X68000 information sponging

[kamiboy]

Are you sure most games that display in 31khz actually render in higher vertical resolution than 240p?

I've heard otherwise, and given the vintage of the hardware I find it hard to believe games actually having assets and actually rendering at 31khz resolutions.

For Arcade ports at least I doubt they redid the sprites at higher resolutions just for the X68000 version. If any of the arcade ports run in 31khz it must most certainly be via line doubling.

I am pretty sure I read that Akumajou runs 240p natively but for some reason only supports being displayed pixel doubled.

I think even for hardware as powerful as the X68000 rendering games at 512x512 is a rather tall order. Its hardware was designed in the 80's and home consoles didnt even hit those resolutions until the 2000's.

Still, I prefer sprite games to have scanlines. Without them they look like PC games which is a look that I do not much care for. So yeah, without a shadow of a doubt I would say line doubled games displayed in 31khz is a majour flaw.

[SuperDeadite]

Well arcade ports were typically redrawn, but there was actual quality thought put into it. For example, Fantasy Zone has 31 and 24 modes. Straight from the manual, in 24khz mode, the game is dot x dot with the arcade original. But of course the arcade was 15khz only.

However, capcom chose not to do this, so all their games default to 15khz. And look icky if forced to 31khz.

The vast majority of X software runs in high-res and does not have a low-res mode. High res was chosen as most monitors are only 15'' and in low res mode, the display can't show the entire picture. In low-res you get much thicker scanlines, but again the size of your display will make a huge difference in the end.

Dracula was made for 31khz. The ps1 version seems to suffer from downscaling imo. Something is very off about it, possibly the hit boxes. For years on the PS1 version, I couldn't get anywhere in "original" mode. On X68k? I beat the game in a week, it's smoother, and controls better for sure. I even went back to the PS1 version recently and again found it much harder.

Finally in low-res some games actually seem to run faster then they normally would. Nemi 90'Kai in low-res makes a hard game borderline impossible. But this also helps some of the more intense games like Geograph and SuperSFII run smoother.

There is a lot of misinformation about this computer (especially in English). My youtube videos often get spammed by rapid fanboys of a certain popular european computer. Jealous morons.

In the end on my original monitor everything looks beautiful and I have no complaints. I've seen people say Daimakaimura is interlaced only, and it's ugly in the DSP 31khz modes, but in it's default 15, it never fails to wow me. If you want true arcade perfection, stick to the arcade games. It's the extras and originals the make this machine worthwhile.

[trap15]

X68000 doesn't render in software, so the only difference as far as software is concerned is how long the H-blank is (for raster effects, and other timing stuff).

X68000 software almost always runs in 31kHz.

[kamiboy]

Trap15, but what native resolution do X68000 games usually run in? If it is 512 vertical as Deadite says then colour me impressed, very, very impressed. That is insane for hardware sold in 1987.

SuperDeadite, I guess by those Euro computer fans you referring the Amiga 500 crowd. I remember the 500 growing up, it was a nice piece of kit, but in my opinion the Japanese developed X68000 library runs circles around the mostly European made Amiga 500 library.

In hindsight about the only thing Amiga 500 games habe going for them is the music, which was often of exceptional quality. They fare far worse in terms of game design.

[trap15]

Usually anywhere between 256x256 and 512x512. Usually just one of thkse two, but Capcom stuff runs at some weird resolution for the most part (384x224 I think, to match CPS1 resolution).

[kamiboy]

Those square resolutions puzzle me. Must make sprite asset production for X68000 games a big hassle. I wonder if the artists compensated for the none square pixels as they drew assets or just didn't care about everything getting stretched horizontally when being displayed.

I have always wondered why consoles games that natively run at a different vertical resolution than what the TV can display look alright anyway. You'd think with the scaling effect there should be all sorts of nasty artefacts, but they look as if being displayed 1 to 1 vertically.

In any regard, I wonder how many sprite based games ran in 512 vertically.

[Ed Oscuro]

natively run at a different vertical resolution than what the TV can display

This is based on a misunderstanding (I used to think the same thing): at all times these games used a 4:3 aspect so it's meaningless to talk about the 1:1 resolution as if it was "running" somehow. Both the screen and a television (say the Chronicles port of Dracula on PlayStation) are 4:3 aspect ratio.

It's maybe different when talking about composing sprites, as you mentioned. I often thought that Simon looked fat in 4:3 aspect for Dracula, compared to 1:1 pixel aspect, but that's just how it looks, regardless of platform.

[trap15]

Those square resolutions puzzle me. Must make sprite asset production for X68000 games a big hassle. I wonder if the artists compensated for the none square pixels as they drew assets or just didn't care about everything getting stretched horizontally when being displayed.

Some did care, some didn't. Definitely game dependent.

I have always wondered why consoles games that natively run at a different vertical resolution than what the TV can display look alright anyway. You'd think with the scaling effect there should be all sorts of nasty artefacts, but they look as if being displayed 1 to 1 vertically.

Because there is no artifacting; it's different sized pixels. Nothing gets "stretched", the pixels are just thinner. The reason they'd have resolutions that aren't 4:3 is to increase resolution. A 512x512 image on a 4:3 screen adds quite a bit of vertical resolution that you wouldn't get from a 4:3 source image.

[Ed Oscuro]

a 4:3 source image.

*square pixel, it's still 4:3 source after all

[trap15]

Yes, yes. You know what I mean

[Ed Oscuro]

Of course, but I think it's worth just reiterating that there is no tiny gremlin inside the PC watching a tiny square monitor. Aspect ratio really doesn't exist outside of the output.

[SuperDeadite]

It's also worth noting that many X68k games use "forced wide screen" modes. For example Cotton, SFII', SSFII, Final Fight, and many other games have a wide mode. But all X68k monitors were 4:3. You actually have to shrink the vertical size pot on the monitor itself to make it go "wide screen." If you run it full screen, it will be improperly stretched. Very easy to see in Cotton, as the spell power icons should be square, if I don't adjust the monitor, they become ugly vertical rectangles. This was really strange at first being a computer, but these kinds of adjustments are normal for any arcade cab owner. And the X68000 was clearly a product of arcade design over that of traditional computers.

As for Japanese computers running in very high resolutions, this shouldn't be so surprising. Kanji is essentially unreadable in 15khz unless you use huge fonts. Therefore even the earliest PC-88s supported 24khz. It was the main reason NEC dominated the buisness world so early.

[kamiboy]

Of course, but I think it's worth just reiterating that there is no tiny gremlin inside the PC watching a tiny square monitor. Aspect ratio really doesn't exist outside of the output.

It exists when the artist is drawing a sprite on their work machine which highly likely had an aspect ratio of 4:3.

I imagine it is very hard to try and draw things on such a setup so they would look most natural stretched on the horizontal direction.

Stuff like this always makes my head hurt thinking about it. I'd say old hardware had so many quirks but I believe even modern games are often rendered at resolutions that are different from what they ultimately get displayed at. Of course with 3D graphics it is much easier to keep things looking consistent. Not so with sprite graphics which by nature have a native resolution.

As for Japanese computers running in very high resolutions, this shouldn't be so surprising. Kanji is essentially unreadable in 15khz unless you use huge fonts. Therefore even the earliest PC-88s supported 24khz. It was the main reason NEC dominated the buisness world so early.

Ah, that little detail I already knew about. But one must remember that there is a world of difference between high res text modes and high res modes for running games. Getting a 512x512 sprite based games with a lot of things going on to run smoothly must be quite a challenge.

[kamiboy]

Because there is no artifacting; it's different sized pixels. Nothing gets "stretched", the pixels are just thinner. The reason they'd have resolutions that aren't 4:3 is to increase resolution. A 512x512 image on a 4:3 screen adds quite a bit of vertical resolution that you wouldn't get from a 4:3 source image.

I don't quite follow you here.

CRT displays have a native vertical resolution of ~480 scan lines. If you take something with a 512 vertical resolution and try to stretch in on 480 lines there should be consequences in the form of visual artefacts. The artefacts should be especially evident if the screen scrolls vertically.

Due to the nature of CRT displays they have no native horizontal resolution though, so you can stretch and shrink the picture horizontally to your heart is content.

Then again maybe I am looking at this from an overtly digital perspective. Maybe analoug technology does not exhibit visual artefacts from a fixed pixel picture being stretched over it.

So far as I understand older game consoles never had a 1:1 correspondence between their native vertical resolution and those 240 scan lines that got put on the TV at the end.

On an LCD anything but a 1:1 correspondence between the source pixel and its destination on the display device looks pretty bad for pixel based graphics such as sprites. It seems that is not the case for CRTs, even though they do technically have a native vertical resolution.

[Ed Oscuro]

It exists when the artist is drawing a sprite on their work machine which highly likely had an aspect ratio of 4:3.

How do you know this? Machines of the era had all kinds of different pixel aspect ratios, the square pixel wasn't standard. Whether or not they worked in the same pixel aspect ratio as the final product was really a choice. Without our finding some reference it's just a guess how it was envisioned at each step of the process (i.e. the artist, the game designer, etc.) and it could change from step to step (for example, plotting out a graphic on graph paper versus punching it in at a terminal.

CRT displays have a native vertical resolution of ~480 scan lines.

How do you know this? X68000 (and similar) machines had unique monitors; I don't know off the top of my head how many scan lines they were designed for, but apparently they could handle it fine without any detail blur. Also, "native resolution" is a term that really only makes sense when talking about fixed-pixel displays like LCD and plasma.

To clear up something that seems to be misunderstood - the number of scan lines on a CRT monitor just represents an upper limit on how much detail (lines) can be displayed, but lines can be drawn basically all over the screen at any point; there isn't really space "between" horizontal lines of resolution in the phosphor pattern. If you don't believe this, go to your cheapest 480i CRT and press your nose up against it; you'll see that there is some space between the vertical lines (but very little), yet the phosphors are applied essentially continuously up and down those lines. The real limits on resolution seem to be the number of vertical lines (which of course delineates a vertical resolution), and the ability of the scanning mechanism to sync with the input signal (and scan out each line of resolution quickly enough). High-resolution displays like the 800 or 900 line Sony BVMs seem to have two purposes - one to have thinner lines of phosphors so that the lines have sharper edges, and another to allow some playing around with overscan or higher resolution images.

[SuperDeadite]

I'm currently viewing this forum on a 21'' CRT monitor from 2001. It will display up to QXGA and looks beautiful, no artifacts. And I think QXGA has has a few more lines then 480.

[Ed Oscuro]

I'm currently viewing this forum on a 21'' CRT monitor from 2001. It will display up to QXGA and looks beautiful, no artifacts. And I think QXGA has has a few more lines then 480.

Well, it's not really clear to be (nor Kamiboy) how many rez lines your 'standard' X68000 monitor had, compared with a CRT display. But yeah there isn't a universal standard (in NTSC land) for how many lines of resolution a set could display - some are better than others in this regard.

Also, notice how it's "480 scan lines" in kamiboy's post - this is misleading or mistaken on two fronts - one, as I already mentioned, there are no hard-and-fast horizontal lines of resolution. Two, those are interlaced lines (in most applications - again, not sure about progressive display from a X68000 if that's what it uses), so the scanning is only 240 fields scanned out per frame.

But even some pro monitors, like this small 9 inch model, don't have 480 lines (this one advertises "only" 420 TV lines, or 420 vertical lines, yet it's a HR Trinitron branded unit). For a simple 480i (or 240p) standard-definition source this is actually still far more than enough.

Actually to be clear I should probably just quote this:

There can be some additional confusion in the horizontal-resolution specification. When refering to fixed-pixel displays [LCD - Ed] or the resolutions of digital-television formats [i.e. HD - Ed], the term "pixel" is correct. When refering to an analog television's horizontal resolution, the term "TV Line" (TVL) is more appropriate. This specification is the number of vertical lines the television can resolve per picture height. If the set is driven by a signal of closely spaced alternating black and white vertical lines, we count the number of visually resolvable lines to determine the set's horizontal resolution. As those vertical lines become more closesly spaced, they eventually turn into a gray blur. The threshold at which the individual vertical lines are still resolved is considered the set's horizontal resolution in TVL. Don't confuse these vertical lines with the horizontal scanning lines of a television. The number of horizontal lines is fixed at 480 in NTSC video, and can by 720 or 1080 in HDTV. (NTSC is actually 525 lines, of which 480 contain picture information. The other 45 lines are not displayed on the screen, and contain synchronization pulses, and information such as Closed Caption data.) The 480 lines you see are called active scan lines.

[Fudoh]

how many rez lines your 'standard' X68000 monitor had

pretty sure that it's a 35.5khz monitor, just like the old PC monitors, 13H at 70Hz, 480p at 60Hz, 600p at 56Hz, 768p at 86Hz (interlaced) max.

For a simple 480i (or 240p) standard-definition source this is actually still far more than enough.

not really, on a full D1 signal (720x480i) you get 720 columns of resolution. A monitor like this won't properly resolve that.

[Ed Oscuro]

13H at 70Hz, [...] 768p at 86Hz (interlaced) max.

Ehhh.

[kamiboy]

Ed, I imagine game pixel artists drew their assets on either PC's or Macintosh computers which were the standard for doing such work as they had industry standard drawing software available for them. These machines almost always came with monitors with a 4:3 aspect ratio.

The X68000 monitor in 15 and 31khz as well as CRT televisions have a "native" vertical resolution of ~480 raster lines which was something that stems from the rasterization process. So far as I understand for 31khz signals the electron beam makes ~480 progressive horizontal swipes across the screen to rasterize a full frame. For 15khz it makes ~240 swipes for 240p, and makes ~480 swipes for a full 480i frame but broken up into two passes, called fields, which each rasterize either the ~240 odd or even raster lines. That is why 480i looks so flickery.

Anyway, I do not care about what resolution the display can "resolve" as that is a subjective matter. what is for certain is how many horizontal swipes the electron gun makes during the raterization proces. That is what I call the native resolution.

Anywaste, my knowledge of how analoug display technology works is limited, so lets leave it at stretching pixel art looks pretty good on CRT's but pretty aweful on LCD's which is a thing to marvel at.

[Fudoh]

For a possible XRGB-4 Micomsoft should consider offering full input support for X68k units, 15/24/31khz all properly processed and output in the same resolution

[kamiboy]

that 24khz signal has always mystified me. What is its vertical resolution? In PC monitor terms what would be the equivalent resolution?

[Ed Oscuro]

Ed, I imagine game pixel artists drew their assets on either PC's or Macintosh computers which were the standard for doing such work as they had industry standard drawing software available for them.

In Japan, you mean? IBM machines didn't gain much share until DOS-V (about 1990). I don't see why you'd go through the trouble of using a Mac when you had this nice shiny X68000 to develop on with everything in Japanese and good graphical capabilities to boot - or NEC PC systems if you wanted. Plenty of graphical packages available for all those series. There is a difference between imagining and knowing - sorry

For a possible XRGB-4 Micomsoft should consider offering full input support for X68k units, 15/24/31khz all properly processed and output in the same resolution

Amen. I was upset to find out the X68000 plug on the XRGB series was just output (because hey, everybody can get one of them, right?)

For the moment VGA monitors negate this need, but eventually it will be necessary to replace most of them.

[SuperDeadite]

LOL why the hell would an X68k artist in 80's use a Mac? The X68k was the graphic editing king in it's day. In fact, from the little bit of real dev info I've found over the years. A lot of Japanese dev houses did most graphical work on X68ks and networked them to various Apple computers for coding. Typically using OS-9 on both machines.


For scaling X68000 signals, Micomsoft already offers the XPC-4 (15/24/31 input ok).

[Ed Oscuro]

For scaling X68000 signals, Micomsoft already offers the XPC-4 (15/24/31 input ok).

Thanks for mentioning this - it's obviously not well known here!

About the use of Macs for coding - was that done to make use of Motorola compilers and similar tools on the Mac? I definitely don't see the purpose of authoring graphics for the X68000 on another platform, although I thought they would code there as well.

I wonder if the XPC-4 is any good for medium resolution arcade games? Oh man...

[Fudoh]

24khz is classically 512x384p at 60Hz. Scud Race and Sega Rally run at this resolution.

Regarding the XPC-4: never had one, but always assumed it was a classic scan converter, very much like the Extron VSC units, meaning the incoming resolutions are output in 480i through composite, s-video and component. Does it upscale as well ?

[Fudoh]

wonder if the XPC-4 is any good for medium resolution arcade games? Oh man..

I've seen 24khz boards running scan converted to 15khz in 480i. Quite ok, but you're looking at upscaling instead, right ?

[SuperDeadite]

The info I have seen on dev work is tiny. But in general JP devs working with Motorola machines adored OS-9. An OS-9 network makes it super easy to link an Apple and an X68K. Graphics were done on the X68K. Mac was mostly just for the MPU code itself.

I have a friend with an XPC-4, he uses it for almost everything (X68000, PC-8801, 8801VA, 9801, FM Towns, and even Amiga). He then outputs it all to his 1080p LCD. He does nothing but rave about it.

[kamiboy]

My knowledge of the X68000 software scene is nonexistent, had no idea there were decent graphics software available for it.

In case that is where the graphics were made, and the graphics software being used ran in a square resolution then the problem of uneven sprite scaling would solve itself.

Of course if the graphics software ran in anything but a square resolution the problem would be no different than if they were using a PC or Mac.

In any case. When using a X68000 emulator I could toggle the screen between being stretched to 4:3 or left square, and square looked best because there was no uneven software scaling being performed. Another reason why playing on the actual hardware hooked to a CRT nets the best result.

About the use of Macs for coding - was that done to make use of Motorola compilers and similar tools on the Mac? I definitely don't see the purpose of authoring graphics for the X68000 on another platform, although I thought they would code there as well.

I wonder if the XPC-4 is any good for medium resolution arcade games? Oh man...

Macs have been the darling of artists and creative folk since the 80's. They were very popular for doing graphic work due to a suit of competent software available for them.

[SuperDeadite]

My friend's XPC-4 quick demos:

http://www.youtube.com/watch?v=qFzqdPdPg0Y
http://www.youtube.com/watch?v=hQV_J9lUcOk
http://www.youtube.com/watch?v=uas1ZNxanh8