Jump to content
pixelpedant

It's downright vexing how good MSX artists were at designing for 9918 graphics

Recommended Posts

On 2/11/2022 at 9:07 PM, Asmusr said:

It's impressive what they have done on the MSX with the 9918A, and I have often thought about porting one of those games.

When considering games to convert to the TI-99/4A, the cool thing about the ZX Spectrum is that a lot of effort has gone into creating disassembled, commented source codes: https://skoolkit.ca/links/ That's what I used to bring Pyjamarama to the TI. (I don't think anyone has proceeded very far in the game, so most of my effort was wasted. 😞)

I have never seen anything like that for other machines, and for the MSX I have never found any source code at all, but please correct me if I'm wrong.

 

R-TYPE is a game we miss on the TI-99/4A. Here's a comparison of MSX and Spectrum. Which version do you think is the best?

 

wh

 

 

Hey @Asmusr , we surely play your games! 🙂 Maybe just Pyjamarama was not the optimal choice to port a game from Spectrum to the TI. For me surely Manic Miner, Ant Attack, Dizzy or Atic Attack would have been more appealing to play on the TI. Nevertheless, your port of Pyjamarama was surely appreciated and is in my FinalGROM, MiSTer, etc. collections. I promise to try it again!  

 

Regarding R-type, the Spectrum version is in various top ten of games for this machine. Looking at the two videos, I'll surely choose this version. It would be great to have the sfx from MSX version...

 

  • Like 1

Share this post


Link to post
Share on other sites
2 hours ago, tmop69 said:

[...] Regarding R-type, [...] It would be great to have the sfx from MSX version...

Also graphics and colors I would add in case 😛
I am sure that, considering what Rasmus already developed for the TI99, an eventual version of R-Type could be still better of these original two :) and maybe a speech added would be wonderful too.
(just for dreamers) 

Edited by ti99iuc

Share this post


Link to post
Share on other sites

I guess anything's possible if you go with F18A support and abandon the 4 sprite limitation.  But for a horizontal shooter with detailed bitmap graphics and massive weapon and enemy graphics, where you can consequently do almost none of that with sprites, I feel like that MSX version does really well for itself. 

 

I have to imagine that, like Space Harrier, it would be a huge undertaking to port well (and comprehensively), due to the sheer volume of enemy and object movement and animation patterns.  There's just such so much *stuff* in there.  Every ten seconds of gameplay (in either game) is "well, that was cool, but now let's make something completely different". 

Share this post


Link to post
Share on other sites

My preference, is always for something developed not only for the F18A because it is not a TI99 but another computer.

At least a double version with possibility to activate the features of the F18A. (the same happens on the nice Scramble for example) 

 

@Tursi also done a sort of fix for the 4 sprite limitation in XB if I remember.

  • Like 1

Share this post


Link to post
Share on other sites
2 hours ago, ti99iuc said:

@Tursi also done a sort of fix for the 4 sprite limitation in XB if I remember.

I think in the end it didn't integrate well with the compiler.

  • Sad 1

Share this post


Link to post
Share on other sites
On 2/9/2022 at 8:40 AM, Tuxon86 said:

If they hadn't declared war on Jack Tramiel, they could've used a CMOS 65xx or they could've gone the Z80 or Motorola 68xx route. But know, Ti was greedy...

I'm pretty sure it was Jack whom declared war on the TI-99/4A and everything else on the home computer market. Tramiel learned all about vertical integration from TI after they used it to nearly destroy Commodore's calculator business in the 1970s. Tramiel purchased MOS technologies to control the 6502 brain-trust and CPU supply, in order to build a computer company with vertical integration, just like one of his competitors -- TI. Commodore sold 6502s to Apple and Atari, and made most of the Atari game cartridge ROMs. I believe they would have happily supplied 6502s to TI for the right price. ;) MOS was shipping CPUs to everyone with money.

 

In the end TI needed to execute around their vertical integration to compete with Commodore. The VIC-20 and C-64 destroyed the competition because they were cheap, better designs, and they were opened up to developers. The TI-99/4 was doomed when the eight bit TI CPU failed to materialize. The decision to shoe-horn the 16-bit 9900 into the TI-99/4 was the death blow. They should have shifted gears and just built a no-shit 16-bit low cost computer around the 9900. They didn't. The shoehorning was permitted because TI had no forward vision (1978) in the personal computer market which didn't yet exist. The TI-99/4 was intended to be a toy computer...a video game with a keyboard. Something closer to a Speak and Spell than a real computer. It was quicker to just shoe-horn in the 9900 than to start over...and it didn't matter because the TI-99/4 was rushed to market as a toy computer.

 

I'd suggest the lack of vision at TI prevented them from being greedy because they couldn't see five years ahead. Had they known they could have produced an industry standard 16-bit personal computer well before IBM. 

Edited by Airshack
  • Like 3

Share this post


Link to post
Share on other sites
4 minutes ago, Airshack said:

I'm pretty sure it was Jack whom declared war on the TI-99/4A and everything else on the home computer market. Tramiel learned all about vertical integration from TI after they used it to nearly destroy Commodore's calculator business in the 1970s. Tramiel purchased MOS technologies to control the 6502 brain-trust and CPU supply, in order to build a computer company with vertical integration, just like one of his competitors -- TI. Commodore sold 6502s to Apple and Atari, and made most of the Atari game cartridge ROMs. I believe they would have happily supplied 6502s to TI for the right price. ;) MOS was shipping CPUs to everyone with money.

 

In the end TI needed to execute around their vertical integration to compete with Commodore. The VIC-20 and C-64 destroyed the competition because they were cheap, better designs, and they were opened up to developers. The TI-99/4A was doomed when the eight bit TI CPU failed to materialize. The decision to shoe-horn the 16-bit 9900 into the TI-99/4A was the death blow. They should have shifted gears and just built a no-shit 16-bit low cost computer around the 9900. They didn't. The shoehorning was permitted because TI had no forward vision (1978) in the personal computer market which didn't yet exist. The TI-99/4A was intended to be a toy computer...a video game with a keyboard. 

 

I'd suggest the lack of vision at TI prevented them from being greedy because they couldn't see five years ahead. Had they known they could have produced an industry standard 16-bit personal computer well before IBM. 

So... It was TI then, as you just wrote, who started the war with Tramiel... Without that fact, he probably had sold 6502 to Ti as well.

Share this post


Link to post
Share on other sites
1 hour ago, Tuxon86 said:

So... It was TI then, as you just wrote, who started the war with Tramiel... Without that fact, he probably had sold 6502 to Ti as well.

He would've sold 6502s to TI if they'd asked. Pretty sure he'd be willing to make a couple bucks on every system TI sold.

 

TI actually CONSIDERED the Z80. But that was shot down because they didn't want their computer to use someone else's architecture.

And then the 9985 didn't make it into production and they were stuck with a "real" 9900 shoehorned into an 8-bit bus, or more delays for an already-late product in a fast-changing market.

  • Like 1

Share this post


Link to post
Share on other sites
8 hours ago, Airshack said:

TI had no forward vision (1978) in the personal computer market which didn't yet exist. 

 

Or maybe

 

"TI remained greedy and afraid of cannibalising its minicomputer business, and never released the software to TMS9900 users, with the result that they ended up a has-been in both the minicomputer and microprocessor businesses."

https://www.fourmilab.ch/documents/marinchip/

  • Like 1

Share this post


Link to post
Share on other sites
1 hour ago, Elia Spallanzani fdt said:

 

Or maybe

 

"TI remained greedy and afraid of cannibalising its minicomputer business, and never released the software to TMS9900 users, with the result that they ended up a has-been in both the minicomputer and microprocessor businesses."

https://www.fourmilab.ch/documents/marinchip/

That was definitely a problem, but not one unique to TI.

And, well... TI at least successfully LAUNCHED a microcomputer, and one that was fairly successful for a time(though certainly their quietly-murdered business microcomputer was the machine they actually needed). Most other companies with minicomputer divisions failed even that, with one very obvious exception. I feel they deserve more praise for overcoming the minicomputer inertia far enough to actually try and sell personal computers.

 

 

The smartest decision made in IBM's microcomputer project was setting the team up on the opposite side of the continent from the rest of the company.  Someone at IBM recognized that this emerging threat to the minicomputer industry was not something they could stop, but IBM's minicomputer division had enough clout to make life very hard for IBM's microcomputer project if they could get anywhere near it.

So while other companies cancelled microcomputer projects because they were "a fad", mandated performance-crippling sacrifices, and forbade them to have 80-column displays to prevent them from being usable in a business environment... IBM released a good(-ish) machine that did a lot of what a business considering a minicomputer would want at a fraction of the price.  (IBM would, of course, bring the PC team back home after launch and place them within reach of the minicomputer division, leading to the AT being crippled in an attempt to protect IBM's minicomputers. Whoops.)

 

Point is... had TI similarly insulated their home computer division from the reach of the dreaded data systems division.... well, we might be talking about the foundation of a standard instead of a largely-forgotten also-ran.

  • Like 3

Share this post


Link to post
Share on other sites

A lot of this stuff, including all the crazy decisions made buy TI, are chronicled in Albright's The Orphan Chronicles. A fascinating read. Names are named and no blushes are spared!

 

TI was greedy - the vertical integration idea is great if you're selling your chips to the wider market. Then you're getting your own chips supplied for cost, and other, external customers are providing the profit margin. However, when they're not selling (and the TMS9900 wasn't) then your chip costs have to be soaked up by the end product that you are selling (the 99/4). If I remember, the 9900 cost TI about $7.50 1979 dollars to sell 'to themselves'. They would have been better saying screw it and just put a Z80 or 6502 in it for a buck.

 

I think there was all sorts of accounting shenanigans also going on to give the illusion that the 4A was profitable, when it actually was not. Also the power supply recall issue (IIRC). Eventually it was all uncovered and the axe fell swiftly. It's all in Albright's book. 

  • Like 3

Share this post


Link to post
Share on other sites
5 hours ago, Willsy said:

Found the book in PDF form on 99er.net - enjoy!

Ah, cool! After years of hearing about it I can finally read it ;)

 

  • Like 2

Share this post


Link to post
Share on other sites
21 hours ago, JB said:

So... It was TI then, as you just wrote, who started the war with Tramiel...

Yes, TI started a war with Commodore but earlier and in a different market -- calculators. They destroyed a number of clients and nearly killed Commodore pre-PET.

 

From the Apple, Atari, Coleco, or Tandy point of view, it was Commodore destroying profit margins in the Home Computer market. A second tech-war which Commodore started and appeared to be the aggressor. Tramiel's race to the bottom in pricing cost even Commodore millions in profits. He spontaneously and unexpectedly cut all Commodore peripheral and software prices by 50% at the 1983 CES. He basically cut off his right arm in order to kill the TI-99/4A which contributed significantly to his termination. Commodore lost millions over that decision which was viewed as a disaster by the Commodore board, especially Irving Gould.

 

Edited by Airshack

Share this post


Link to post
Share on other sites

I like this computer because it has allowed me to do both hardware and software that wasn't planned from the beginning. Like the 64 K 16-bit contiguous RAM in the console and implementing bitmap graphics in the UCSD p-system. It's not larger than that you have a real chance to understand what it does, within certain contexts.

Share this post


Link to post
Share on other sites
18 hours ago, PeteE said:

I've found another gallery that has some really nice MSX artwork.  Some of the shading and dithering is incredible.

This is some great art, but the screenshots are from emulators. So the old ones would look different on a CRT, and new ones could be perhaps done differently if made on, and for, a CRT.

Edited by youxia

Share this post


Link to post
Share on other sites
On 2/25/2022 at 7:57 AM, youxia said:

This is some great art, but the screenshots are from emulators. So the old ones would look different on a CRT, and new ones could be perhaps done differently if made on, and for, a CRT.

So I discovered that some of my favorite interlaced-style bitmaps were, in fact, converted using a program, made by a person in this MSX forum thread.  I converted some of his examples into a viewer program cartridge, as well as some from the previous linked gallery, so you can see for yourself on CRT.

 

bitmap18.bin bitmap28.bin

  • Like 5

Share this post


Link to post
Share on other sites

To be clear, I was just being pedantic seeing as CRTs and their inimitable "look" are one of my fixations. I did not of course mean to belittle the artistic merits of your examples in general - they are indeed amazing. MSX is one of my favourite micros now, I got the bug for it last year and must exercise some strong will to not buy more than the two that I already have.

 

And the use of its colour palette is quite astonishing, seeing as it's not my favourite one - I much prefer Spectrum/Amstrad/Atari colours. But MSX games look much better to me (colour wise) than the somewhat similar C64 ones.

 

I've tried to capture some of that composite magic for my CRT SCR$ collection, it's damn hard though.

  • Like 1

Share this post


Link to post
Share on other sites
3 hours ago, PeteE said:

So I discovered that some of my favorite interlaced-style bitmaps were, in fact, converted using a program, made by a person in this MSX forum thread.  I converted some of his examples into a viewer program cartridge, as well as some from the previous linked gallery, so you can see for yourself on CRT.

Interesting... you can get a similar result, though not identical, by changing the error distribution in Convert9918. Just move all the error down to the second scanline, with none on the current. I also changed the error mode from accumulate to average, which reduces the amount that the error can travel. It's not exactly the same - his converter strictly manages even/odd scanlines, and Convert9918A doesn't have a concept of that except in the ordered dithers. (Which are not currently configurable). But I think the increased flexibility gives a better result. There's more color mixing, especially on the curve of the planet. Although, the choices of color will vary, since my settings and color selection algorithm are going to be different from his.

 

image.thumb.png.617a0199708030ec0395463ae42c3e73.png

 

I have lost a lot of love for horizontal dither as well, except for the ordered dither. I do feel that provides the best compromise on the 9918A. Especially if you're talking about viewing on a CRT with composite video. Ordered dither essentially does exactly what that person describes in their thread (the concept of mapping to pseudo colors), but does it in 2D instead of 1D. There's still some color clash, of course, but the regular dot pattern does a good job of masking it. Note this conversion still uses average on the error mode.


image.thumb.png.0f8b251d773ca62f4d5ef87eebb3efdb.png
 

  • Like 4

Share this post


Link to post
Share on other sites

As a kid, I felt the lack of certain colors like brown, orange, pink flesh.. I used dark red for brown or orange. Red cowboy hats, or leather armor, just never looked right. Yellow for faces was ok as a convention, but thin yellow arms were a color-burst disaster.  On my Geneve, I redefined dark red as a good brown, to revisit my sprite intentions.

 

But looking at those MSX images, my eye is fooled into thinking I see pink flesh where the color is actually grey, but framed by magenta. I wonder if that holds true on a CRT? I liked:

 

55964 Trek

34986 Hunter,

55625 Genie v1 (v2 is 55632) 

35115 Phantis,

but 57264 Oban didn't do it.

 

I'm pleased! But questioning the fairness of showing emulated pixels, not affected by CRT.  

Actually, looking again, it's not working as much for me.  I turned up the brightness since then?

 

On its own, Sunset City 50551 is stunning.

 

Similar perceptual tricks in dithering.  @Tursi's Mario above dithers white and reds in the face.  But since it has a color gradient on the nose, my brain expects the shadowed red gradient to lead to fleshy pink.

 

54202 Amazonia puts most on a black background, which I expect would be credible.

 

Share this post


Link to post
Share on other sites

Thanks @Tursi, I had tried convert9918 to emulate the interlaced-lines effect, but I didn't figure out the right error distribution.  Thanks for that.  You are also right that order4 dither look s pretty good too, so I made it into a cartridge too.

mariodc.bin

  • Like 3

Share this post


Link to post
Share on other sites

From what I heard, the Z80 and TMS9918 woodgrain prototypes that were made ran circles around the 9900.  They only exist in memories at this point.

 

The Tomy Tutor and 99/2 both did decently architecturally for having basic chipsets.  

  • Like 1

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...