MSX1 Video Board: F18A 1.3

Yeah, we can align 32 sprites on the same scanline
rotating/ordering/and any other tricks go away finally

pratically this gem offer, sprite caps that no other vdps can offer (apart the v9990)

I agree with Sander. SMS mode 4 is a great addition.

It offers a lot of new stuff to play with but still has this its limitations. I recently started coding for mode 4 just out of curiousity and I really like it. Very easy to master and it still feels like MSX ;)

I would really like to have such a chip in my MSX.

there is a test of the replacement with q-bert where is clear that the F18A show all 32 sprites on scanline

I read on his site that he is planning/toying with amongst others hardware scroll and bitmap layer, so it may end up equal or even better then v9958? (since it will have more sprites/scanline, more colors)

But that was blogpost was made over a year ago on his site so I dont know what the current plans or specs are.

But indeed, it looks very promising. I should get a msx1 and mod it with this new chip if the specs are worth the effort

If you want this extension to be adopted by a lot of people, you need software. Support for the SMS could be an additional drive for non programmers to buy this card. Your argument about the sprite flickering in SMS is not valid 'cause like the MSX modes this flickering could easily be removed in FPGA hardware. If this thing can do everything but has no software for it you'll get the same thing as with the GFX9000. Plus, this is not a device you can plug in and out, you have to wreck/modify an existing MSX. Also not something a lot of people are willing to do (or can do for that matter).

And for those who don't want to wreck their MSX or want to spend a lot of money on the PlaySoniq, I'm still thinking of releasing a low cost DIY kit to build your own Franky card. Drop me an e-mail if you're interested.

Luckely other don't share you view and are already developing for it. More people probably, than developers for this announced card. With a big enough FPGA the Mode 4 could be supported and still have enough space for all the other things you mention so what is exactly the problem here?

If you modify the MSX1 VDP completely, what's left of the MSX? And another thought: if making a complete new msxish standard I would rather go for completely new MSX in FPGA like the Orbit with the specs mentioned than a half new / half modified MSX. Just thinking.

If you want less limitations without falling in the exotic GFX9000, why not code new games for MSX2+ ?
What Kay Nishi wanted to do was not MSX1 but at least MSX2, but unfortunaly some chips were not ready in 1983. So the real MSX is not MSX1, but MSX2(+).

I find interesting is a cartridge with an FPGA with a similar SMS graphic mode, with the palette of MSX2 and other enhancements ( including a VGA connector). Also it would include an emulated sound chip, similar to AY, with 4 +4 channel,  hardware envelopes and some forms of basic waves (or a SCC with 4 +4 channels, and envelopes).

The main problem is the time it takes to develop software and hardware.  Who we are in the MSX scene, we do as a hobby and only have little free time.
We have to think of ways to simplify development.

( I hope you understand me. My English is terrible! .

@wolf, @mars2000:

The point is not making sw for v9958 or v9990. Obviously one can do whatever.
The point is the difference:
Saying the v9990 have tile mode like TMS is saying that a car and a bycycle are equals. They both have wheels.
Really a v9990 is very different. Compared to others chip, It's the chip that have the larger differences in programming. All is different: registers, behaviour, 16 bit nametable, pattern format, sprites. 
Also coding for v9958 is the same of coding for SMS. It's different. To make the things clear, i summarize what i think of the VDP chips available on msx:

1) TMS: the old crappy MSX videochip, would be better if some limitations were removed.
2) V9938: a half/half evolution, that have the chance to fix TMS issues, but failed in most areas, something making worst the situation
3) V9958: a half evolution, little more than V9938, limitations are still here. the blitter is somewhat slow, making unsuitable for bitmap management
4) SMS VDP: another half evolution, perhaps oriented &nbsp of course, the SMS is a console not a computer ) to games, does not add more, because still limitations are here, 8x16 sprites, 8 sprites/scanline, rather limited pattern table.
I see this chip as a path to play sms games on msx, nothing more.
5) F18A: finally the right path: no limitations, pratically the lowest impact (almost the same) in programming, because is  similar to TMS, but without limitations. Plus one can drive a decent VGA monitor and not a old TV set.

Seems like we all have our own ideas about what we need/want.
Atm I am having a blast playing around with the SMS mode 4 features (which add a lot more compared to MSX modes ).

After beeing an MSX1 only oriented developer I have finally reached the next stage to experiment with other VDP's.
The biggest issue is making gfx without the color clash limitation. I am working on 4 projects each for a different screen mode (screen2,mode4,screen0 and screen4) each utilizing their strong points and working around their weaknesses.

I'm hoping that the F18A will drop in price. It's still a bit pricey.

About the sprite limit; Are there people actually disabeling the limit on their emulator when playing games?

....Atm I am having a blast playing around with the SMS mode 4 features ....

Cannot imagine if you were working with a v9990. A real party.....
Try with a VIC-II, it's even more funny.

Well, I don't want VGA. I'd love new CRTs...

Huey, I wonder when you'll get to P1 mode of V9990 Have fun!

I think coders think that v9990 is for megaproductions and they fail in the attempt with result of unfinished games.

They have to try with "simple" games or msx1 or msx2 games with a plus that normal msx can't offer, but no try arrive to the limit of the card.

Having gfx9000 and Playsoniq, who wants buy another similar vdp?

You can create a Street Fighter 2 in gfx9000 but Pong too.

@guantxip
I think coders think that v9990 is for megaproductions and they fail in the attempt with result of unfinished games.Partly, but in the 90's MSX2 was the default machine. Coders made games like Shrines of Enigma, the Pumpkin Adventures and other Umax RPGs, Dass, Frantic, Nosh, Troxx, Dix, all kinds of Tetris clones, Arc, Magnar, Akin, Blade Lords. Some games are relatively small, others are big.

A few posts above this one I already mentioned that any MSX1 game could be converted to G9k since they're tile based. and ok ok, there will be different registers to deal with, but it doesn't require a complete redesign in terms of artwork and AI/gameplay.

Point is: the MSX2 videochip doesn't make things more troublesome for large games. In some situations it'd make things even easier, think of scroll registers compared to manually updating tiles on MSX1 - Malaika comes to mind. Think of creating tiles with a true artistic mindset, rather than constant calculating how to create diagonal, round, and colourful objects with two colours per 8 pixels. And sure, there is a difference between an MSX2 RPG and an MSX1 puzzle game, yea.. that's logical, but the MSX1 videochip has nothing to do with that. I mean, shall we compare the development time (and cancellation risk) of an MSX2 puzzle game to Konami's Shalom? Our Infinite MSXdev'xx games that have been cancelled were big games, really big games. They were cancelled, not because of the MSX1 VDP, but because the games were too complex (would've been complex on MSX2 too), and because I couldn't come up with a storyline. Yet, we wanted to create these games because we wanted them to create - and we like a challenge. These problems I think are far more important these days:

• Teams have less time for large projects due real-life things.
• Somehow, people like a challenge (or maybe they even need a challenge to move their asses), and the past decade (or most of it) was flooded with MSXdev'xx contests. If you wanted to compete in a game contest, there wasn't much choice. And don't get me wrong: they've been fun, but it more or less ruined MSX2 gaming..
• A significant amount of people have moved to cross development, not just for code, but for graphics and music too. This actually invites project inflation because things are so easy. In Age (DD-graph) you were creating one page, in Photoshop you could be working on a large image with a height of 512 pixels (two pages!). All this unlimited editing richness inflates game concepts and complicates its development because you end up with so many loose files, 'make' batch files, converters, editors, you name it, that projects soon run out of hand. It's an issue we (Infinite) want to cure with Polka2, being an integrated all-in-one editor.

@Mars
It's also partially the case for MSX2/2+ coders and it's one of the main reasons why the Passion MSX2 contests were not a success.May I remind you that your promotion/communication activities surrounding these contests have been like a Greek tragedy? (at least I hope that's what you mean with the other reasons in this 'one of the main reasons' term) Because I've all sorts o' website alerts I was notified soon enough that these contests had been announced on PassionMSX, but other than that, there was no word from you. I've seen nothing from your side to hype up the contest, to boost discussions, to talk people into competing. Heck, you didn't even make a banner for it, I did, primarily for MRC, but someone at PassionMSX 'took it' and used it - and later on MSX Café took it too. News? Nothing. In earlier years, Viejo_Archibeardo at least submitted news to us when there was a new MSXdev'xx entry (or he'd bump me on IRC about it), or even Pitpan did, dunno, it was a long time ago. At least there was communication, both via our news submit and via our forums.
Now, I can imagine you wouldn't feel like doing all that, being in your loud 'ignore MRC' years, but the consequence was that, apart from me reporting about contest news (whatever news there was), you've been on your own all the time. All between the four walls of your own website, and while I don't intend to take any credit, I do believe we (MRC) can reach more people. In the end, I truly wonder what this silence has really meant for the PassionMSX2 contest. At least I did whatever I could to support it, by writing about it, by making a banner for it, by referring to it when there was a discussion about the lack of MSX2 games, by referring to it when there was a discussion about MSX2 specifications in the MSXdev'xx contest, and by competing in it. (I don't ask any credits for all that, btw)

I simply don't believe it's all because of the VDP. Manbow 2 is a fairly recent MSX2 game (and norakomi has more MSX2 games coming up), Goones 'R' Good Enough is recent as well. A little bit down memory lane there's the TNI version of Guru Logic (though not finished yet) and Artrag is busy with his raycasting adventure. There *is* MSX2 activity! Perhaps the new MSXdev contest at MSXBanzai manages to streamline higher generations of MSX computers and extensions.. we'll have to see what happens with that.

Anyway, all this lies in the past now, I don't intend to make it all bigger than necessary.

Dunno... I think Pengo was MSX2... And the new Paxanga software is MSX2... somebody told me.

just 2 remarks for wolf_ :

1. What you've said about the banner is not correct :

- SuperCobra had created a logo for the contest and we used a resized version as banner on the PassionMSX website, this even before you created your banner for the first edition of the contest
- The same for the second edition, but with as difference that after some months, you (or someone else in the MRC crew) have removed the banner from the MRC rotating banner system. Maybe a try to get some reaction from me, but you should have known it was useless.

2. I will not answer for the rest of your post. Time will tell if the alternative contest launched on the MSX Banzai website will be or not successful. At least on PassionMSX forum, there was place for discussions and comments about projects and released games (but it was almost not used by coders and fans). On MSX Banzai, there is no any place to make even a comment !

Because that's what happens when people talk.

What Kay Nishi wanted to do was not MSX1 but at least MSX2, but unfortunaly some chips were not ready in 1983. So the real MSX is not MSX1, but MSX2(+).Wink

No matter wat Nishi wanted to do but what he finally did, what he had at our homes. No matter the reasons or the limitations that brought us the MSX as it is: saying that the real MSX is not MSX(1) is just a blasphemy and a complete distortion of reality itself. Most of us drawn their first pixels with it, played the first notes, wrote their very first lines of code and enjoyed the best games from the system with our good old MSXes. And yep, I've noticed the smiley at the end of the quote, but a funny face does not always mean you can write such a tremendous nonsense and be left without a reply. MSX is MSX. Face it.

What I do remember from the eighties, I do remember that back then, I did read that MSX was developed to be cheap but good computer so everyone around the world could buy it and so japanese computers could conquer the world. I do understand that there were many people who would like to buy the best computers with good (=better than MSX(1)) graphics and sound chips, but the reality was that not everyone was rich enough or wanted to waste money for such thing or to waste their money to computer at all. Please remember,  not everybody wanted computer in early eighties. Yes, I know that many people bought Commodore 64 even it was not cheap at all. Commodore 64 costed 1000U$D when it was announced, you can try to imagine how much more your decent MSX2 or MSX2+ may have costed in 1983..... and I am not talking about 16Kb or 32Kb MSX2, but decent one in 1983, just like someones are dreaming. Do you remember what happened to most very expensive very good computers in the eighties? They were mostly NOT bought at all, because of their way too high price, no matter that they had literally arcade perfect graphics and/or were perfectly cabable to make CAD-like stuff at home...... Sure, there surely could have been people who would had bought decent MSX2 or MSX2+ in 1983, but majority would had bought something they really could afford to buy. Yes, nowadays more people can buy whatever they want to buy, but things (and people) were different thirty years ago. But now I let the topic continue again.

In the eighties, when I read about MSX, it was many times told, that japanese electronic giants behind MSX chose Texas Instruments´ graphics chip, because it was trusted and good chip and it did exist already.

Maybe its just time to move your ass instead of the boo hoo examples... maybe the truth is a bit more complex... there are enough gfx creators still alive to create the most colorfull gfx for some nice games... maybe the brick walls of critism are to high to ever create nice gfx for a good looking 9k game... would it be possible to create a brand new v9k v2.0 to make it also backwards compatible?? i did serval atemps in creating new v9k gfx... the last project was a 9k shooter... some liked it and some gave stupid critism like hey that looks like plastic panels... i think the biggest problem is spoiled minds that create those high brick walls.... we should keep in mind even a v9k (v9990) has huge bariers we cant break down if critism stays at this level.... i think the biggest bottleneck is making the gfx for v9990 wich takes a lot of love and time... on the programing side the v9990 was more easy because of the new things it could achieve!.
So maybe if we drop some overated wishes we could all step on the same boat and make some good msx stuf...
all new hardware fails is there is no support for it!! i still love MSX and everyone who puts time in it to achieve new levels of MSX-nirvana

What Kay Nishi wanted to do was not MSX1 but at least MSX2, but unfortunaly some chips were not ready in 1983. So the real MSX is not MSX1, but MSX2(+).Wink
Well, to me it just means MSX2 would be a 100% japanese computer (with japanese components or VDP)... that's all.

Of course it was better than MSX1 (due the new chips and options).

Hey guys,

Interesting thread.  I have not had much feedback from the MSX community on the F18A, so I was not sure about the interest.  Sorry for not having an up to date feature list posted on my site, too much to do, very little time.  Here is the updated list of what is done, and will be done for the final release:

Main F18A V1.3 Features

Outputs standard 640x480@60Hz analogue VGA
Direct hardware pin-compatible replacement for the 9918A, 9928, and 9929 VDPs
Software compatible with the 9918A VDP
All four original 9918A graphics modes
80-column text mode
Per-tile color in text modes
Sixty-four programmable color registers with a selection from a palette of 4096 colors
Enhanced color modes for 2, 4, or 8 colors for tiles and sprites
Tile priority over sprites on a per-tile basis
Extra attribute information per tile: flip x, flip y, transparent, and palette select
Support for 30 rows in Graphics Mode I for an NES compatible display size (32x30 tiles)
Enhanced name table configurations to support scrolling
Horizontal and vertical scroll registers
Horizontal scan line interrupt
No per-line sprite limit, so all 32 sprites can be displayed on the same horizontal line
Sprite linking (move multiple sprites by updating a single sprite's coordinates)
Enhanced sprite collision detection information
Sprite size, 8x8 or 16x16, selectable per sprite
Extra attribute information per sprite: flip x, flip y
No CPU to VDP read or write speed limitations (up to 25MHz CPU clock speed)

Original 9918A Features Not Reproduced

External sync
External video input
Composite output
50Hz operation
4K VRAM addressing

A few thoughts as I read through the comments here:

1. Why VGA?  Because pretty much every monitor since 1987 can display 640x480 analoge VGA.  Also, HDMI is very complicated and requires licensing.  I was looking to keep things simple so it would actually get done instead of being vaporware for another 2 years.  Also, I believe analog VGA is a subset of HDMI, so cheap adapters can probably solve that problem for a long time.

2. Why no 9938 modes?  A few reasons here, but the main ones are cost, board size, and power.  FPGAs don't have gobs of internal RAM until you reach the > 1M gates sizes, which are simply too expensive.  The 250K gate FPGA I used has enough internal RAM for the original 16K, plus some extra for the scan line buffers I need to drive the VGA.  External RAM would mean a larger board, more complicated design, more power, a larger FPGA, more cost.  I designed the F18A to directly replace the original VDP, so I also have to be concerned with not using more power than the original VDP, otherwise I could easily overload the host system.  It also has to physically fit in the socket of the original VDP, which on some systems is a very limited space.

With 16K you can get a 2-color bitmap in 6K, or a 4-color in 12K, but then you have the problem of pushing pixels vs. tiles, which is why the 9938 had the built in commands.  It seemed to make more sense to enhance the tile modes rather than make sacrifices in a bitmap mode.

Finally, the 9938 had an extra mode pin which gives it 4 ports to the host CPU vs. the 2 for the 9918A.  That means there is no physical way to make a 9938 compatible VDP that works in a system designed to work with a 9918A (without modifying the original system that is).  My goal was to not require modification to the original system, and without the extra mode pin the F18A cannot be software compatible with the 9938, so I didn't try.

There is a philosophical argument about enhancing these old systems that seems to be common among all the groups (MSX, Atari, Apple, TI, CV, etc.)  With today's cheap technology, microcontrollers, FPGAs, etc. it is possible to enhance these old systems beyond our wildest dreams.  We can make them do everything we ever wanted and more.  However, where do you stop?  At some point the system quits being an MSX (or whatever) and turns into something else.  At that point, you are along in the world with a computer that no one gives a crap about.  I did not want the F18A to become unused and unavailable.

Everyone here is bound by one thing, the MSX system.  Everyone has their own story, their own history, their own reason for having an affinity with the machine.  If you add an uber modern video processor the the MSX, you don't have an MSX any more.  Plus, you unbalance the original machine with the new hardware and it becomes hard or unreasonable to use.  Last, no one else has one and there is no software for it, thus you are again alone in the universe.

The main goal of the F18A was to make it easier to connect the 99/4A to a modern monitor because I really hate using a TV and the original VDP quality is very bad.  Along the way I had a chance to "fix" some of the stuff that always bothered me, and add a few things I always wanted (a few more colors, hardware scrolling support, etc.)  I tried to keep the new enhancements in line with the capability of the original system so the features were usable.

Where possible I made things compatible with the 9938, like the 80-column mode (except I purposely don't support the "blink" ability.)  Also, since I knew there was not enough RAM to make a good bitmap mode, I looked around for alternatives and found the NES "mode".  I never knew the technical details about the NES and I was blown away by the video specs.  I hope TI licensed the 9918A to Nintendo because the NES pretty much copied graphics mode I, with some enhancements.  The main differences are the 30-rows vs. 24, more colors per tile and sprites, and some hardware support for scrolling.  So, I added all that to the F18A and even went a little further.  It is hard to deny the quantity of software written for the original NES, so it seemed like a good thing to emulate (and even enhance a little).

To me the F18A is a way to have fun with my classic computer of choice.  It gives me the extra stuff I always wanted, and hopefully will entice some others to mess around with it and make some cool stuff.  Being from the U.S., I never had an MSX (or even heard of them until 3 or 4 years ago) and I really wish it had been my first computer.  However, it was not, the 99/4A was my first system.  The MSX software is simply awesome, and really shows what the 9918A is capable of.  The 99/4A always had crappy blocky graphics that TI put out in their software, and everyone accepted and emulated that.

I played Nightmare on my HX-10 for the first time a few months ago, and was totally blown away.  The 99/4A never had anything like that - not even close.  The BOLD demo too, simply amazing that I'm looking at a system with a 9918A in it.  I really hope the F18A finds a home in the MSX community and that people use and push it the same way they did the 9918A.

The MSX is what it is, and no new hardware add-on will change that.  The F18A simply tries to bring a little new life to our classic systems, and provide us with a little more fun when using them.

Unfortunately this has changed in the last few years and more and more monitors are appearing without analogue VGA support, particularly at higher resolutions. My 4K monitor has only DisplayPort and HDMI, and one of my 2.5K monitors has only USB-C (I'm not sure exactly which protocols it supports for alt mode) and HDMI.

Well, you don't need HDCP, so you don't need to license that, and so long as you don't claim your port is "HDMI" or use their logo you don't even need to license HDMI.

That said, yeah, it's considerably more complex than analogue VGA, and if I were making such a thing even today I probably wouldn't go with HDMI unless someone dropped an easy to integrate FPGA HDMI core in my lap.