Support NeoGAF

[polybius80]

Like if the PS2 or game cube did use per pixel lightning ever....

yes they did, I even linked a paper about how to achieve it in PS2 and it was used in hitman blood money for example

silent hill games in that era were amazing they applied all sort of effects and improved each new one
https://abload.de/img/fakebumpbzf4y.gif


konami was great and cared about their games back then

 

[nkarafo]

Like if the PS2 or game cube did use per pixel lightning ever.... Silent Hill 2 did it on Xbox and it looks noticeable better specially with the flash light.

The PS2 original still has those great looking real time soft shadows though. And the XBOX has a less complex looking fog effect.

 

[Fafalada]

PS2/RRV: Ok, cool.
PSP: WOW

I've seen PS2 and PSP at their respective TGS shows just before hw-launch - and this was the two times I recall significant changes to launch products...
PS2 R5 was the most impressive thing shown on the floor by a sizeable margin. The game we actually got 6 months later was - not so much.
PSP screen was also one of the most WOW moments (especially playing things like RR). Retail units 3 months later were again - much less so.

The screen downgrade got some - minimal - coverage, although I am pretty sure Sony never acknowledged it (it got talked about in dev community a bit more though - early devkits were also different). R5 downgrade was even more under-the radar, and it wasn't just IQ. It always looked like the ridiculously short dev-cycle really negatively impacted the end-product, but unlike Tekken Namco never bothered touching it again for some reason.

If all it is, is upscaling the image

It's not - it's setting output to progressive - just like DC does, but it has to be toggled in software.
There are 'some' games that can only be displayed interlaced (or if you force them into progressive you'll get 240P upscaled), but that applies only to 60fps games, and only subset of such games that forced interlaced output buffer.
Out of those - a further smaller subset only rendered half/240 lines (majority were still rendered full-height - 480, but downscaled to 240 on scanout). Ie. the total number of games on PS2 that only rendered half-frame pixels is quite small.

In any case both are beaten by the dreamcast and xbox.

I don't think anyone argued about that - the topic was specifically about the fact both eDram machines made similar compromises on render-resolution, and it was clearly not for performance reasons on either of them.
There was a good reason why PS2 was originally specced to have 8MB eDram - the designers knew full well that to target all supported resolutions (especially HDTV outputs - which graphic-chip was intended to do) they needed extra memory, but alas - cost cutting. But hey - we eventually got (true)1080i support on some video content/DVDs...
And GC - well the documentation literally recommended to render 2-passes when using AA if you wanted full-frame - the Frame-buffer was quite small there too.

What would it take on future modding of existing games or newer indie developments to put DC lighting at PS2 or at least PSP level?

'Lighting' means using more compute - in that era that meant operations per pixel (the fillrate), and FP math (operations per vertex).
PS2 for all its faults kept up decently on the amount of math it could perform 'per-pixel' (peak games hit as high as 40 ops/pixel, and even at launch it was easily above the average games at the time that did 2-3), and while it was a lot more cumbersome to meaningfully use that compared to GC/XBox - even XBox had to work a bit to do that much work per-pixel.
As for FP/vertex math, the field was clearly stacked XBox > PS2 > GC > DC.

 

[Geometric-Crusher]

Like if the PS2 or game cube did use per pixel lightning ever.... Silent Hill 2 did it on Xbox and it looks noticeable better specially with the flash light.

splinter cell Pandora Tomorrow Multiplayer ?

 

[Geometric-Crusher]

You want an example of a 32 bit game ported to DC and actually rebuilt around DC capabilities: SOUL CALIBUR. And by the way, Soul Calibur was built originally on System 12, a PS1 Pro, and it´s probably the game which takes that system board beyonds it´s limits, even over TTT (there is a reason SC never reached PS1). This is a game which in its base is more advanced than any PS1 elite game, and it actually turned into a real full DC game and at 60 fps.

It's the same principle, I just avoided mentioning fighting, racing and sports games cause are easier to reach 60fps
let's take Soul Reaver, despite being a early game is very competent, Illbleed from 2001 has an advantage in the polygon budget the purple girl has 400 more polygons than Raziel and the game displays more polygons per frame, however it is a walking simulator, the stages are large but use fog, all characters have shadows (it's rare to see more than two in a scene) and the frames continue to oscillate like in Soul Reaver maybe more.
Therefore, there was an improvement in the 2001 game compared to the 1999 game, when we talk about games like Dead or Alive 2, Sonic Adventure 2 we are talking about the crème de la crème of Dreamcast games and not the average.

 

[PaintTinJr]

the better path is to employ more aggressive LOD systems and apply some effects to the very minimum and in a controlled way, some genres are more permissive than others too, not having shaders is not necessarily an impediment for complex effects more passes can be used among other things for effects............. yes I know DC is particularly limited in this too, but being creative can do wonders after all when it comes to real time graphics in games all effects are basically smoke and mirrors and there are effects in DC games that were almost unused like the bump mapps effecs put them here and there and you can get an improvement

imagine soul calibur with an effect like this in the floor and some walls

For me, ultimately the DC's framebuffer being RGB_565, limiting it 65K colours(65,536) or lighting shades, kills a lot of the reasons to use bump mapping with the disproportionate fill rate cost. But I agree about looking at other paths, and as backwards looking as it might sound, I suspect being a console the CPU probably had better access to framebuffers without stalling framerate - that my experience of 20years ago with a ThinkPad R40's ATI Radeon 7500 mobile GPU was - and I would think there might be a path to better than per vertex lighting, by doing a low resolution render buffer on the GPU, say 160x120 and then modifications interlaced on the CPU (so 80x60 per field) and then use it as a texture to shade/modify(blend with) a flat shaded interlaced game render on the GPU.

Or alternatively do render targets on the GPU that render with index values at each pixel, like a deferred renderer, but then use the CPU or 2D GPU rasterization features where possible to do the fragment part of colouring in the picture with the right shading. As a full alternative Outcast from 1999 ran fine on my Pentium2 350mhz at 512x384@30 IIRC, and was pretty much just on CPU rendering, so a 200Mhz Dreamcast CPU doing something similar but adopting a clipmapping system for increasing voxel sizes at distance, and rendering those large distance voxels as polygons might work too.

 

[polybius80]

For me, ultimately the DC's framebuffer being RGB_565, limiting it 65K colours(65,536) or lighting shades, kills a lot of the reasons to use bump mapping with the disproportionate fill rate cost. But I agree about looking at other paths, and as backwards looking as it might sound, I suspect being a console the CPU probably had better access to framebuffers without stalling framerate - that my experience of 20years ago with a ThinkPad R40's ATI Radeon 7500 mobile GPU was - and I would think there might be a path to better than per vertex lighting, by doing a low resolution render buffer on the GPU, say 160x120 and then modifications interlaced on the CPU (so 80x60 per field) and then use it as a texture to shade/modify(blend with) a flat shaded interlaced game render.

Or alternatively do render targets on the GPU that render with index values at each pixel, like a deferred renderer, but then use the CPU or 2D GPU rasterization features where possible to do the fragment part of colouring in the picture with the right shading. As a full alternative Outcast from 1999 ran fine on my Pentium2 350mhz at 512x384@30 IIRC, and was pretty much just on CPU rendering, so a 200Mhz Dreamcast CPU doing something similar but adopting a clipmapping system for increasing voxel sizes at distance, and rendering those large distance voxels as polygons might work too.

yes, the machine has its problems but also the fundamental problem is "what you want to run on it?" then there are problems and some stuff that can be done that go in the strengths of the machine so it is not a good approach to "can it run bump mapping" or can it do "phong shading" before thinking what game you want to do because maybe you can make a competent game with what you have, yes DC is very limited compared to subsequent consoles but also has some nice things I wont mind having inferior versions of games at least its better than "no version of the game", so even if some bump maps are added no matter if they doesnt look good is a good adition

vertex light is not bad as long as has a good amount and distribution of triangles, the DC has problem with light but as said the effects are basically smoke and mirrors so if you can make something convincing then its fine, other things that can be done in DC is to minimize vertex transforms, lot of DC games use discreet shapes to form character(you know like mario64 or FF7 where character are made from parts) it is better for performance to have big chunks of triangles that doesnt change other than moving, scaling and rotating the whole group so you can have complex space ships or cars and scenery with a good or even huge amount of triangles without that much problem, and what you save from CPU in not doing vertex transforms you gain for other stuff and if we make the game in cell shading then we can save in color needs

 

[PaintTinJr]

yes, the machine has its problems but also the fundamental problem is "what you want to run on it?" then there are problems and some stuff that can be done that go in the strengths of the machine so it is not a good approach to "can it run bump mapping" or can it do "phong shading" before thinking what game you want to do because maybe you can make a competent game with what you have, yes DC is very limited compared to subsequent consoles but also has some nice things I wont mind having inferior versions of games at least its better than "no version of the game", so even if some bump maps are added no matter if they doesnt look good is a good adition

vertex light is not bad as long as has a good amount and distribution of triangles, the DC has problem with light but as said the effects are basically smoke and mirrors so if you can make something convincing then its fine, other things that can be done in DC is to minimize vertex transforms, lot of DC games use discreet shapes to form character(you know like mario64 or FF7 where character are made from parts) it is better for performance to have big chunks of triangles that doesnt change other than moving, scaling and rotating the whole group so you can have complex space ships or cars and scenery with a good or even huge amount of triangles without that much problem, and what you save from CPU in not doing vertex transforms you gain for other stuff and if we make the game in cell shading then we can save in color needs

I agree, but Cell shading is effectively per pixel lighting, because the banding/discontinuity fx is done in the fragment shader typically. Or is there a vertex version I'm unaware of, or did I misunderstand and you were meaning it would be done on the CPU, given its low frequency colour variation would be less process intensive?

 

[SomeGit]

For cell shading on the Dreamcast you do it in 2 passes, you render the game as normal in the first pass then apply another texture using a very low depth for the second pass. For the outline use the inverted hull technique.

 

[polybius80]

I agree, but Cell shading is effectively per pixel lighting, because the banding/discontinuity fx is done in the fragment shader typically. Or is there a vertex version I'm unaware of, or did I misunderstand and you were meaning it would be done on the CPU, given its low frequency colour variation would be less process intensive?

well cell shading can be ambiguous with its meaning it references the black lines of cartoon characters but people use that term to whatever games that look like a cartoon or comic even if they dont use the lines so you can have very simplistic approach or a very complex one but people will still relate the concept to your game you dont require pixel light or anything special in fact you can almost use no textures in them there is also some toon shading with light elements to look as other type of cartoon styles but even in PS1 we had cell shading games and probably even if you add vertex light it will still look cool



lot of people say this is cell shading for example


it is not but still it looks great, basically for the general public cell shading = "looks like a cartoon or comic" its interesting how you dont really require a specific technicality just an art direction and some tech to make it possible to get a good product

 

[PaintTinJr]

For cell shading on the Dreamcast you do it in 2 passes, you render the game as normal in the first pass then apply another texture using a very low depth for the second pass. For the outline use the inverted hull technique.

I agree that at the time that would pass as cartoon effect, but cell/toon shading in today's parlance does requires banding the interpolation of the normal vector at each fragment.

I know I'm being a bit mean on that front, but anything less than windwaker's effort is misleading imo. In fact, I never played it, but I'm not sure if PS2 Herdy Gerdy actually qualifies at being toon "shaded" either.

 

[SomeGit]

IIRC wind waker also did it in a similar way for the models, not 100% sure on the backgrounds. XIII also is another example of it.

Yes it’s crude by todays standards, but it looked very unique at the time.

 

[PaintTinJr]

IIRC wind waker also did it in a similar way for the models, not 100% sure on the backgrounds. XIII also is another example of it.

Yes it’s crude by todays standards, but it looked very unique at the time.

No, Windwaker definitely does the banding in a fragment shader, and even uses an additional system, probably a simplified particle system, in the same algorithm idea of Valkyria Chronicles painterly fx.

AFAIK Windwaker had a working version on N64 and it was the lack of quality in the Cell Shading that was why it was held back for the Gamecube's more powerful and more advanced texture hardware.

 

[SomeGit]

Fragment shader? Sure man what ever you say so

 

[PaintTinJr]

Fragment shader? Sure man what ever you say so

Are you joking or just not familiar with cube games and cube hardware?

We discussed the water fx in Super Monkey Target 1 & 2 at length and that can't be done without shaders. Nor can Waverace's water, or metroids visor and ball fx, so much so that it led to Ubershaders on Dolphin.

Ubershaders: A Ridiculous Solution to an Impossible Problem

When you're playing your favorite game on Dolphin with a powerful computer, things should run fairly well. The game is running full speed, there are no graphical glitches, and you can use your favorite controller if you want. Yet, every time you go to a new area, or load a new effect, there's...

dolphin-emu.org

Pretty much every visually impressive fx on a cube game is done with vertex and fragment shaders together. This was the gulf between the PS2/Cube/Xbox to the Dreamcast.

 

[SomeGit]

It seems like you are not familiar with the GC GPU and it’s API if you think it’s a fragment shader, hell you are saying vertex shaders when the GC has a fixed function vertex pipeline.

But this is the man that wants to run DLSS on PS2, I shouldn’t be surprised.

 

[polybius80]

Pretty much every visually impressive fx on a cube game is done with vertex and fragment shaders together. This was the gulf between the PS2/Cube/Xbox to the Dreamcast.

no, only Xbox had shaders, PS2 and GC had ways to manipulate data using some very powerful hardware, GC had the TEV to mix textures add some multipass and you can get some very interesting effects, PS2 is just a multipass monster and has the VU's which have lot of math power and are programmable, both consoles have hardware that can can do what shaders do or some of the things a shader do but they are not shaders in the sense of writing your shader in something like GLSL compiling and executing so it wasnt simple to use, when they were being developed tech was advancing very fast but wasnt well stablished like today they identified graphical needs and tried to prepare for the future, MS just used their money to build a custom PC with what nvidia had at the time it was very powerfull but very constly I remember some articles at the time that they were losing lot of money for each console, they literally bought their place in the industry and wasnt cheap

 

[PaintTinJr]

It seems like you are not familiar with the GC GPU and it’s API if you think it’s a fragment shader, hell you are saying vertex shaders when the GC has a fixed function vertex pipeline.

But this is the man that wants to run DLSS on PS2, I shouldn’t be surprised.

You clearly didn't read about the Uber"shaders", of if you did you are just arguing in bad faith when they are semantically the same.

Consoles are very different. When you know the precise hardware you are going to run the game on, and you know that the hardware will never change, you can pre-compile GPU programs and just include them on the disc, giving your game faster load times and more consistent performance. This is especially important on older consoles, which may not have enough memory for or possibly even the capability to store shaders in memory. Flipper, the GameCube GPU, is the latter.

While it has some fixed-function parts, Flipper features a programmable TEV (Texture EnVironment) unit that can be configured to perform a huge variety of effects and rendering techniques - much the same way that pixel shaders do. In fact, the TEV unit has very similar capabilities to the DirectX 8 pixel shaders of the Xbox! It was so flexible and powerful that Flipper was reused as the Wii GPU (redubbed Hollywood) with few modifications. Unfortunately for us though, the TEV unit is designed for the game to configure and run TEV configurations immediately when an effect is needed. There is no preloading of the TEV configurations whatsoever, since the TEV unit doesn't have the memory for that.

That instantaneous loading is the source of all our problems. Dolphin has to translate each Flipper/Hollywood configuration that a game uses into a specialized shader that current GPUs can run, and shaders have to be compiled, which takes time. But the TEV unit doesn't have the ability to store configurations, so GC/Wii games must configure it to render an effect the instant it is needed, without any delay or notice. To deal with this disparity, Dolphin's only option is to delay the CPU thread while the GPU thread and the video driver perform the compilation - essentially pausing the emulated GC/Wii. Usually the compilation will take place in under a frame and users will be none the wiser, but when it takes longer than a frame, the game will visibly stop until the compilation is complete. This is shader compilation stuttering. Typically a stutter only lasts a couple of frames, but on really demanding scenes with multiple compiling shaders, stutters of over a second are possible.

Until a shader cache has built up, Metroid Prime 3 is quite painful.
As the first emulator to emulate a system with a highly programmable GPU at full speed, Dolphin has had to go it alone at tackling this problem. We implemented shader caching so if any configuration occurred a second time it would not stutter, but it would take hours of playing a game to build a reliable cache for it, and a GPU change, GPU driver update, or even going to a new Dolphin version would invalidate the cache and start the stuttering all over again. For years, it seemed like there was nothing more we could do about shader compilation stuttering, and many wondered if it would ever be solved...

And the official Nintendo pdf tech sheet I have for the cube stored on an old CD-R somewhere in the shed/Garage even states Opengl shaders support IIRC.

GameCube Architecture | A Practical Analysis

An in-depth analysis that explains how this console works internally

www.copetti.org

And here's the quote from the practical analysis specifically about the Cell shading.

During the same time, PC graphics cards were starting to discard fixed-function pipelines in favour of shader cores (units that run small programs which define how pixels are operated). Flipper still contains a fixed-function GPU, however, by including components such as the TEV unit, one could argue that Nintendo provided their own shader-like solution.

I guess one of the best examples of games that exploited this new capability is The Legend of Zelda: Wind Waker which implements a unique colour/lighting technique known as Cel shading to make its textures look cartoonish.

 

[PaintTinJr]

no, only Xbox had shaders, PS2 and GC had ways to manipulate data using some very powerful hardware, GC had the TEV to mix textures add some multipass and you can get some very interesting effects, PS2 is just a multipass monster and has the VU's which have lot of math power and are programmable, both consoles have hardware that can can do what shaders do or some of the things a shader do but they are not shaders in the sense of writing your shader in something like GLSL compiling and executing so it wasnt simple to use, when they were being developed tech was advancing very fast but wasnt well stablished like today they identified graphical needs and tried to prepare for the future, MS just used their money to build a custom PC with what nvidia had at the time it was very powerfull but very constly I remember some articles at the time that they were losing lot of money for each console, they literally bought their place in the industry and wasnt cheap

I had a ps2 linux kit and ran shader assembly examples on the GPU back in the day, and was learning shaders on an Opengl 2.0 PC card at the same time, and the PS2 had more shader feature flexibility, even if it was essentially GPU driver microcode being ran.

 

[SomeGit]

My man posting a paragraph that contradicts what he is saying, brilliant. I’d like to see those OpenGL shader documents from Nintendo.

Yes I read the ubershaders, I’m not sure if you did.

 

[polybius80]

I had a ps2 linux kit and ran shader assembly examples on the GPU back in the day, and was learning shaders on an Opengl 2.0 PC card at the same time, and the PS2 had more shader feature flexibility, even if it was essentially GPU driver microcode being ran.

never used one, but afaik the VU ae programmed in assembler you used assemble or a library that hides that complexity in something more like C?


I always wanted a PS2 linux kit, I remember it was very fast and there were incredible demos for VU, I remember a youtuber many years ago that ran quake2 without any problems despite not having any form of optimization just compile the engine and run very impressive

here are some demos entirely in VU


super impressive stuff

 

[Fafalada]

Like if the PS2 or game cube did use per pixel lightning ever.... Silent Hill 2 did it on Xbox and it looks noticeable better specially with the flash light.

Per-Pixel flashlights were used across plenty of PS2 games:
GhostHunter, Echo Night Beyond, SH0 (PSP/PS2), SH Shattered Memories - PS2/PSP/Wii (that'd qualify for GCN too) to name some of the more notable examples.

I'd expect there were more GCN games with it too just don't know off the top of my head.

 

[Alexios]

Some random average Dreamcast games nobody cares about (making finding decent footage pretty hard) did nifty tricks at times. You could probably make a pretty convincing flashlight using something like this with nicer meshes/textures for subtler edges combined with other DC game effects:

​

It looks like it properly illuminates ahead rather than obviously overlay a texture as in Tokyo Xtreme Racer 2 (though it illuminates other cars there), 4 Wheel Thunder & San Francisco Rush 2049. V-Rally's lighting seems more real but definitely not as smooth/per pixel even though it's (mostly) 60fps.

Sega Rally 2's equivalent seems even more step-like though & MSR's is so subtle it's barely there. Le Mans would have benefitted from any one of these techniques as well, it has a truly lovely headlights-flashing-the-camera effect and other cool stuff but the track is 100% unaffected by headlights.​

That's using DCs equivalent of stencil volumes - it's decent enough and cost effective (it saves on processing any geometry math from an actual spot-light) but you can't really do soft edges or attenuation of any sort this way - it's just on/off switch for every pixel in the volume. The technique has seen 'some' use on other consoles - in NPR/cell-shaded games where the art-style actually dictates hard edges/no attenuation of colors.

I'm not saying make the volume subtler (though it could perhaps have more polygons to look smoother/rounder when not used by every opponent car but just one game character), just overlay a texture with some faked flashlight style rings as in Tokyo Xtreme Racer 2 and/or different color accents, add a lens flare when looking towards the source and other similar stuff any random game does in cheap methods. The rest of the game's lighting still works (all these tracks have different tod/weather versions to play btw) so the regular spot lights or baked lighting also illuminate the backgrounds and cars as seen in the Las Vegas or whatever part of the Nevada track or the first tunnel in the first video, so it seems doable. For a game with real time tod like Shenmue they would disable it at the bright parts of the day and indoor areas that have fixed bright lights (toggleable or not) if it would clash with the look (Ryo would just go "I don't need the flashlight now") etc., basically just put some more attention to detail, program more associated behaviours and functions that don't come naturally from it being a true light source/system but additional entities/stuck on art assets.

You keep saying it didn't really ever work/look right but it looks/works pretty good here. Even if it's a little bit extreme (just make the volume smaller) it's still better than the texture only and other approaches in mentioned titles even without this extra polish they could take the effort to do in this or other titles with this effect. Sure it's a bit harsh and stark, but so are other assets. It fits. We're talking late 90s/early 00s here, texture resolutions, polycounts and effects weren't exactly peak level even when going for a realistic art style (which I think this game mostly does even with the fake cars etc.).

 

[Fafalada]

PS2 is just a multipass monster and has the VU's which have lot of math power and are programmable

The modern equivalent of VUs are MeshShaders - right down to working in 'meshlets'.
Eg. for Doom 3 style volumetric shadows, Xbox would use CPU processing and multiple passess with Vertex Shaders to extrude, cull and cap the shadow volumes.
On PS2 the whole process ran in single pass entirely on the VU.

 

[Fafalada]

Some random average Dreamcast games nobody cares about (making finding decent footage pretty hard) did nifty lighting tricks. You could probably make a pretty convincing flashlight using something like this with nicer meshes/textures for subtler edges combined with other DC game effects

That's using DCs equivalent of stencil volumes - it's decent enough and cost effective (it saves on processing any geometry math from an actual spot-light) but you can't really do soft edges or attenuation of any sort this way - it's just on/off switch for every pixel in the volume. The technique has seen 'some' use on other consoles - in NPR/cell-shaded games where the art-style actually dictates hard edges/no attenuation of colors.

 

[Lysandros]

never used one, but afaik the VU ae programmed in assembler you used assemble or a library that hides that complexity in something more like C?


I always wanted a PS2 linux kit, I remember it was very fast and there were incredible demos for VU, I remember a youtuber many years ago that ran quake2 without any problems despite not having any form of optimization just compile the engine and run very impressive

here are some demos entirely in VU


super impressive stuff

Real time ray tracing with a single vector co-processor with 16 kb of memory, impressive indeed.

 

[Geometric-Crusher]

What would be a high level, medium level and low level Dreamcast game? some criteria are the polygon budget, texturing, frame rate and effects.

High

Dead or Alive 2 Limited Edition is certainly the best technology shown on the Dreamcast. DOA2 is possibly the limit, this means that DC would not be able to make more advanced fighting games than DOA2, only sidegrades with some degrees of inferiority.
Test Drive Le Mans 24 hours this game stands out, on the Dreamcast there are other equally enjoyable racing games but test drive le mans is such a fabulous achievement that for years it fueled the urban legend of 5M polygons, when in fact the game has 1M as peak.
NBA 2k2 NFL 2k2 they are basically the doa2 of the sport, 60fps, nice.
Shenmue 1/2 these games are very advanced.
Ikaruga amazing gfx this game is on the edge of the line that separates a regular Dreamcast game from those mentioned here.

Low

low level games in general are 5th generation games ported to the Dreamcast without major graphical upgrades games like The Grinch, Star Gladiator 2, Tech Romance, Creatures 2, street fighter Alpha 3, Guilty Gear X (all these common 2D games) puzzle games as Cleopatra Fortune.

Medium

All games in the range starting with the N64 ports like Star Wars Episode 1 Race and Shadowman, going through Arcade Model 2 games like Daytona and ending with the big five high-level games above.

Dreamcast was a very easy console to program, it reached its maximum power since launch, you can see that Sonic Adventure 2 has the same polygonal budget as Sonic Adventure, Sega just changed the exploration theme in favor of a more on-rails game, let's say , however the computational power used is the same.

 

[Freelance Yakuza]

VMU > Memory Card

 

[Xaeroxcore666]

Well, it's not our fault if GTA3 developers and others say the Dreamcast is more capable than what we thought... .

- Soulcalibur was released 8 months after the launch, imagine Namco doing another fighter in 2003 with a 3 years of development and not a 1 year arcade port...
- ... bUt tHe DrEaMCast hAS aLMosT rEached ITs LimITs wIth SoULcalibur 1!

All we know is that it's a capable console for a 5 years life cycle (98-2003). We never expected Gamecube/Xbox level...

Amen

What would be a high level, medium level and low level Dreamcast game? some criteria are the polygon budget, texturing, frame rate and effects.

High

Dead or Alive 2 Limited Edition is certainly the best technology shown on the Dreamcast. DOA2 is possibly the limit, this means that DC would not be able to make more advanced fighting games than DOA2, only sidegrades with some degrees of inferiority.
Test Drive Le Mans 24 hours this game stands out, on the Dreamcast there are other equally enjoyable racing games but test drive le mans is such a fabulous achievement that for years it fueled the urban legend of 5M polygons, when in fact the game has 1M as peak.
NBA 2k2 NFL 2k2 they are basically the doa2 of the sport, 60fps, nice.
Shenmue 1/2 these games are very advanced.
Ikaruga amazing gfx this game is on the edge of the line that separates a regular Dreamcast game from those mentioned here.

Low

low level games in general are 5th generation games ported to the Dreamcast without major graphical upgrades games like The Grinch, Star Gladiator 2, Tech Romance, Creatures 2, street fighter Alpha 3, Guilty Gear X (all these common 2D games) puzzle games as Cleopatra Fortune.

Medium

All games in the range starting with the N64 ports like Star Wars Episode 1 Race and Shadowman, going through Arcade Model 2 games like Daytona and ending with the big five high-level games above.

Dreamcast was a very easy console to program, it reached its maximum power since launch, you can see that Sonic Adventure 2 has the same polygonal budget as Sonic Adventure, Sega just changed the exploration theme in favor of a more on-rails game, let's say , however the computational power used is the same.

Interesting approach. Agree with most of it but have some comments on:

High: i'd add also Headhunter, Ferrari F355, Quake 3, Propeller Arena and Ecco, Gundam Federation vs Zeon (also some Shumps like Under Defeat, Psivariar 2 and Castle of Shikigami II, amongst others), Code Veronica, 18 Wheeler, Rez, Sega Sports Jam and Crazy Taxy 2. All of them are way above most part of DC catalogue.

You say DOA 2 hits DC limit and thats true probably, but Esppiral is showing basically to the world that even that game has room for improvement and that it was pretty comparable to lotta PS2 3D fighter.

On Medium i'd also add all Atomiswave 3D game ports.

On Low i know we are talking about polygon budget, but Guilty Gear X was probably the full 2D most advanced game on that console. Its sprites are way above anything former gen. The only game i'd say surpass it, is Atomiswave Hokuto No Ken...Damn that game looks awesome! Too bad cannot be paused.

 

[Geometric-Crusher]

Interesting approach. Agree with most of it but have some comments on:

High: i'd add also Headhunter, Ferrari F355, Quake 3, Propeller Arena and Ecco, Gundam Federation vs Zeon (also some Shumps like Under Defeat, Psivariar 2 and Castle of Shikigami II, amongst others), Code Veronica, 18 Wheeler, Rez, Sega Sports Jam and Crazy Taxy 2. All of them are way above most part of DC catalogue.

You say DOA 2 hits DC limit and thats true probably, but Esppiral is showing basically to the world that even that game has room for improvement and that it was pretty comparable to lotta PS2 3D fighter.

On Medium i'd also add all Atomiswave 3D game ports.

On Low i know we are talking about polygon budget, but Guilty Gear X was probably the full 2D most advanced game on that console. Its sprites are way above anything former gen. The only game i'd say surpass it, is Atomiswave Hokuto No Ken...Damn that game looks awesome! Too bad cannot be paused.

I need to explain my concept of medium game and high game. You notice that almost all Dreamcast games have a clean look and good textures, so they will all be wonderful in this regard, so what separates a high-level game from a medium-level one is not the beauty we see on the screen but the percentage of system use, that said.

Under Defeat I have a preference for Ikaruga but it's entirely possible that Under Defeat is on the same level or perhaps a step above.

Headhunter It's very simple, words from a programmer who worked on it , the game moves from 8 to 15k polygons per frame and the lighting (although beautiful within the game's art) is flat and almost non-existent. Source Beyond3D.

Ferrari F355 This could be without a doubt because it is a game that pushes over 1.5M polygons when using the third-person mod I chose Le Mans because it's unbelievable so many vehicles, with dynamic shadows on all of them, medium lighting (above HeadHunter, below RECV) reflections from the brake lights on the wet track. It's not just polygons per second but polygons per frame.

Propeller Arena amazing for sure but again what separates a high-level game from a medium-level one is not the beauty we see on the screen but the percentage of system use.

Quake 3 can you believe this game does between 8 and 16k per frame at 20/30fps to 320k peak 480k poly/s ? It impresses with its textures and lighting effects, it's a shame it's not a solid 30fps.

On Medium i'd also add all Atomiswave 3D game ports

I agree


I think that from the moment the list of high level games increases it naturally becomes average. the high level needs to be the highest among the others, it is not possible at same time for Sonic Adventure, Rayman 2, Ecco The Dolphin, Super Magnect neo etc. you need to select the highest among them imo.

 

[Esppiral]

What would be a high level, medium level and low level Dreamcast game? some criteria are the polygon budget, texturing, frame rate and effects.

High

Dead or Alive 2 Limited Edition is certainly the best technology shown on the Dreamcast. DOA2 is possibly the limit, this means that DC would not be able to make more advanced fighting games than DOA2, only sidegrades with some degrees of inferiority.
Test Drive Le Mans 24 hours this game stands out, on the Dreamcast there are other equally enjoyable racing games but test drive le mans is such a fabulous achievement that for years it fueled the urban legend of 5M polygons, when in fact the game has 1M as peak.
NBA 2k2 NFL 2k2 they are basically the doa2 of the sport, 60fps, nice.
Shenmue 1/2 these games are very advanced.
Ikaruga amazing gfx this game is on the edge of the line that separates a regular Dreamcast game from those mentioned here.

Low

low level games in general are 5th generation games ported to the Dreamcast without major graphical upgrades games like The Grinch, Star Gladiator 2, Tech Romance, Creatures 2, street fighter Alpha 3, Guilty Gear X (all these common 2D games) puzzle games as Cleopatra Fortune.

Medium

All games in the range starting with the N64 ports like Star Wars Episode 1 Race and Shadowman, going through Arcade Model 2 games like Daytona and ending with the big five high-level games above.

Dreamcast was a very easy console to program, it reached its maximum power since launch, you can see that Sonic Adventure 2 has the same polygonal budget as Sonic Adventure, Sega just changed the exploration theme in favor of a more on-rails game, let's say , however the computational power used is the same.

The psx was also easy to program for, even more so than the Dreamcast , now compare Tekken 1 to Tekken 3, Rage Ricer to Ridge Racer Type 4, Toshinden to Soul Blade, Busby to crash 3, Final Fantasy VI to IX, and the list goes on, sorry your argument is just not valid.

 

[Geometric-Crusher]

The psx was also easy to program for, even more so than the Dreamcast , now compare Tekken 1 to Tekken 3, Rage Ricer to Ridge Racer Type 4, Toshinden to Soul Blade, Busby to crash 3, Final Fantasy VI to IX, and the list goes on, sorry your argument is just not valid.

Bubsy 3D is a single-buffered open-world PS1 game running at one of its highest resolutions of 512x480 twice than regular games like Rayman 2.
don't underestimate Bubsy3d it's a bad game but it's the only full 3d adventure game on ps1 to run in this resolution, in 1999 IS internal section 640x480 the others to achieve this were fighting games. Crash is quite nice but are we going to deny that they are on rail?
Soul Blade is more pleasing to the eye but again what separates a high-level game from a medium-level one is not the beauty we see on the screen but the percentage of system use.
the system usage in Toshinden is incredible for a launch game they made a character with translucent clothing just to demonstrate the Playstation's capabilities, it reached its limit at launch, let's hear the numbers
Q: What are the technical specs of the game?

A: Takara proudly proclaims 90,000 polygons per second on the CD and
booklet cover. In a magazine interview, it was stated that the game runs
at 30 frames per second, so it shifts about 3000 polys in a single frame.
Each character uses about 800+ polys, with texture-mapping, gouraud
shading, multiple light sources and transparency effects.

Q: What about glitches/bugs/slowdowns?

A: Only inevitable but very slight polygon sorting glitches (due to no
Z-buffering) and occasional noticeable slowdown when the characters are
really up close to the camera. Especially so in Kayin's stage where there
is a real-time screen display (picture in picture in picture...)

Interestingly, there are more polygons per frame than Tekken 3
the characters have almost same number of polygons however Toshinden has 3d backgrounds
resolution 640x240 while Tekken 3 has a higher resolution 384x480.

Tekken 1's objective was not to compete with graphics but to be the first 3D fighting game at 60fps on consoles (Toshinden, virtua fighter Remix 30fps)

Ridge Racer 640x480 60fps on arcade 240.000 poly/s 320x240 30fps
"Considering the fact that this was our first development for PlayStation and the development period was only six months, we could not take a risk and decided to have a locked 30fps," Sakagami says
more impressive thing is that the PS1 achieved 60fps in the version that came with RR-T4

sorry your argument is just not valid.

Honestly my thesis remains, it is not possible to improve everything at the same time, sorry

 

[nkarafo]

The psx was also easy to program for, even more so than the Dreamcast , now compare Tekken 1 to Tekken 3, Rage Ricer to Ridge Racer Type 4, Toshinden to Soul Blade, Busby to crash 3, Final Fantasy VI to IX, and the list goes on, sorry your argument is just not valid.

You are comparing some very early "1st gen" PS1 games with it's latest, best looking ones. And you also use Bubsy for some reason, as if it represents your average early PS1 game and not a completely shitty one? What exactly is your argument here?

The best looking DC games are not the equivalent of these early games you are mentioning. Test Le Mans was released 2.5 years in DC's life. Now let's see some better representatives of PS1 games that were released that late in PS1's life: Crash Bandicoot, Resident Evil, WipeOut 2097, Tobal 2. These games are not the absolute highest tech demos for the console but they are still a good indication of what it can do. The DC lasted 3 years so it did reach this level of software development. This argument about the DC not having shown everything it could do because it died too early is true, but Sega fans tend to exaggerate it a lot.

 

[Xaeroxcore666]

I need to explain my concept of medium game and high game. You notice that almost all Dreamcast games have a clean look and good textures, so they will all be wonderful in this regard, so what separates a high-level game from a medium-level one is not the beauty we see on the screen but the percentage of system use, that said.

Under Defeat I have a preference for Ikaruga but it's entirely possible that Under Defeat is on the same level or perhaps a step above.

Headhunter It's very simple, words from a programmer who worked on it , the game moves from 8 to 15k polygons per frame and the lighting (although beautiful within the game's art) is flat and almost non-existent. Source Beyond3D.

Ferrari F355 This could be without a doubt because it is a game that pushes over 1.5M polygons when using the third-person mod I chose Le Mans because it's unbelievable so many vehicles, with dynamic shadows on all of them, medium lighting (above HeadHunter, below RECV) reflections from the brake lights on the wet track. It's not just polygons per second but polygons per frame.

Propeller Arena amazing for sure but again what separates a high-level game from a medium-level one is not the beauty we see on the screen but the percentage of system use.

Quake 3 can you believe this game does between 8 and 16k per frame at 20/30fps to 320k peak 480k poly/s ? It impresses with its textures and lighting effects, it's a shame it's not a solid 30fps.


I agree


I think that from the moment the list of high level games increases it naturally becomes average. the high level needs to be the highest among the others, it is not possible at same time for Sonic Adventure, Rayman 2, Ecco The Dolphin, Super Magnect neo etc. you need to select the highest among them imo.

Now i totally get it. I´m also on that Beyond 3D "Yes, but how many polys" thread. Don´t know if you have watched recently, we found out there is a Pachinko version of VF4, with real time rendered scenes with sort of interaction, and visually looks under regular Naomi 2 VF4...Seems like it runs on normal Naomi or the Reglation 7 Dreamcast made for that Pachinko machines.
Also wow...So if Q3 has that low poly budget, why didnt they made it run 60 fps? I mean, yes textures, lighting and other effects were pretty "next gen" at the time for that game...To the point with it, with a geometry budget a little above regular 32 bit era game it manages to show generational leap on visuals. Also, how would DC handle another Q3 engine games, like the MoH and Jmes Bond ones, if on multiverse EA would had developed for it?

The psx was also easy to program for, even more so than the Dreamcast , now compare Tekken 1 to Tekken 3, Rage Ricer to Ridge Racer Type 4, Toshinden to Soul Blade, Busby to crash 3, Final Fantasy VI to IX, and the list goes on, sorry your argument is just not valid.

But i do get his point: probably by 1995 Namco reached PSX peak performance with those games, but due to optimization, smarter usage of resources and better dev tech they were able to round around those limitations and make games that made a huge visual leap over the years, despite running on the same hardware. So same applies for DC or any other console, with SA 2, they managed to make a better looking game, which is equally demanding as first part...You are also testing and overcome DOA2 limits...The Atomiswave games also made something similar: Faster than speed runs on Sega´s GT engine, an d despite being inferior to, for example Toky Extreme Racer, game looks leagues above Sega GT. In fact, just few DC games really pushed it to the the limit and also showed impressive real "128 bit gen" worth visuals, most part of the catalogue are 5th gen enhanced ports, 1999 Cheap PC ports or games built under 32 bit dev logic...So even taking Ferrari or Le Mans, or Shenmue, or DOA2 tech could be viable to make more DC games that can look more on par to what was the rest of the gen (obviously with cutbacks).

 

[Zathalus]

Not going to read through this entire thread but of course the Dreamcast was a powerhouse when it launched. It was a significant step up from the PS1/N64/Saturn. It's equally true that it wasn't quite as powerful as consoles that launched years later, but that doesn't mean it was a full member of the 6th generation of consoles. Regarding power isn't it generally agreed that DC<PS2<GC<Xbox? That is overall ranking with certain features from one console being better then others. Such as the PS2 being really good at Pixel fill rate. Such was the nature of technology at the time that no single console could claim to be best at everything.

 

[Fafalada]

So if Q3 has that low poly budget, why didnt they made it run 60 fps?

Q3 polygon budgets being low was well known I though? (entire levels being the budget of a single character in fighting games on consoles).
But that was by design - rendering was still built around idea of minimizing overdraw and geometry visible as the main concern - which was actually opposite to evolution of contemporary 3d acceleration (especially subsystems found in consoles) where polygon throughput was at the forefront(so it was often more efficient to ignore visibility concerns and just draw more - faster). Basically the code of those old PC engines was very suboptimal for console purposes - and expecting ports to fix such design issues in the codebase is unreasonable of course.

I'm not saying make the volume itself subtler (though it could perhaps have more polygons to look smoother/rounder when not used by every opponent car as seen here but just one game character instead), just overlay a texture on top of it with some faked flashlight style rings as in Tokyo Xtreme Racer 2

The flashlight rings would be 'ok' but tbh I never liked the look when experimenting with it. Overlaying a texture on the entire volume would require projected texture in which case you might as well just do that and skip the volume - as that's the bulk of compute/render cost (projecting geometry transforms + multiple polygon passes).
I've experimented with this approach at some length back in the day when looking for a 'cheap' way to render car-headlights. In the end - it just didn't really work - even with filtered volumes to get softer edges, it just kind of looked off if the scene had 'realistic' lighting/materials.
That said - one thing that it might have been useful for could be in combination with vertex light (like that Sega Rally example) make the volume fit tighter and use it to 'cull' away the visible vertex edge-artifacts. That still requires the real light calculations - but it could work to make it look more like 'per-pixel' approach, without multiple geometry passes/projected textures.

 

[PaintTinJr]

The psx was also easy to program for, even more so than the Dreamcast , now compare Tekken 1 to Tekken 3, Rage Ricer to Ridge Racer Type 4, Toshinden to Soul Blade, Busby to crash 3, Final Fantasy VI to IX, and the list goes on, sorry your argument is just not valid.

What you are saying is true, but within the context of Ps1/DC to-the-metal that your comment is mixing ease of programming with versatility and hardware abstraction, this isn't equal IMO, and exemplary final results of games like Iss Pro Evolution 2(with brilliant animation/gameplay of 22 players) or Colin McRae Rally 2.0 versus early efforts of fun games like Olympic Soccer, etc were possible mostly because of the to-the-metal, of ps1 and less about the time to first triangle ease of programming.

Given that you have been building the unofficial DC remaster of DoA2, how close to the metal can you go? Or is it pretty much the same high level, heavily abstracted Opengl ~1.4 API situation?

 

[Fat Frog]

The best looking DC games are not the equivalent of these early games you are mentioning. Test Le Mans was released 2.5 years in DC's life.

Test Le Mans was released 1.9 years in DC's life. (DC: Nov 27th 98/ Le Mans :Nov 8th 2000)

The Dreamcast died after 2 years 3 months and a few days:

November 27th 98 (end of the year...) - March (early) 2001.

88-91 It's not even the release of Sonic 1 on the genesis lol.(june 91)

Esppiral is goddam right

 

[nkarafo]

Test Le Mans was released 1.9 years in DC's life. (DC: Nov 27th 98/ Le Mans :Nov 8th 2000)

The Dreamcast port was released in March 2001 according to wiki.

And how is Esppiral right? They mention games that were either launch titles or released within a few months after the PS1 launch... But you don't seem so eager to be as precise with the dates and numbers in this case.

2+ years is enough for software to advance to a level that can represent the hardware.

 

[Alexios]

2+ years is enough for software to advance to a decent level for any machine.

There are "decent" games for DC, but way more would have been done. If DOA2 or Le Mans, just a couple games, were never released (or were crappier) for any reason you'd adjust DC potential expectations to way lower levels (like Sc/SR2) and expect folks to agree just as you do now saying this or that is close enough. The industry poured way less resources into DC so it'd get there slower than PS1 which kept advancing way past the first couple of years so all it takes is a couple games to be missing for that to happen. With a longer life way more could have been done, even by fewer devs.

As for esppiral's work, it's still the DOA2 engine, he's testing its limits and doing things it was made to do but for whatever reason weren't implemented in that ever updated game, it's not like he has the source code or resources to optimize it further, improve problem areas and outdo whole Tecmo.

 

[nkarafo]

There are "decent" games for DC, but way more could have been done, as was done for PS1. If DOA2 or Le Mans, just a couple games here and there, were never released you'd all adjust DC potential expectations to way lower levels and expect folks to agree in the exact same way you now want to say this or that is close enough to the absolute top it could ever achieve. The industry poured way less resources into DC than PS1 which still kept advancing way past the first couple of years so all it takes is a couple games to be missing for that to happen. With a longer life more could be done.

Yes, more could be done but i don't know what exactly are you expecting? How much of the DC's potential do you think the best looking released games tapped? 10%? 50%? 90%? Is there even a way to be sure or are we going to argue over assumptions?

My initial argument was the DC lived way past it's launch/first batch games era. It's not like the difference between it's best looking released game and the hypothetical "maxed out" later game would be like comparing Bubsy 3D VS Crash Bandicoot 3, like the person i responded to claimed.
Like i said in that post, the argument about the DC having more untapped potential stands, but Sega fans tend to stretch this to the limit when it comes to DC.

 

[Fat Frog]

The Dreamcast port was released in March 2001 according to wiki.

Le Mans 24 Hours (video game) - Wikipedia

en.m.wikipedia.org

"

• NA: 8 November 2000[4]" (copy paste from Wikipedia... March 2001 is the jap version of the OG Western game...)

1:52
You can even see the usual 2000 copyright on the title screen.

 

[Alexios]

I'm not assuming anything, just going by logic. I won't give you any % to give you room to argue over assumptions, you people want to assume it's so close it'd basically make 0 or 0.1 perceivable difference (because Bubsy and Toshinden were in fact amazing after all of all reasons/excuses to make such claims) yet it's made huge difference for the vast majority of systems that got to benefit from such (and the short lifespan of the Dreamcast already made a difference for it as well even with far fewer developers both invested in and capable of exploiting it to any decent degree). Whatever

 

[nkarafo]

I'm not assuming anything, just going by logic. I won't give any % to give you room to argue over assumptions, you people want to assume it's so close it'd basically make 0 or 0.1 difference when it's made huge difference for the vast majority of systems that got to benefit from such. Whatever

So your issue is "us people" who assume it's so close it would barely make any difference, no more than 0.1% (wow that's very precise for someone who "doesn't want to give any %") even though i never posted anything that warrants this response.

But it's OK if someone claims the existing best looking DC games use so little of the machine that are basically the equivalent of the launch games or the worse looking ones of another console.

So yeah, whatever

 

[Fat Frog]

but Sega fans tend to stretch this to the limit when it comes to DC.

Probably because many Sony boys said for years that the DC cannot really do much more than Soulcalibur. (before realizing decades later with polycount that DOA2 is already a huge improvement)...

All we want is seeing more Dreamcast games with modern tools like the ps2.

For instance, G Geometric-Crusher mentioned the impressive Fight Night 2 probably thinking it was using the shit load of polygons of the PS2...
Then, it appeared that it was in fact due to modern tools. (there weren't 15k polygons per boxer but 8k very well used.)

 

[Roronoa Zoro]

In any case, those two years of beautiful visuals and fantastic games, were glorious.

It might have been a half step to the experiences the PS2, GC, and Xbox would eventually provide, but it was a wonderful one, and I wouldn’t change a thing.

I mean I'd change the sales...

 

[Geometric-Crusher]

sorry but the theory of the short life of the Dreamcast is total bullshee imo, Sega pulled the plug in 2001 but many games continued to come out (SA2, HeadHunter, Soldier of Fortune, stunt gp etc) then Dreamcast gained extra life through the arcade Naomi 1 and Atomiswave. Ranger Mission 2004 is superior to similar games on Dreamcast, for all intents and purposes Dreamcast has not only completed its cycle but is also one of the most explored consoles.

 

[Alexios]

Ranger Mission is not superior to HOTD2 or CM with its Behind Enemy Lines level low polygon enemies and drab backgrounds (though admittedly it does a half decent job aping the camera and snappy gameplay of Sega lightgun games, it's far from any of them in real quality, visually or otherwise) and most Atomiswave games past the DC's lifespan are cheap Sammy shovelware for mom and pop stores going for the cheapest, crappiest machines/games they can get their hands on and often don't even compare to Model 2 greats, lol. Dynamite Cop looks better than Demolish Fist, Manx TT & Motor Raid don't seem a whole gen behind Maximum Speed and Faster than Speed has some slick neon art assets but is clearly low budget, crappy and 30fps (without MSR tracks' polygonal complexity to make up for it) so it doesn't top any of the half decent and above racers on Dreamcast, not even Rush 2049. None of the later fighters, finished or otherwise, top anything good on the DC either, whether going for full 3D or cartoony looks like Project Justice (itself a very modest game but still better realized than the Rumble Fishes). The arcades had some life left but this stuff was crap.

Maybe it's possible on the Dreamcast with 30fps, reduced draw distance and some cuts in alpha effects.Each game has a reason, the dev's choice to make Gungriffon this way was to explore long draw distances without fog at 60fps. In the later stages there is fog, but with an artistic purpose. Gungriffon Blaze is part of the initial wave of the ps2 with a Dreamcastish look.

We cannot underestimate 60fps, the Dreamcast cannot reach 60fps even in some advanced 5th generation games.

shadowman 30fps (looks amazing)
Soul Reaver 60 (with frame drop)
star wars ep 1 race 30fps
Tony Hawk pro Skater 30fps
Re-Volt 30fps with code to reduce draw distance
Rayman 2 60fps ( almost no geometry increase)
Dino Crisis (3D engine) 30fps
RE3 and AitD 60fps
Vigilante 8: 2nd Offense 60fps

As we can see, some games have had an upgrade in polygon count, becoming full Dreamcast games, the character Tony Hawk has 1,500 tris and Soul Reaver Raziel 1,222 tris. Shadowman was also upgraded so 60fps was not possible.

Floigan Bros
Illbleed (with frame drop)
KAO the Kangaroo
Outtrigger
Sonic Adventure 2
Spawn: In the Demon's Hand
Super Magnetic Neo
Unreal Tournament 60fps fake
Soul Fighter
Pen Pen TriIcelon

As we can see, some games despite being complete Dreamcast have one or another intelligent trade off Sonic Adventure 2 holds 60fps better in the on-rails stages, Kao The Kangaroo has only one enemy at a time in the scene, Soul Fighter has shadows but collisions are bad, Spawn also doesn't have shadows like in Shadowman, alpha simulations are simple, Outtrigger is basically Spawn. Sonic 2 looks the best

the best and most impressive Dreamcast games are 30fps so these early ps2 games running at 60fps was a huge achievement in those days.

Great logic, Toukiden 1 is 30fps on any beast PC, so any beast PC can't handle slightly upgraded PSP games at 60fps!

GTAIV still has performance problems more severe than way more impressive games so any beast PC can't handle PS3 era jank, you can't ever fault a dev/port/game for an unoptimized result, only the platform!

You just used one of the few 480p examples on PS2 (despite huge issues with it) to decide it negates the 480p advantage of Dreamcast's almost every game and now you use any example of a badly performing/looking game/port on Dreamcast as its best scenario and ignore its actual greats?

Dreamcast had plenty 60fps and 30fps and any fps games that look well beyond anything on previous platforms and wouldn't look out of place or like a different generation next to any PS2 great, even if they wouldn't reach its absolute best (just as plenty worthwhile PS2 games don't reach its top 1% in the same way and stuff like Persona 4 look closer to Sakura Wars 3 & 4 and Skies of Arcadia than Final Fantasy XII), in every stage of its short life. It's been demonstrated here more than enough to go back to such ignorant, pure, distilled bullshit.

The Dreamcast had the excess power needed to emulate (and enhance with double resolution, texture filtering, improved performance and whatever else) PS1 with a half baked and quickly killed off emulator app, not just handle proper ports with a few minor upgrades.

Going by that "logic" (personally I don't think beautiful, polished, stylish 60fps games like F355 or even Daytona USA lose to any 30fps equivalent like Le Mans for not having all its best effects, as the 60fps has a visual factor too, same for MGS2 vs 3 on PS2 even if Kojima clearly chose how to advance with the series at that point with lower fps for higher fidelity, same for Crazy Taxi 2 being one of the most impressive Dreamcast games improving on the original with way less pop in, etc.) the best and most impressive games on any 3D platform ever are generally lower fps since it's always been a trade off between fidelity (in various ways, resolution, polycounts, scale of game design, effects, whatever) and performance, the fuck is your point? You just name dropped GTAIII as so beautiful on PS2 and I showed you its fps, now you prefer the plainest looking 60fps game on it instead to push the opposite narrative? Even current platforms demonstrate this with "more impressive" fidelity vs performance modes available. Or just PC games always having options that will increase some visual aspect and reduce performance. It's how this 3D shit works!

PS: stuff like Shadowman get single freaking digit fps on PS so between that, the resolution & other upgrades it's a huge performance gulf, even if far from DC's best as it never got a high budget/talent/effort exclusive of the sort. This is what some effort did for ports, never mind real next gen titles:

​

It's not like PS2 or even GC always fared better in instances such multi version games existed, see Rush 2049 and Hydro Thunder for example, where you can both fault them for having a game also on last gen platforms yet not achieving constant 60fps and having a worse version than the Dreamcast.

AitD in your list is on PS2 too. Clearly devs were waiting for PS2 before moving on from the last generation, another hit for Dreamcast, hence there are less examples of real multiplatform games that are on all of these systems. Most DC ports to others were hardly enhanced if not worse though

 

[Xaeroxcore666]

sorry but the theory of the short life of the Dreamcast is total bullshee imo, Sega pulled the plug in 2001 but many games continued to come out (SA2, HeadHunter, Soldier of Fortune, stunt gp etc) then Dreamcast gained extra life through the arcade Naomi 1 and Atomiswave. Ranger Mission 2004 is superior to similar games on Dreamcast, for all intents and purposes Dreamcast has not only completed its cycle but is also one of the most explored consoles.

Sorry my man, but cannot agree that. There was no commercial reasons to put the pedal to the metal on an already dead console. Yes, Sega gave DC an almost entire year of extra life on the west, nut all of them were projects started on 2000 or even 99, and also the console kept receiving few games at japan until 2007, but were how many? like few shumps per year. Yes, they took advatange of the hardware, but that was easy because 1. most of them also were released on Naomi 2. Is way easier, faster and cheaper to max out the console on games with fixed top view on rails than doing it on a proper 3D game. Atomiswave games were all low budget developments and the bigger franchises (The SNK ones) were all 2D. Yes, Force Five and Faster than Speed maybe showed the advantages of developing for DC years after its commercial demise (specially Force Five), but they are still low budget games (Force Five isnt even finished) which are technically and visually under the best DC could get on it actual commercial short lifespan. For example, Confidentail Mission stomps Ranger Mission. In other hand, all desktop PlayStations, for example, were taken to the very limit and way beyond because they had more years to know how to overcome hardware limitations and the investment on I+D to achieve greater results because they all have been profitable systems with longer lifespan. Do you think anyone on the first 2.5 years of lifespan of PS1 and PS2 could had thought games like RR 4, GT2, Driver 2 ort Vagrant Story would had been possible on PS1? Or GOW 2, Burnout Revenge, Black and even a Gamecube powerhouse like RE 4 were going to be possible on PS2? I insist in maybe you could be right on Le Mans, F355, both Shenmues and DOA 2 took DC to the very limit, but even rounding around them you could still reach the very limit but obtain greater graphical results, sometimes even with clever art direction, more than just tech pushed beyond, like like the ones we have witnessed on other consoles.. We haven´t got something like that on DC and we really hope through community at last is going to be possible.

 

[polybius80]

For instance, G Geometric-Crusher mentioned the impressive Fight Night 2 probably thinking it was using the shit load of polygons of the PS2...
Then, it appeared that it was in fact due to modern tools. (there weren't 15k polygons per boxer but 8k very well used.)

what makes a game look good(other than art direction) are the effects, the more effects you use it can affect the number of triangles that is why the number of triangles is not necessarily a good metric against another game since you have to make concessions depending what you want to show in your game there are things that you wont get just with high polycounts, not only depends how are you using those triangles, what if you want particles? what if you want water that reacts with movements? what if you want bump maps? o simply what if you want more than a couple lights in the scene?, I have heard that doa2 in DC has a debugg mode where you can play with lights, afaik the game uses 1 light per character and one for the scene you add more than that and the framerate decreases just like the cutscenes are 30 fps as it has more lights(probably it doesnt require 30 but vsync forces it) so even doa2 has to make concesions when it wants more, wants more light and vsync? then you can only display half the triangles per second as before, that is the concession

an it will be until you find a way to do better

I dont know exactly what tools you are talking but sure better tools help in development and experience get better result so naturally later games in general have better effects than the early ones but dramatic improvements tipically require something in the machine that have not been exploited and now you are using it, wasnt used correctly before or clever ways of using it are found and this last one is particularly important, for example sonic R in saturn use effects the machine really cant do but with clever ways of using the hardware and using part of the hardware you dedicate to it then it becomes possible




so yea you can use improved effects in DC games but there is a cost so it doesnt help to make exagerations about what is possible or disregard another game just because another game has this or that characteristic, its better to see what is available and how can be used