The Saturn Early Launch in North America - Help Me Understand...

[axel]

At the time I used to have SEGA Pro, SEGA Force, Advanced MegaDrive Gaming, Mean Machines, EDGE, MeanMachines SEGA, MEGA, Sega Magazine and those were every month without fail. Saturn coverage started in 1993 with Mean Machines SEGA doing an exclusive report and then in early 1994 all the mags were buzzing with all the latest Saturn talk. After EDGE 9 June 1994, I don't think anyone could have been in any doubt the Saturn was coming to Japan for the Fall of 1994 and if you did doubt it, the June 1994 Japanese Toyko Game Show made it so clear, it was coming and you better start saving That show was widely covered by the gaming press because it was't just SEGA bring out new Hardware, SONY, NEC and SNK were all getting ready to launch new systems in the fall of 1994 too. It actually was a great time and I used to look forward to each new mag every month.

Yeah, I remember the press talk of Saturn having MS make its OS, how the video chips would be changed, how at one stage it was even meant to have voice control and so on. I didn't like SOJ's silly move to say it was a 64 Bit system due to twin SH-2, but I guess it was a bit of simple PR and a simple counter. SOJ even kept the 64 Bit tag, in the game flyer with the launch Saturn ( I still have it )

Who else was around to make hardware for SEGA? and that could would have put a system back by years. Too much is also made of BC too, did anyone care that the Snes never offered BC? and I doubt it was what held back sales of the N64, never mind I doubt many really used the Mega Drive BC . SEGA needed to make a clean break, but it still felt it could milk and win a 32bit war, with an Add-on to a 16-Bit system, I was always more in favour of the Jupiter plan. No one would then be left behind and development kits and developer tools/resources could be shared (unlike the 32X)

Speaking of that mind, that is one area Sega America/Europe called right and did a lot of good for the Saturn development costs to developers, by looking to dump the programmer box that SOJ supplied and used its Cart Devbox with Cross Products official development kit, which meant you could use a production Saturn. To SN system's credit, they did that months before anyone with their Psy-Q system.

Well yeah, if you were reading 10 magazines a month then of course you would have a different view than 99.9% of consumers. There's no way I could keep up with all that.

The only possible evidence I see of MS working on an OS for the Saturn (other than the Dreamcast obviously) was that a couple of Pocket PCs came out circa 1996 running WinCE on SuperH chips. The HP 300LX for example has an SH3 at 44 MHz with 2 MB of RAM. So it would not surprise me if WinCE could have run on the Saturn... but what would have been the point? It was stuff like that that made me question where exactly Sega thought they were going with all this.

I agree a clean break was probably a good idea but releasing so much hardware so close together just isn't feasible, there's a reason no other company tried it. I think we've discussed this in other threads though.

[gamevet]

Well yeah, if you were reading 10 magazines a month then of course you would have a different view than 99.9% of consumers. There's no way I could keep up with all that.

The only possible evidence I see of MS working on an OS for the Saturn (other than the Dreamcast obviously) was that a couple of Pocket PCs came out circa 1996 running WinCE on SuperH chips. The HP 300LX for example has an SH3 at 44 MHz with 2 MB of RAM. So it would not surprise me if WinCE could have run on the Saturn... but what would have been the point? It was stuff like that that made me question where exactly Sega thought they were going with all this.

I agree a clean break was probably a good idea but releasing so much hardware so close together just isn't feasible, there's a reason no other company tried it. I think we've discussed this in other threads though.

If you read any of the gaming magazines of that time, you'd be hard pressed to come up with a gaming magazine that didn't have screen shots or games, of news about those upcoming consoles. I read about and saw screenshots for the 32X, Saturn and PlayStation games and hardware a good 6 months to a year before their release. I was interested in the 32X for a hot minute, until a month or 2 later when news about the Saturn and some preliminary gaming screenshots started leaking in the likes of EGM and Gamefan magazines.

[Leynos]

I mean SEGA turned down SGI and they could have worked with them and a CPU could have looked into NEC. They loved Yamaha for sound. SEGA had options for hardware.

[TrekkiesUnite118]

I mean SEGA turned down SGI and they could have worked with them and a CPU could have looked into NEC. They loved Yamaha for sound. SEGA had options for hardware.

We have no idea what state the SGI hardware was in when it was shown to Sega. When Nintendo went with it they decided it needed more time in the oven and ended up delaying the N64 until late 1996, which that wasn't an option for Sega. And when we look at the time line for when this was being proposed the Saturn hardware was getting close to being finalized (I believe this was late 1993 when this supposedly happened?). So it was too late to completely change design at that point, and as Gryson has pointed out, they needed to release in 94/95.

As for CPUs from NEC, the SH-2s are the least controversial part of the Saturn's design. It's actually a very good CPU for the time. In many ways it's better than the PS1's CPU in that things like Multiplication and Division instructions take significantly less cycles to complete than they do on the PS1. The reason the PS1 pulls a head in 3D calculations though is because it has the GTE, while Saturn has a 2nd SH2 for that kind of stuff.

Again, I don't think the Saturn's hardware was as bad as people make it out to be. It was powerful enough to be acceptably competitive for it's generation. The real issue was more with how poorly it's launch was handled in the west combined with the mess that was the 32X. Those two things made it start in the hole and it never was able to recover. Poor launch software leads to poor sales. Poor Sales leads to less developer support. Less developer support leads to more poor software. Rinse and repeat.

[Gryson]

For developer insights you might want to listen to some of Pandamonium's interviews he's done with developers working on Saturn launch titles. Some of them knew of the surprise launch as early as I think February? Others didn't know until after it happened. But either way they all say it added a lot of pressure on them to get their games out as soon as possible. So some of them were released unfinished, or missed the launch window entirely coming out around the time of the original planned launch instead. So I'd say that definitely had an impact on the perception of the Saturn's library at the time, as well as an impact on what working with Sega was like for developers.

This is one of the standard narratives around the surprise launch that I don't think holds up under scrutiny.

First, let me bring in some of what Melf has said in his book, Playing at the Next Level:

The decision to release the [Saturn] early, which is still largely attributed to Sega of Japan, though there are conflicting versions of the story, doomed the machine from day one... The surprise launch left software publishers in the lurch without completed titles, and the meager few that were released arrived alongside too few Saturn consoles to be effective.

The May launch doomed the console. Those are big words. And yet... still nobody can produce any clear evidence of this.

Here's the critical fact: Only one Western-developed game was released on the Saturn before September 1995 - Bug!

To go a bit further, Ghen War was the only Western-developed game released in September 1995 (that I know of). All of the others came months later. If developers were feeling pressured to meet the early launch date and that resulted in unfinished releases, that would have happened regardless, because they failed to meet the September launch date.

Here's what Joe Miller said when Melf asked him about the effects of the early launch:

Sega-16: But on the software side, [the early launch] didn’t have an effect on what you were working on?

Joe Miller: No, it didn’t have an effect on what we were working on. It obviously made it more important that we resolve thorny challenges – and we had many... The early launch meant that we had to accelerate all of that in the pipeline as well – all the cross-development tools, all the documentation – and get developers ramped up 6, 8, 12 months earlier than we had expected or intended.

As I've mentioned before, I strongly suspect that two things are being conflated here:

1) The decision to release the Saturn in Sept. 1995
2) The decision to do a limited release in May 1995

Recall that SOA had not committed to a clear Saturn release prior to the launch of the 32X. The original plan was for SOA to focus entirely on the 32X (as I've documented here).

Once the 32X failed, then SOA was quick to announce a September release date. I suspect people like Joe Miller were referring to this "early release", rather than the May 1995 release. Notice how Miller says they had to "get developers ramped up 6, 8, 12 months earlier than we had expected or intended." There are only four months between May and Sept. The issue he's talking about is that they weren't expecting to release the Saturn until much later, into 1996. The failure of the 32X messed up everything.

The failure of the Saturn is blamed on 1) the surprise May release, and 2) the hardware. These points conveniently shift blame away from SOA.

The reality is much more complicated:

1. SOA was already having problems with game development during the latter half of the Genesis era.

They spent millions upon millions of dollars to create Sega Multimedia Studio, which was then unable to put out any products aside from Jurassic Park CD and eventually Wild Woody.

The situation at STI was reportedly bad. Here's another quote from Melf's book:

Craig Stitt recalled that he began work on another project shortly after Sonic Spinball was finished. It had been authorized to go into full production and lingered for an agonizing 14 months before being shelved. Stitt was given artists who had zero computer knowledge or had no video game playing experience, and his best programmer was fired not long after work began. Even so, Stitt considered himself lucky. According to him, there were people at STI who went several years without shipping a complete product.

The number of quality products coming out of SOA during this era was very low.

2. The 32X was a huge fiasco.

Saturn developers thought the Saturn would come out later than it did in the West. The concurrent release of consoles made consumers skeptical. Third parties became wary of Sega.

3. The slump in the game industry had a huge negative effect on SOA and their ability to put out software.

SOA reduced its staff in 1995 to just a fraction of what it had before - from 900 in 1993 to just 350 in 1995. Staff were inexperienced, the hardware was new, budgets had been lowered, and there was less support.


All of these points contribute to the same outcome: SOA was unable to put out quality software for the Saturn. It had to rely on Japanese titles, which went against SOA's entire strategy. In the end, it's the software that determines whether a console succeeds or fails.

Note: I'm not trying to play some kind of blame game here. I think we have to acknowledge, though, that the entire idea of the May release dooming the Saturn comes from one person who is trying very hard to shift blame away from himself. That just distorts the true history, so it's important to acknowledge and be skeptical about.

[TrekkiesUnite118]

I think both sides of that story could be true with the surprise launch. From the Developer interviews we've heard, it seems some devs got the memo early, others didn't. Some devs did try to release as early as possible (Virtua Racing ended up releasing sometime in the summer, when originally planned for September), other devs just decided to go "Fuck it we'll release it when it's done then" (SimCity2000 was originally to be a launch title but ended up releasing a month later than the original September launch).

So it could be that some of the developers closer to Sega were more prepared for it (Bug! for example), and as you get to more third party developers it becomes more of a crapshoot of if they knew about it or not, if they were able to cope with it, etc.

[Gryson]

Some devs did try to release as early as possible (Virtua Racing ended up releasing sometime in the summer, when originally planned for September), other devs just decided to go "Fuck it we'll release it when it's done then" (SimCity2000 was originally to be a launch title but ended up releasing a month later than the original September launch).

Virtua Racing wasn't released until November:

https://segaretro.org/Press_release:...SEGA_SATURN(TM)

I'm just not seeing the evidence that these developers were ready to release even by September.

Did they feel pressured? Did games get released unfinished? No doubt. I'm sure that would have happened even without the May release.

I don't see a bit of evidence to suggest any of this doomed the Saturn, though.

Now, if you want to make the argument that SOA was not forthcoming with developers about a 1995 release for the Saturn - that's something I could see. I do think that caught people off guard, since the narrative until then was so focused on the 32X.

[TrekkiesUnite118]

Virtua Racing wasn't released until November:

https://segaretro.org/Press_release:...SEGA_SATURN(TM)

I'm just not seeing the evidence that these developers were ready to release even by September.

Did they feel pressured? Did games get released unfinished? No doubt. I'm sure that would have happened even without the May release.

I don't see a bit of evidence to suggest any of this doomed the Saturn, though.

Now, if you want to make the argument that SOA was not forthcoming with developers about a 1995 release for the Saturn - that's something I could see. I do think that caught people off guard, since the narrative until then was so focused on the 32X.

Must have gotten it mixed up with a different game then that got released in the summer. But it is definitely mentioned in some of those interviews from Pandamonium where the devs flat out say the launch came out of no where to them, and they were getting pressure from Sega to release as soon as possible.

I do agree though that I don't think the surprise launch on it's own doomed the Saturn. I do think it was just another wrench thrown at it though. Ultimately I think what doomed it the most was Sega of America's lack of interest in the system and instead focusing on the 32X.

[Gryson]

But it is definitely mentioned in some of those interviews from Pandamonium where the devs flat out say the launch came out of no where to them, and they were getting pressure from Sega to release as soon as possible.

As I mentioned above, is it possible they are referring to the Sept. 1995 release date as coming out of nowhere to them? That's something I could see. It seems a lot of these early development projects began around mid-1994, and at that time nothing firm had been decided about when the Saturn would be released in NA.

[Gryson]

We have no idea what state the SGI hardware was in when it was shown to Sega. When Nintendo went with it they decided it needed more time in the oven and ended up delaying the N64 until late 1996, which that wasn't an option for Sega. And when we look at the time line for when this was being proposed the Saturn hardware was getting close to being finalized (I believe this was late 1993 when this supposedly happened?). So it was too late to completely change design at that point, and as Gryson has pointed out, they needed to release in 94/95.

As for CPUs from NEC, the SH-2s are the least controversial part of the Saturn's design. It's actually a very good CPU for the time. In many ways it's better than the PS1's CPU in that things like Multiplication and Division instructions take significantly less cycles to complete than they do on the PS1. The reason the PS1 pulls a head in 3D calculations though is because it has the GTE, while Saturn has a 2nd SH2 for that kind of stuff.

Again, I don't think the Saturn's hardware was as bad as people make it out to be. It was powerful enough to be acceptably competitive for it's generation. The real issue was more with how poorly it's launch was handled in the west combined with the mess that was the 32X. Those two things made it start in the hole and it never was able to recover. Poor launch software leads to poor sales. Poor Sales leads to less developer support. Less developer support leads to more poor software. Rinse and repeat.

All good points. Also about the SH-2: We know Sega negotiated a very good deal on these, to the point where Hitachi said they weren't profiting on them.

About the SGI hardware:

I came across this quote from one of Square's FF7 programmers who worked with both the PlayStation and the N64 prototype hardware at SGI:

I kind of had a suspicion that things weren’t going too well for the 64 at that point, because … one of my responsibilities … was to write performance applications that compared how well the 64 fared against the prototype [PlayStation]. And we’d be running parallel comparisons between the [PlayStation] where you’d have a bunch of 2D sprites bouncing off the screen and see how many polygons you could get within a 60th of a second. And even without any kind of texturing or any kind of lighting, it was less than 50% of what you would be able to get out of the [PlayStation]. Of course, the drawback of the [PlayStation] is it didn’t really have a z-buffer, so you’d have these overlapping polygons that you’d have to work around so that you wouldn’t get the shimmering [look]. But on the other hand, there was no way you’d be able to get anything close to what FF7 was doing [on PlayStation] on the 64 at that time.

There was actually this one trip that [Nintendo] organized for me, [main programmer Ken] Narita-san, a few other lead devs who were working on the battle portion for the Final Fantasy 6 [Siggraph] demo at that point. … I think Nintendo had been getting signals from Square saying, you know, “Your hardware isn’t up to snuff. Not only in terms of raw 3D performance, but in terms of storage.” And they said, “We’re gonna fabricate this brand new chip,” which was supposed to have a bunch of hardware improvements to get a little bit more performance. Which, my suspicion is they probably just repeated that verbatim from SGI, and I think there was, in general, a disconnect between SGI and Nintendo in terms of what they were expecting the hardware to do. SGI was probably talking more along theoretical lines of what the hardware would be able to do, and they were trying to make it general purpose so that it wasn’t just a 3D rendering machine. But Nintendo had certain specific performance metrics that had to be met, but I don’t think those were communicated well to SGI.

The wires just — they weren’t in sync there. So they sent us down to Mountain View, and I took all that code I was writing for the Shoshinkai to run on the [latest prototype] hardware there. And it didn’t really change in terms of performance.

This is exactly the reason why Hideki Sato did not want to take the hardware out of their control. Not only were there risks with production scale and deadlines, there was also uncertainty about whether they'd be able to implement changes easily to suit their needs.

[axel]

If you read any of the gaming magazines of that time, you'd be hard pressed to come up with a gaming magazine that didn't have screen shots or games, of news about those upcoming consoles. I read about and saw screenshots for the 32X, Saturn and PlayStation games and hardware a good 6 months to a year before their release. I was interested in the 32X for a hot minute, until a month or 2 later when news about the Saturn and some preliminary gaming screenshots started leaking in the likes of EGM and Gamefan magazines.

Right, it wasn't that upcoming consoles were a total mystery. It was that so much hardware was coming out so quickly it was clear there were going to be winners and losers. Nobody wanted to end up with the next 3DO, Jaguar or 32X.

[TrekkiesUnite118]

As I mentioned above, is it possible they are referring to the Sept. 1995 release date as coming out of nowhere to them? That's something I could see. It seems a lot of these early development projects began around mid-1994, and at that time nothing firm had been decided about when the Saturn would be released in NA.

It's entirely possible, but some specifically mention it being around May. I think probably the story that really emphasizes how disorganized the whole situation was at the time is the Virtua Racing Story. The keypoints from the interview are the following:

1) Time Warner was given it because Sega of America was stretched too thin to do it themselves.
2) They weren't given any source code, just some of the 3D Renders for the tracks and some of the car models.
3) When they asked for code/more info Sega of America just gave them an Arcade Cabinet to study.
4) The devs weren't even aware there was a 32X port being developed to leverage code from and never saw anything from it.
5) They never had their own devkit to use. They had to lease time to use one at Sega of America after hours. Apparently their slot was around Midnight and they had to take their compiled builds into Sega to test.

Basically it explains why the Saturn port of Virtua Racing is so odd. Part of me wonders if Sega of Japan was even aware that this is what was happening with the port.

[Gryson]

It's entirely possible, but some specifically mention it being around May. I think probably the story that really emphasizes how disorganized the whole situation was at the time is the Virtua Racing Story. The keypoints from the interview are the following:

1) Time Warner was given it because Sega of America was stretched too thin to do it themselves.
2) They weren't given any source code, just some of the 3D Renders for the tracks and some of the car models.
3) When they asked for code/more info Sega of America just gave them an Arcade Cabinet to study.
4) The devs weren't even aware there was a 32X port being developed to leverage code from and never saw anything from it.
5) They never had their own devkit to use. They had to lease time to use one at Sega of America after hours. Apparently their slot was around Midnight and they had to take their compiled builds into Sega to test.

Basically it explains why the Saturn port of Virtua Racing is so odd. Part of me wonders if Sega of Japan was even aware that this is what was happening with the port.

I think some important points about game publishing are being misunderstood in the video (I just watched it).

First, Chuck Tolman admits that he has no idea how the game came about. There was a lot of speculation, but we can figure most of this out through knowledge of how the industry works.

Saturn VR was not an outsourced project. Sega would absolutely publish the game themselves if that were the case. Outsourcing development to second party developers was the bread and butter of SOA. I mean, Sega was a publisher. That was where the money was. Everything about this case was different.

Here, Time Warner has clearly licensed the game property from Sega to publish the game themselves (I mean, that's no mystery - it says so in big letters on the title screen). All Sega is getting out of this are licensing fees and royalty fees. There were probably conditions on how the property was handled, as there usually are with licensing deals, which is why Tolman said they occasionally got approval for changes from Sega. But this was not in any way, shape, or form a Sega project.

I strongly suspect Sega did not want to touch Virtua Racing for the Saturn - otherwise, they would have developed it in Japan. The game was too strongly associated with the Genesis and 32X, and it was not as fresh as the racing games they were pushing on the Saturn. But SOA wasn't about to turn down a third party publisher who wanted to take the risk while also boosting the launch library.

It's also no surprise whatsoever that Time Warner would not be given the source code. 1) It was probably incomprehensible for anyone who didn't work on the original team and would have required a lot of support, and more importantly, 2) Sega wanted to protect the source code since it was potentially more valuable than the licensing deal. As far as I know, these kinds of licensing deals never included source code (see, for example, Ghouls 'N Ghosts on the Genesis, licensed from Capcom - no source code was provided). Therefore, the licensing deal only included 3D models.

Also, your statement about the dev kit is not correct. Tolman said that at first they didn't have one and had to use Sega's, but they got one towards the end of 1994.

I think all of this is important to point out because your post and the video suggest Sega didn't know what they were doing. I don't think that's the case at all. This game had nothing to do with Sega. This was 100% a Time Warner project. Sega's name is nowhere on it aside from acknowledgement as property copyright holder.

[Team Andromeda]

Well yeah, if you were reading 10 magazines a month then of course you would have a different view than 99.9% of consumers. There's no way I could keep up with all that.

The only possible evidence I see of MS working on an OS for the Saturn (other than the Dreamcast obviously) was that a couple of Pocket PCs came out circa 1996 running WinCE on SuperH chips. The HP 300LX for example has an SH3 at 44 MHz with 2 MB of RAM. So it would not surprise me if WinCE could have run on the Saturn... but what would have been the point? It was stuff like that that made me question where exactly Sega thought they were going with all this.

I agree a clean break was probably a good idea but releasing so much hardware so close together just isn't feasible, there's a reason no other company tried it. I think we've discussed this in other threads though.

It's not about reading 20 mags, the point was anyone one of them was talking of the Saturn and looking to get any sort of info of it and what it could do, along with the trademark artist impression of what the machine may look like
You had to be living under a rock not to know of the Saturn or that it was coming out reading any major gaming mag in mid-1994. I also don't know how much real truth was in MS helping SEGA or the silly talk of voice controls.

And if not for the 32X, SEGA wouldn't have been releasing main hardware close together, the Mega Drive, Mega CD, Game Gear and Master System were hardly new kids on the block by late 1994 and 1995. Though I've always said in 1994 SEGA should dropped all in-house support for its systems and just focus on the Saturn (and of course the Arcade)

I mean SEGA turned down SGI and they could have worked with them and a CPU could have looked into NEC. They loved Yamaha for sound. SEGA had options for hardware.

NEC had their only system to worry about in 1994 and their system CPU was even more unpowered than a Saturn. It's widely rumoured that Yamaha worked with SEGA on the VDP1/2 chips and that's to overlook any major move to a brand new GPU and CPU would have delayed the project massively and it's not like going late with better spec, helped the Cube or XBox against the PS2.

Sure SEGA could have gone with SGI, but that would have meant a delay to the Hardware and SEGA going in 1996 and by then SONY would have already cleaned up the world over and SEGA going late would then also have to take on Nintendo too. I can't see that working out myself, not that I was ever that impressed with N64 hardware. It had a horrible display and was quite weak when it came to sound, screen res and 60 Fps support, it was said to cost NCL a fortune and NCL couldn't wait to drop its SGL contract and some developers said it development environment was more difficult and more costly, to that of the Saturn.

It's not always as easy to say go late and get a more powerful chip. They weren't many around that could match SONY tech in 1993/4


------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------


I rather just go on what was in SEGA control and budget. The 32X should have been killed and even if SEGA needed or wanted to go early the wrong date was chosen and I can't understand why the May date was picked, when Late July or Aug would have been much better and I still say it was a fatal mistake to release Daytona USA like it was, those graphics was a massive open goal for SONY, never looked to get a Joe Montana game on the system, even it was a quick upgrade from the Mega Drive version with smooth scaling and better speech or looked to get an improved version of Sonic CD on the system early in.

[Leynos]

You do realize the N64 CPU was NEC right?