XXXXXX

54 results
1
2

5598

Forum Posts

0

Wiki Points

0

Followers

Reviews: 53

User Lists: 0

Poll 20 Years 20 Days Since the Launch of the Dreamcast. What went wrong? (60 votes)

Poor hardware. 3%

Poor software. 3%

PS2 is what went wrong. Too much for Sega to overcome in terms of market share and mindshare. 67%

Other. 27%

It's been 20 years and 20 days since the Original Dreamcast launch. As the below video points out Sega had an actually pretty good launch.

They could have sold couple of hundred thousand more according to Sega in Japan if it weren't' for NEC not being able to provide PowerVR 2 GPU for it's initial launch. But other than that it's North America sales were pretty good. It actually had a much, much better lineup than the launch of the PS2. It had games like the stellar Soul Calibur with it's awesome graphics and game play, NFL2K (which was great) and many other tittles. It also had a descent GPU. Some people think that choosing the Voodoo 2 over the PowerVR 2 was what caused the Dreamcast to fail. But Voodoo 2 lacked 32-bit graphics and it didn't had 2D graphics support you would have to get another chipset for 2D graphics, which would have increased the cost of the Dreamcast in my opinion. While the PowerVR 2 had 32 bit graphics (making games look more vibrant and more colorful vs Voodoo 2's 16-bit graphics) and it had support for 2D which made it ideal fit for Dreamcast's $199.0 launch price tag. Sure the PowerVR 2 wasn't going to beat the nVidia's TNT2 or ATI (now AMD's) the original Rage Fury back in 1999 but PowerVR 2 held it's own for the time it was released (originally in 1998).

Aside form the the graphics, it had pretty stellar lineup. It had 18 different launch lineup including Sould Calibur as mentioned, but also other such as House of the Dead 2. While PS2 had a piss poor lineup.

Based on the documentary it seems Sega's failure seems to go back to the Sega Saturn launch. That entire generation Sega made piss poor decisions. It didn't embrace 3rd party developers like Sony did which resulted in Sony outgunning Sega in the games department with high quality games. While Sega focused on only 1st party titles such as Panzor Dragon and the Sonic franchise. Which isn't nearly enough to hold against Sony's onslaught of 1st Party titles likes Crash Bandicoot and many great 3rd party titles. Also, Nintendo had better 1st party titles like Donkey Kong 64, Mario Cart 64 and so on.

It's shocking how low market share Sega had in the console market. It was down to 5% compared to Sony's 60% and Nintendo's 30%. Yes, Sega had only 5% of the console market in 1999 in North America.

But it still it makes me wonder despite doing everything right with the Sega Dreamcast launch, good hardware with good graphics, good launch line up compared to the joke launch of PS2 line up of games and being 1 year ahead of the PS2 and two years ahead of the Gamecube. Sega was even smart enough to include a modem for multiplayer which helped usher in online gaming on the consoles we see today. What caused it to fail?

I think it all comes down to the Sony and their successful Playstation. Sony just opened up too big of a lead in terms of market share and gamer's mindshare. Once Sony launched the PS2. Sega couldn't sell enough to bring back the momentum back to them despite having a good console launch. In other words the 5 year gap between the launch of the Sega Saturn and it's piss poor lineup of games and it's subsequent 5% market share in North America was too much for Sega to overcome.

But it still makes wonder what Sega could have done with the Dreamcast despite doing everything they did, correctly, in my estimation, to bring them back from the dead.

What do you guys think? What caused Sega to fail in the console market? What could they have done differently?

• 5 years ago

7840

Forum Posts

0

Wiki Points

0

Followers

Reviews: 1

User Lists: 0

#51 Howmakewood

Member since 2015 • 7840 Posts

No true system sellers, sports games

5 years ago

20698

Forum Posts

0

Wiki Points

0

Followers

Reviews: 219

User Lists: 0

Member since 2005 • 20698 Posts

@uninspiredcup said:

@Jag85 said:

@Xtasy26 said:

@Jag85 said:

Actually, DC games were pushing way more polygons than PC games at the time. This was partly due to the PowerVR CLX2's highly efficient tiled rendering architecture, which eliminated overdraw and enabled a higher effective fillrate. But the bigger reason was the Hitachi SH4 CPU, which was a powerful geometry processor for its time. It was the SH4+CLX2 combo that made the Dreamcast a polygon powerhouse for its time.

It's worth noting that the DC's arcade counterpart, the Sega Naomi, cost nearly $2000. And yet the DC only sold for $200. In other words, Sega was selling the hardware at a loss. They were hoping to offset the hardware loss with software profit, hence why they put so much effort into creating a stellar software lineup. But unfortunately, Sega failed to sell enough software to offset the hardware loss, for whatever reasons.

I am not going to deny that DC games looked even better than some PC games. Which was one of the reasons I found it impressive being a PC gamer. Yes, the PowerVR titled architecture is very efficient. But it was still slower than high end PC GPU's like the TNT2 Ultra at the time shows as the review of the Neon 250 (which was the desktop version of the PowerVR 2 chip used in the Dreamcast) below shows.

https://www.eurogamer.net/articles/neon250#comments

Also, the use of titled rendering had its draw backs such as blurrier images in far distances.

With respect to the Sega Naomi, it would cost a lot more. You have to consider the cost of the monitor, everything inside, and if it's a racing title you would have to include steering wheel, pedals, or if it's a shooter arcade game (guns) and it's tracking function when shooting.

I am not a surprise that it would be selling at a loss. The desktop version of the GPU was going for like at least $130 - $150 alone. Consoles typically sell at a loss in it's beginning cycle. Bad thing for Sega was as you mentioned they didn't sell enough software and it didn't have scale of sales to make up for the console losses.

The Dreamcast's polygon-processing was faster than high-end PCs upon release. The most intensive PC game of '99, Quake III Arena, went up 10,000 polygons per scene. Dead or Alive 2 on the Dreamcast went up to 70,000 polygons per scene, almost an order of magnitude greater than Quake III. In terms of raw polygon performance, the DC was unrivalled upon release.

Quake III had to be downgraded for the Dreamcast, PC version looked better with loaded significantly faster loadtimes and frame-rate, overall a superior experience.

Although for a PC-to-Console port, it was still very good, as back then even porting a competent Doom was hard.

I was referring to the PC version of Q3, not the DC version. The PC and DC versions both pushed a similar number of polygons, around 10-15 K polys/scene, but with the DC version capped at 30 FPS.

As a general rule of thumb, ports are usually inferior to the original. This was especially true for both PC-to-DC ports and DC-to-PC ports, due to PC games being made for a standard pixel-rendering architecture and DC games being made for an alternative tiled-rendering architecture.

To get the most out of the DC, games need to be built from the ground-up around a tiled-rendering architecture. An example of such a FPS/TPS for the Dreamcast (and Naomi arcade system) is Outtrigger. It was an early "hero shooter" of sorts (e.g. like Overwatch). The graphics in Outtrigger look more detailed than Q3, it pushes a lot more polygons per scene, and it runs at a smooth 60 FPS.

5 years ago

5598

Forum Posts

0

Wiki Points

0

Followers

Reviews: 53

User Lists: 0

Member since 2008 • 5598 Posts

@Jag85 said:

@Xtasy26 said:

@Jag85 said:

Actually, DC games were pushing way more polygons than PC games at the time. This was partly due to the PowerVR CLX2's highly efficient tiled rendering architecture, which eliminated overdraw and enabled a higher effective fillrate. But the bigger reason was the Hitachi SH4 CPU, which was a powerful geometry processor for its time. It was the SH4+CLX2 combo that made the Dreamcast a polygon powerhouse for its time.

It's worth noting that the DC's arcade counterpart, the Sega Naomi, cost nearly $2000. And yet the DC only sold for $200. In other words, Sega was selling the hardware at a loss. They were hoping to offset the hardware loss with software profit, hence why they put so much effort into creating a stellar software lineup. But unfortunately, Sega failed to sell enough software to offset the hardware loss, for whatever reasons.

I am not going to deny that DC games looked even better than some PC games. Which was one of the reasons I found it impressive being a PC gamer. Yes, the PowerVR titled architecture is very efficient. But it was still slower than high end PC GPU's like the TNT2 Ultra at the time shows as the review of the Neon 250 (which was the desktop version of the PowerVR 2 chip used in the Dreamcast) below shows.

https://www.eurogamer.net/articles/neon250#comments

Also, the use of titled rendering had its draw backs such as blurrier images in far distances.

With respect to the Sega Naomi, it would cost a lot more. You have to consider the cost of the monitor, everything inside, and if it's a racing title you would have to include steering wheel, pedals, or if it's a shooter arcade game (guns) and it's tracking function when shooting.

I am not a surprise that it would be selling at a loss. The desktop version of the GPU was going for like at least $130 - $150 alone. Consoles typically sell at a loss in it's beginning cycle. Bad thing for Sega was as you mentioned they didn't sell enough software and it didn't have scale of sales to make up for the console losses.

The Dreamcast's polygon-processing was faster than high-end PCs upon release. The most intensive PC game of '99, Quake III Arena, went up 10,000 polygons per scene. Dead or Alive 2 on the Dreamcast went up to 70,000 polygons per scene, almost an order of magnitude greater than Quake III. In terms of raw polygon performance, the DC was unrivalled upon release.

The Neon 250 only represents a fraction of the Dreamcast's power. It was heavily downgraded compared to the Dreamcast's PVR CLX2, significantly reducing its tiled-rendering capabilities. The Neon 250 was essentially a low-end GPU in the PVR2 series. And more importantly, the PVR2 was designed around the Dreamcast's Hitachi SH-4 CPU, the real source of the DC's power. The SH-4 was pushing 1.4 GFLOPS of 3D geometry computation, which was unrivalled by PCs up until the PIII 800 and GF256. It's the SH4+PVR2 combo that made the DC a polygon powerhouse, not the PVR2 alone.

Tiled rendering has no real limitation on draw distance. The only real limitation it has is when it comes to translucency effects. That's where the DC's fillrate takes a hit. But when dealing with just opague polygons, the DC has a very high effective fillrate for its time (equivalent to well over 1 GPixels/s).

The Sega Naomi base arcade hardware cost $2,000, without any peripherals. And even that $2,000 was considered low by arcade standards, compared to the Sega Model 3 which cost far more. Its CPU and GPU appear to be slightly faster than the DC versions. But much of the cost went towards the memory. It has much more RAM than the DC. And it stores games on solid-state ROM memory for fast access times, rather than discs. Either way, the DC was definitely sold at a significant loss margin. The manufacturing cost of each DC must've been much higher than what it sold for at retail.

Neon 250 was a downgraded version (as mentioned in the article). It would have been interesting to see how a full fledged version of the Neon 250 with the same specs as the Dreamcast version would have performed against the TNT 2 Ultra. I remember reading about PowerVR 2 on Boot magazine during the heydays of the 3D Graphics War in Boot magazine (now Maximum PC). I do think it would have beaten the TNT most definitely. I would have to see how it would have performed against the TNT2 Ultra. I have my doubts because if it was indeed better than it why didn't Imagination Technologies release a full blown version of it to compete with TNT 2. Were they that stretched with respect to manufacturing it or did they made an internal decision that it wouldn't have been competitive? Why launch a cut-down version? GeForce launched in October of 1999 (couple of weeks ago was in fact it's 20th Anniversary). Plenty of time between 1998 and fall of 1999 to get a full fledged PowerVR 2 out the door IMO. But one thing for sure is that it would definitely been slower than GeForce 256.

5 years ago

20698

Forum Posts

0

Wiki Points

0

Followers

Reviews: 219

User Lists: 0

Member since 2005 • 20698 Posts

@Xtasy26 said:

@Jag85 said:

The Dreamcast's polygon-processing was faster than high-end PCs upon release. The most intensive PC game of '99, Quake III Arena, went up 10,000 polygons per scene. Dead or Alive 2 on the Dreamcast went up to 70,000 polygons per scene, almost an order of magnitude greater than Quake III. In terms of raw polygon performance, the DC was unrivalled upon release.

The Neon 250 only represents a fraction of the Dreamcast's power. It was heavily downgraded compared to the Dreamcast's PVR CLX2, significantly reducing its tiled-rendering capabilities. The Neon 250 was essentially a low-end GPU in the PVR2 series. And more importantly, the PVR2 was designed around the Dreamcast's Hitachi SH-4 CPU, the real source of the DC's power. The SH-4 was pushing 1.4 GFLOPS of 3D geometry computation, which was unrivalled by PCs up until the PIII 800 and GF256. It's the SH4+PVR2 combo that made the DC a polygon powerhouse, not the PVR2 alone.

Tiled rendering has no real limitation on draw distance. The only real limitation it has is when it comes to translucency effects. That's where the DC's fillrate takes a hit. But when dealing with just opague polygons, the DC has a very high effective fillrate for its time (equivalent to well over 1 GPixels/s).

The Sega Naomi base arcade hardware cost $2,000, without any peripherals. And even that $2,000 was considered low by arcade standards, compared to the Sega Model 3 which cost far more. Its CPU and GPU appear to be slightly faster than the DC versions. But much of the cost went towards the memory. It has much more RAM than the DC. And it stores games on solid-state ROM memory for fast access times, rather than discs. Either way, the DC was definitely sold at a significant loss margin. The manufacturing cost of each DC must've been much higher than what it sold for at retail.

Neon 250 was a downgraded version (as mentioned in the article). It would have been interesting to see how a full fledged version of the Neon 250 with the same specs as the Dreamcast version would have performed against the TNT 2 Ultra. I remember reading about PowerVR 2 on Boot magazine during the heydays of the 3D Graphics War in Boot magazine (now Maximum PC). I do think it would have beaten the TNT most definitely. I would have to see how it would have performed against the TNT2 Ultra. I have my doubts because if it was indeed better than it why didn't Imagination Technologies release a full blown version of it to compete with TNT 2. Were they that stretched with respect to manufacturing it or did they made an internal decision that it wouldn't have been competitive? Why launch a cut-down version? GeForce launched in October of 1999 (couple of weeks ago was in fact it's 20th Anniversary). Plenty of time between 1998 and fall of 1999 to get a full fledged PowerVR 2 out the door IMO. But one thing for sure is that it would definitely been slower than GeForce 256.

The fully-fledged PVR2 was licensed by Sega exclusively for their systems. That's why the PC got a cut-back version with the Neon 250, as Sega had the exclusive rights to the fully-fleged PVR CLX2 for the Dreamcast and Sega arcade systems.

It's not possible for a PVR2 on its own to achieve the same specs as the Dreamcast, since the real source of the Dreamcast's power came from the Hitachi SH-4 (CPU). It's the SH-4 that can process up to 1.4 GFLOPS of 3D polygon geometry (more than even the GF256's T&L unit) and then the PVR2 renders those polygons on screen. Without the SH4+PVR2 combo, you're not going to get the real Dreamcast experience.

Nevertheless, if we were to strip out the PVR2 from the DC and put it into a PC, its tiled-rendering architecture would mean its performance on PC varies from game-to-game, depending on how the game is designed. The PVR2 is designed around only rendering what's on screen and eliminating overdraw. Which means that the PVR2 excels with games that consist mostly of opague polygons (which don't require overdraw), with little-to-no translucent polygons. In scenes like these, the PVR2 has a very high effective fillrate exceeding even the GF256, let alone the TNT2U. But when it comes to scenes with a lot of translucent polygons (e.g. particle effects or water effects, which require plenty of overdraw), that's where the PVR2's fillrate takes a hit, with the TNT2U having the advantage there.

5 years ago

54 results
1
2

Jump to Top

Please Log In to post.

Use your keyboard!

ESC

Log in to comment