Comparison of 6th generation game console hardware

[Black_Tiger]

High res Saturn games like VF2, Decathalete, DoA, etc look silky smooth and radically different from 30 fps games.

[kool kitty89]

High res Saturn games like VF2, Decathalete, DoA, etc look silky smooth and radically different from 30 fps games.

More silky smooth than the smoothest Dreamcast/PS2/GC/Xbox/Wii games running on an SDTV?

30 FPS, and I do mean a solid 30 FPS with no slowdown whatsoever, is quite smooth too, and interlacing can trick the human eye to make things seem smoother in some respects too. (just look at the CRT screen up close and the interlacing should be obvious)

Now, I do recall a few games actually using half high-res mode, ie normal verticla res, but double width (and 8-bit color depth), and those would indeed allow 60 (or 50) FPS.

Most of the "high res" Saturn games are also 60FPS. I think KoolKitty is hung up on the interlacing. SDTVs only actually update at 29.97 FPS, but a 60FPS input displays one frame on lines A and the other on lines B. I always thought 60FPS games looked better on SDTVs than they do on progressive scan TVs or VGA monitors.

Not sure about 60 FPS looking better on an SDTV . . . blurrier, yes, and I don't just mean the lower precision beam or dot pitch of the shadow mask, but the very high persistence phosphor SDTVs used to mask flicker. (hence why 60 Hz on a good low-persistence VGA monitor is much more flickery than 240p on a TV . . . and why interlaced resolutons look super flickery and nasty on VGA monitors)
OTOH, most CRT HDTVs have similar high persistence phosphor too (needed for 1080i), so that should be similar in any case.

240p is still interesting for the 15.7 kHz native CRT situation though, especially if you're talking about any sort of display that's using digital scaling/filtering rather than simple line doubling for 240p to 480p conversion. (and given that the actual cases for really nice, direct line doubling like that is uncommon -particularly if the console/computer doesn't support doing that internally)
Actually, I wonder how Wii games running in 240p (in SD mode) would compare in 480p mode, I was going to look into that at one point, but never got around to it. (granted, it's still going to be somewhat apples to oranges since you probably won't have a multi-sync monitor that goes down to 15 kHz -that would be REALLY nice though )

Isn't it a requirement to use high-res mode that your game render at 60fps otherwise you'll get really screwed up graphical errors?

Also I'm pretty certain that Sega themselves said that games like VF2, Decathalete, etc. ran in high res mode at 60fps. Yeah, the Screen output will be an interlaced image, but it can still render internally at 60fps. After all plenty of Dreamcast games do this if you are using composite, it's not like the frame rate for Sonic Adventure 2 drops to 30fps if you're not using VGA.

Unlike DC, the Saturn renders interlaced internally since the framebuffers aren't large enough (and use fixed 1024x256 byte organization) to allow a full frame, so each framebuffer uses a single field, thus requiring 60 field per second rendering to avoid tearing artifacts. (PS2 and GC can render natively interlaced too iirc, which is why not all games have 480p support)

As for cases of internally rendering at 480p/60, you still lose half the temporal resolution when going to interlaced . . . but instead of losing every other frame, you loose every other 1/2 a frame and also get some combing artifacts due to mismatched fields/frames being interlaced. (more noticeable during fast action)
Actually frame limiting a 480p game to 30 FPS will reduce/elliminate those combing artifacts though, similar to natively rendering at 480i. (you still get SOME combing since you still update 1/2 a frame -1 field- at a time, but you only get those errors for every other field rather than every single field as with direct conversion from 480p/60 -where every single field is from a different frame)

Granted, actual interlaced video RECORDINGs have the combing problem too, since they're recorded one field after the other rather than in pairs. (both a problem with analog and digital video, but more noticeable in digital due to the higher effective horizontal resolution and lack of blurring -albeit it's also more noticeable on high quality/professional video tape for black and white video, and you do indeed see more visible combing artifacts in some old TV shows shot with tape -namely a few shows in the 1960s that were around when VTR based production was common, but the transition to color wasn't yet universal -especially for early seasons of lower budget shows)

[sheath]

More silky smooth than the smoothest Dreamcast/PS2/GC/Xbox/Wii games running on an SDTV?

30 FPS, and I do mean a solid 30 FPS with no slowdown whatsoever, is quite smooth too, and interlacing can trick the human eye to make things seem smoother in some respects too. (just look at the CRT screen up close and the interlacing should be obvious)

The difference between 60FPS games and 30FPS games, even back in the 8-bit and 16-bit days, is quite obvious on SDTVs. More below.

Now, I do recall a few games actually using half high-res mode, ie normal verticla res, but double width (and 8-bit color depth), and those would indeed allow 60 (or 50) FPS.

Dead or Alive, Virtua Fighter 2, and Decathelete are reputedly 60FPS. I am pretty sure this is only in the floors/backgrounds, as the characters defeinitely do not have 60 frames of animation.

Not sure about 60 FPS looking better on an SDTV . . . blurrier, yes, and I don't just mean the lower precision beam or dot pitch of the shadow mask, but the very high persistence phosphor SDTVs used to mask flicker. (hence why 60 Hz on a good low-persistence VGA monitor is much more flickery than 240p on a TV . . . and why interlaced resolutons look super flickery and nasty on VGA monitors)
OTOH, most CRT HDTVs have similar high persistence phosphor too (needed for 1080i), so that should be similar in any case.

Okay, here is how I have always tested for 30FPS versus 60FPS. So far it has held up for me. Find a scene where the whole background moves or where you know an object is going to move across the screen. Fix your eyes on a detail on that object or background. If the detail blurs as it pans it is 30FPS, if it stays solid/clear it is 60FPS. If it jumps across the screen it is sup 30FPS. 60FPS makes everything sharper especially in motion on SDTVs.

[TrekkiesUnite118]

To my knowledge sheath 3D games don't really have "frames of animation" in the literal sense. They instead perform animation by changing the coordinates of the 3D model to simulate movement. So no, you wont have 60 individual frames of animation, but that doesn't mean you can't be running at 60fps or rendering at 60fps. So if I'm rendering at 60fps (as in Virtua Fighter 2) every 1/60 of a second I'm drawing a new frame to draw on the screen, and in drawing that new frame I'm seeing if there's any updates to the positions of the 3D Models I need to make and making them. So the difference between say 30fps and 60fps in a 3D game is that you are doing those position updates every 1/30 of a second instead of 1/60 of a second. So you either move twice the distance you would have at 60fps which results in the movement being choppier, or you move the same distance resulting in the movement looking slower.

So yeah, the characters in VF2 do not have 60 individual frames of animation, but they do render and move at 60fps.

And a better test for 60fps would be to use some emulator that can properly count the frames.

[Team Andromeda]

What do you mean by block joints?

VF 2 in the Arcade like Tekken 1 and 2 and Last Brox have bloocky joints (no much smoothness on the arms ect )



While on the Saturn VF 2 the shading on the characters arms and joints is far smoother and far more human like

Saturn high res games, or specifically the 640/704x448/480/512 stuff, by definition can't be 60 FPS

Decathlete runs in the Saturn High Res mode 704x480 for every part of the game and also at a rock solid 60 fps

Also I'm pretty certain that Sega themselves said that games like VF2, Decathalete, etc. ran in high res mode at 60fps. Yeah, the Screen output will be an interlaced image, but it can still render internally at 60fps

Yes they do and all Saturn High Res games are interlaced

More silky smooth than the smoothest Dreamcast/PS2/GC/Xbox/Wii games running on an SDTV

Depends on if any of those games are running at 30 fps . I mean it's very easy to tell that JSRF on the XBox is running at 60 fps compared to JSR 30 fps on the DC.

[

[TrekkiesUnite118]

Oh you mean the visible polygons? That's actually because the Saturn ports lack any and all lighting the arcades had. So the lighting effect on each individual polygon makes the individual polygons of each model more visible.

So if the Saturn port had proper lighting you'd see those individual lines too. You can see this in the PC port if you turn lighting on and off. You see similar issues in Fighting Vipers and Fighter's Megamix on any character that uses more texture mapping than gouraud shading.

[Team Andromeda]

Oh you mean the visible polygons? That's actually because the Saturn ports lack any and all lighting the arcades had. So the lighting effect on each individual polygon makes the individual polygons of each model more visible.

.

It's nothing to do with the Lighting as Last Brox on the Saturn suffered from the blocky joints just like VF 2 and Last Bronx in the Arcades . Its just seems that AM#2 for Saturn VF 2 and indeed Namco with Tekken 3 on the PS went for far smoother looking arms and joints

[sheath]

http://www.gamepilgrimage.com/VF2PClight.htm

[stu]

Looks better than Model 2 quality to me.

The points I was trying to get across was:

1. That comparing a 300 dollar home system to a arcade system costing 10,000 dollars plus or a $2000.00 PC are both a pointless and irrelevent comparisons.

AND

2. That saying that the Saturn was better at Model 2 conversions than the PS1 cannot be proven considering that the PS1 only received 1 Model 2 conversion.

I personally think that the conversions that both systems got were pretty darn good for the cost of the systems.
As for VC2, it certainly looks a good conversion although I have to admit I've never played either the Saturn or Arcade version so I'm really not going to pass judgement on whether its "better than Model 2 quality" or not. Youtube videos aren't really representative considering how low grade the video quality is. However I can see some differences in the graphics (the textures are better on the arcade version).

Looks like I'll have to track down a copy of the Saturn version and/or find an arcade with VC2 so I can check it out.

[sheath]

Okay, really, I did not say that the Saturn was better at Model 2 conversions by default. It is evident in the software that Model 2 conversions were a greater priority in Saturn software. How is this so confusing?

[kool kitty89]

The difference between 60FPS games and 30FPS games, even back in the 8-bit and 16-bit days, is quite obvious on SDTVs. More below.

For 240p yeah, it is noticeable, especially for a fast paced game (slow scrolling will be less noticeable, particularly since you might not even scroll every frame due to resolution limitations)

Dead or Alive, Virtua Fighter 2, and Decathelete are reputedly 60FPS. I am pretty sure this is only in the floors/backgrounds, as the characters defeinitely do not have 60 frames of animation.

Those are all interlaced games though, I'm talking about ones that are 240p but still high res horizontally. (I could have sworn there was a winter sports themed game that did this)

Interlaced rendering is still 60 FIELDS per second, and rendering just as much as 60 frames per second at 240p, but it's still technically 30 frames per second max as far as 704x480 goes.
Same goes for the VDP2 BGs, since they're going to be interlaced too . . . though in those cases, it could actually be 60 FPS 480p screens being interlaced/"combed" down, so more like 480p native games on later systems. (you could also potentially use high res BGs and lower res VDP1 layer scaled up, which might even include the 31 kHz progressive scan support for VGA/hivision -not sure if any games actually supported that though)

Okay, here is how I have always tested for 30FPS versus 60FPS. So far it has held up for me. Find a scene where the whole background moves or where you know an object is going to move across the screen. Fix your eyes on a detail on that object or background. If the detail blurs as it pans it is 30FPS, if it stays solid/clear it is 60FPS. If it jumps across the screen it is sup 30FPS. 60FPS makes everything sharper especially in motion on SDTVs.

In my experience, it's the opposite. 30 FPS (or lower) makes motion sharper and more distinct, while 60 FPS stuff will be somewhat blurred/blended and smoother due to the image persistence on the CRT screen. (talking CRT SDTVs here)

Games with 60 Hz scrolling are indeed noticeably different though . . . 60 Hz animation is even more dramatic in games of that era (80s/early 90s) given how much slower/choppier typical animation was. (Crusader of Centy almost feels weird because of how smooth some of the animation is, and that 360 degree turning animation was especially noticeable -particularly as the character turns around rather than clipping through himself and instantly facing the opposite direction)
The difference is even more obvious in an emulator and a good LCD screen. (same thing for 60 Hz flicker effects -they look almost perfectly translucent on an LCD in an emulator)



Though, that said, you could still have 2 different ways of rendering to a native interlaced display:
1. would be rendering complete interlaced frames (using the same set of vertices to render both fields of the interlaced frame), or
2. render each field as an independent frame, effectively simulating what you'd get from a 480 native 60 FPS renderer that was down converted to 480i (like 60 FPS dreamcast/xbox games).

And those would apply to any interlaced native framebuffer renderers, be it Saturn, PS2, GC, etc.

The trade-offs for case 2 would be needing double the vertex performance (same as a solid 60 FPS 240p game) and having combing/blur artifacts since the fields won't be paired/matched, and instead having each independent. So, you effectively have the same physical AND temporal resolution as 240p, just with overlapping fields rather than distinct frames.

It's sort of the same argument as running a game with vsync off for the case of a game that runs at a solid 120 FPS on a 60 Hz native display, except the screen is nearly divided every other scanline rather than split mid-screen into solid bands/chunks.

So, technically, yes, in some respects it could still be 60 FPS rendering (more definitively if you're loose with what that "F" in FPS actually denotes), but at the same time, you're still never going to get 60 Hz rendering speed AND 704x480 FRAMES. (ie you could have 30 Hz rendering and 704x480i frames OR 60 Hz rendering of non-matched interlaced fields)


Personally, I'd argue doing 704x240p/60 would actually look nicer in some respects, but then again, you can make the very same argument for actual interlaced TV/video content. (which, in some cases is actually true as well, especially very rapid action stuff . . . 240p/60 will indeed look better than 480i, or single field rendering of interlaced digital video too, same goes for 1080i as 480i too -something I realized when taking a digital video editing class a few years back, and also realized my frustration when I could use QuickTIme Pro to view interlaved video as single-fields, but couldn't actually render the video as 540p instead of 1080i -which looked a hell of a lot better than blurry/deinterlaced 1080p or native 1080i . . . I was so flustered by that, that I failed to realize the potential merits of actually rendering the project to 1080i native and having a definitive copy of it I could later play around with . . . )

To my knowledge sheath 3D games don't really have "frames of animation" in the literal sense. They instead perform animation by changing the coordinates of the 3D model to simulate movement. So no, you wont have 60 individual frames of animation, but that doesn't mean you can't be running at 60fps or rendering at 60fps. So if I'm rendering at 60fps (as in Virtua Fighter 2) every 1/60 of a second I'm drawing a new frame to draw on the screen, and in drawing that new frame I'm seeing if there's any updates to the positions of the 3D Models I need to make and making them. So the difference between say 30fps and 60fps in a 3D game is that you are doing those position updates every 1/30 of a second instead of 1/60 of a second. So you either move twice the distance you would have at 60fps which results in the movement being choppier, or you move the same distance resulting in the movement looking slower.

So yeah, the characters in VF2 do not have 60 individual frames of animation, but they do render and move at 60fps.

It depends. Some 3D games use jointed 3D models that can be manipulated (articulated/posed), while others use animated 3D models with each model having multiple 3D frames of animation. (basically a whole new "model" for each frame)
That's the way Quake animates its enemies and objects (hence why they're choppy regardless of rendering speed -15 FPS iirc), Quake 2 as well (but at a higher framerate), and quite a few other games too. (off the top of my head, Elder Scrolls Adventures: Redguard animates that way too)

Animating that way has the advantages of avoiding odd texture distortion and model clipping or holes, but with the major trade-off of requiring more memory to store all those "frames" of animation, and thus also a lower framerate if RAM capacity is too limited.

And a better test for 60fps would be to use some emulator that can properly count the frames.

It's really easy, just halt the emulator and use frame advance . . . at least that's how I do it in fusion. (it's the pause and insert keys iirc)

[zyrobs]

If you compare VF2 arcade to VF2 Saturn, you can see that the Saturn version is extremely poor - it lacked a lot of animations (wind blowing Jackys hair), had zero lightning, the few fluff animations had almost no collision detections, the character models are way less detailed, the textures are practically trash, and the backgrounds are static images. Yeah, it runs in higher resolution, but the textures are so awful that it doesn't really matter.

It's nothing to do with the Lighting as Last Brox on the Saturn suffered from the blocky joints just like VF 2 and Last Bronx in the Arcades . Its just seems that AM#2 for Saturn VF 2 and indeed Namco with Tekken 3 on the PS went for far smoother looking arms and joints

No, it is the lightning. The Saturn version doesn't have any, the arcade has flat shading.

Tekken 3 looks pretty blocky to me.

Dead or Alive, Virtua Fighter 2, and Decathelete are reputedly 60FPS. I am pretty sure this is only in the floors/backgrounds, as the characters defeinitely do not have 60 frames of animation.

They run in 60fps, but interlaced. Both the characters and the backgrounds. Except for the rotating playfield, which is always standard res.

Now, I do recall a few games actually using half high-res mode, ie normal verticla res, but double width (and 8-bit color depth), and those would indeed allow 60 (or 50) FPS.

On the Saturn it's called Double Density mode, it is double the horizontal resolution. You can make it use double density, interlaced, both, or neither. VF1/Remix, and Winter Heat uses double density but not interlaced mode. Winter Heat is the sequel to Decathlete, and it needed extra polygon power to draw more detailed polygonal backgrounds due to the winter setting (Decathlete just had a flat field and little else). It looks really good. I think the title screen and credit roll of Waku Puyo Dungeon used interlacing but no double density, oddly enough.
Toshinden URA uses both double density and interlacing, but only the background bitmaps are hi-res, the rotating playfield is standard res, and the polygons are... double density (technically not the correct term since interlacing and dd modes only apply to the VDP2; here the VDP1 is likely using a lower framebuffer size).

[Barone]

For 240p yeah, it is noticeable, especially for a fast paced game (slow scrolling will be less noticeable, particularly since you might not even scroll every frame due to resolution limitations)

For 480i it's as noticeable as for 240p for all games that I've played so far.

In my experience, it's the opposite. 30 FPS (or lower) makes motion sharper and more distinct, while 60 FPS stuff will be somewhat blurred/blended and smoother due to the image persistence on the CRT screen. (talking CRT SDTVs here)

I have to agree with sheath here. It's actually exactly like he describes and the opposite to what you say.
More like:

Personally, I'd argue doing 704x240p/60 would actually look nicer in some respects

Surely not in terms of jaggy/pixelation, since x480 games look noticeably better than x240 for both Saturn and PS1.

Quake 2 as well (but at a higher framerate)

Quake II on the N64 clearly shows less frames for enemies when compared to the PC/PS1 versions.

If you compare VF2 arcade to VF2 Saturn, you can see that the Saturn version is extremely poor - it lacked a lot of animations (wind blowing Jackys hair), had zero lightning, the few fluff animations had almost no collision detections, the character models are way less detailed, the textures are practically trash, and the backgrounds are static images.

I agree only with the parts in bold. Cool that you have brought them up though, since TA just "forgot" to mention stuff like the use of static images when overhyping about it.

No, it is the lightning. The Saturn version doesn't have any, the arcade has flat shading.

Yep.

Tekken 3 looks pretty blocky to me.

Same here. DOA looks much better IMO. IDK why so many people say good things about Tekken 3 graphics while it's nothing really special for the time IMO.

On the Saturn it's called Double Density mode, it is double the horizontal resolution. You can make it use double density, interlaced, both, or neither. VF1/Remix, and Winter Heat uses double density but not interlaced mode. Winter Heat is the sequel to Decathlete, and it needed extra polygon power to draw more detailed polygonal backgrounds due to the winter setting (Decathlete just had a flat field and little else). It looks really good. I think the title screen and credit roll of Waku Puyo Dungeon used interlacing but no double density, oddly enough.
Toshinden URA uses both double density and interlacing, but only the background bitmaps are hi-res, the rotating playfield is standard res, and the polygons are... double density (technically not the correct term since interlacing and dd modes only apply to the VDP2; here the VDP1 is likely using a lower framebuffer size).

I'd like to know what can be said about the PS1 side of things.

[zyrobs]

Same here. DOA looks much better IMO. IDK why so many people say good things about Tekken 3 graphics while it's nothing really special for the time IMO.

Tekken 3 looks extremely detailed to me, better than anything I've seen on the Saturn for sure. Higher detail models, better textures, and lightning effects too. Plus, Namco knows how to actually animate characters.

DOA definitely is the best looking 3d fighter on the Saturn, though. The Playstation port had the characters turned into lifeless dolls.

[spiffyone]

Spot on. I've never seen in the point and why people feel the need to compare consoles to PC's

Well, they really shouldn't, considering video game home consoles and computers/PCs exist in different product markets. Even if focusing on the game side of things for computer/PC, it and home consoles are still in different markets. User base is somewhat different (PC skews older on average), typical usage scenarios quite different (desk vs. couch, inches away from smaller but higher res monitor vs. feet away from larger but lower resolution TV, and even average play times are different), and basic stock user interface is very different (kb/m vs. gamepad; yes, gamepads can be used on PCs, but the games themselves are designed primarily around the interface devs are sure everyone has in that market: kb/m).

On that same train of thought, it's also odd and silly to directly compare portable game systems with home consoles, and either one with mobile platform devices. So-called "gamers" do it, of course, but more confounding is that analysts and folks within the industry do it as well, and they should really know better. The idea of a "merger" of product markets has never taken hold in the industry, nor has the idea of "one market killing the other" (even the downturn of the arcade market had very little to do with the idea of home consoles becoming powerful enough to make arcades obsolete).

I meant it took until the likes of Screamer and 3DFX cards for the PC to really catch up with the consoles.

Wasn't there an NV1 based card out in '95? If so, right there we had a card that, combined with a good PC CPU and enough RAM, caught up/surpassed what the consoles could do, at least on paper.