Support NeoGAF

[thicc_girls_are_teh_best]

It didn't though, not at all, DOA2 DC characters beat them all in polycounts and every Tekken after TTT had lower polygon models too (and also less painstacking detailing to stuff like the different specular settings to account for different materials etc.), as did the impressive VF4 ports.

And even so in DOA2 you can get 3 characters on screen in the tag mode, in TTT one first disappears. It did top Sc as it's a really modest Dreamcast game tech wise, it was their first (and last) big DC/next gen game after all and paved the way for TTT in terms of character modeling etc.

Well when I was saying "more detailed" I didn't just mean in terms of poly counts. But also things like the textures, the color depth, outfit designs and artistic choices. The DOA2 characters might have more polygons (just going off what you said), but Tekken 5's characters look noticeably better in spite of that. It's a combination of things.

Modelers simply got better at utilizing less polygons for better form (and not wasting polygons for unseen things like 3D teeth in TTT). VF3tb DC character polycounts are similar to the PS2's VF4 ports (and eclipse Sc's) yet the game looks far more crude, since it's an older 1996 pioneer.

That's not the only reason; the art style is also less sophisticated and realized than VF4's, and I think that accounts for more on the overall impact than some initially realize.

The western release just traded it for extra blur. People always love major widespread flaws in Sony systems but trash every imperfection elsewhere, from the polygon and texture warping and gaping geometry seams on PS1 to the blurry, aliased, low res & low quality video output of 2.

TBF, I don't excuse or pretend those flaws exist. PS1's texture warping is very infamous because they cut corners going with fixed integer math for the GTE instead of floats. But it's also really funny in hindsight to see how some of the biggest games that gen (any platform) did similar things for calculating polygons, geometry, collision etc.

For example, Super Mario 64.

Soulcalibur was impressive at the time but not an example for the age, it was a rushed port (only a few months dev, check the release dates of the arcade original...and the DC release.)

DOA 2 is a better example (1999 Naomi game):
Tekken 4 has lower character polycount than DOA2:

According to Beyond3D, DOA2 characters are almost 3X Soulcalibur...

Tekken 5 is pretty cool as i said but nothing breathtaking. Of course, backgrounds are more detailed than DOA2 since there are multiple layers in DOA2:
Reduce DOA's level size and you'll get more details as well...
Produce a new Dreamcast DOA with 2004 tools and you'll again have graphic improvements. The guy on youtube is already enhancing DOA2 with better textures, higher polycount, new lights, effects, normal mapping and he has a lot of new ideas

I'm not downplaying T5's graphics, i was just seeing great graphics since 1996 Scud Race... So, i found the T5 art pretty cool (waterfalls, Treasure stage and the following T6 on ps3 was pretty lackluster) but nothing mindblowing. It's good but it was...2005.

FWIW I'm sure you could squeeze more out of Tekken 5 on PS2 as well, there just seem to be less hobbyists interested in doing that type of stuff because the common POV is PS2 was a "well-realized" system in terms of games making strong use of its potential, whereas the Dreamcast isn't, since it died so early commercially.

So that motivates hobbyists to push one over the other well after the fact and I think sometimes people mistake that as meaning the underdog (DC in this case) not only had so much gas left in the tank, but other systems (like PS2) were just completely maxed out. I don't think that's necessarily true even for systems like PS2, they just probably have less headroom for improvements on games by hobbyists than a Dreamcast does.

I think there’s some revisionist history going on…I also always had a top end PC my whole life but when the Dreamcast launched, it was a leap. I was excited to show it off because people were slack jawed when they saw it. My friend’s mom thought NFL2K was a live football game at first glance. I loved me some PC gaming, but nothing in 99 touched what Dreamcast was putting out.

Yeah, I don't think anything on PC was touching Dreamcast in '98 or even '99. There was nothing on the market in terms of combined visuals, tech, and scope that was as much a total package as Sonic Adventure was, for example. Especially considering that game was only two years after Super Mario 64, and released the same year as OoT and Banjo-Kazooie.

Also people gotta keep in mind Shenmue dropped in (very late) '99, and that was also in a league of its own vs. basically every other game in the market at the time. Including on PC.

The first screenshots from the DC blew me away when I saw them in a magazine back in the day. Virtua Fighter 3 seemed like HD for me. Played Crazy Taxi to death later when I got the DC.

To answer the question about power: Could Resident Evil 4 be ported?

Probably. But with less geometry and worst lighting than the PS2 version, and lower particle effects. Cleaner output image and potentially sharper textures for certain background objects, though.

Both it and the PS2 version would still be noticeably worst than the GameCube one though, visually. At times the GameCube version seemed a generation ahead of PS2's, but a lot of that may've also been due to the PS2 version being a bit rushed. For example Silent Hill 3 has much better textures than PS2 version of RE4, so better textures were possible.

 

[TGO]

Mid gen refresh? I can't understand this. Dreamcast games looked amazing compared to the prior generation and only got better as it went on. A much much bigger upgrade than what I'm seeing with the current gen today relative to it's last gen. The consoles of that generation also came out years apart unlike the the current gen and the last one. GC and Xbox came out 3/4 years after the DC but are part of the same generation.

I think the DC never got a chance to show what it could do and it had a lot of ports from the previous gen which probably why I see it like that
And while the arcade ports were fantastic and beyond the 32-Bit era, I don't think they were on par with what was to come.
But like I said, the DC didn't get a chance to truly shine.

 

[Geometric-Crusher]

I know it's hard work but could someone post screenshots of the real console between DOA 2 DC, Tekken 4 and VF4 at similar stages.

Thanks

 

[Geometric-Crusher]

To answer the question about power: Could Resident Evil 4 be ported?

I have an almost identical question.
Could Perfect Dark on N64, Zelda Majora's Mask, King Sol's Mines stage in Indiana Jones run without the expansion pack?

 

[Esppiral]

Dead or Alive 2 character models have more polys than VF4 (ps2), Soul Calibur 2 or any Tekken Game on the Ps2, the Game IS truly underrated and I Will never understan how Soul Calibur got the crown for the BEST graphics in the system with 1/3 of the polycount for characters and super simple 3D backgrounds... Also, Dead or Alive has ALWAYS 2 stages loaded in ram at any given time (wich restrics the scope of the levels) and It streams data while playing unlike other fighters than load between fights doa 2 IS doing much more work in the background,

Not to mention the cloth physics, some characters have up to 8 individually physically based elements between, cloth, hair, ribbons chains etc, no other Game did that that gen except for the secuels to Doa2 and that accounts for processing power, fuck doa characters even have modeled 3D Nails....

 

[Fafalada]

I have an almost identical question.
Could Perfect Dark on N64, Zelda Majora's Mask, King Sol's Mines stage in Indiana Jones run without the expansion pack?

When asking "could it be ported" the answer is almost always yes.

If you doubt it, remember that psp received a port of DOAX2.
And that wasn't the only ps360 port, there were many others.

 

[Fat Frog]

thicc_girls_are_teh_best
We won't agree but fair enough for Tekken 5 etc .

But, price aside (a good argument for the Dreamcast) and even if graphics would have been a real problem compared to a few big boys from Sony, Namco, Konami,
Dreamcast would have had a help from the Skies for the following years:

HD TV and PC monitors...

On a HD TV, Dreamcast has a beautiful image, PS2 a crappy one.

Dreamcast IQ is time proof. It's reversed for the PS2

 

[cireza]

On a HD TV, Dreamcast has a beautiful image, PS2 a crappy one.

Yeah, SEGA was way ahead of the competition as usual. Picture quality is very good and even the Wii (released 8 years later) has a worse picture on HD TV...

 

[Esppiral]

To put things in to perspective...

Julia Tekken 5 (ps2) 4.165 polys Kasumi costume 5 Dead or alive 2 (Dreamcast) 9.374 polys.

Let's see where those extra polys went....

Eyeballs, Tekken 112 polys, Doa 2 316 polys.

Ears are fully modelled in Doa 2 so...

Tekken 62 polys, Doa 2 512 polys...

Full face

Tekken 1472 polys, Doa2 2526 polys...

Both Hands

Tekken 580 polys Doa2 1430 polys, heck even the nails are modelled in DOA 2.....

Torso ( I was expecting bigger numbers on the DOA 2 model tbh xD)

Tekken 5 - 536 polys Doa 2 - 876 polys.

Lowest polytcount model I've found in Dead or Alive 2 is 7k+ for Genfu, Lowest for Tekken 5 is 4k+ for Kazuya...

Extra gift, the Wind Mill stage from Soul Calibut 2 running on the Dreamcast, old wip, lacking collissions and super crappy capture, it wasn't meant to be shown publicly but after reading so many stupid things about the DC capabilities here it is.



Pretty good for a "Nintendo 64 Pro" or psx gen mid gen refresh....

 

[digital_ghost]

My first console was a Genesis, then dreamcast. I was blown away by the graphics. I remember when I bought home shenmue and was floored. Same with sonic adventure, and evolution. Had so many good memories with that system.

My next system was an Xbox and I got to play jet set radio 2 and shenmue 2. Really felt like a dreamcast 2.

 

[Alexios]

awesome stuff

Would it be possible for someone to just hotswap VF3tb models with PS2 VF4 models since they're similar in polycounts (choose the simplest costumes too, whatever) or are the assets so completely different it'd be way too much work to bother? I'd sure play more VF3 like that if it was possible, lol.

 

[64bitmodels]

Dreamcast wasn’t that impressive - compared to Half Life

Half Life? The game that looks no better than an N64 game most of the time?

 

[Dr. Wilkinson]

It was more powerful than anything that was currently available. But obviously a day late and a dollar short compared to what would come with PS2 in 2000 and certainly what would come in 2001 with GameCube and Xbox.

 

[PaintTinJr]

To put things in to perspective...

Julia Tekken 5 (ps2) 4.165 polys Kasumi costume 5 Dead or alive 2 (Dreamcast) 9.374 polys.

Let's see where those extra polys went....

Eyeballs, Tekken 112 polys, Doa 2 316 polys.

Ears are fully modelled in Doa 2 so...

Tekken 62 polys, Doa 2 512 polys...

Full face

Tekken 1472 polys, Doa2 2526 polys...

Both Hands

Tekken 580 polys Doa2 1430 polys, heck even the nails are modelled in DOA 2.....

Torso ( I was expecting bigger numbers on the DOA 2 model tbh xD)

Tekken 5 - 536 polys Doa 2 - 876 polys.

Lowest polytcount model I've found in Dead or Alive 2 is 7k+ for Genfu, Lowest for Tekken 5 is 4k+ for Kazuya...

Extra gift, the Wind Mill stage from Soul Calibut 2 running on the Dreamcast, old wip, lacking collissions and super crappy capture, it wasn't meant to be shown publicly but after reading so many stupid things about the DC capabilities here it is.



Pretty good for a "Nintendo 64 Pro" or psx gen mid gen refresh....

That's really great info, but IMO it doesn't prove the point because what LoD level in game is being used could easily invert those numbers, and even the on screen geometry size impacts the culling of non-visible polygons - the last comparative imagines of geometry rendered in the games' typical side on render would have the second model self occlude more of its geometry - rendering less - than the elongated left side imagine even if LoDs weren't used on either game.

AFAIK the reason why the Dreamcast wasn't very powerful, and was the same flaw with the OG Xbox, was that at a time of Zbuffering performance and precision being the technological leap, the Dreamcast was 4bit/8bit depth IIRC from what I read on here a few years back, so still relying heavily on the CPU for finer hidden surface removal than the PS2 or Gamecube, and the PS2 was a full 32bit Zbuffer, and the Gamecube like ATI GPUs of the time was super fast at16bit, and normal speed at 24bit - and the OG Xbox getting it all wrong and doing a W-buffer to cheat on its inability to have adequate bandwidth to compete with a quality zbuffer.

Fighting games don't really stress the zbuffer meaningfully, and only stress the fill-rate heavily when lots of graphics fx are used, which is probably why the tech specs of the consoles quoted flat polygon and full lit shaded polygon rates separately, which IIRC the Dreamcast fighting games lacked shading/fx compared to the PS2/GC versions, especially when you consider Dural in VF4.

You also mentioned the cloth physics in DoA2, but AFAIK that was all CPU based at the time, unlike some of the PS2 stuff in open world games such as SotC, which I believe was simulated on the Emotion Engine.

 

[Fat Frog]

Dreamcast fighting games lacked shading/fx compared to the PS2/GC versions, especially when you consider Dural in VF4.

Yeah, we know that PS2 has better SFX...

Nonetheless, these SFX goodness are temporary during our gameplay sessions.
(MGS2's rain was great but indoor graphics were nothing special...)

On the other hand, Dreamcast IQ is a permanent benefit (unlike permanent PS2 alliasing bugging our eyes).


That's why the new generation (millenials and zoomers) cannot decide which one has the more pleasant graphics :

Dreamcast vs PS2 graphics

OK WAIT, I KNOW PS2 WAS MORE POWERFUL, I KNOW Before you get to that My question is more specific What did you prefer in this sense. Dreamcast and its ultra sharp, clean but low poly/low geom/lower spec graphics or PS2 with flickery/blurry but higher poly/better "technically" graphics? I...

www.resetera.com

50/50

It's not my fault if the young gamers love Dreamcast crisp image .

T-i-m-e p-r-o -o-f it is

 

[PaintTinJr]

Yeah, we know that PS2 has better SFX...

Nonetheless, these SFX goodness are temporary during our gameplay sessions.
(MGS2's rain was great but indoor graphics were nothing special...)

On the other hand, Dreamcast IQ is a permanent benefit (unlike permanent PS2 alliasing bugging our eyes).


That's why the new generation (millenials and zoomers) cannot decide which one has the more pleasant graphics :

Dreamcast vs PS2 graphics

OK WAIT, I KNOW PS2 WAS MORE POWERFUL, I KNOW Before you get to that My question is more specific What did you prefer in this sense. Dreamcast and its ultra sharp, clean but low poly/low geom/lower spec graphics or PS2 with flickery/blurry but higher poly/better "technically" graphics? I...

www.resetera.com

50/50

It's not my fault if the young gamers love Dreamcast crisp image .

T-i-m-e p-r-o -o-f it is

I didn't click the purple link, but that will be the same idiotic situation of people preferring TV store mode crushed blacks with no fog on Xbox 360 vs Ps3 games in head to heads. In a technical discussion a baseline of viewers being knowledgeable about what they are discussing is surely a requirement to discuss in good faith, no?

 

[ManaByte]

Pretty good for a "Nintendo 64 Pro" or psx gen mid gen refresh....

Anyone calling it that is:

1. Trolling
2. Never owned or played a DC

 

[Esppiral]

That's really great info, but IMO it doesn't prove the point because what LoD level in game is being used could easily invert those numbers, and even the on screen geometry size impacts the culling of non-visible polygons - the last comparative imagines of geometry rendered in the games' typical side on render would have the second model self occlude more of its geometry - rendering less - than the elongated left side imagine even if LoDs weren't used on either game.

AFAIK the reason why the Dreamcast wasn't very powerful, and was the same flaw with the OG Xbox, was that at a time of Zbuffering performance and precision being the technological leap, the Dreamcast was 4bit/8bit depth IIRC from what I read on here a few years back, so still relying heavily on the CPU for finer hidden surface removal than the PS2 or Gamecube, and the PS2 was a full 32bit Zbuffer, and the Gamecube like ATI GPUs of the time was super fast at16bit, and normal speed at 24bit - and the OG Xbox getting it all wrong and doing a W-buffer to cheat on its inability to have adequate bandwidth to compete with a quality zbuffer.

Fighting games don't really stress the zbuffer meaningfully, and only stress the fill-rate heavily when lots of graphics fx are used, which is probably why the tech specs of the consoles quoted flat polygon and full lit shaded polygon rates separately, which IIRC the Dreamcast fighting games lacked shading/fx compared to the PS2/GC versions, especially when you consider Dural in VF4.

You also mentioned the cloth physics in DoA2, but AFAIK that was all CPU based at the time, unlike some of the PS2 stuff in open world games such as SotC, which I believe was simulated on the Emotion Engine.

Doa 2 does not use LODs I canguarantee you that.

Also The DC didn't have co processors or a t&l unit to offload work like the Ps2, so everything being calculated on the CPU is even more impressive

 

[Geometric-Crusher]

When asking "could it be ported" the answer is almost always yes.

If you doubt it, remember that psp received a port of DOAX2.
And that wasn't the only ps360 port, there were many others.

I understand that ''ported'' means keeping the core of the game. RE4 is not possible on Dreamcast except maybe ''reimagined''

 

[Fat Frog]

I didn't click the purple link, but that will be the same idiotic situation of people preferring TV store mode crushed blacks with no fog on Xbox 360 vs Ps3 games in head to heads. In a technical discussion a baseline of viewers being knowledgeable about what they are discussing is surely a requirement to discuss in good faith, no?

It's not about knowledge and technical discussion, it's just about graphic preferences:

Some of them preferred the more advanced polycount, SFX on the PS2, some others the crisp IQ of the Dreamcast.

It's just pleasantely surprising to see how popular Dreamcast graphics are through time... (it's even more impressive since most of them were developped with 90's tools !)

 

[Geometric-Crusher]

To put things in to perspective...

Julia Tekken 5 (ps2) 4.165 polys Kasumi costume 5 Dead or alive 2 (Dreamcast) 9.374 polys.

can you look Fight Night 2 and Ultraman PS2 ?

 

[MadchesterManc]

To put things in to perspective...

Julia Tekken 5 (ps2) 4.165 polys Kasumi costume 5 Dead or alive 2 (Dreamcast) 9.374 polys.

Let's see where those extra polys went....

Eyeballs, Tekken 112 polys, Doa 2 316 polys.

Ears are fully modelled in Doa 2 so...

Tekken 62 polys, Doa 2 512 polys...

Full face

Tekken 1472 polys, Doa2 2526 polys...

Both Hands

Tekken 580 polys Doa2 1430 polys, heck even the nails are modelled in DOA 2.....

Torso ( I was expecting bigger numbers on the DOA 2 model tbh xD)

Tekken 5 - 536 polys Doa 2 - 876 polys.

Lowest polytcount model I've found in Dead or Alive 2 is 7k+ for Genfu, Lowest for Tekken 5 is 4k+ for Kazuya...

Extra gift, the Wind Mill stage from Soul Calibut 2 running on the Dreamcast, old wip, lacking collissions and super crappy capture, it wasn't meant to be shown publicly but after reading so many stupid things about the DC capabilities here it is.



Pretty good for a "Nintendo 64 Pro" or psx gen mid gen refresh....

It's surprising the Tekken devs manage to get most of the way there with less than half the poly count. It's probably just a case of different priorities given the characters are only part of the scene, be interesting to see a similar comparison with stage poly counts between the two games. Played DOA2 Hardcore a few months back on PS2 and the stages on that are looking a bit dated these days, especially when compared to Tekken 4/5 on the same system

 

[Geometric-Crusher]

I've never understood why there's so much excitement about DOA2 in the Dreamcast version, I mean it's amazing what they achieved on the Dreamcast.

characters ~10,000 stage ~18,000 (most are 10,000). By forcing the engine in tag mode it is possible to do 47,000 for a short while, 640x480.

For comparison, Tekken 3 has 2,000 in the entire scene and Soul Blade has something like 3,000, the ps1 limit, N64 in Mace the Dark Age makes up to 7,000 but when this happens the game becomes power point. Overall it's 3,000 at 30fps.

 

[Esppiral]

I understand that ''ported'' means keeping the core of the game. RE4 is not possible on Dreamcast except maybe ''reimagined''

That is exactly what they did with the ps2 version... The Game has like 3x times less polygons on ps2, unlike popular belief It wasn't a bad Port they customized the whole Game to fit on the Ps2...

 

[Fafalada]

I understand that ''ported'' means keeping the core of the game. RE4 is not possible on Dreamcast except maybe ''reimagined''

And I'd disagree - that's why I pointed to 360->PSP ports (which mostly all retained the core of the game), a wider gulf by large margin than even XBox -> DC.
This isn't to say you're getting 1:1 perfect conversions - but for Internet to call it 'perfect' you need cycle-level accurate emulation, no port ever lived up to that.

Here's a few more contentious ports that are possible for people to piss about:
Doom3 -> PS2/GC
Halo -> DC
GT3 -> DC
PG2 -> PS2

characters ~10,000 stage ~18,000 (most are 10,000). By forcing the engine in tag mode it is possible to do 47,000 for a short while, 640x480.

Worth noting this puts it at 3M polys/s which was a very average cross-platform number for PS2 generation (decidedly not high end in any way, and yes most crossplats on GC/XBox games ran around the same - including those with bad framerate).

 

[Fat Frog]

That is exactly what they did with the ps2 version... The Game has like 3x times less polygons on ps2, unlike popular belief It wasn't a bad Port they customized the whole Game to fit on the Ps2...

It also seems that cutscenes are prerendered on PS2.

It's okay to have downgraded versions ...

 

[PaintTinJr]

Doa 2 does not use LODs I canguarantee you that.

Also The DC didn't have co processors or a t&l unit to offload work like the Ps2, so everything being calculated on the CPU is even more impressive

You'll need to provide me something technical to show the actual rendered front facing polygon count per second for each game to convince me of that claim, because the onscreen fidelity in comparison to its peers is lower in what I'm seeing rendered.

In your first image comparing the full models in blender, the left side image has more undulation, so the polygon model counts aren't proportional to the displayed geometric fidelity in the overall image comparison you've chosen .

I suspect their is a lot of unseen polygons under the large polygon scarf and large occluding arm bands in DoA2, meaning that geometry doesn't ever pass culling, and the ears and eyes that aren't both seen at most times in standard gameplay will also be completely culled in both games, meaning the discrepancy for those parts is actually nothing. Personally I think there's something wrong with the numbers given the polygon areas are pretty similar in the main body parts, but without access to the models myself, to check, I couldn't say yet what the issue could be, if there is one.

As for the co-processor issue, it isn't more impressive, because the cloth doesn't have any gameplay impact, it is just eyecandy, and eyecandy that gain provides large polygon occluders to pre fill a stencil/depth buffer and cull more smaller polygons from reaching the rasterization stage. Doing more stuff on a strong CPU was more one foot in the old 2D console rendering paradigm, and at the time even a decent PC with MMX and multi-core struggled to do justice to the work the EmotionEngine was doing for physics and animation of many characters simultaneously in Pro Evolution Soccer games to make it fun to play on a PC, which was clearly way beyond what a DC could do on its CPU, if decent PCs were struggling.

 

[PaintTinJr]

It's not about knowledge and technical discussion, it's just about graphic preferences:

Some of them preferred the more advanced polycount, SFX on the PS2, some others the crisp IQ of the Dreamcast.

It's just pleasantely surprising to see how popular Dreamcast graphics are through time... (it's even more impressive since most of them were developped with 90's tools !)

Is it though? it is normal observed human perception, and I put in the TV showroom mode comment because that is the human psychology that was adopted by Sega when choosing their games' colour palettes for arcade games that needed to draw you in and put real coin per minute in the slot. Why would you expect anything different?

 

[Geometric-Crusher]

Worth noting this puts it at 3M polys/s.

Dreamcast really can do 3M polys/s at peak (or Tech Demo ) but I don't think that 1 sec in tag mode can be considered a 3M polys/s game, I mean normal numbers in game are 30 ~ 37 which puts DOA2 at 1,8M to 2,3M which is incredible because the nature of DC is to operate at 1M.
1,5M polys/s makes a game extremely advanced on the platform.
Dead or Alive 2 (Dreamcast) is the real world game closest to its limit, not only pushing polygons but the technical set of the work.

 

[Panajev2001a]

Also The DC didn't have co processors or a t&l unit to offload work like the Ps2, so everything being calculated on the CPU is even more impressive

The Dreamcast had one vector unit in the CPU and PS2 had two of them and they were strong sure… I doubt launch a lot of PS2 software made really good use of the VUs actually…

 

[Esppiral]

You'll need to provide me something technical to show the actual rendered front facing polygon count per second for each game to convince me of that claim, because the onscreen fidelity in comparison to its peers is lower in what I'm seeing rendered.

In your first image comparing the full models in blender, the left side image has more undulation, so the polygon model counts aren't proportional to the displayed geometric fidelity in the overall image comparison you've chosen .

I suspect their is a lot of unseen polygons under the large polygon scarf and large occluding arm bands in DoA2, meaning that geometry doesn't ever pass culling, and the ears and eyes that aren't both seen at most times in standard gameplay will also be completely culled in both games, meaning the discrepancy for those parts is actually nothing. Personally I think there's something wrong with the numbers given the polygon areas are pretty similar in the main body parts, but without access to the models myself, to check, I couldn't say yet what the issue could be, if there is one.

As for the co-processor issue, it isn't more impressive, because the cloth doesn't have any gameplay impact, it is just eyecandy, and eyecandy that gain provides large polygon occluders to pre fill a stencil/depth buffer and cull more smaller polygons from reaching the rasterization stage. Doing more stuff on a strong CPU was more one foot in the old 2D console rendering paradigm, and at the time even a decent PC with MMX and multi-core struggled to do justice to the work the EmotionEngine was doing for physics and animation of many characters simultaneously in Pro Evolution Soccer games to make it fun to play on a PC, which was clearly way beyond what a DC could do on its CPU, if decent PCs were struggling.

Are you talking about front face culling or LODs... Anyways if you don't want to believe even with facts the discusión ends here.

 

[PaintTinJr]

Are you talking about front face culling or LODs... Anyways if you don't want to believe even with facts the discusión ends here.

Both. The back facing culled polygons, or occluded polygons I've explained how they don't get rendered, meaning the model is correctly built, but isn't reflective of what real geometry will get rendered per frame, compared to the left hand models.

As for LoDs, you can't just give it, "trust me bro!", especially at a time when aggressive LoD transitions were used in most games because polygon rendering was still in small digits and any savings were huge.

 

[Esppiral]

Both. The back facing culled polygons, or occluded polygons I've explained how they don't get rendered, meaning the model is correctly built, but isn't reflective of what real geometry will get rendered per frame, compared to the left hand models.

As for LoDs, you can't just give it, "trust me bro!", especially at a time when aggressive LoD transitions were used in most games because polygon rendering was still in small digits and any savings were huge.

Ok the Dreamcast is a pile a trash, next question?

Jesus, everything you say also apply to the Tekken models, yet you only seem to care about the Doa 2 models...

Oh by the way I know the game does not use Lod's because

1. I have eyes.
2. The models can be ripped with Naomilb from the game files and THERE ARE NO LODS
3. You can also rip the models using Ninja Ripper and will always get the same polycount no matter what because.. the game does not uses LODS.

When you Rip the models with Ninja Ripper the aren't culled BTW, they give the same polycount as when you extract them directly from the game files.

Do you have the same questions about Tekken? I can answer them too, or you don't care?

 

[PaintTinJr]

Ok the Dreamcast is a pile a trash, next question?

Jesus, everything you say also apply to the Tekken models, yet you only seem to care about the Doa 2 models...

Oh by the way I know the game does not use Lod's because

1. I have eyes.
2. The models can be ripped with Naomilb from the game files and THERE ARE NO LODS
3. You can also rip the models using Ninja Ripper and will always get the same polycount no matter what because.. the game does not uses LODS.

When you Rip the models with Ninja Ripper the aren't culled BTW, they give the same polycount as when you extract them directly from the game files.

Do you have the same questions about Tekken? I can answer them too, or you don't care?

I already knew the models you were using didn't have LoDs, but many games for memory reasons would calculate them on the fly, or pre-calculate temporarily in a scratch memory. If you've built the model sections in a regular way, you can programmatically generate lower LoDs. The only way to check for sure would be using the emulators with debug shaders - and without tessellation enabled - to get them to report actual polygon numbers that were passed to the fragment shader per frame.

 

[Geometric-Crusher]

That is exactly what they did with the ps2 version... The Game has like 3x times less polygons on ps2, unlike popular belief It wasn't a bad Port they customized the whole Game to fit on the Ps2...

there is a difference between customized and reimagined by reimagined I mean games like

F1 Racing Championship ubisof DC,PS2
4x4 evo 2 xbox, ps2
Space Station Silicon Valley ps1, n64

Same name but different games. RE4 (ps2 version) there is nothing lacking in terms of gameplay. You seem to suggest that the GC can push more polygons than the Playstation 2 and that the difference between the versions was due to limitations in the PS2 and not due to the devs' choices.

 

[Esppiral]

there is a difference between customized and reimagined by reimagined I mean games like

F1 Racing Championship ubisof DC,PS2
4x4 evo 2 xbox, ps2
Space Station Silicon Valley ps1, n64

Same name but different games. RE4 (ps2 version) there is nothing lacking in terms of gameplay. You seem to suggest that the GC can push more polygons than the Playstation 2 and that the difference between the versions was due to limitations in the PS2 and not due to the devs' choices.

It is exactly what I am implying lol the ps2 limitations made them redo all the assets do you think they did It by pleasure or what?

 

[Fat Frog]

It is exactly what I am implying lol the ps2 limitations made them redo all the assets do you think they did It by pleasure or what?

THE PLAYSTATION 2 HAS UNLIMITED POWER.

HOW DARE YOU !

 

[PaintTinJr]

It is exactly what I am implying lol the ps2 limitations made them redo all the assets do you think they did It by pleasure or what?

The PS2 used the RenderWare middleware engine for many games because of the difficulty in programming for the system, not because it lacked performance, and the middleware itself didn't lend itself to all game types because of the memory situation with the PS2, but it was the cheapest to license fully capable engine AFAIK, and that was the reason EA bought the company between the PS3 reveal and launch IIRC.

I'm pretty sure that game is one that uses RenderWare. As similar example is Super Monkeyball deluxe by Tt Games, which as a 3 way faceoff is worst on Xbox, because despite the 800x600 resolution has no depth cueing fog (because of the W-buffer) giving a Ulysses 31 cartoon optical illusion puzzle in Monkey Target 1 & 2 where it is impossible to rationalize floor from sky, or whether something is in front or behind when of similar size on screen.

On PS2 it has the roughest image precision because of the custom primitive assembly software to render polygons, but it has the best draw distance for Monkey Target, best fog, same water shaders and textures as the original gamecube versions of 1 and 2, and only slightly lower polygon counts than the other two, and despite the woeful PS2 pad deadzone configuration it was still relatively faithful to the cube versions in 4 player split, and unlike the outsourced Tt Games of deluxe, the GC versions of 1 & 2 are AM2, and spectacular in every way. But had the PS2 version been by AM2, it would have superior or equal to the cube versions.

Comparing custom developed Metal Gear Solid: Twin snakes on Cube - probably using unbadged Unreal Engine- and Snake Eater on PS2 shows how much more advanced the PS2 hardware was when fully used, in much the same way that Winning Elven 6 Final Evolution(japan name for PES) on GC - probably using unbadged Unreal Engine - is brilliant in every way possible, except it isn't technically as good as the PS2 versions because the hardware was less powerful.

So the end result being that custom solutions on each hardware produced results closer to their technical capability, and the PS2 was very capable versus the gamecube that shared a lack of pack-in HDD and limited RAM.

 

[TNT Sheep]

On the PS2 it has the roughest image precision because of the custom primitive assembly software to render polygons, but it has the best draw distance for Monkey Target, best fog, same water shaders and textures as the original gamecube versions of 1 and 2, and only slightly lower polygon counts than the other two, and despite the woeful PS2 pad deadzone configuration it was still relatively faithful to the cube versions in 4 player split, and unlike the outsourced Tt Games of deluxe, the GC versions of 1 & 2 are AM2, and spectacular in every way. But had the PS2 version been by AM2, it would have superior or equal to the cube versions.

Doesn't the PS2 version of Super Monkeyball Deluxe run at half the framerate? For that reason alone I would not rank it higher than the Xbox version.

 

[Team Andromeda]

I don't think most Model 3 games were that profitable in general, because most operators just stuck with the Model 2. And once the NAOMI was ready, those who would've upgraded to Model 3 setups just skipped it altogether and went to the NAOMI because it was cheaper.

In a lot of ways the NAOMI is to Model 3 what the Wii was to GameCube: a repackaging of prior (more expensive) tech becoming cheaper thanks to Moore's Law and component prices dropping over time. Though NAOMI had more extensive changes to it than the Wii did, in terms of underlying tech & specifications.

It was basically a very different architecture aiming for general throughput performance of the Model 3.

I seem to remember the AM teams posted a loss for the 1st time in their history (at that time) due to the low profit margins on Model 3 games. Model 3 was a great board but it was just too expensive for a lot of Arcades to make a decent profit from. Sega Rally 2 did well and it was so many Arcades but Daytona USA 2 wasn't such a hit and I bet that helped with SEGA's thinking not to port the game to the DC.

 

[Fat Frog]

Doesn't the PS2 version of Super Monkeyball Deluxe run at half the framerate?

Correct !

So much technical blablabla to in the end "forget" the essential flaw:

Half the framerate for the PS2 port.
How convenient !

THE PLAYSTATION 2 IS THE MOST POWERFUL SYSTEM EVER MADE !

 

[PaintTinJr]

Doesn't the PS2 version of Super Monkeyball Deluxe run at half the framerate? For that reason alone I would not rank it higher than the Xbox version.

That's such as disingenuous comment to make, when in 4 player split monkey target 1 and 2 - like me and my friends have played since the Japanese launch of Cube - the frame-rate is clearly equal, and that the Xbox versions aren't remotely playable and would definitely cause seizures by its broken depth perception, make it a non-entity for comparison. Even at the xbox's higher res, it represents less than 1/3 of the fillrate demands of the other two versions, because fog crippled frame-rates back in the day, as did depth buffering at FP32, as it is on PS2. If matched on the Xbox, and even the capable Cube the depth would have crippled frame-rates too -, both of which had a smaller FOV because of the draw distance - and then because the Xbox couldn't do the water shaders either, which would have eaten even more fillrate, the comparison with the Xbox version really is a non-starter.

 

[polybius80]

Dead or Alive 2 character models have more polys than VF4 (ps2), Soul Calibur 2 or any Tekken Game on the Ps2, the Game IS truly underrated and I Will never understan how Soul Calibur got the crown for the BEST graphics in the system with 1/3 of the polycount for characters and super simple 3D backgrounds... Also, Dead or Alive has ALWAYS 2 stages loaded in ram at any given time (wich restrics the scope of the levels) and It streams data while playing unlike other fighters than load between fights doa 2 IS doing much more work in the background,

Not to mention the cloth physics, some characters have up to 8 individually physically based elements between, cloth, hair, ribbons chains etc, no other Game did that that gen except for the secuels to Doa2 and that accounts for processing power, fuck doa characters even have modeled 3D Nails....

I dont think DoA2 is an underrated game but people have preferences and SC is considered the better game

polycount is not the only metric in a game from the era, you cannot simply take the polycounts of a character compare to other game's character and declare what have better graphics just by that, if we compare master chief model's polycount to other games character then halo is a bad looking game?, doa looks better than halo? or halo looks better than halo2 since MC have more triangle in the first game?, does baldurs gate is above all other games because the sexy elvens in the tavern have more triangles in one of their tits than entire 3d character models from the past gen or you have a dragon boss with 20k triangles with each scale modeled? to say something looks better it is a mixture of perception and effects and things involved and some things have more impact in some people, since you mentioned VF4 that game uses for example a deformation of the ground to simulate snow and sand behaviour, if people see a fight in DoA2 in a scene with snow vs a VF4 fight in a scene with snow, VF4 looks better not only the deformations but the light and shadows and how dynamic the scene looks like the one with lots of people on it, you cannot blame people if they see that and think it looks better, each effects have its cost too so even technical savy people can think one is better than the other because they recognise the difficulty of the effects involved for example its more impresive to have triangles deformed from their positions than a huge amount of static triangles so a game with more particles is better looking game for some people just because of that

 

[Esppiral]

I dont think DoA2 is an underrated game but people have preferences and SC is considered the better game

polycount is not the only metric in a game from the era, you cannot simply take the polycounts of a character compare to other game's character and declare what have better graphics just by that, if we compare master chief model's polycount to other games character then halo is a bad looking game?, doa looks better than halo? or halo looks better than halo2 since MC have more triangle in the first game?, does baldurs gate is above all other games because the sexy elvens in the tavern have more triangles in one of their tits than entire 3d character models from the past gen or you have a dragon boss with 20k triangles with each scale modeled? to say something looks better it is a mixture of perception and effects and things involved and some things have more impact in some people, since you mentioned VF4 that game uses for example a deformation of the ground to simulate snow and sand behaviour, if people see a fight in DoA2 in a scene with snow vs a VF4 fight in a scene with snow, VF4 looks better not only the deformations but the light and shadows and how dynamic the scene looks like the one with lots of people on it, you cannot blame people if they see that and think it looks better, each effects have its cost too so even technical savy people can think one is better than the other because they recognise the difficulty of the effects involved for example its more impresive to have triangles deformed from their positions than a huge amount of static triangles so a game with more particles is better looking game for some people just because of that

It is not just polycount Doa 2 is more adavnced and better looking than SC I never understood the praise SC gets, and don't get me wrong I LOVE It, but technically is meh.

 

[polybius80]

Oh by the way I know the game does not use Lod's because

1. I have eyes.
2. The models can be ripped with Naomilb from the game files and THERE ARE NO LODS
3. You can also rip the models using Ninja Ripper and will always get the same polycount no matter what because.. the game does not uses LODS.

When you Rip the models with Ninja Ripper the aren't culled BTW, they give the same polycount as when you extract them directly from the game files.

Do you have the same questions about Tekken? I can answer them too, or you don't care?

an average 3d fighting game doesnt use LoD in their characters so is not really that important if they do, but it is true there are problems that can arise when ripping 3d models and using them to compare betwen games like double tirnalges if they doesnt make backface culling, also that kind of programs like ninja ripper appear to take the models from the buffers and not necesarily what is visible so there are stuff that they capture that is not part of the frame for example try capturing from marvel ultimate alliance and you will get the whole map of the level instead of what is being displayed on screen, it is recomendable to use emulators to count polycounts and only use rippers to compare 3d models(after cleaning them)

 

[SomeGit]

That's such as disingenuous comment to make, when in 4 player split monkey target 1 and 2 - like me and my friends have played since the Japanese launch of Cube - the frame-rate is clearly equal, and that the Xbox versions aren't remotely playable and would definitely cause seizures by its broken depth perception, make it a non-entity for comparison. Even at the xbox's higher res, it represents less than 1/3 of the fillrate demands of the other two versions, because fog crippled frame-rates back in the day, as did depth buffering at FP32, as it is on PS2. If matched on the Xbox, and even the capable Cube the depth would have crippled frame-rates too -, both of which had a smaller FOV because of the draw distance - and then because the Xbox couldn't do the water shaders either, which would have eaten even more fillrate, the comparison with the Xbox version really is a non-starter.

The better looking version at twice the framerate is a non starter. Least delusional PS2 fanboy

 

[PaintTinJr]

The better looking version at twice the framerate is a non starter. Least delusional PS2 fanboy

That you can't actually play at all, whether 60fps single player or 30fps in split screen with friends. Believe me when I say the PS2 version of target 1&2 is infinitely better, because it is a faithful copy of the cube version with all "essential" features like fog, and would still be the case if comparing 1080p120 to 576p30

 

[SomeGit]

That you can't actually play at all, whether 60fps single player or 30fps in split screen with friends. Believe me when I say the PS2 version of target 1&2 is infinitely better, because it is a faithful copy of the cube version with all "essential" features like fog, and would still be the case if comparing 1080p120 to 576p30

The only thing that makes SMBD unplayable on Xbox is the dead zone issue, but that’s also on the PS2 version.

And the Xbox version has fog what the hell are you talking about? Can you show, with pictures, the essentials that are missing?

 

[Lysandros]

PS2 was a very complex system architecturally with plenty of calculating power, notoriously difficult to program for. It wasn't 'maxed out' in comparison to Dreamcast, far from it. Dreamcast was a remarkably efficient and easier to use design, its main advantage was image quality not graphic complexity. Things like physics, lighting, geometry transformation, transparencies etc. was not the machine's forte. And no, it would of course never 'catch' PS2 if it existed until the end of this generation, this is a pure fanboy fantasy. The gap in raw power was just too big. It would continue to produce graphically less advanced games with a cleaner looking image, that's about it. To answer the original question; yes for its time of release it was a very capable and well architected machine.

 

[Esppiral]

an average 3d fighting game doesnt use LoD in their characters so is not really that important if they do, but it is true there are problems that can arise when ripping 3d models and using them to compare betwen games like double tirnalges if they doesnt make backface culling, also that kind of programs like ninja ripper appear to take the models from the buffers and not necesarily what is visible so there are stuff that they capture that is not part of the frame for example try capturing from marvel ultimate alliance and you will get the whole map of the level instead of what is being displayed on screen, it is recomendable to use emulators to count polycounts and only use rippers to compare 3d models(after cleaning them)

The 3D Ripper mention was an example, as I said the models are decrypted from the Game files themselves.... So 1:1 what they are, also shouln't that apply to the Tekken models I posted too? Seems that the only thing bothering people IS the higher polycount on the models compared to Tekken....