brain_stew said:
The actual LCD panel used is most likely just a standard 800x480 LCD panel and even if it isn't, an 800x240 screen is hardly low end for a ~3.5" inch screen and its sheer obscurity will make it more expensive.
If it's 800x480 then Nintendo is ripping people off by not adding a mode to let games actually use the full resolution. I could see them doing that for a previous handheld where the GPU is closely coupled to scanline generation but not here.
Nothing being used in the 3DS is obscure by virtue of it being in the 3DS. That's an expected order of at least tens of millions, that's probably more than enough to offset obscurity tax.
I wouldn't say the iPod has a "significantly better GPU" the results speak for themselves and the PICA200 in the 3DS is no slouch, heck including dual ARM11s probably isn't any less expensive than a single cortexA8 (it isn't 1ghz in the iPod but clockspeeds are irrelevant to costing).
It's going to be less expensive if they didn't include L2 cache. Not including L2 would be another good reason to go with 2x266MHz instead of one of something with a much higher clock, because mobile CPUs don't scale well into those ranges without L2, at least for typical mobile RAM latencies.
The GPUs seem pretty different and hard to compare. Here Nintendo is getting an extreme benefit of having game developers trying much harder on it, and have less platform overhead going about.
brain_stew said:
The iPod probably does have a higher BOM than the 3DS (but not by the huge degree some seem to suggest, its a device that is cost reduced in damn near every area compared to the iPhone) and with that you could say its $230 pricepoint is a much better deal but at the end of the day all of this is academic anyway.
I too think that it's exaggerated how much they're saving vs Apple - Nintendo is likely paying premiums on the screen and is shaving smaller amounts in areas they could probably afford not to in (for their asking price). Namely CPU, RAM, and on-board flash.
But I don't think Apple is pricing nearly as "aggressively" as you do either; why would they? The perceived cost of iPhone w/o contract is misleading. What is iPod Touch really competing against that forces low margins? It's making App Store revenue, but probably not nearly as much as Nintendo makes from software and otherwise it's just stealing iPhone sales. Apple has probably got this stuff down cheaper than we realize, but I imagine Nintendo does as well.
brain_stew said:
Consumers don't just value by raw specs and how expensive the BOM of any given device is. If the 3DS can offer $250 worth of portable entertainment then it'll be successful.
Maybe. There's a threshold for how much people are willing to pay for gaming platforms. Correlate PS3 sales vs PS3 price. This is especially true for Nintendo handhelds which have a much lower average target audience age, meaning a lot of them are being bought by parents. If it costs too much the parents just aren't going to buy it, period, no matter how amazing it looks.
brain_stew said:
The perfect example of this is Apples other portable device, the iPhone. That thing has more than a 300% margin yet it sells by the bucketload because it offers something that consumers want.
The comparisons with iPhone's "price" seem totally unfair to me. Most people are buying it as subsidized through a carrier; we could argue what the equivalent value is here but the real point is that I don't think Apple is selling to carriers at prices nearly as high as they're being sold standalone. These prices are probably marked up artificially to get people to buy subsidized.
Gotcha.
By the way, I'm assuming software emulation isn't happening with the 3DS- so does is it include the DS hardware? ( would that be powering the lower screen?)
It's possible, or it could be something in between. I imagine that it has to at least have the entire DS 2D and 3D hardware, maybe with some RAM blocks reused from somewhere else for the VRAM. That stuff can't be done very accurately on a GPU like PICA200 and the CPUs aren't powerful enough to do it in software.
The DS ARM9 code could be running directly on one of the ARM11s, likely with some glue logic supplying the DS address space and ARM9 visible peripherals. ARM7 code could be ran on the other ARM11, possibly with some code modifications making less glue logic necessary. DS games only used a few different ARM7 binaries, which the 3DS could store modified versions of on its flash without taking up too much space.