Why did Nuon fail?

[mattley]

.....But lay off the badgers

Bit harsh on the ones who've devoted their whole working lives.....


But I think it might just have been a good idea that didn't quite work. One of the things that may have gone against it is that if you say to a manufacturer, this is the cheapest DVD chipset and Oh look it's got a games engine built in, the first instinct of the (blinkered, after all this was over spec and that makes buying departments uncomfortable) buyer is gonna be OK how cheap can you do it without the games bit.

[Dave]

I think it was really a demand issue. Here were some of the problems:

1. It wasn't a gaming system.

Everyone who wanted to play video games had a PlayStation, N64, or Dreamcast at the time. Nuon offered nothing new to this market except for a couple games, only two or three of them any good (Tempest 3000, Iron Soldier 2, and Merlin Racing).

2. Special DVD features were a waist.

The Monkey Bone DVD and a couple others had special Nuon-only features! Why wouldn't you want that? Because they were pointless and studios didn't want to sped extra dough developing Nuon-enhanced discs when only 1% of 1% of their purchasers would have benefited from them. Mass-merchants like movie studios wanted to sell discs to the highest number of people, meaning aiming their offerings at the lowest common denominator.

3. Little 3rd/2nd/1st party support.

I never once saw an ad for a Nuon on TV. The first time I saw one in a store was at Best Buy when they were on clearance. That's when I bought mine.

4. Pricing/Launch Issues

Already covered in other posts.

5. 99% of people just want a simple DVD player.

The best selling DVD player in the U.S. for the past several years has been the bottom-of-the-line Apex DVD player (now around $40.00 new). It costs nothing, has four buttons on its front (Play, Stop, Reject, Power), and my grandmother can use it. In fact, she does! It plays movies easier than a VCR (no re-winding), and they look great on your standard piece-of-crap television that everyone has. VERY few people have HDTV's and home theatre setups.

6. When DVD players were hot, PS2 and Xbox had them built in.

This kind of put the nails on the coffin. When PS2 and Xbox came out, game systems had DVD players built into them for almost free; there was no need to buy another DVD player, let alone one that's a gaming system too.



If VM Labs somehow inked deals with every DVD-distributor in the world to include the Nuon chipset in their players, the format may have had a slight chance. Instead, this never happened, and the Nuon failed miserably. Kudos to the VM Labs crew trying the nifty Trojan horse approach! Unfortunately, very few DVD makers let that horse into their kingdoms.

[Riff]

From a developer's point of view I would point out that the hardware lacks an RGB framebuffer (which they had, but took out!), and only has 4K of local code memory to work with and not enough bandwidth.

You can claim that it has no graphics hardware, but it does. There are four media processors. One processor runs the main game code and the other three processors *are* the graphics hardware. The Aries architecture supports bilinear addressing, color indexing and mipmapping in hardware as well as z-buffer comparisons and format-conversion in the DMA controller. The instruction set allows 128-bit vector addition, subtraction, and 16x16 vector multiplies and dot products. The rendering library code acts as the firmware. The Nuon is not lacking in this area.

The platform is awesome for 2D based games where there isn't a lot of per-pixel processing. The problem is they didn't want any of those games. The graphics libraries are also far from optimized. MML2D is the slowest, most convoluted mishmash of code *ever* and while the other libraries are decent, they all attempt to work for all pixel modes in every configuration imaginable. The best performing library is the sprite library due to its concentration on 32-bit mode. Too bad VMLabs was too busy turning away anyone would wanted to do 2D games! The GCC compiler is also guilty of churning out awful code. It works, but it doesn't parallelize worth a crap. If you want fast code, you're forced to use assembly.

The Aries3 might have done a lot towards saving the Nuon, had it been available sooner. The Aries3 nearly quadrupled the amount of local memory per processor and the clock speed was doubled to 108 Mhz. Good, but not great. It is still best at power games heavy on graphics, low on processing. The processor have a huge amount of instruction 'bandwidth' (seven instructions per cycle), but the resource units are so limited that most of the time the use efficiency is near 12% (one instruction per cycle). If I re-worked the processor, I would have replaced the ALU and MUL units such that they could both perform ALU and MUL instructions. This would have significantly improved execution of game code.

So, in my opinion, their downfall was a combination of hardware deficiencies and bad business decisions. If you're a no-name company, you don't have enough power to convince developers to work with hardware that takes a large amount of effort or radically different programming methodology in order to get it to do what you want at an acceptable speed. You most certainly do not have enough power to turn away offers from people willing to develop games for your machine if your existing software library count is zero.

[Skah T]

Much could be said about the technical aspects of the chip (with hindsight being 20/20).

However I believe the main problem was the market crash in late 2000. That dried up nearly all possible funding and caused the slow spiral death that killed the company. (In fact, the bankruptcy proceedings for VML are *still* going on to this day. I get periodic mailings from the lawyer firm handling it.)

We had countless meetings where we'd be told of a number of prospects for a few million here or there, but hardly any of them panned out. What we needed was one investor to come in and pony up a large chunk of cash. That would have triggered a bunch of other smaller investors to kick in money (no one wants to invest a small amount when they're the only one doing it).

Without money we couldn't woo developers, do the desired amount of advertising, or get new OEMs on board.

So yeah, we could say that the VML made mistakes in design or marketing, but the heart of the matter is running out of cash. The crash happened too early in the company's life cycle.

One could argue that decisions could have been made earlier to mitigate the effect of the crash (for example, sacrificing control to get more VC funding), but remember this was a time when money was flowing out of everywhere and Yahoo had a stock price over 100.

[Snakey]

I think it was really a demand issue. Here were some of the problems:

I think, as some other people did, you are missing the point here.

They were not trying to sell it to people who wanted a gaming system. They were trying to make it a standard feature of DVD players. The idea being that once everyone had one of these, they then might be interested in buying software for it.

Same goes for special DVD features. No, nobody is going to go out and buy a new player for this stuff. But if their player can do it already, then maybe they'll buy the "enhanced" DVD.

Same goes for advertising. None needed. Most players would include Nuon, because Nuon was cheap. That also covers the point about "most people just want a simple DVD player". These players would also include Nuon - again, because it was the cheapest DVD chipset there was. It didn't matter if the user wasn't going to use it. It was still cheaper for the manufacturer, and the user - and VML still got paid for it.

As for PS2 and Xbox - well, they were much later.

As Mattley said, and from what the VML guy told me, it really did seem like a good idea, that should have worked - but for some reason, didn't.

[EDIT] - ooh - Skah-T - you worked there ? What did you do there ?

[Skah T]

They were not trying to sell it to people who wanted a gaming system. They were trying to make it a standard feature of DVD players. The idea being that once everyone had one of these, they then might be interested in buying software for it.

Makes perfect sense to me (still does in fact). However it's a right pain to explain it to consumers and OEMs so that they get it. They hear games and immediately start comparing it to consoles. Very frustrating.

The best example I can think up is to ask the person what their parents or other non-gaming relative would buy for a DVD player. Would they get a PS2? No, they'd get a dedicated DVD player. *That's* the consumer that we wanted.

ooh - Skah-T - you worked there ? What did you do there ?

Snakey, I was on the hardware side. I joined after Aries1 was mostly done so I wasn't responsible for designing anything on the chip. I mostly helped with verification. Towards the end I did some design work with another guy on a new block that was going to be part of Aries4 or 5.

[Dave]

Note: I said Monkey Bone was a Nuon DVD in my last post. That's wrong. I meant to say Bedazzled, both horrible Brendan Fraser movies.

As Mattley said, and from what the VML guy told me, it really did seem like a good idea, that should have worked - but for some reason, didn't.

My 5th point is the main reason it failed. Nobody wanted/needed one. They were trying to push a souped up DVD format that wasn't necessary. DVD's already did what people wanted; movies with better picture quality and sound. Nuon DVD's simply added extra special features that were not necessary. It essentially made DVD menu's/features look a bit better. Most people who watch DVD's today don't even bother to look at the special features/menu's anyways. They want to simply push play and watch.

Money was a huge issue. If you throw money at something, it'll eventually become successful (Xbox, cough, cough). The question remains, would have it become profitable? Possibly. If it was funded for several more years, hard-core DVD nuts may have considered it being their number one choice for a DVD player. But this required one important factor: Movie studio's would have had to support the format. This would have required extra money sticking Nuon content into their DVDs. If less than 1% of their customers care about this, they're not going to spend a heck of a lot. Your top selling DVD's today, selling for under $10 in Wal-Mart's bargain bin (yes, Wal-Mart has a bargain bin ), don't even have any extra content at all. Studios simply rip the film to DVD's, press the discs, and sell them. You're lucky to even see commentary on them.

The idea was certainly neat and I don't want to talk down to the guys/gals at VM Labs. I believed in it when I first heard about it circa 1996-1997. I loved their Mario Killer image that ran on Next-Generation's cover. In hindsight, you can nitpick the little faults of just about anything. The Trojan horse idea was cool and might have worked, but we all know the end result. I have and like my DVD-N501 quite a lot! Other than the horrid arrow selector on the remote (which I swear will break within the next year), it's a great little DVD player. On top of that, T3K rocks!

[Snakey]

My 5th point is the main reason it failed. Nobody wanted/needed one.

The Trojan horse idea was cool and might have worked

See, thats the point right there. Nobody wanted/needed one - not important. VML didn't expect anybody to go out and buy a player for the Nuon capability. Get all manufacturers to put it in anyway, because it cost them less money to do so - and bingo. Everyone suddenly has one.

So it wasn't because nobody wanted one. As Skah T says, it was because manufacturers didn't seem interested, and funding was lacking. Bit crap really, you'd think they'd jump at the chance to cut costs.

As for people not putting Nuon content on DVDs - again not important. It was intended as a free, extra feature in the player. No content, doesn't matter. Nuon still there, VML still getting paid. As more players DID have the chip in then it would make more sense to add features. And of course if these boxes ended up in the homes of 40% or more of DVD owners, more software would have appeared.

[dr_memory]

So it wasn't because nobody wanted one. As Skah T says, it was because manufacturers didn't seem interested, and funding was lacking. Bit crap really, you'd think they'd jump at the chance to cut costs.

Well, I think the key thing here is that choosing the Nuon chipset ended up not being a cost-cutting choice! Nuon got caught by Moore's law: by the time the final silicon was actually available, every far-eastern boxmover and their cousin (*cough* Apex *cough*) had figured out that a $5 AMD, Intel or ARM chip would do the job just as well.

[Dave]

So it wasn't because nobody wanted one. As Skah T says, it was because manufacturers didn't seem interested, and funding was lacking. Bit crap really, you'd think they'd jump at the chance to cut costs.

Well, I think the key thing here is that choosing the Nuon chipset ended up not being a cost-cutting choice! Nuon got caught by Moore's law: by the time the final silicon was actually available, every far-eastern boxmover and their cousin (*cough* Apex *cough*) had figured out that a $5 AMD, Intel or ARM chip would do the job just as well.

That's pretty much everything right there. Sticking Nuon chips into DVD players isn't free, and no manufacturer was going to increase the production costs of their DVD players. This would have led to higher prices just to add some chip that added no marginal value to the DVD player. Customers were not willing to pay extra for a feature that they didn't see a use for. If they demanded it, it would have been an entirely different story.