PS2 vs Dreamcast Graphics

[Cloofoofoo]

So this afternoon I did a ton of IGN research on my tablet.

1. I will look at NHL 2K, it was the first 6th gen hockey game and has developer polygon statistics.

2. Even in early 1999, there was one developer that decided it was right to be realistic about the PS2: https://ca.ign.com/articles/1999/02/...n-2-cg-quality



Although I will still be sticking with Jason Rubin's 9 million PPS claim as the best reference.

3. I might do a NFL 2K1 vs Madden 2K1 comparison. We can finally settle this debate.

4.The Frogger series had a ton of sequels on PS2/GameCube/Xbox which also would've been on Dreamcast if it survived the North American PS2 launch. I will look at these.

9mpps does seem to be the most realistic as i havent found a that goes over 10mpps. On average games seem to run around 2.5 to 6 mpps. With quite a few like the zone of enders or gta 3 running basically below 1mpps most thr times.

Its gonna be super hard to compare seeing how nulldc gives extremely bloated numbers, twice to even three times the actual polygon count. I personally suspect madden probably does around 2x polygon count.

[Yharnamresident]

So after 10 months of figuring out how to use these emulators, I'll show you guys what I can create now.

That baseball game will be the first game in my new format, for whatever reason. Also this post will be around 350 MBs of data so it may take a while to load. This post took hours to create so I hope you guys like it.



PuroYakyuu Team Wo Tsukurou (1999, Dreamcast)


This was an early Dreamcast game created by Smilebit. It was released in early August 1999, so about the prime days of the N64. This game has a lot of charm, but a huge language barrier so I'm only looking at it from a graphics perspective.





One thing thats very noticable about this game is the fancy vertex shading everywhere.



Boss has something to say.





I absolutely love this balloon scene.





Gameplay, batting scene.



Different team. Notice the gloss map on the face mask.



Throwing it to home base. Some particle effects in the form of grass.



Ball in the air. Lots of fancy vertex shading.


So in terms of polygons this game isn't doing anything impressive. Even as a mid-1999 game. I feel they should've tried to increase the polycounts by around 33%.

Average: 600,000 polygons per second




Team select. 950,000 PPS(this should've been what the main game would run at).




Balloon scene. One of the highest spikes in the game at 1.2 million PPS.



A balloon and passenger is 2000 polygons.




600,000 PPS.



Batter is 1600 polygons.




600,000 PPS



Pitcher is 1600 polygons. So all high-LOD players are 1600 polygons.




So this game does use LODs. Even with tons of players on-screen, its only 450,000 PPS.



Low-LOD pitcher is 500 polygons.



Batter is same 500 polygons.


If anyone is gonna quote this, please snip a couple pictures.

[Yharnamresident]

9mpps does seem to be the most realistic as i havent found a that goes over 10mpps. On average games seem to run around 2.5 to 6 mpps. With quite a few like the zone of enders or gta 3 running basically below 1mpps most thr times.

Its gonna be super hard to compare seeing how nulldc gives extremely bloated numbers, twice to even three times the actual polygon count. I personally suspect madden probably does around 2x polygon count.

Ok you're going with 2x polygon count, I'm going with 1.5x polygon count with better lighting but interlaced while NFL 2K1 has better textures and progressive scan. SegaAMD will probably go with 3-4x polygon count.



So I do actually got a post planned for the PS2 20th anniversary. But right now its time for some bedtime.

[Cloofoofoo]

So after 10 months of figuring out how to use these emulators, I'll show you guys what I can create now.

That baseball game will be the first game in my new format, for whatever reason. Also this post will be around 350 MBs of data so it may take a while to load. This post took hours to create so I hope you guys like it.



PuroYakyuu Team Wo Tsukurou (1999, Dreamcast)


This was an early Dreamcast game created by Smilebit. It was released in early August 1999, so about the prime days of the N64. This game has a lot of charm, but a huge language barrier so I'm only looking at it from a graphics perspective.





One thing thats very noticable about this game is the fancy vertex shading everywhere.



Boss has something to say.





I absolutely love this balloon scene.





Gameplay, batting scene.



Different team. Notice the gloss map on the face mask.



Throwing it to home base. Some particle effects in the form of grass.



Ball in the air. Lots of fancy vertex shading.


So in terms of polygons this game isn't doing anything impressive. Even as a mid-1999 game. I feel they should've tried to increase the polycounts by around 33%.

Average: 600,000 polygons per second




Team select. 950,000 PPS(this should've been what the main game would run at).

Balloon scene. One of the highest spikes in the game at 1.2 million PPS.



A balloon and passenger is 2000 polygons.


600,000 PPS.



Batter is 1600 polygons.


600,000 PPS


Pitcher is 1600 polygons. So all high-LOD players are 1600 polygons.


So this game does use LODs. Even with tons of players on-screen, its only 450,000 PPS.



Low-LOD pitcher is 500 polygons.



Batter is same 500 polygons.


If anyone is gonna quote this, please snip a couple pictures.

The team select model is what used for the homerun derby for both pitcher and batter. Over all it runs much higher than the main game.

These figures from nulldc? If so that means the game is running 2 to 3 times less geometry than those numbers.

[Yharnamresident]

The team select model is what used for the homerun derby for both pitcher and batter. Over all it runs much higher than the main game.

If you know a way of telling me how to access that, I will re-download the ISO and look at it.

These figures from nulldc? If so that means the game is running 2 to 3 times less geometry than those numbers.

Naw dude, its fine. Just because Agartha might be giving you inaccurate numbers doesn't mean every game does.


So I will start working on my PS2 20th anniversary post.

[Cloofoofoo]

Speed Punks its the PSX absolute graphic peak and can pass for a initial DC game, put some texture filter here and there and the magic will work.



Gorgeous game, no N64 game can touch this graphics level!

If you know a way of telling me how to access that, I will re-download the ISO and look at it.



Naw dude, its fine. Just because Agartha might be giving you inaccurate numbers doesn't mean every game does.


So I will start working on my PS2 20th anniversary post.

It's all games though. You even proved it with soul reaver thinking raziel 3k when he's only 1.2k. All games we tested HD massively bloated polygon counts. I even game you an example of around \ over 3 games that had debug polygon counters and it still doubled it when extracting via nulldc. So it's not just agartha like you claim. It's broken on nulldc.

For the baseball game can't help you because I am not on my PC right now, all you have to do is go through menu, very easy to figure it out from there. I'll see mybe in a few I'll take a screenshot.

[Cloofoofoo]

If you know a way of telling me how to access that, I will re-download the ISO and look at it.



Naw dude, its fine. Just because Agartha might be giving you inaccurate numbers doesn't mean every game does.


So I will start working on my PS2 20th anniversary post.

Scroll through the options till you see the option that has a bat icon next to it. then you will pick a stadium and baseball derby will begin. The game is considerably higher batter and pitcher models.So they can obviously stress the engine further, makes me wonder if more players/ a.i. impacts performance heavily.



Speaking of high polygon counts on the ps2, not sure if iposted this in the past but f1 racing 2001 : 91,242 tris per frame x 60 fps = 5.4 mpps. Pretty high, air blade pushes more though. still finding games that push beyond 100K tris per frame gets hard even on the ps2.

[Team Andromeda]

Studio Liverpool made the PS2 sing with 2001 I was very impressed with it, shame the PS2 screen res made the game look a bit messy.

What's the Polygon output of Hundred Swords on the DC . I really loved the look of the game and was gutted than it wasn't brought to the west on the DC (Just onthe PC) Also it was one of those games that SEGA was going to use on its Net@ service linking SEGA Arcade, home and players at its Net@ centres all in one SEGA Japan was so forward thinking at the time, such a shame it never worked out

[SegaAMD]

still finding games that push beyond 100K tris per frame gets hard even on the ps2.

See : F1 2000, Dark Alliance, Spy Hunter, Crash Cortex, Need for speed Hotpursuit 2, Blood Rayne, NBA 2003. Tak power of juju, Metal Arms Glitch, Lord of the ring third age, prince of persia warrior within, crash team racing, x-men official game, Tomb Raider legend, oo7 From Russia love.

source: https://www.sega-16.com/forum/showth...raphics/page16

[Cloofoofoo]

See : F1 2000, Dark Alliance, Spy Hunter, Crash Cortex, Need for speed Hotpursuit 2, Blood Rayne, NBA 2003. Tak power of juju, Metal Arms Glitch, Lord of the ring third age, prince of persia warrior within, crash team racing, x-men official game, Tomb Raider legend, oo7 From Russia love.

source: https://www.sega-16.com/forum/showth...raphics/page16

some of those numbers seem iffy at best. know for a fact from my own test soul calibur 2 hovers around 33k tris per frame at 60 fps which around 2 million polygons per Second.

Further researching the devs for soul calibur 2 mentioned each character is 4k tris and the stages are 20k tris. Which in total is around 28k tris ler frame, slightly under 2 million polygons per Second, seems doable on the dc even. Seems pcsx2 + zero gs are very accurate.

[Yharnamresident]

It's all games though. You even proved it with soul reaver thinking raziel 3k when he's only 1.2k. All games we tested HD massively bloated polygon counts. I even game you an example of around \ over 3 games that had debug polygon counters and it still doubled it when extracting via nulldc. So it's not just agartha like you claim. It's broken on nulldc.

For the baseball game can't help you because I am not on my PC right now, all you have to do is go through menu, very easy to figure it out from there. I'll see mybe in a few I'll take a screenshot.

Let me post my Ready 2 Rumble Boxing: Round 2 comparison and you'll see what I mean.

Scroll through the options till you see the option that has a bat icon next to it. then you will pick a stadium and baseball derby will begin. The game is considerably higher batter and pitcher models.So they can obviously stress the engine further, makes me wonder if more players/ a.i. impacts performance heavily.

Appreciate it, it does look a lot higher poly.

Also yes you may be onto something, I noticed theres a scene with a ton of players on-screen so they dropped the frame rate to 30 FPS.

Speaking of high polygon counts on the ps2, not sure if iposted this in the past but f1 racing 2001 : 91,242 tris per frame x 60 fps = 5.4 mpps. Pretty high, air blade pushes more though. still finding games that push beyond 100K tris per frame gets hard even on the ps2.

Ratchet and Clank is one of them, I estimate the game is hovering around 3-7 million PPS.

See : F1 2000, Dark Alliance, Spy Hunter, Crash Cortex, Need for speed Hotpursuit 2, Blood Rayne, NBA 2003. Tak power of juju, Metal Arms Glitch, Lord of the ring third age, prince of persia warrior within, crash team racing, x-men official game, Tomb Raider legend, oo7 From Russia love.

source: https://www.sega-16.com/forum/showth...raphics/page16

They might be getting duplicate polycounts. I have to delete the duplicate polygons with every PS2 capture I take.


So my PS2 20th anniversary post is all done, if you guys wanna help me out by helping this thread reach the next page.

[SegaAMD]

They might be getting duplicate polycounts.

the games I mentioned I measured using Dolphin 4.0 and I found it very accurate, Sonic Adventure 2 is very close to the Dreamcast version and Burnout 2 is very close to the PS2 version. the others should also follow the same trend remembering if the PS2 does less poly it's because the gpu is doing other effects, since the PS2 version is superior in most of these multiplats.

Note that they are multiplat games if the GCN version was above 100k per frame in any of them, then the PS2 version will float in the same league.

[Yharnamresident]

the games I mentioned I measured using Dolphin 4.0 and I found it very accurate, Sonic Adventure 2 is very close to the Dreamcast version and Burnout 2 is very close to the PS2 version. the others should also follow the same trend remembering if the PS2 does less poly it's because the gpu is doing other effects, since the PS2 version is superior in most of these multiplats.

Note that they are multiplat games if the GCN version was above 100k per frame in any of them, then the PS2 version will float in the same league.

SegaAMD are you still around? We gotta get this thread to the next page so I can post the motherload.

Anyone could help me out. Why don't you answer this question, what was your personal experience with the various 6th gen consoles? Did you get a Dreamcast or PS2 in elementary or high school?

[Leynos]

I was in HS with Dreamcast. Also just posting to bump to the next page.

[Yharnamresident]

I was in HS with Dreamcast. Also just posting to bump to the next page.

Nice work, person thats not Segata anymore.


Alright everyone, hold still...