General News

UME (Universal Machine Emulator) combines the features of MAME and MESS into a single multi-purpose emulator. The project represents a natural course of development for the emulators which already share large amounts of code and is part of an ongoing effort to unify development efforts and provide a single emulation platform for users and developers alike.

As an end user this means that the software provided here is not only capable of emulating arcade machines like the baseline versions of MAME, but in addition can emulate a large number of home computers and consoles from across the world using the very same code, developed by the very same team of developers.

0.148 Windows binaries (32-bit and 64-bit) (Self Extracting 7-zip) (all MAME / MESS tools included, both 32-bit and 64-bit versions in tool32/tools64)

The source is identical to that found on mamedev.org

General Notice

As of 0.148 the MESS binaries are offered as an optional extra to the MAME base package on mamedev.org, while this isn’t a fully merged solution like UME it is a step in the right direction towards having MESS recognized as part of the MAME family, which is one of my goals with UME.

I hope in the future we can see them offered as a single package, and eventually merged, spawning the ‘classic’ builds (Arcade-only emulator without any of the pinballs, gambling stuff) alongside a complete build with everything (Arcade + Home) Some people remain unconvinced over this even if it is already the direction in which development is heading. Doing that would give a ‘best of both worlds’ scenario, with the main release showing our dedication as a team to the documentation and preservation of everything we can (a unified voice of what I hope UME shows right now, rather than it just coming from me) along with the Classic version for people who really don’t want some of the recent ‘bloat’ and prefer it to simply emulate arcade video games and nothing else. Everybody would win, I think?

What’s New

You can read the various whatsnew files on mamedev.org
From MAME, From MESS

Points of Interest

There are some really good additions this time around, as well as bug fixes for some long term lingering bugs (such as the control redefining problem for some of the Consoles) Preliminary disk support has also been restored for systems like the CPC (although protected disks won’t work yet until support is improved dramatically) as well as the ability to change disks with softlisted software working again (although not with 7z compreed files due to some pending debate and the need to establish an actual .diff format for writable disks, rather than trying to modify the content of the original files, which you really don’t want to be doing for softlisted media anyway!)

The Rise of the Robots proto made it in, so as I’ve mentioned, that’s if you like trainwrecks that’s no doubt one to study. Mahnattan for the Deco Cassette is there too, along with a couple of Deco Cassette bug fixes I made to support it. Grasspin too, in a working state even if there are a few things to be ironed out.

From MESS the PV1000 has been promoted to working, a curiously underpowered system from Casio to which some not terrible ports were done, although the limitations of the hardware are quite apparent even from those.

HLSL stability fixes have also been put in place, so it now copes better with games doing dynamic resolution changes, which previously were a very good reason to keep HLSL off altogether. If you like HLSL you won’t want to be using anything pre 148 because those fixes are essential for a significant number of popular drivers / systems. The vector games may still have issues, so you might want to create a vector.ini with HLSL turned off, although it has been said a more fancy HLSL vector renderer is in the pipeline, something I’m sure fans of AAE will be glad to hear.

A word of caution, the -mt multithreading option was turned on by default, on my system that still causes issues with noticable random graphic corruption / flickering in several drivers if running unthrottled, if you experience anything like that you’ll probably want to turn it off; the performance benefits are negligible to say the least and I was really tempted to flip it back off for UME (but wanted to stick to the baseline code)

One of the other potential highlights is the Microvision support, also known as the ‘very first handheld’ Unfortunately it seems like nobody added any kind of Software List for it, which makes knowing what are valid images to run on it for testing rather tricky. I’ll probably revisit that one in one of the u updates if one gets added. The cartridges all contain only an MCU, which acts as the CPU, although with a resolution of only 16×16 monochrome pixels as you can guess it is going to be a very limited system!

Once again proving the projects are an international effort and just not about the popular systems everybody knows there has been work on a number of relatively obscure systems including the Hanimex PENCIL II, a system mostly known in Australia (although apparently manufactured in Hong Kong) It’s more or less an MSX style machine, although again there is no software listed and it apparently does not have the capability to load in any software at the moment (it was placed in the wrong section of the messnew, it isn’t actually marked as working yet) Still, it’s one to watch if some dumps turn up, otherwise you can play with the built in basic.

(Stay tuned for more) (sorry, not been too well these days)

I’ve been holding off doing an update on this one for a couple of days because there are still a few issues to iron out, however the visuals seem OK now even if I remain unconvinced over the implementation. Anyway, on to the update.

Italy has been a treasure chest for rare PCBs turning up for a while now and what we have here is another one. Still on a high from finding Planet Probe, the duo of Alberto Grego and ‘Zabanitu’ found what they thought was a regular Jaleco ‘Saturn’ PCB, however on booting it they found out it was something much, much more interesting. What they actually found they were in possession of was a game called ‘Grasspin’ running on the same hardware, a game which there is apparently no record of in Japan, or anywhere else for that matter, a piece of history completely forgotten until that moment in time.

Now, there are likely several reasons why there are no traces of the game in the history books. The most likely of those reasons is that it’s a prototype, an unreleased game, something which failed it’s location test and was shelved; although you still hear about many of those and we have screenshots from plenty of long-lost games falling into that category so that alone really isn’t reason enough alone for it to be completely unheard of.

The second contributing factor requires a little more digging. The games Jaleco released on the ‘Blueprint’ / ‘Saturn’ board were not developed by Jaleco. We know from credits in the roms that a company called Zilec were involved in the production of the games. Zilec were a UK company and produced a number of conversions, as well as a selection of games exclusively for Jaleco. Jaleco also oversaw the arcade game Dingo, produced by Ashby Computers. What all this has in common is that it represents the history of a company better known back then as ‘Ultimate Play the Game’ or today as ‘Rare’ If you look in the sound rom for Grasspin you’ll see credit given to “CHRISTOPHER STAMPER,JOHN LA” (which I’m guessing should be JOHN LATHBURY, as credited in Saturn, but has been truncated)

Now, assuming those credits aren’t just a leftover it would strongly indicate that that game was also developed by Zilec (and the main character sprite reminds me a lot of Blue Print) The default high score table does throw some doubt on this because the names all appear to be Jaleco staff rather than Zilec staff, but it is rather traditional for this to be the case. Anyway, if we go on the logic that the game was developed in the UK, and tested in Italy it’s possibly it never even reached the Japanese shores at all, despite Jaleco’s involvement. That doesn’t explain why nobody remembers it here, but we’ve never been as thorough with record keeping it would seem whereas the Japanese even have high score lists for games many consider to have never been released!

As mentioned at the start, there are a few peculiarities with the emulation, first off there seems to be a minor comms issue between the two CPUs (the existing hookup for Blue Print / Saturn is apparently wrong) this causes you to always have infinite lives, because it thinks that dipswitch is always on (the dipswitches are read via the AY sound chip on the sub-cpu side then passed to the main CPU) It also seems to access video ram in a rather unusual way, often writing the upper tile attribute bits to the line after they’re needed, but a dump of the video RAM from the PCB seems to show everything in the right place. For the first reason the game is still tagged as NOT WORKING, although it is playable.

The game, it’s an a decent concept, but not really too well implemented and probably shows the weakness of the hardware a little too much. The whole thing centers on the novelty of being able to ‘spin’ the grass (the lines which make up the maze) by pressing a button, so yes, while the name ‘Grasspin’ may seem like a nonsensical word at first it’s simply a pun on the core game concept. Every rotatable piece of the maze rotates at the same time, so you have to carefully plan when you flip based on your position, the enemy position and where things will be as well as which paths are open after you rotate.

That’s where the flaws start to appear, the biggest issue is that it’s not always obvious where anything is going to end up. Every rotatable piece has arrows on it showing the direction of rotation which should in theory help, but in reality if you get pushed or not doesn’t always feel obvious and the whole rotation effect is done with an absolute minimal number of frames of animation (due to hardware limits) so the visual effect isn’t very good and quite often you end up pushing yourself into an enemy or an enemy into you or ending up on completely the opposite side of a barrier than you expected after feeling you you transported through it rather than being pushed.

The frustration continues because the red grass near the top of the maze will push you further, and kill you and/or any enemies that happen to be standing in the way when you rotate it… sometimes.. and that’s the problem again it’s never really obvious if it’s safe or not. I’m sure if you played the game enough you’d get used to it, but that feels more like learning your way around the flaws of the game than playing it.

Now inevitably there will be times when you get trapped, and on some lines of the maze are little life rafts, if you grab one of these and continue walking you’ll float down a river to the bottom of the playfield as an emergency escape, be warned tho, turn around for a second and you’ll lose the life raft, at which point jumping in the water is fatal, another odd design choice. Not to mention that the bottom of the playfield isn’t always safe and you can end up with just as many enemies congregating there making the whole escapade pointless.

In all this I’ve failed to mention what you actually need to do to finish a level. It’s actually rather simple, down the sides of the level are 6 flags, marked A-F, collect them all and you move on to the next level, collect them all in order and you get a bonus. The various bugs are what make this game look prototype-like and there’s another one when entering the areas containing the flags; if you flip the maze while entering you’ll momentarily start walking straight through the outer wall until the game realizes something bad is happening and moves you back.

Most enemies abide by the movement allowed by the maze pattern, but one of the cheaper enemies is a little Cupid / Angel thing floating around, slowly moving in diagonals and homing in on you, quite often you’ll end up trapped waiting for enemies to move, and dying because this thing gets to you before you have an escape route. Later levels / harder difficulties also add some kind of ball which just spawns somewhere near you, bounces towards you and kills you if you stand around without moving for too long, very cheap.

Level progression is almost non-existent, the mazes don’t change at all, just palette swap, and enemies become greater in number and speed as you go through the levels. In that sense it’s no different to a lot of classic games, but it doesn’t feel very rewarding because the game just starts to feel very cheap after a few levels, especially when as I’ve mentioned at the start it can be quite hard to judge where you’ll end up at times.

The sound / music is pretty decent for a game of the period, although without doing a video it’s not something you can grasp from the screenshots alone, and I don’t want to do a video until the remaining issue with the dipswitches has been ironed out.

I’ve no idea what the game-over / high-score sequence is like because as mentioned earlier you currently have infinite lives.

The game was definitely never destined to be a classic, although I’d say it was memorable so it does surprise that nobody seems to remember it at all, not even in location test form, but that’s just the way things are sometimes.

Not a great deal to report here, but I just thought I’d mention that Manhattan has been improved since the last batch of screenshots and the background scrolling is much more correct now. I’ve also done some general purpose cleanups.

The reference videos we have (YouTube) show the colours cycling between levels, that doesn’t happen yet in MAME, and to be honest I can’t see any registers for it, I’m starting to wonder if it’s a different version. I’ve asked Charles who dumped the cassette if he can test the version we have, but we’ll have to see. It could just be a missing emulation feature, considering the clown-car credit lists often associated with Deco Cassette the driver is still in rather bad shape but I’m running out of places to look!

While on the subject of the Cassette system it does seem to have been designed primarily with the ‘Highway Chase’ game in mind, and in many senses the feature-set of the hardware is very similar to the standalone Mad Alien / Highway Chase hardware, but seemingly a bit weaker because the actual Highway Chase released on the Deco Cassette system is a worse looking, and rather glitchy game if the references we have are to be considered accurate; it even has bizzare pink/blue desert scenes and water scenes with green water, despite the correct palettes being right there! Furthermore the dividing line in the road isn’t present in any of the reference videos despite the writes for it and emulation of it in the driver?!

Anyway, the background ’tilemap’ is just one of those things that turned out to be very close to Mad Alien, instead of being one simple tilemap it’s actually a left ‘edge’ and right ‘edge’ (which can have independent scrolling to bring them closer together / further apart, while sharing a vertical scroll register for the overall vertical movement speed, clearly designed to draw a road capable of becoming wider / narrower..) Furthermore, there are multiple sections to the tilemap and there is a register to control looping of the top part only, or the bottom part only, or doing a transition between the two, again very much designed to do different sections of a road, allowing for tunnels and the like; that feature was missing hence Manhattan not looping the correct section of the tilemap properly. Fixing this also fixed the annoying background looping in Super Astro Fighter (I also fixed the colours there, the driver was failing to actually store the written palette values in RAM, but then expecting to be able to read them back)

Clearly it was a cheap system for low-budget games, trying out ideas, and creating new revisions etc. without breaking the bank, but that does sometimes make it hard to know if bugs in the driver are real game bugs, or actual emulation bugs, especially when you’re not 100% sure your reference material is the same version as the dumped version. Sometimes there are obvious clues that things differ, the Astro Fantasia videos we have show a completely different main boss, the Graplop videos we have seem closer to the Starcade version with an actual Graplop title drawn whereas the ones we have are the Cluster Buster variant, and what seems like a very early version of Graplop before any polish was added.

I should have another previously unknown 80s game to show here in a couple of days thanks to the same guys who uncovered Planet Probe, but more on that when it happens :-)

I was looking a bit at the protection on the Pgm 3-in-1 but haven’t been able to make any logical conclusions about some of the ops, even what should in theory be the simple ones controlling the movement of the cards. I’ll have to do some hardware tests.

Charles MacDonald and The Dumping Union managed to get another one of the rare Deco Cassette games secured, and dumped. This time it’s ‘Manhattan’

A video of the original game was uploaded to YouTube a few months back, taken from a Japanese arcade where the game was actually still being operated (the same place often have rare classic period games for a short period of time) but beyond that sightings of the game had been few and far between, it apparently did make the US shores back in the day but was quickly replaced.

To be honest, it’s not hard to understand why, first of all the game lacks any kind of attract mode, it just cycles the title logo sequence over and over, something which I consider to be a deathwish for any arcade title (surprisingly Nam1975 got away with it, although the CD version oddly rectifies that for an environment where it doesn’t matter) If you can’t see the gameplay, why would you risk playing the game? At first I thought this might be an emulation glitch, but if you look carefully at the video it does the same, just starts drawing the title logo again after it does the colour cycle and has displayed the DECO logo.

Then there is the secondary problem, it isn’t really obvious what you’re meant to do, arcade games should be obvious without having to read the instructions. In this case I managed to waste several attempts simply dying because I didn’t understand that you had to press a button when your character falls on the trampoline, failure to do so means you get no height at all and results in near instant death and one life lost, no second chances or prompts.

Finally, the actual game. Once you’ve been launched up to a certain height, which depends on when you press the button, all you really do is fall down while attempting to steer your character towards the bounce pads (balconies..) at the side to gain points, and possibly a bit of extra height depending on how fast you’re going, but losing most control over your character each time you do so. To end a stage safely you must always land on the bottom most pad, which generally means you can’t be going too fast by the time you reach that point. Obviously there are things to avoid too, a feat much trickier each time your character goes spinning out of control; glancing the edge of a platform also does you no good at all. Some of the pads also contain letters to spell the word ‘DECO’ landing on the same one twice also seems to end the stage. It’s an original concept, but I’m not convinced it works as a game. I might be missing something, I’ve not quite understood the meaning of the word ‘Bound’ which is displayed (maybe a target score you have to reach to finish later levels?) but there really doesn’t seem to be much more to it.

The final nail in the coffin has to be the sound, the whole thing sounds like a cross between Galaxian and a fire alarm, it’s truly nauseating.

Of course, it’s good to see it emulated, these Deco Cassettes remain some of the more fragile and ‘at risk’ of the arcade games, I think they also gave Data East a cheap way experiment with ideas at a low cost, hence the various evolutions of gaming concepts / code and alt versions of the same things you find turning up (there are still missing versions of Graplop / Cluster Buster / Flying Ball for example, which just seems to be the evolution of a single game) (personal note, cgraplop2 doesn’t seem to let you coin up / start at the moment, investigate)

Currently the emulation has some issues, it’s most likely these are down to incomplete emulation of the Deco Cassette video system, rather than a bad dump, they affect the rendering of the building graphics at the side of the playfield which is rather obvious in the screenshots below compared to the video above, and even moreso in action. Several other Deco Cassette games have similar issues, which is why I’m inclined to think it’s a video issue, not a bad dump; the only cassette game I’m rather suspicious of when it comes to the dump is “Super Astro Fighter” due to it erasing the game palette, although that could be a CPU or protection issue of some sort (stupid problem, this driver is old and a bit crap.) Anyway, what this means is I should take a look at the video emulation at some point in the future if nobody beats me to it.

MAME screenshots

So thanks again to Charles MacDonald and The Dumping Union, the only work I’ve put into this myself is translating their analysis of the protection dongle scrambling into actual MAME code so that the cassette loads in MAME, but because they’d already done all the hard work that part was easy enough :-)

In other news Phil B. submitted his driver for Rise of the Robots, the failed prototype of the overhyped pre-rendered piece of junk fighting game from the 90s which was meant to sell computers / consoles and change gaming forever, so this Christmas / New Year could well be an enjoyable one if you like to witness rare train wrecks ;-)

*additional Manhattan notes*

After playing it a bit more, and actually figuring out what is what (with the sound turned off) the game is a little different to how I describe. The main crime remains a lack of an attract mode / how to play instruction screen / some prompts which would have made the gameplay clear, but essentially your ability to land on any given platform depends upon the speed you’re traveling at, except for the lowest one which you can always land on as long as you hit it. That’s what the ‘bound’ text is all about, it means that the current speed will cause you to bounce, not land. Still, the game is quite shallow, a lot depending on the moment you hit the launch button (the exact right time and you’ll launch high enough to land on the roof for some bonus points) There also doesn’t seem to be much in the way of level progression aside from some more annoyingly placed obstacles which, including ones in the center of the playfield you’ll need to avoid on the way up too.

It remains an interesting concept, and an important part of Data East history, and possibly with a bit more polish could have lasted a bit longer as a more popular game, but without a more depth or addictive qualities I don’t think it was ever destined to be a classic. I think if somebody did a more modern remake it might be possible to make more of a game of it but it feels more like it should be a bonus round of a larger game than an actual title in it’s own right.

One recent surprise was a previously unknown PGM multi-game cartridge showing up with Smitdogg / The Dumping Union, containing Photo Y2k2 as well as 2 poker games.

It’s an official release from 2004, the same year several other games were re-released, probably as some kind of stock clearance exercise.

Being a 2004 release it does mean it’s almost certainly the fully secured ARM / 027A type, so can’t be fully dumped, it also lacks external ROM, so we have no way of running our own code on it directly even if an exploit was found.

Being a game without an external ROM the protection routines are typically weaker tho, more ‘functional’ than ‘most of the game runs on the ARM’ like several others which means it might be possible to simulate it, although this builds on what already existed with the Photo Y2k2 protection, and that hasn’t been done yet either!

In addition to the protection from Photo Y2k2 it appears to have some extra bits to get the write offsets of the background tilemap (used in the intro) which is exactly the same as Knights of Valor (PhotoY2k doesn’t even have any background tilemap roms, so obviously it wasn’t used on that) There is also protection controlling the movement of the cards (not understood) for the poker game in both the attract demo, gameplay, and game select screens. Selecting a game results in the same value out of range crash as Photo Y2k2.

I’m hesitant to write ANYTHING about PGM here, because as soon as I do I seem to get a flood of annoying SPAM to my email, but at the same time I do like to keep people informed, I’ll just keep hoping the spammer goes away for good.

I’m not sure of the exact title of the game, the internal string calls it ‘Flash 3-in-1′ the way the letters have been decorated on the logo makes me think another Logic / Puzzle King, but clearly the primary game is Photo Y2k2 with this just being used to clear old stock mask roms of that game. I say that because the PGM system has a max capacity of ’16MB’ for the bitmap (sprite) shape data, PhotoY2k2 was already using the full 16-meg area with 2 MASK roms (2x8MB) Instead of making new mask roms to add the data needed for the extra poker games they’ve instead used the original pairing, and used an additional rom to act as a patch / overlay over the last 1MB of those roms. This to me indicates they had an excess stock of the original mask roms they needed to use it.

In terms of regions it looks like this was only released in China, Taiwan and Hong Kong because the region warning screens for all other possible regions are incomplete or just corrupt. As usual the region is provided by the protection device tho.

Anyway, screenshots from the non-working emulation.

I’ve also uploaded the original hardware reference video Smitdogg provided (it would be nice if he threw all his vids on YouTube even if the quality is lower than the direct downloads they complement them well) Note how the cards get dealt out during the attract, then spin around in a circle, all that movement is controlled by the protection commands. I’m not embedding it, because it’s not an emulation video.

Smitdogg’s thread about it can be found here

One of the cool things about UME / MESS is the way it unites the various ports of games under one emulator, which can be interesting to study. Take Pacman, here you have the classic arcade

Now, when you’re running MAME you have that, and only that (yes I know there are remakes and the Namco Anniversary one is really quite neat, but that’s an aside)

When you add MESS, you can look in the Software Lists and find many more versions of Pacman listed, some official, some unofficial / unlicensed, but all of them trying to be the same game. There are probably many more versions out there too, the lists aren’t complete by any means!

Some versions are good, and play close to the original, some are terrible and a million miles apart (the notorious A2600 version) Maybe it’s just my nerdy / geeky side coming through, but having an emulator capable of this makes for an incredible adventure, being able to see the limitations of each machine by studying the emulator code, and then being able to see what they came up with.

Of course Pacman isn’t the only game ported to lots of platforms, plenty of the other classics were for example, Mappy, see the original arcade

and some ports… (I was a bit hesistant over putting the Super Cassette Vision one because it runs too fast to play, but it ‘looks’ ok)

Some platforms got ports of more obscure games instead, for example the Casio PV-2000 got .. Super Pacman, under the title Mr. Packn (no idea why they used that title, it’s a licensed port)

and it wasn’t only Namco games which received a lot of ports, Nintendo was popular back then too, with the Amstrad CPC getting what is usually considered one of the best home ports (there is a graphic glitch in MESS on the lives bar tho)

Now it has to be said that all of these were in an era where the arcades were vastly more powerful than the home systems so obviously the home versions do end up looking inferior, but from a technical point of view many are actually much more interesting due to their creative use of the hardware available.

All in all it’s reasons like this why I consider the MESS component of our project to be possibly the most fascinating bit (and definitely where the future is) that’s why I offer it as part of the MAME project with my UME binaries. MESS might not be perfect for all systems, and it does have a learning curve, but even with systems I’m mostly unfamiliar with I’ve managed to boot up, run, and take snapshots of many games even if a couple of the systems did take a little figuring out and digging up some basic usage instructions for or plugging virtual joysticks into slots before I had any controls.