% of the MSXDEV's games are absolute crap. If developers can't still design and develop a good MSX1 game (easier than G9k, according you), how could they create good games for this monster?

Yes, easier, not easy. As specs increase so does the todo of game developers. I mean, don't you know the golden rule of money? Machines break down just right when you have fresh new money, and a new machine costs exactly all the money you've just got. Same principle applied on gaming. Increase the specs and suddenly people dare to make bigger games, with more images, more animations, bigger maps, etc. etc. And in a way, it's perfectly normal to have such a mindset. Otherwise people would be making simple games which are in reality just MSX1/MSX2 games with more fancy graphics. That may be a thing to do, but it wouldn't really do justice to the capabilities of such a new chip. So, that's why I said that to do a very proper MSX1 game is easier to do than a game which fully squeezes the last drop from a G9k (let alone a VSU).

And yes, there may be only a handful o' groups/individuals who control the MSXdev food chain in terms of quality. Then again: it's all about - the more the better!

I think playability or quality may be different things in his point of view. Manic Miner is very playable and very nice, but also heavily outdated (call it 'retro' if you wish) when it comes to artwork. The game doesn't compare to -say- Nemesis 2, Shalom etc. It's all purely what one's anticipating on. In any case, he thinks 80% is crap, which is his own view on the thing. Someone else may indeed think that 80% is good. But if each year you're waiting for the next Konami-quality game, then yea; there may be only 1, 2 or 3 entries that may qualify for that. Not that it matters, as long as we together!