• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Between the Dreamcast, GameCube, and Xbox, which console do you think was the best?

Between the Dreamcast, GameCube, and Xbox, which console do you think was the best?

  • Dreamcast

  • GameCube

  • Xbox


Results are only viewable after voting.

Gambit2483

Member
I think most gamers consider that the Xbox got the best of the Sega love post DC. Many of the Sega games on the Xbox had Dolby Digital 5.1 Surround Sound, 480p, 16:9 and a few games like Sega GT Online displayed in 16:9/720p.

I'll give the GameCube the Sonic Adventure games and Skies of Arcadia Legends, as they were great.

Panzer Dragoon Orta, GunValkyrie, Sega GT 2002, Sega GT Online, enhanced Shenmue 2 (came with Shenmue The Movie), Jet Set Radio Future, Otogi I&II, Outrun 2 and Outrun 2006 were all amazing.

The Xbox launched with two DC sequels, Project Gotham Racing and Dead or Alive 3. Enhanced versions of Dead or Alive (Saturn) and DOA2 (DC) were also released as Dead or Alive Ultimate and included online fighting. Project Gotham Racing 2 is arguably the best racing game of the gen.

The XB port of PSO EP. I&II was superior to the DC and GC as well, as it added voice chat and Dolby Digital 5.1 surround sound.

The House of the Dead 3 on the Xbox also included The House of the Dead 2 from the DC.

The port of Crazy Taxi to the GameCube and PS2 was just a strait port done by Acclaim. Crazy Taxi 3 on the Xbox was done by Hitmaker/Sega, included enhanced versions of Crazy Taxi and the Small Apple from Crazy Taxi 2, along with the new Las Vegas map from Crazy Taxi 3.


And if I has to choose between the Gamecube, it's exclusives, and DC ports or the Xbox, it's exclusives and it's DC ports, I'd still easily pick the Gamecube.

Outside of Dead or Alive 2 and (maybe) House of the Dead 2, every other DC port I'd want is either also on Gamecube or ONLY on GameCube.

A lot of those games you listed were SEGA exclusives on Xbox, not DC ports, which is what I was talking about. Though tbf, Xbox did get a few awesome SEGA exclusives like Shemue 2, DOA 3 and JSR Future.
 
Gamecube was the best party console.
What Is It Reaction GIF by Nebraska Humane Society

It did not have SingStar, Guitar Hero, Dance Dance Revolution, Twisted Metal: Black, or Star Wars: Battlefront II. Were you hosting parties for five year olds?
 
Last edited:

REDRZA MWS

Member
It depends on the library that you like. Dreamcast has a ton of very cool Japanese titles that are Japan only, for example.

But yeah, Xbox OG was a very interesting console when it came out and especially so after it got cracked.
I loved them all. the NFL 2k games were ridiculous.But OG Xbox was wa way better machine and early on had ridiculous games, and way before their time built in Ethernet foe Xbox live, most powerful hardware at the time, and a built in HDD.
 

SF Kosmo

Al Jazeera Special Reporter
Dreamcast run was short, and a lot of the best stuff got ported to other platforms, but it was dense. I can't think of many other times when there were so many games I wanted to get every month. So it will always be something of a Golden Age for me.
 

StueyDuck

Member
I loved my dreamcast to bits, there are games on there that still feel wholly unique, plus the vmu.

All that being said.

The answer GameCube, where dreamcast has some great games the GameCube had a plethora of bangers. I mean how can you compete with melee, sunshine, wind waker, Luigis mansion, fzero, metroid prime, double dash etc. And that's just the amazing mainstream games, then on top of that you still had some of the best off brand Nintendo games too.

Resi 4, eternal darkness, REmake, twin snakes, star fox adventures etc

All in that tight compact caboose
 
Last edited:

Magik85

Member
Dreamcast will always be special for me.
But i gotta vote for original Xbox. Back then it almost seemed like it was Dreamcast 2 with so many sequels form Sega console.
Plus its technical advantage was so huge that some games looked half gen ahead if PS2.
 

PaintTinJr

Member
Xbox had 20 giga FLOPS while the PS2 had 6.2 giga FLOPS, so on paper the xbox GPU was faster. ...
GFLOP/s isn't rendered triangles. Here's the extracts from each of those consoles' wiki pages with some important numbers bolded/enlarged - and this comparison isn't taking account of the PS2 having pipeline/shader versatility that just wasn't there on PC styled GPUs at Opengl 1.5/DX8 of the Xbox OG, that had one foot still firmly in fixed path T&L

Xbox OG
Technical specifications
Main article: Xbox technical specifications
The Xbox CPU is a 32-bit 733 MHz, custom Intel Pentium III Coppermine-based processor. It has a 133 MHz 64-bit GTL+ front-side bus (FSB) with a 1.06 GB/s bandwidth. The system has 64 MB unified DDR SDRAM, with a 6.4 GB/s bandwidth, of which 1.06 GB/s is used by the CPU and 5.34 GB/s is shared by the rest of the system.[59]

Its GPU is Nvidia's 233 MHz NV2A. It is capable of geometry calculations for up to a theoretical 115 million vertices/second. It has a peak fillrate of 932 megapixels/second, capable of rendering a theoretical 29 million 32-pixel triangles/second. With bandwidth limitations, it has a realistic fillrate of 250–700 megapixels/second, with Z-buffering, fogging, alpha blending, and texture mapping,[60] giving it a real-world performance of 7.8–21 million 32-pixel triangles/second

PS2

Performance
Floating point performance: 6.2 GFLOPS (single precision 32-bit floating point)
FPU 0.64 GFLOPS
VU0 2.44 GFLOPS[13][14]
VU1 3.08 GFLOPS (Including internal 0.64 GFLOPS EFU)
Tri-strip geometric transformation (VU0+VU1): 150 million vertices per second[15]
3D CG geometric transformation with raw 3D perspective operations (VU0+VU1): 66–80+ million vertices per second[9]
3D CG geometric transformations at peak bones/movements/effects (textures)/lights (VU0+VU1, parallel or series): 15–20 million vertices per second[15] [16] [17] [18]
Lighting: 38 million polygons/second
Fog: 36 million polygons/second

Curved surface generation (Bézier): 16 million polygons/second [15]
Image processing performance: 150 million pixels/second
Actual real-world polygons (per frame): range of 500–600k at 30 FPS, 250–300k at 60 FPS [17]
Instructions per second: 6,000 MIPS (million instructions per second)[19]

Graphics processing unit
Parallel rendering processor with embedded DRAM "Graphics Synthesizer" (GS) clocked at 147.456 MHz
279 mm² die (combined EE+GS in SCPH-7500x: 86 mm², 53.5 million transistors)
Dedicated connection from and to EE and VU1 via GIF
Programmable CRT controller (PCRTC) for output
Video output resolution: Variable from 256×224 to 1920×1080[27] [28]
NTSC (interleaved/progressive scan): 256 x 448/224, 320 x 448/224, 384 x 448/224, 512 x 448/224 or 640 x 448/224
PAL (interleaved/progressive scan): 256 x 512/256, 320 x 512/256, 384 x 512/256, 512 x 512/256 or 640 x 512/256
VESA: 640 x 480, 800 x 600, 1024 x 768 or 1280×1024 pixels
DTV: 720 x 480 (480p) or 1920 x 1080 (1080i)
4 MB of embedded DRAM as video memory
48 gigabytes per second peak bandwidth
Texture buffer bandwidth: 9.6 GB/s
Frame buffer bandwidth: 38.4 GB/s
eDRAM bus width: 2560-bit (composed of three independent buses: 1024-bit write, 1024-bit read, 512-bit read/write)
Pixel configuration: RGB:alpha, 24:8, 15:1; 16-, 24-, or 32-bit Z-buffer
Display color depth: 32-bit (RGBA: 8 bits each)
Pixel/Texel pipelines: 16 (unified) [21] [29] [28]
Raster setup & execution consists of the entire block of 16 pixel pipes being involved in every stage of drawing a frame, in parallel, being equipped to handle processing functions such as fogging, texture mapping, AA and more, cycles split between all pipes [28]
On their own, the 16 pipes output/process 16 pixels/cycle (1 pixel/pipe), giving a maximum throughput of 2400 megapixels/sec, at the GS's max clock speed, which includes 32bit pixels & all the systems basic alpha-blending, Z-buffering & filtering operations [28]
When needing to do texture mapping, the 16 pipes output/process 8 pixel/cycle & 8 texels/cycle, per each pipe, in parallel, for a maximum throughput of 1200 megapixels & 1200 megatexels [28]
The hardware's native fogging & AA uses additional cycles which reduces the overall throughput; but AA & fogging can be done as a post effect, using full-screen passes or sprites, and VRAM imaging operations, using then just a part of a said 2d vfx budget, as opposed to lowering the ceiling, perpetually, from the prior said textured fill-rate figure [30] [21] [9]
Overall pixel fillrate: 16 × 147 Mpix/s = 2.352 gigapixel/s
1.2 gigapixel/s (with Z-buffer, alpha, and texture)
With no texture, flat shaded: 2.4 Gpix/s (75,000,000 32-pixel raster triangles)
With 1 full texture (diffuse map), Gouraud shaded: 1.2 Gpix/s (37,750,000 32-bit pixel raster triangles)
Texture fillrate: 1.2 Gtexel/s
Sprite drawing rate: 18.75 million/s (8×8 pixels)
Particle drawing rate: 150 million/s
Polygon drawing rate: 75 million/s (small polygon)
50 million/s (48-pixel quad with Z and A)
30 million/s (50-pixel triangle with Z and A)
25 million/s (48-pixel quad with Z, A and T)
16 million/s (75-pixel triangle with Z, A, T and fog) [28]


Graphics Synthesizer as found in SCPH-390xx
GS effects include: Dot3 bump mapping (normal mapping),[31] [21] mipmapping, spherical harmonic lighting,[32] alpha blending, alpha test, destination alpha test, depth test, scissor test, transparency effects, framebuffer effects, post-processing effects, perspective-correct texture mapping, edge-AAx2 (poly sorting required),[9] bilinear, trilinear texture filtering, multi-pass, palletizing (6:1 ratio 4-bit; 3:1 ratio 8-bit), offscreen drawing, framebuffer mask, flat shading, Gouraud shading, cel shading, dithering, texture swizzling.

Multi-pass rendering ability
Four passes: 300 Mpixel/s (75 Mpixels/pass) [21]
 
Last edited:
GFLOP/s isn't rendered triangles. Here's the extracts from each of those consoles' wiki pages with some important numbers bolded/enlarged - and this comparison isn't taking account of the PS2 having pipeline/shader versatility that just wasn't there on PC styled GPUs at Opengl 1.5/DX8 of the Xbox OG, that had one foot still firmly in fixed path T&L

Xbox OG


PS2



Graphics processing unit
Parallel rendering processor with embedded DRAM "Graphics Synthesizer" (GS) clocked at 147.456 MHz
279 mm² die (combined EE+GS in SCPH-7500x: 86 mm², 53.5 million transistors)
Dedicated connection from and to EE and VU1 via GIF
Programmable CRT controller (PCRTC) for output
Video output resolution: Variable from 256×224 to 1920×1080[27] [28]
NTSC (interleaved/progressive scan): 256 x 448/224, 320 x 448/224, 384 x 448/224, 512 x 448/224 or 640 x 448/224
PAL (interleaved/progressive scan): 256 x 512/256, 320 x 512/256, 384 x 512/256, 512 x 512/256 or 640 x 512/256
VESA: 640 x 480, 800 x 600, 1024 x 768 or 1280×1024 pixels
DTV: 720 x 480 (480p) or 1920 x 1080 (1080i)
4 MB of embedded DRAM as video memory
48 gigabytes per second peak bandwidth
Texture buffer bandwidth: 9.6 GB/s
Frame buffer bandwidth: 38.4 GB/s
eDRAM bus width: 2560-bit (composed of three independent buses: 1024-bit write, 1024-bit read, 512-bit read/write)
Pixel configuration: RGB:alpha, 24:8, 15:1; 16-, 24-, or 32-bit Z-buffer
Display color depth: 32-bit (RGBA: 8 bits each)
Pixel/Texel pipelines: 16 (unified) [21] [29] [28]
Raster setup & execution consists of the entire block of 16 pixel pipes being involved in every stage of drawing a frame, in parallel, being equipped to handle processing functions such as fogging, texture mapping, AA and more, cycles split between all pipes [28]
On their own, the 16 pipes output/process 16 pixels/cycle (1 pixel/pipe), giving a maximum throughput of 2400 megapixels/sec, at the GS's max clock speed, which includes 32bit pixels & all the systems basic alpha-blending, Z-buffering & filtering operations [28]
When needing to do texture mapping, the 16 pipes output/process 8 pixel/cycle & 8 texels/cycle, per each pipe, in parallel, for a maximum throughput of 1200 megapixels & 1200 megatexels [28]
The hardware's native fogging & AA uses additional cycles which reduces the overall throughput; but AA & fogging can be done as a post effect, using full-screen passes or sprites, and VRAM imaging operations, using then just a part of a said 2d vfx budget, as opposed to lowering the ceiling, perpetually, from the prior said textured fill-rate figure [30] [21] [9]
Overall pixel fillrate: 16 × 147 Mpix/s = 2.352 gigapixel/s
1.2 gigapixel/s (with Z-buffer, alpha, and texture)
With no texture, flat shaded: 2.4 Gpix/s (75,000,000 32-pixel raster triangles)
With 1 full texture (diffuse map), Gouraud shaded: 1.2 Gpix/s (37,750,000 32-bit pixel raster triangles)
Texture fillrate: 1.2 Gtexel/s
Sprite drawing rate: 18.75 million/s (8×8 pixels)
Particle drawing rate: 150 million/s
Polygon drawing rate: 75 million/s (small polygon)
50 million/s (48-pixel quad with Z and A)
30 million/s (50-pixel triangle with Z and A)
25 million/s (48-pixel quad with Z, A and T)
16 million/s (75-pixel triangle with Z, A, T and fog) [28]


Graphics Synthesizer as found in SCPH-390xx
GS effects include: Dot3 bump mapping (normal mapping),[31] [21] mipmapping, spherical harmonic lighting,[32] alpha blending, alpha test, destination alpha test, depth test, scissor test, transparency effects, framebuffer effects, post-processing effects, perspective-correct texture mapping, edge-AAx2 (poly sorting required),[9] bilinear, trilinear texture filtering, multi-pass, palletizing (6:1 ratio 4-bit; 3:1 ratio 8-bit), offscreen drawing, framebuffer mask, flat shading, Gouraud shading, cel shading, dithering, texture swizzling.

Multi-pass rendering ability
Four passes: 300 Mpixel/s (75 Mpixels/pass) [21]
FLOPS are easy to calculate and understand, but I'm not a game developer, so I won't pretend that I have an understanding of the strengths and weaknesses of GPUs architecture, to be able to say definitively which console was the faster one.

PS2 could well be more powerful in some areas, or with theoretical numbers, but I prefer to let the games speak for themselves:

PS2 graphics

4.png


Xbox graphics

Clipboard02.png


6.png


Clipboard08.png



Dolphin-2019-07-09-02-35-09-23.png


FFFE07-D220190706174756274.png


pcsx2-2019-07-10-22-18-42-63.png


Clipboard09.png


pcsx2-2019-07-10-22-34-55-84.png


Clipboard15.png


All games made primarily for the PS2 hardware (Black, Burnout, MGS2, Silent Hill 2) could run on xbox without big compromises and that wasnt the case when xbox games were ported to PS2, therefore I have no reason to believe that PS2 was faster and more capable than xbox hardware.


z9YXDmT.png


btecDvv.png


RPAB0RK.jpeg

iOQtNIn.png


0FwqJQq.png

Clipboard03.png


fotLPRT.png

15.png

3.png

5.png

riddickscr8large.jpg


yUEYpwI.png

8EQYc4X.png

Clipboard09.png


FFFE07-D220051122140739245.png

FFFE07-D220051122135954948.png

FFFE07-D220051122143054548.png

FFFE07-D220051122144120848.png

FFFE07-D220051122143331108.png

pcsx2-2019-07-10-22-34-55-84.png

FFFE07-D220051122123158271.png

FFFE07-D220051122124557160.png


This is PS3/360 level of effects at a lower resolution. I have never seen such insanely impressive graphics on my PS2. You seem to be impressed by the graphics of Shadow of the Colossus, but that game had very simple graphics compared to xbox games.

8937702-shadow-of-the-colossus-playstation-2-this-is-one-of-the-biggest-.png


Screenshot-20240817-172933-Samsung-Internet-2.jpg


Screenshot-20240817-173217-Samsung-Internet-2.jpg


Xbox also run games at 480p, so the image was sharp even on HDTV.

20190712-201014.png

20190713-192304.png

20190712-201722.png

20190712-202133.png


20190712-204026.png

20190713-112725.png

20190713-113210.png

20190713-113104.png

20190713-112741.png

20190713-161220.png


20190713-161413.png


20190713-122035.png


20190713-121540.png

20190713-121543.png


20190713-121614.png
 
Last edited:
Interesting quotes from developer who written code for PS2 / xbox / GC.


having actually written code for all 3 platforms in that generation that's my assessment.
You could get pretty good initial performance out of GC, but usually that was pretty much it, no amount of dicking around would get you more polygons or more pixels.
PS2 was a pain in the ass poor implementations were really bad.
Xbox was basically underutilized because no one could be bothered.
You might get prettier pixels out of a gamecube than a PS2, but usually only because you couldn't be bothered trying to figure out have to make the PS2 produce the better imagery.
PS2 was often the "lead SKU" at big publishers because of the installed base, Xbox was a version you had to do, in most cases you could write a simple version of your renderer and just drop the assets on Xbox and they would usually run faster. So you'd increase texture quality and call it done.
Usually when you dropped it on gamecube it would run slower and you'd have no memory left, so you downsample to make things fit, figure out how you could use ARAM without crippling performance and ship it.
If you wrote an XBox exclusive with no intention of ever shipping on PC, and you actually spent time optimizing there was a lot of performance to be had, usually most titles were CPU limited because then the polygon indices had to be copied into the GPU ring buffer (which wasn't actually a ring buffer). If your app was pushing a lot of geometry it could literally spend 60% of it's time doing nothing but linear memory copies.
It was possible to place jumps into the ringbuffer, to effectively "call" static GPU buffers, but it was tricky to get right because of the pipeline and the fact you had to patch the return address as a jump into the buffer so you'd have to place fences between calls to the same static buffer.
If you did this however you could trivially saturate the GPU and produce something much better looking.
On GameCube the biggest issue is it was just had pathetic triangle throughput, the 10M polygons per second (I don't remember the real number) assumes you never clip or light anything.
GameCube was DX7 class hardware for the most part, albeit a more fully featured version than ever shipped in a PC. The GPU just wasn't very fast.
As I said it's real benefit was the memory architecture and I still feel it was over engineered.
On the whole it wasn't a bad machine, but I wouldn't have said it was "more powerful than PS2)

It seems the Xbox console was so fast that it could run PS2 assets without any optimisation. However, I was surprised to learn that the GC wasn't much faster than the PS2.
 
Last edited:

REDRZA MWS

Member
GFLOP/s isn't rendered triangles. Here's the extracts from each of those consoles' wiki pages with some important numbers bolded/enlarged - and this comparison isn't taking account of the PS2 having pipeline/shader versatility that just wasn't there on PC styled GPUs at Opengl 1.5/DX8 of the Xbox OG, that had one foot still firmly in fixed path T&L

Xbox OG


PS2



Graphics processing unit
Parallel rendering processor with embedded DRAM "Graphics Synthesizer" (GS) clocked at 147.456 MHz
279 mm² die (combined EE+GS in SCPH-7500x: 86 mm², 53.5 million transistors)
Dedicated connection from and to EE and VU1 via GIF
Programmable CRT controller (PCRTC) for output
Video output resolution: Variable from 256×224 to 1920×1080[27] [28]
NTSC (interleaved/progressive scan): 256 x 448/224, 320 x 448/224, 384 x 448/224, 512 x 448/224 or 640 x 448/224
PAL (interleaved/progressive scan): 256 x 512/256, 320 x 512/256, 384 x 512/256, 512 x 512/256 or 640 x 512/256
VESA: 640 x 480, 800 x 600, 1024 x 768 or 1280×1024 pixels
DTV: 720 x 480 (480p) or 1920 x 1080 (1080i)
4 MB of embedded DRAM as video memory
48 gigabytes per second peak bandwidth
Texture buffer bandwidth: 9.6 GB/s
Frame buffer bandwidth: 38.4 GB/s
eDRAM bus width: 2560-bit (composed of three independent buses: 1024-bit write, 1024-bit read, 512-bit read/write)
Pixel configuration: RGB:alpha, 24:8, 15:1; 16-, 24-, or 32-bit Z-buffer
Display color depth: 32-bit (RGBA: 8 bits each)
Pixel/Texel pipelines: 16 (unified) [21] [29] [28]
Raster setup & execution consists of the entire block of 16 pixel pipes being involved in every stage of drawing a frame, in parallel, being equipped to handle processing functions such as fogging, texture mapping, AA and more, cycles split between all pipes [28]
On their own, the 16 pipes output/process 16 pixels/cycle (1 pixel/pipe), giving a maximum throughput of 2400 megapixels/sec, at the GS's max clock speed, which includes 32bit pixels & all the systems basic alpha-blending, Z-buffering & filtering operations [28]
When needing to do texture mapping, the 16 pipes output/process 8 pixel/cycle & 8 texels/cycle, per each pipe, in parallel, for a maximum throughput of 1200 megapixels & 1200 megatexels [28]
The hardware's native fogging & AA uses additional cycles which reduces the overall throughput; but AA & fogging can be done as a post effect, using full-screen passes or sprites, and VRAM imaging operations, using then just a part of a said 2d vfx budget, as opposed to lowering the ceiling, perpetually, from the prior said textured fill-rate figure [30] [21] [9]
Overall pixel fillrate: 16 × 147 Mpix/s = 2.352 gigapixel/s
1.2 gigapixel/s (with Z-buffer, alpha, and texture)
With no texture, flat shaded: 2.4 Gpix/s (75,000,000 32-pixel raster triangles)
With 1 full texture (diffuse map), Gouraud shaded: 1.2 Gpix/s (37,750,000 32-bit pixel raster triangles)
Texture fillrate: 1.2 Gtexel/s
Sprite drawing rate: 18.75 million/s (8×8 pixels)
Particle drawing rate: 150 million/s
Polygon drawing rate: 75 million/s (small polygon)
50 million/s (48-pixel quad with Z and A)
30 million/s (50-pixel triangle with Z and A)
25 million/s (48-pixel quad with Z, A and T)
16 million/s (75-pixel triangle with Z, A, T and fog) [28]


Graphics Synthesizer as found in SCPH-390xx
GS effects include: Dot3 bump mapping (normal mapping),[31] [21] mipmapping, spherical harmonic lighting,[32] alpha blending, alpha test, destination alpha test, depth test, scissor test, transparency effects, framebuffer effects, post-processing effects, perspective-correct texture mapping, edge-AAx2 (poly sorting required),[9] bilinear, trilinear texture filtering, multi-pass, palletizing (6:1 ratio 4-bit; 3:1 ratio 8-bit), offscreen drawing, framebuffer mask, flat shading, Gouraud shading, cel shading, dithering, texture swizzling.

Multi-pass rendering ability
Four passes: 300 Mpixel/s (75 Mpixels/pass) [21]
The OG Xbox had a GPU too. Nvidia I believe. I ADORED my PS2, but big 3rd party games looked and played better on Xbox. You forgot the console worlds first dedicated broadband online Ethernet port built in, and it’s HDD.
 

REDRZA MWS

Member
Xbox for me

Most powerful console of that gen, pushed and standardized a bunch features(ie internal storage, broadband internet requirement). Cool UI, also could rip CD's and play music in some games (Amped)

Halo1/2 also pushed a bunch of things on consoles, ie dual-stick fps, autosaves, matchmaking
Fact of the matter is PlayStation” FANS, will always poo poo or dismiss anything and everything positive. It’s a lost cause here for any Xbox owner. Notice I never said “FAN”. Real gamers enjoy video games period. The brand Fanism never once affected me. Boot up, grab your controller, and have fun.
 

PaintTinJr

Member
FLOPS are easy to calculate and understand, but I'm not a game developer, so I won't pretend that I have an understanding of the strengths and weaknesses of GPUs architecture, to be able to say definitively which console was the faster one.

PS2 could well be more powerful in some areas, or with theoretical numbers, but I prefer to let the games speak for themselves:

...
I'll repeat the same thing again, with different words. Your 20GFLOPs vs 6 GFLOPS is not representative of rendering, as shown by how in real terms PS2 was able to render far more than the Xbox theoretical limit, which was already a 1/3rd bigger than the maximum achieved with the hardware with any setup.

Your pleasant but pointless splurge of images just confirms what I said in my very first reply in this thread, and though the xbox images look perfectly fine, it is mostly texturing with bump mapping and dialled back fx we are looking at on Xbox, but because the textures are probably at least 4x bigger (due to S3TC), might be as much as x8 bigger due to RAM and HDD the difference in the visuals of the texturing alone makes it hard for you to actually see what you are looking at in terms of graphics fx.

Your comparison choice of these two images you choose doesn't show what you think it shows, as that isn't shadow mapping being used

6.png
Clipboard08.png


/ps
I suspect your poor choice of images to showcase PS2 SotC will trigger at least someone here on gaf to show the game in its true light
 
Last edited:

BlackTron

Member
Melee, Monkeyball, Waverace all in and around launch to wait for ...Double Dash, Mario Party, 1080 Avalache, Kirby's Air ride, etc, such an amazing local multiplayer console IMO

I gotta admit the only other one I had at the time was MonkeyBall, but yeah Melee was a megaton game and still being played competitively eons later.

Gamecube was also better than PS2 graphically. So was Xbox.
 

PaintTinJr

Member
I gotta admit the only other one I had at the time was MonkeyBall, but yeah Melee was a megaton game and still being played competitively eons later.

Gamecube was also better than PS2 graphically. So was Xbox.
They were - but not on Konami's PES/WE - but in Xbox's case, definitely not more powerful, and certainly not more versatile which was the subtle differences being discussed.

Quality texturing with bump mapping and baked lightmap lighting go a long way against a console struggling to texture map at reasonable texture sizes.

edit
Visuals and performance in MGS2 is also a massive win for PS2, fighting Rex with all that particle fx stutter from the missiles on the Xbox version should never have got through Xbox certification.
 
Last edited:

RobRSG

Member
They were - but not on Konami's PES/WE - but in Xbox's case, definitely not more powerful, and certainly not more versatile which was the subtle differences being discussed.

Quality texturing with bump mapping and baked lightmap lighting go a long way against a console struggling to texture map at reasonable texture sizes.

edit
Visuals and performance in MGS2 is also a massive win for PS2, fighting Rex with all that particle fx stutter from the missiles on the Xbox version should never have got through Xbox certification.
To be fair, the output from Konami outside of Silent Hill 4 was an afterthought. But the performance and graphics at display has to do with the porting effort, not with the technical prowess of the Xbox.
 

BlackTron

Member
They were - but not on Konami's PES/WE - but in Xbox's case, definitely not more powerful, and certainly not more versatile which was the subtle differences being discussed.

Quality texturing with bump mapping and baked lightmap lighting go a long way against a console struggling to texture map at reasonable texture sizes.

edit
Visuals and performance in MGS2 is also a massive win for PS2, fighting Rex with all that particle fx stutter from the missiles on the Xbox version should never have got through Xbox certification.

Literally every version of Sonic Adventure released after Dreamcast had ruined effects. Dreamcast better than every future console confirmed? This is a desperate argument...nitpicking effects from ports lol.
 

PaintTinJr

Member
To be fair, the output from Konami outside of Silent Hill 4 was an afterthought. But the performance and graphics at display has to do with the porting effort, not with the technical prowess of the Xbox.
It isn't, the fillrate in worst cases for xbox could drop to less than a 1/4 of the reliable PS2 available fillrate of PS2, and at best was only 2/3rds of the PS2. MGS3 even required frustum changes because the PS2 was using the full zbuffer precision to draw depth with a wide FOV. Konami were just top tier with EmotionEngine animation and physics and fx, and the xbox even being a mid-gen release with double the ram and a HDD wasn't as powerful to run Konamis games better than PS2.
 
I'll repeat the same thing again, with different words. Your 20GFLOPs vs 6 GFLOPS is not representative of rendering, as shown by how in real terms PS2 was able to render far more than the Xbox theoretical limit, which was already a 1/3rd bigger than the maximum achieved with the hardware with any setup.

Your pleasant but pointless splurge of images just confirms what I said in my very first reply in this thread, and though the xbox images look perfectly fine, it is mostly texturing with bump mapping and dialled back fx we are looking at on Xbox, but because the textures are probably at least 4x bigger (due to S3TC), might be as much as x8 bigger due to RAM and HDD the difference in the visuals of the texturing alone makes it hard for you to actually see what you are looking at in terms of graphics fx.

Your comparison choice of these two images you choose doesn't show what you think it shows, as that isn't shadow mapping being used

6.png
Clipboard08.png


/ps
I suspect your poor choice of images to showcase PS2 SotC will trigger at least someone here on gaf to show the game in its true light
This is what the game looks like on PS2. Actually it looks even worse because you are looking at an emulator screenshot (the game runs at 480i on the real PS2). Xbox screenshots are taken from x360 BC. This is the same (training) level, but the PS2 version had levels that were rebuilt (probably due to hardware limitations). The Xbox version had much more impressive shadows because the game used shadow buffers technology (Geforce 3 feature), allowing developers to render detailed shadows at relatively low cost. The PS2 version had dynamic shadows as well, but very little compared to the Xbox version because they were more expensive to render, and still not as detailed.

8678678.png


Clipboard20.png

21.png


Clipboard06r.png

4.png


Clipboard02.png


11.png


Clipboard03.png

5.png


Clipboard05.png


86.png


Clipboard06.png


27.png


Clipboard09.png


43.png


Clipboard11.png


5235.png

Clipboard12.png


4544.png


Clipboard17.png


78.png


Clipboard01.png


5467.png


Clipboard13.png

You want to tell me that all those DX8 hardware features (S3TC, T&L, shaders, EMBM) were not really needed? If this is the case, why did Sony decide to use Nvidia GPU in PS3 if all these features were useless? Maybe PS2 could (theoretically) achieve better results in software, but it seems developers were not smart enough to go that route. On the Xbox, however, extremely impressive effects were easy to implement and developers had plenty of resources to use them (RAM memory). I loved my PS2 console dude, I even sold my PC back then to buy it, but when I saw how the Xbox games looked like, I thought I was looking at the next generation of graphics. Xbox games had much bigger levels with more details, better lighting, shadows and textures. You try to suggest xbox could render way less polygons compared to the PS2, but games like DOA3 looked extremely detailed to me and I dont remember being so impressed with PS2 character models even in games like Tekken 5 or Ghost Hunter (one of the best looking PS2 games).



20190713-161413.png
btecDvv.png



Far Cry Instincts on Xbox had large open levels with hundreds of trees, lots of grass and dynamic shadows on top of that. I have never seen such an absurd amount of detail in a PS2 game. It's baffling how anyone could actually believe that PS2 had the upper hand when it came to polygon budget, when no PS2 game looked anywhere near as detailed.


far-cry-instincts-20050411104625824.jpg


far-cry-instincts-20050511113621866.jpg



far-cry-instincts-20050511113625897.jpg


Edit - I found DEFINITIVE confirmation that the PS2 was unable to reach the polycounts found on the wiki page in real games. Sony's extensive sampling with the Performance Analyzer found the maximum sustained on PS2 was 7.5 million polys a second, a far cry from 70Million, and average was half of that.


Here's a discussion about this on the beyond3d forum. The link for the Sony PDF is no longer active, but you can still find it here:


What's more I also found developer comments about Xbox vs PS2. According to "ERP" (he wrote games on Xbox / PS2 and GC) on the beyond3d forum, the Xbox was faster and better in every single way except for particle rendering.

e6lnTJC.jpeg
 
Last edited:

Brigandier

Gold Member
Xbox by a fucking landslide, It had a shitload of great games and superior online gaming experience with Halo 2 , Ghost Recon, Rainbow Six and Splinter Cell leading the way, It was so easy to modify and transfer games from PC to Xbox hdd via a cross over cable it was such a class machine.

Let's not forget KOTOR, Ninja Gaiden, Rallisport, Project Gotham 2,
Crimson Skies, Panzer Dragoon Orta and many more.

I loved the DC I'm a big DC fan but the original Xbox is a huge part of gaming for me, That was absolutely the best era for Xbox live.

The Gamecube I genuinely don't care about a single game on the console.
 
Last edited:

Dane

Member
Very tough, the Dreamcast had plenty of quality software even on its short time, but the Xbox also did have its sequels, Tom Clancy games that were much better on it, some strong exclusives and games generally ran better. My vote goes for the Xbox by a slim margin.

This is what the game looks like on PS2. Actually it looks even worse because you are looking at an emulator screenshot (the game runs at 480i on the real PS2). Xbox screenshots are taken from x360 BC. This is the same (training) level, but the PS2 version had rebuilt levels (probably due to hardware limitations). Xbox version had much more impressive shadows, because the game used shadow buffer technology, so developers could render detailed shadows at reasonable performance penalty. The PS2 version had dynamic shadows as well, but very little compared to the Xbox version, and not as detailed.

4.png


Clipboard02.png

5.png


Clipboard05.png


86.png


Clipboard06.png


27.png


Clipboard09.png

You want to tell me that all those DX8 hardware features (S3TC, T&L, shaders, EMBM) were not really needed? If this is the case, why did Sony decide to use Nvidia GPU in PS3 if all these features were useless? Maybe PS2 could (theoretically) achieve better results in software, but it seems developers were not smart enough to go that route. On the Xbox, however, extremely impressive effects were easy to implement and developers had plenty of resources to use them (RAM memory). I loved my PS2 console dude, I even sold my PC back then to buy it, but when I saw how the Xbox games looked like, I thought I was looking at the next generation of graphics. Xbox games had much bigger levels with more details, better lighting, shadows and textures. You try to suggest xbox could render way less polygons compared to the PS2, but games like DOA3 looked extremely detailed to me and I dont remember being so impressed with PS2 character models even in Tekken games. Can you tell me which PS2 game the xbox


20190713-161413.png
btecDvv.png
The PS2 could do normal mapping, Matrix Path of Neo had it, and Hitman Blood Money almost did as even one programmer did a paper on it. But the hardware support + more power for these effects made it much easier on the Xbox, while the PS2 was stronger with non Pixel Shader effects. Some games like the aforementioned Matrix had botched Xbox and PC versions due to the lack of time to port these effects properly despite their hardware supporting it, it was a common thing back then with PC games.
 
Last edited:
Very tough, the Dreamcast had plenty of quality software even on its short time, but the Xbox also did have its sequels, Tom Clancy games that were much better on it, some strong exclusives and games generally ran better. My vote goes for the Xbox by a slim margin.


The PS2 could do normal mapping, Matrix Path of Neo had it, and Hitman Blood Money almost did as even one programmer did a paper on it. But the hardware support + more power for these effects made it much easier on the Xbox, while the PS2 was stronger with non Pixel Shader effects. Some games like the aforementioned Matrix had botched Xbox and PC versions due to the lack of time to port these effects properly despite their hardware supporting it, it was a common thing back then with PC games.
Yes, hardware support was a big advantage because developers did not have to look for ways to create similar graphical effects in software (and software usually require more HW resources). More RAM was also a very important factor, because RAM allowed developers to do all these effects on the Xbox on a very large scale, and that's what made the Xbox games look so next-gen compared to the PS2. The Xbox Classic made me realize how important it is to have good lighting in games. You can have millions of polygons, but without good lighting and shadows, the graphics will still look flat.
 
Last edited:

PaintTinJr

Member
This is what the game looks like on PS2. Actually it looks even worse because you are looking at an emulator screenshot (the game runs at 480i on the real PS2). Xbox screenshots are taken from x360 BC. This is the same (training) level, but the PS2 version had levels that were rebuilt (probably due to hardware limitations). The Xbox version had much more impressive shadows because the game used shadow buffers technology (Geforce 3 feature), allowing developers to render detailed shadows at relatively low cost. The PS2 version had dynamic shadows as well, but very little compared to the Xbox version because they were more expensive to render, and still not as detailed.

8678678.png


Clipboard20.png

21.png


Clipboard06r.png

4.png


Clipboard02.png


11.png


Clipboard03.png

5.png


Clipboard05.png


86.png


Clipboard06.png


27.png


Clipboard09.png


43.png


Clipboard11.png


5235.png

Clipboard12.png


4544.png


Clipboard17.png


78.png


Clipboard01.png


5467.png


Clipboard13.png

You want to tell me that all those DX8 hardware features (S3TC, T&L, shaders, EMBM) were not really needed? If this is the case, why did Sony decide to use Nvidia GPU in PS3 if all these features were useless? Maybe PS2 could (theoretically) achieve better results in software, but it seems developers were not smart enough to go that route. On the Xbox, however, extremely impressive effects were easy to implement and developers had plenty of resources to use them (RAM memory). I loved my PS2 console dude, I even sold my PC back then to buy it, but when I saw how the Xbox games looked like, I thought I was looking at the next generation of graphics. Xbox games had much bigger levels with more details, better lighting, shadows and textures. You try to suggest xbox could render way less polygons compared to the PS2, but games like DOA3 looked extremely detailed to me and I dont remember being so impressed with PS2 character models even in games like Tekken 5 or Primal (one of the best looking PS2 games).



20190713-161413.png
btecDvv.png



Far Cry Instincts on Xbox had large open levels with hundreds of trees, lots of grass and dynamic shadows on top of that. I have never seen such an absurd amount of detail in a PS2 game. It's baffling how anyone could actually believe that PS2 had the upper hand when it came to polygon budget, when no PS2 game looked anywhere near as detailed.


far-cry-instincts-20050411104625824.jpg


far-cry-instincts-20050511113621866.jpg



far-cry-instincts-20050511113625897.jpg
What is absurd is that we are specifically talking about performance (at a technical level) not game results by feelings, and unless you are going to re-write facts on the wiki pages of both consoles, it is factually true to say that the PS2 could render more polygons and had higher rasterization capability(fillrate), ie it was more performant than the Xbox, even if the vast majority of multiplatform games looked more pleasing on Xbox with higher resolution and much higher quality textures and lightmap textures.

I know that factual point doesn't sit well with you, but facts are unfortunately facts, are they not?

As for your claim about not needing S3TC, I didn't say that, PS2 would definitely benefit from modern tooling and someone adding in the ability to use BC7 textures blocks. And just for clarity, the Nvidia RSX on PS3 used Opengl with 2.1 level features in Opengl ES form, not DirectX, and certainly not defunct DX8 features/techniques.
 

TNT Sheep

Member
What is absurd is that we are specifically talking about performance (at a technical level) not game results by feelings, and unless you are going to re-write facts on the wiki pages of both consoles, it is factually true to say that the PS2 could render more polygons and had higher rasterization capability(fillrate), ie it was more performant than the Xbox, even if the vast majority of multiplatform games looked more pleasing on Xbox with higher resolution and much higher quality textures and lightmap textures.
There is a way to look at performance in a game. It's called framerate, which is also usually higher or more stable in the Xbox versions of multiplatform games.

For whatever reason this fact is conveniently left out in comparisons.

In any case, I think we are going off-topic again.
 
Last edited:
What is absurd is that we are specifically talking about performance (at a technical level) not game results by feelings, and unless you are going to re-write facts on the wiki pages of both consoles, it is factually true to say that the PS2 could render more polygons and had higher rasterization capability(fillrate), ie it was more performant than the Xbox, even if the vast majority of multiplatform games looked more pleasing on Xbox with higher resolution and much higher quality textures and lightmap textures.

I know that factual point doesn't sit well with you, but facts are unfortunately facts, are they not?

As for your claim about not needing S3TC, I didn't say that, PS2 would definitely benefit from modern tooling and someone adding in the ability to use BC7 textures blocks. And just for clarity, the Nvidia RSX on PS3 used Opengl with 2.1 level features in Opengl ES form, not DirectX, and certainly not defunct DX8 features/techniques
The PS3 console used a different API, but the GPU itself had all the GeForce hardware features built in (the most important ones were pixel and vertex shaders). You tried to say that shaders were not really needed on the PS2 because that console could achieve even better FX effects in software. If that's the case I wonder why developers still use shaders when there are better ways to achieve FX effects in software.

According to you limited RAM had no bearing on versatility and performance of "reality synthesizer" (that was PS3 GPU BTW., I think you wanted to say "Graphics Synthesizer"). Your own words:

Limited RAM and lack of pagefile (temporary RAM storage) has no bearing on the versatility and performance of the Reality Synthesizer and Emotion Engine capabilities with polygons and shaders in the PS2 to render the fx in Shadows of the Colossus.

You tried to downplay the importance of RAM when it comes to FX effects, but if developers wanted to use all those nice looking texture effects and shaders, they needed RAM for that obviously. Even if the PS2 had all DX8 hardware capabilities and developers could easily implement these effects in their games, they would still not use them nearly as often because of RAM limitations on the PS2.

As for s3tc, it was an ASIC extension feature available to all on PC, and could have been easily implemented on the Reality Synth via Shader assembly had Sony bought a patent license from new owner Nvidia, ironically, the PS2 today could implement the much newer modern Block compression formats used by Oodle/Kraken that are supersets of S3TC, so it wasn't a performance issue why PS2 didn't use S3TC.

Then you said that S3TC could have been implemented on "reality sysynth" (again, I think you wanted to say "graphics syntheziser") if sony sony bought a patent and implement modern block compression formats used by oodle/kraken. You thought the PS2 hardware was so versatile that it could do anything in software. You want to use modern tools and knowledge to overcome PS2 limitations, but back in early 2000 this knowledge was not available, so S3TC was extremely important to developers. They had a much higher texture budget as a result of using S3TC, and none of them could travel into the future to realise that they could compress textures using other methods.

Lots of the real-time shadow mapping you think the Xbox OG did was either static shadow map lookup, because the shadow remains the same position to the car like in PGR, regardless of the relationship of the car and the sun light source,
I think the position of the car shadows was changing in PGR2, because in all these photos I can see the car shadows on different sides. Whether they were dynamic or static, I think they looked amazing. In GT4 on the PS2 cars had simple blob shadows.

PGR2

20190713-112725.png

20190713-113210.png

20190713-113104.png

20190713-112741.png


GT4

171.jpg

716.jpg


It's interesting to note that the car models in GT4 had very limited interiors, so developers used dark windows to hide the lack of detail. Cars in PGR2 have detailed interiors and also much sharper reflections (you could clearly see all the building details reflected in the cars).

it was more performant than the Xbox, even if the vast majority of multiplatform games looked more pleasing on Xbox with higher resolution and much higher quality textures and lightmap textures.

I'ts funy how carefull you choose you wording. You are only willing to admit that multipplatform games looked better on Xbox, but you are not willing to admit that this was also the case for exclusives. Nothing on the PS2 can match the graphics of Far Cry, DOA3, Halo 1/2, Riddick, Doom 3, and when xbox game (like Ghost Recon 2, or Splinter Cell) was ported to PS2 downgrade was so huge that it was no longer the same game anymore (smaller and reworked levels to fit the PS2's limited RAM, limited shadows and lighting effects because they were done in software and more expensive because of that). However, it has been possible to port some of the best looking PS2 games (Burnout 2-4, Black, SH2, MGS2) to the Xbox with minimal downgrades. The PS2 had better fillrate, so it could render rain in MGS2 without slowdowns (the game was still playable BTW.) and a little bit more fog in SH2. The Xbox still however had better lighting (the flashlight lighting in SH2 was so blocky on the PS2) and textures. Some ports like GTA3 and VC were even remastered on xbox. The were some small slowdowns during rain effects because of slower fillrate, but overall assets (more detailed character and car models, and even trees), effects (shading, dynamic car reflections), texture quality, draw distance were improved. What's interesting is that when it wasn't raining the Xbox was running these games at locked 30fps, while the PS2 was constantly dropping fps (maybe these were "max payne bullet time" effects?).

You think the PS2 had the upper hand when it comes to polygon budget, but how is that possible that supposedly slower xbox could run PS2 games without any polygon counts reduction? If PS2 would be really faster, that would be necessary. Xbox could run PS2 assets with 2x the framerate (some multiplatform games run at 30fps on PS2 and 60fps on xbox, for example Red Faction 2). Let's also not forget some games like tony hawks pro skater 4, or true crime even run at 720p on xbox.

You said I'm rewriting facts because it is factually true to say that the PS2 can render more polygons, however tech articles comparing the two consoles (anandtech article for example) and the insights of the developers paint a different picture. EA developer who write a code for CMR4 on PS2 and xbox said they could render more 2D grass on the track edges on the PS2, but xbox could push more polycounts. Other game developer also said he could port port PS2 assets to xbox without any optimization and xbox would still achieve higher framerate. The Xbox has been under-utilised in multiplatform games because of that brute force approach and xbox games still looked and run better (at least that was the rule). It seems Xbox could push PS2 polycounts without any problems dude. I'm not a developer, but it seems to me all these xbox DX8 features (T&L, S3TC, vertex and pixel shaders, EMBM, shadow buffers) really matered despite you trying to downplay them, so xbox didnt had to waste it's resources and performance by doing things is software, so in the end it had more power left to push more fx effects but also higher polycounts with the exception of fillrate intensive effects (rain, 2D grass). In some games the fillrate disadvantage showed (for example small dips in MGS2, but the game still ran at over 30fps and was a joy to play).
 
Last edited:

BlackTron

Member
What is absurd is that we are specifically talking about performance (at a technical level) not game results by feelings, and unless you are going to re-write facts on the wiki pages of both consoles, it is factually true to say that the PS2 could render more polygons and had higher rasterization capability(fillrate), ie it was more performant than the Xbox, even if the vast majority of multiplatform games looked more pleasing on Xbox with higher resolution and much higher quality textures and lightmap textures.

"Friends, if you wish to judge capabilities, please. Take my expert advice, and pay no attention to the moving image on your screen. Look at this pile of text instead. Now you are no longer absurd. You're welcome."
 

BlackTron

Member
You think the PS2 had the upper hand when it comes to polygon budget, but how is that possible that supposedly slower xbox could run PS2 games without any polygon counts reduction? If PS2 would be really faster, that would be necessary. Xbox could run PS2 assets with 2x the framerate (some multiplatform games run at 30fps on PS2 and 60fps on xbox, for example Red Faction 2). Let's also not forget some games like tony hawks pro skater 4, or true crime even run at 720p on xbox.

Sorry waaay too logical. We should be comparing numbers from wiki pages like the pros do
 

PaintTinJr

Member
There is a way to look at performance in a game. It's called framerate, which is also usually higher or more stable in the Xbox versions of multiplatform games.

For whatever reason this fact is conveniently left out in comparisons.

In any case, I think we are going off-topic again.
Not that type of anecdotal game performance: Hardware performance indifferent of specific games.

People keep falsely stating the Xbox was the most powerful, when all technical metrics(polygon+lighting+fog + physics/inverse kinematic + fillrate) individually and combined that can be measured of graphical capability and flexibility don't support that in the Wikipedia info or the actual physical hardware, and top tier software from Konami like MGS2, MGS3, PES don't support that alternate narrative anecdotally, either.

This is a separate differentiation from saying whether the general opinionated consensus of presentation of Xbox games felt better, which isn't being argued against, because textures make up the vast majority of those gen games IQ, and HDD + double RAM and 4:1 ratio of texture compression via the s3TC patented tech gave Xbox a clear advantage in making games look good just with texturing alone compared to many PS2 multiplatform games, and cube was clearly not as weak as Xbox for fx, but had the s3tc feature so had good texture quality too, as shown by games like TimeSplitters 2 being best on Cube.
 

PaintTinJr

Member
Sorry waaay too logical. We should be comparing numbers from wiki pages like the pros do
If using the term "powerful" like so many have, then yes, understanding what that actually means in hardware, rather than being impressed with a lot of flat texturing with scant fidelity fx and saying more powerful, doesn't cut it in such a discussion.

Carmack's email conversation with Nvidia engineer talking about shadow stencil volumes working on PS2, but then the Xbox version didn't have that brutally bandwidth demanding feature is pretty definitive.

Back in those days, zbuffering/stencilling, fog, lighting had the same type of performance impact as RT does today.
 

EruditeHobo

Member
I'm not tech-y at all... but it seems clear to me he's not talking about your assessment of the quality of a game's visuals? So posting a slideshow of screens to compare "graphics" seems to completely miss the point of what he's saying.

But what do I know, I'm just trying to follow along...

Anyways, the answer remains Xbox.
 
Last edited:

BlackTron

Member
Not that type of anecdotal game performance: Hardware performance indifferent of specific games.

People keep falsely stating the Xbox was the most powerful, when all technical metrics(polygon+lighting+fog + physics/inverse kinematic + fillrate) individually and combined that can be measured of graphical capability and flexibility don't support that in the Wikipedia info or the actual physical hardware, and top tier software from Konami like MGS2, MGS3, PES don't support that alternate narrative anecdotally, either.

A console is the sum of its parts which is partially why Gamecube punched above its weight as a $99 box and could trade blows with Xbox. But that's also why trying to state raw power with factoids from wikipedia is so disengenious. It's like comparing teraflops today as if they indicate the better system, which is why it was common to think Series X would beat PS5 performance during the hype phase.
 

King Dazzar

Member
Dreamcast was amazing for a couple of years. Up until when the Xbox released at end of 2001, I moved across to that. But my fondest gaming memories from the two are with the Dreamcast for sure. Xbox was great too, but was two years later...
 
  • Like
Reactions: KU_

bobone

Member
Although I disagree with the question; the Dreamcast was dead before either the Gamecube or Xbox launched. Hard to consider that the same generation of console. Most of the multiplatform games on the Dreamcast were on PS1 and N64, not Gamecube and Xbox.
But anyway

I was a huge fan of the Gamecube. The Capcom exclusives sold me on the console when I was trying to decide which to get that Christmas.
Add in Smash Bros, great Zelda games, Pikmin, Lugis Mansion, and legitimately good 3rd party support for a Nintendo console.
I know they considered it a failure. But to me the Gamecube was Nintendo at its peak.
 
A console is the sum of its parts which is partially why Gamecube punched above its weight as a $99 box and could trade blows with Xbox. But that's also why trying to state raw power with factoids from wikipedia is so disengenious. It's like comparing teraflops today as if they indicate the better system, which is why it was common to think Series X would beat PS5 performance during the hype phase.
Exactly, a console is the sum of its parts. RAW performance can be deceiving, because if there are bottlenecks somewhere, real games will never use that power. The PS2 was VRAM limited for sure and you need RAM to load more detailed models.

It's interesting to note that the PS2 was actually beaten in polycounts by the Dreamcast, despite the popular belief that the PS2 was faster. I have a feeling that more texture memory may have had something to do with it.

 

PaintTinJr

Member
The PS3 console used a different API, but the GPU itself had all the GeForce hardware features built in (the most important ones were pixel and vertex shaders). You tried to say that shaders were not really needed on the PS2 because that console could achieve even better FX effects in software. If that's the case I wonder why developers still use shaders when there are better ways to achieve FX effects in software.
I didn't say that at all, but your understanding of graphics tech is only in DirectX brand names, so it is understandable you'd think that, when I was discussing Playstation 2 using shader assembly instead - it is like the low-level ATI/AMD/Nvidia graphics card driver language that high level APIs(opengl/DirectX/Cg) get translated to - and isn't developer friendly like high level GLSL, HLSL, Cgfx, but that assembly was capable of producing all the DX8 fx and more.

Unlike these old semi fixed path GPUs in the Xbox and Cube that had to pipeline in one direction from Vertex to Fragment(Pixel) shader stages, and were semi defunct by Opengl 2.0/Dx9 and completely defunct by Opengl 2.1/Dx10 when geometry shaders was added between the Vertex and Fragment stages, PS2's EmotionEngine and Gs were already able to implement those accelerated techniques and were free to run as an asynchronous bi-directional pipelines to allow reuse of calculations across vertices and fragments in ways that even Opengl2.1/DX10 couldn't, hence why the comments from Carmack about Shadow Stencil Volumes on PS2 did validate some of the PS2 complexity in exchange for performance and versatility.
According to you limited RAM had no bearing on versatility and performance of "reality synthesizer" (that was PS3 GPU BTW., I think you wanted to say "Graphics Synthesizer"). Your own words:



You tried to downplay the importance of RAM when it comes to FX effects, but if developers wanted to use all those nice looking texture effects and shaders, they needed RAM for that obviously. Even if the PS2 had all DX8 hardware capabilities and developers could easily implement these effects in their games, they would still not use them nearly as often because of RAM limitations on the PS2.
Actually, PS2 was capable of Deferred rendering too, even in a forward renderer fx are only VRAM intensive if producing separate render targets(backbuffer images) or sampling lots of high quality resident textures at once, and although not a lot (32MB) SotC, PES, GT4, MGS2. MGS3 show that 32MB was just enough when the PS2 by its design was a StreamProcessor (ie data streamed in from 8-9GB of dual layer DVD as needed)
Then you said that S3TC could have been implemented on "reality sysynth" (again, I think you wanted to say "graphics syntheziser") if sony sony bought a patent and implement modern block compression formats used by oodle/kraken. You thought the PS2 hardware was so versatile that it could do anything in software. You want to use modern tools and knowledge to overcome PS2 limitations, but back in early 2000 this knowledge was not available, so S3TC was extremely important to developers. They had a much higher texture budget as a result of using S3TC, and none of them could travel into the future to realise that they could compress textures using other methods.
Nvidia had litigated 3DFX out of business with help from a massive hundreds of million dollar loan from Microsoft and 3DFX until bought by Nvidia for $1 and the settlement IIRC owned the former S3 company and its patented S3TC, so Sony/Toshiba not adding S3TC as an extension to PS2's EE or GS was purely because the feature is simple maths and bit logic, but this was also at a time when there was no legal precedence to license patents at fair market rate, hence why 3DFX were put out of business, so licensing the patent might not have been possible for Sony. Gamecube had the patent because it was available to ATI via existing deals with 3DFX, or a sharing of Opengl Extensions with Nvidia.
I think the position of the car shadows was changing in PGR2, because in all these photos I can see the car shadows on different sides. Whether they were dynamic or static, I think they looked amazing. In GT4 on the PS2 cars had simple blob shadows.

....
It's interesting to note that the car models in GT4 had very limited interiors, so developers used dark windows to hide the lack of detail. Cars in PGR2 have detailed interiors and also much sharper reflections (you could clearly see all the building details reflected in the cars).



I'ts funy how carefull you choose you wording. You are only willing to admit that multipplatform games looked better on Xbox, but you are not willing to admit that this was also the case for exclusives. Nothing on the PS2 can match the graphics of Far Cry, DOA3, Halo 1/2, Riddick, Doom 3, and when xbox game (like Ghost Recon 2, or Splinter Cell) was ported to PS2 downgrade was so huge that it was no longer the same game anymore (smaller and reworked levels to fit the PS2's limited RAM, limited shadows and lighting effects because they were done in software and more expensive because of that). However, it has been possible to port some of the best looking PS2 games (Burnout 2-4, Black, SH2, MGS2) to the Xbox with minimal downgrades. The PS2 had better fillrate, so it could render rain in MGS2 without slowdowns (the game was still playable BTW.) and a little bit more fog in SH2. The Xbox still however had better lighting (the flashlight lighting in SH2 was so blocky on the PS2) and textures. Some ports like GTA3 and VC were even remastered on xbox. The were some small slowdowns during rain effects because of slower fillrate, but overall assets (more detailed character and car models, and even trees), effects (shading, dynamic car reflections), texture quality, draw distance were improved. What's interesting is that when it wasn't raining the Xbox was running these games at locked 30fps, while the PS2 was constantly dropping fps (maybe these were "max payne bullet time" effects?).
GT4 is still the basis of the excellent GT7(and former HD -> 5Pro-> 5-> 6 and GT mobile on Vita), today and it is far more than rendering boxes with interiors and soap-on-a-rope physics and handling. I'm pretty sure PGR2 isn't the basis of Forza, and if it is, the comparison between today's GT7 and Forza simulators might suggest something.
You think the PS2 had the upper hand when it comes to polygon budget, but how is that possible that supposedly slower xbox could run PS2 games without any polygon counts reduction? If PS2 would be really faster, that would be necessary. Xbox could run PS2 assets with 2x the framerate (some multiplatform games run at 30fps on PS2 and 60fps on xbox, for example Red Faction 2). Let's also not forget some games like tony hawks pro skater 4, or true crime even run at 720p on xbox.

You said I'm rewriting facts because it is factually true to say that the PS2 can render more polygons, however tech articles comparing the two consoles (anandtech article for example) and the insights of the developers paint a different picture. EA developer who write a code for CMR4 on PS2 and xbox said they could render more 2D grass on the track edges on the PS2, but xbox could push more polycounts. Other game developer also said he could port port PS2 assets to xbox without any optimization and xbox would still achieve higher framerate. The Xbox has been under-utilised in multiplatform games because of that brute force approach and xbox games still looked and run better (at least that was the rule). It seems Xbox could push PS2 polycounts without any problems dude. I'm not a developer, but it seems to me all these xbox DX8 features (T&L, S3TC, vertex and pixel shaders, EMBM, shadow buffers) really matered despite you trying to downplay them, so xbox didnt had to waste it's resources and performance by doing things is software, so in the end it had more power left to push more fx effects but also higher polycounts with the exception of fillrate intensive effects (rain, 2D grass). In some games the fillrate disadvantage showed (for example small dips in MGS2, but the game still ran at over 30fps and was a joy to play).
Polygons generate fragments that consume fill-rate, so that's not surprising if you cut fillrate heavy fx on xbox that leaves more limited fill-rate for polygons at increased resolution, it can look superior with quality texturing, but work done is significantly less.

CMR4 is still based on its original PC/PS1 engine for hidden surface removal, of which the PS1 didn't have a Zbuffer, so the game is hardly fill-rate bound on Xbox, as is the worst case scenario for Xbox OG rendering.

You keep mentioning DX8 fx like they were impressive features by themselves, which they weren't say compared to shadow stencil volumes, and even if we said they were pretty cool, they were always used in ways that was highly controlled to not make fill-rate a bottleneck.

They aren't used like the full scene FX in Metroid Prime with the visor or the Water in Waverace or done like the full scene layered fx on Colossi in SotC which Xbox definitely couldn't handle when the zbuffer is being used fully because of the scene scale, lot of geometry and camera freedom, and fog is being added on top of the fx too. It is a full workout with the animations and inverse kinematics too.
 
I didn't say that at all, but your understanding of graphics tech is only in DirectX brand names, so it is understandable you'd think that, when I was discussing Playstation 2 using shader assembly instead - it is like the low-level ATI/AMD/Nvidia graphics card driver language that high level APIs(opengl/DirectX/Cg) get translated to - and isn't developer friendly like high level GLSL, HLSL, Cgfx, but that assembly was capable of producing all the DX8 fx and more.

You keep mentioning DX8 fx like they were impressive features by themselves, which they weren't say compared to shadow stencil volumes, and even if we said they were pretty cool, they were always used in ways that was highly controlled to not make fill-rate a bottleneck.
Who knows, with today's knowledge some talented developers might be able to achieve similar FX effects on the PS2 using software-based workarounds, but back then they obviously couldn't. The 6th generation of consoles has already ended, and not a single PS2 game could come close to the FX effects seen in Xbox exclusives like Riddick or Splinter Cell games. Developers on the PS2 did the best they could, yet the lighting and texture effects were clearly very limited compared to the Xbox games, as my Splinter Cell 1 comparison clearly shows. Some PS2 games like Tekken 4 or God Of War 2 had some nice water effects, but water in Xbox games was still on another level. Either the PS2 was not as capable as you think, or the developers were clueless and should learn from you how to implement FX on the PS2.

According to your wikipedia PS2 was able to render up to 70 million polygons, but real games like jak and daxter 2 were peaking at around 15 millions polygons per second. I can see a big difference between what Wikipedia says and what the developers have achieved.

The PS2 not only had problems running Xbox ports, but even 1998's Half Life 1. Moders ported this game to the Xbox and it ran at 60fps. Even Dreamcast could run old PC games better than the PS2.

The Xbox had a state-of-the-art NV2A GPU. This console was released in 2001, but had similarities with the Geforce 4 (2x vertex shaders and memory architecture improvements) released in 2002. When the Xbox came out, I had a high-end PC at the time (Athlon 1.2GHz + Geforce 3) that could run PC games even at 1600x1200 (Quake 3, for example, ran at about 100fps at 1600x1200x32). I was able to run the PS2 ports (MGS2, Colin McRae 3, GTA3, Tony Hawks Pro Skater 4) without any problems, but the Xbox ports did not run very well. Splinter Cell 1 run at sub 30fps on my PC, so I was very impressed with xbox console. Soon I sold my PC and started gaming on consoles till 2007. Having played hundreds of games on both consoles, I simply cannot believe that the PS2 was faster. Your cherry picked wikipedia info suggests that the PS2 had the upper hand when it came to polycounts, but actual games painted a different picture. Even Dreamcast could surpass PS2 polycounts in number of games, yet alone Xbox.
 
Top Bottom