Hi Ethin.
Sounds intriguing. But I'll admit that would take some serious technical knowledge on your part. Just to build a console could cost thousands even millions.
On the other hand if your talking about a virtual console that would run on a computer it could work. But if your going to write this from scratch this could take years especially if it's a solo project. You would have to write a screen reader, a program to play sounds and plenty more things like that. If you don't believe me there's a topic in the developers room where camlorn and I discussed building an operating system.
I do have one question if you built this system what could you offer that hasn't already been offered on windows or mac?

The difficulty/desirability/viability are all in a grand struggle, here.
The simplest way would probably involve a Raspberry Pie and... Linux or something.

So... to answer the question:
1. Yes, I want one.
2. But I don't expect it to work out very well without lots of effort and luck and good promotion. A good platform is tricky to build, but then you need to have games for it, and you need the price to be reasonable. For a simpler example, see Aprone's Se Munkey.
3. Don't ask me for what I want no one else will want it that way bwahahaha (Translation: I get along better with retro style limitations and optimization problems and ridiculous exploits that aren't the least bit human readable. Also I wanna a tactile display. I got my hands on a PS4 controller last week, and discovered it has a touch screen in the middle. And it's not a Senseg touch screen, so bleargh.)

Ok, so let's pretend we all know what we're doing for a minute and completely ignored my warning in my third list item, because now you have me thinking about Accessible Console Optimizations again.

There are multiple layers to think about. There's the surface-level i/o and performance details that make up the user experience; that's the end goal.
Let's ask these questions: what input/output techniques, channels, devices, etc should be supported? Would there be any graphics support at all? How about a WiFi or ethernet connection for online games? Are you going to use an existing audio system to handle the question of stereo/mono/surround sound (5.1? 7.1? Which manufacturer?)/headphones/microphones/etc?
How about controllers? Existing USB devices (where do you get the drivers?) or something unique to this console (seems more like your aim, and in a situation where I know it could be done, I'd say do that one, but reality is not gentle, as mistresses go)? Would you want to do anything unique with their design? Maybe multiple designs, even, to accommodate other disabilities or play styles?
What about haptic feedback? Is there vibration? (Ahem: Immersion sounded cool for this at one point)? How about my idea of "The standard Playstation/XBox controller has two seemingly useless legs; can we put an actuator in both of them, and vibrate in stereo?" Can we put a touch screen in the middle like the PS4, but add Tixels or Bending Wave Actuators for tactile feedback? Can we stick a 4:2 Braille Display in the middle instead? Can we make it support the Novint Falcon?
(Psst: a homebrew device similar to Senseg's Tixels would probably be rather cheap, if lacking all the fine-tuning they've doubtless come up with in their years of development *cough*. Even easier if the touch area can be opaque.)

Ok, enough optimism. Let's worry about the internals.

For something like this, you'll either need to build off an open source OS or something similar (*points to Android and Raspberry Pie*), or learn how to build and include everything you'll need in C/C++ (I'd say Purebasic sounds decent for this, but I know nothing about the licenses or OS-specific dependencies involved)
Audio Games can get away with a lot of corner-cutting in terms of detail. If you want to avoid supporting graphics (or at least, avoid supporting graphics better than those in Swamp), all you have to worry about efficiency-wise is your audio engine.
Custom hardware would be awesome, but impractical, seeing as the single most impressive thing you could do here is build an audio processor to rival modern 3D GPUs, at which point the console is just icing. (Resurrecting the dead 3D audio hardware would be spectacular enough a feat of techno-necromancy that you would immediately be revealed as a terrible wizard.)

Actually, the hardware design for the lowest level stuff here is actually kinda fun to think about. Mostly because there's probably a correct answer and I don't know it so get to compare what I know about today's devices to what I know about 20-30 year old devices.

Ok, ok, this post is a long mess, and I need to go pretend I know anything about the processor-level stuff. So let's focus on the middle level: the part where game developers work. If games need to be designed for the console specifically, that presents several issues, but opens up the option of a nice high-level API to simplify access to the hardware.
On the other hand, you could just write bindings to the major programming languages out there for any specific hardware, and let nature take its course.

I like the "all custom, all the time" idea, but it's ridiculously impractical.
The opposite end--just wrap an open source OS like Android and throw in accessibility features--is kinda boring and pointless without the i/o being especially fun (see: the Braille2go for an example of something similar getting stuck in development hell).

OK, I'ma go try to rewrite all this and actually think it through more and hopefully come back with something more coherent.

(Ah, missed where you said you were building on Debian. Well, and I was working on my previous post when post 3 came in. Anyway, hopefully some of this makes sense? I think I'm jumping the gun by a lot.)

OK, take 2, with headings:

Haptics

What I want is a super awesome haptic feedback controller that manages to retain the basic controller shape that's been around since PC and PS1 took the SNES controller and made it twice as thick. Or possibly an alternate scheme with a touch output device and a controller that can function one-handed. Or, even more insane, a tactile display on top of something like a braille keyboard (as in, most of the buttons ina  row for easy switching between viewing and input). These are horribly impractical and I think I might be a minority or a minority in this regard.
I don't know if stereo vibration is practical in a non-clunky controller. I'd like to see it attempted, if it hasn't been already. It has the advantage of being pretty straightforward and intuitive compared to the other haptics techniques no one has managed to commercialize that I want to play with. But practicality, man.
I don't know what experience you have with this sort of thing, but I get the impression that jumping into something innovative on the haptics end would be crazy enough to derail the whole thing, so practically speaking, it should probably stay minimal for now. If this happens and you can stick with it, though, I'd at least suggest setting it up so that the possibility of supporting arbitrary haptic devices is available in case of future products.

I could summarize the i/o methods that I know about, if desired, but I don't want to get too far off topic.

Audio/video

Would this be intended as a TV-based device? Handheld? Other?
There are loads of audio output devices: traditional line-out headphones, USB headphones, random speakers, surround sound systems, televisions, anything intermediate to a television (I have no idea how these work anymore; I haven't had to hook up a console that way on anything newer than a VCR). All of these can have any number of channels. Heck, you might even find a use for supporting MIDI devices (disclaimer: is probably joking about that one. Probably.)
So, how many ports does it have? USB, line-in/line-out (microphone support?), HDMI? ... Dare I say, RF?
How about video support? If you can have at least enough graphics support to be interesting to the sighted, it might be more popular. I don't mean in the sense that sighted gamers would be especially into it; just being able to play with sighted friends and family would make it more attractive to more players.
(under-the-hood possibility: if there are any libraries for game developers to use, and there is graphics support, some defaults that draw basic shapes appropriate to the library would be better than nothing.)

(Might you include some sort of MIDI driver, with support for higher quality sound fonts? VST? RAM and disk space are cheaper now than they were 20 years ago, but if you're building physical products on the budget you can get from Kickstarter and hoping to get them built and distributed, you'll find that everyone likes to pay less. Might be irrelevant if said features wind up taking up more resources than recorded music, without resorting to lower quality synths.)
(Ooh, on that note, synthesizers. Because everyone likes to play with synthesizers! ... Hardware Synthesizers? Hmm.)

I'm imagining the ideal setup being similar to mainstream consoles, maybe with better headphone support. So the most interesting details in this regard are more under-the-hood.

Overall Concerns

I pause here, to reiterate the market problem. This would be cool in theory, but in order for it to work, there need to be plenty of games, and plenty of people to play them, otherwise you built a cool toy that will spend most of its time sitting on a shelf as a monument to your ability to build cool junk.
Kickstarter is awesome because it is a clever commitment device: "I want to support this project, but only if it succeeds, so instead of giving you my money, I promise to give you my money only if enough people likewise promise that the project is at least fully funded."
Game Consoles need to overcome the same problem: they need developers to create games for the console, but the developers need there to be an audience. The latest Playstation / XBox / Nintendo console are guaranteed enough of an audience that game publishers can start playing market politics by picking which platform they like most. A console designed for blind people by an individual, or even a couple people, doesn't get that luxury. So, for this project to become reality, you need to line up loads of incentives: backers need to know that you will develop, produce, and distribute the console, and that there will be enough games for it to keep it active; developers need to know that it will get made and distributed, and that there will be enough gamers to warrant the effort; and you need backers, gamers, and developers to support you (and for your hardware efforts to succeed, preferably quickly).
That requires lots of things to go right. I'd still like to see it happen, but the hard part isn't likely to be building a prototype.

Back to the hardware...

It seems best to arrange the processors and their components based on operations that need to be done quickly and in parallel.
A possibility would be to dedicate a core (or in certain cases, a dedicated processor) to a specific category: one for each audio, video, controller i/o, and other game-related processing. In especially complex cases, multithreading for AI mightn't hurt.
This seems like overkill. Are we going to need a high-end GPU? Probably not, but does that mean it should be avoided? Eh...
If you're making a blind people console, though, it seems like a given that the audio system needs to be impressive. To that end, dedicating a core/processor to audio seems like an obvious step.

The complexity needed for processing controllers depends entirely on how the controllers are made. Let's assume we're going with something modern, and that we want to keep track of what's going on with minimal latency. I don't see this taking up a whole core by a longshot, but I'd still feel better giving it as little interference as possible. If touch screens, or complex haptics are involved, they might require more complicated on-the-fly calculations, so I'd lump all those into one. But if you want to stop with a generic controller, probably leave it easy to offload game-specific processes into that one.

So that just leaves the processor for game instructions.

Ok, that's great and all, but then comes the part of going from a game (loaded from... disk? USB? Does this thing come with a hard drive? SSD? SD Card? Internet connection?) to instructions for the processor(s). ... Hm, good luck with that.

Summary

If you're not just attaching some fun periferals to a Raspberry Pie, things get kinda crazy.
- Making something like this for a community, rather than just yourself, involves loads of coordination problems. Devs, gamers, financial backers, and your own abilities to make the hardware happen all need to know the others are going to do their part, in order for it to happen.
- I think video support is a good idea, even if it isn't "OMG COD12 looks so real I'm going to shoot my friends every day until the next COD comes out!"-quality.
- Audio output is complicated. Don't try to build it from scratch, if you can avoid doing so.
- I want more haptics than you! [1][2][3][4]
- I have no idea what the interface for devs would look like, but I like a mix of ridiculously low level with ridiculously high level.

With the ps4 including tts in its next update, I think this would be an interesting project and I would definitely follow it... but I have doubts as to the games that will be available and how big of a market there would be. I will help in whatever ways I can though, if this does indeed get off the ground!

Well, I need enough money to purchase a Raspberry Pi. Its only about $130.00, so that should be plenty. I don't need the $10000.00 unless you actually want me to build that. Thanks for the good suggestions, though. I'll speak to aprone and see what he thinks.

Hi, its actually been a while since I thought of this project, because I abandond it quite a while ago as a plane waist of time, and a pure nightmair to get it working. The idea originally started from techmaster20, which I at first ignored but then thought about giving it a try, going so far as to at least look into software development for it. Their are menny problems with this however. I don't have the money to even get the parts for such a console, nore the technical knowhow to actually assemble it and make sure the processer doesn't explode in my face, lol. Secondly, while in therey, I could easyly see how this could be done using a linux operating system or kernel, after intensive thought and investigation, in practice it wouldn't be possible. You may as well make a minny computer and be done with it instead, which is part of the reason it was abandond. @guitarman when I said sound panning wasn't on linux, I ment for pure basic, not for every audio component lol. Its also been suggested I consider ios, I could actually easyly see a snow race mobile working on it, but again, the ishue here is the lack of resources, the lack of a mac to develop it on at least that's what i'm assuming and the fact that cross platform is more difficult than it sounds. I know, mason and I have tried it with dmnb and dm 2, and it was a massive fail on so menny levels we eventually said screw it and gave it up as a waist of time.

Hello.
Hmmm, well I don't know what a raspberry pie is but all this talk about it has me interested. Does it have accessibility features? I will look into this for my computer collection.
@Danny, thanks for setting me straight on that. I looked at the language reference and sure enough sound panning isn't supported on linux which is pretty silly. You know I have a mac. If you wrote code for a game to run on mac or iOs I could compile and test it with x-code. Just let me know the offers always open.
@Ethin, you might also want to talk to aprone about graphics support. Since he is sighted you would have someone to test the visual part of the system. If this system could run a game that is graphical but also an audio game you would get a lot of sighted people interested in the project.
@Simba, well there's stuff going around that sony is developing a screen reader for the ps4 so we'll see what happens. I wouldn't urge microsoft to do anything only because I don't think they would actually do anything to help the blind. If you don't believe me look at the pile of crap that is narrator lol.

I think the last raspberry pi I tried was about $130.00. I'll check again.
Yes, with a Quad Core CPU, 1GB RAM, 900MHz Board Clock Speed, 40 GPIO Pins, 4 x USB Ports, 4 Pole Stereo Output, HDMI Port, 10/100 Ethernet, and Micro SD Card Slot, its $30.00; add the extras, and we get a total of about $144.94. This includes the RASPBERRY-MODB+-512M; RPI NOIR CAMERA BOARD; ENCLOSURE, BLACK, RPI B PLUS; SAMSUNG MEMORY, MICROSD, NOOBS, JAVA, 8GB, B+ COMPATIBLE; MCM AC/DC CONVERTER, 11W, 5V, 2.2A; and PIFACE
IO EXPANSION BOARD, RASPBERRY PI.

Ah. Okay, ,I thought you were talking about the pi itself, not the many other components you'd need.

Okay, I have a number of comments on this subject so I'll try and address as many as I have time to address at the moment.

First, as for the issue of panning sounds on Linux it is absolutely possible. It all depends on what programming language and audio API the developer happens to be using.. I have been able to successfully pan sounds using SDL Mixer, FMOD Ex, and OpenAL so I know for certain it is possible. I also have written some demos in Java using its sound mixer and it also can pan sounds. So its not any restriction of the OS but whatever language and API someone happens to be using. If someone finds they can not pan sounds on Linux then they need to look for different languages and APIs since they are using the wrong tools for the job.

As for a game console for the blind I think it is an interesting concept in theory, but foresee a number of technical issues in implementing such a thing and bringing it to market. As has been mentioned I do think something like a Raspberry Pie system might be a good platform to begin building such a console on, but realize that means one will have to rely on game APIs like SDL for building games. While SDL is doable its nowhere in the same class as DirectX or XNA so I can foresee someone perhaps wanting to build something more advanced if they want to create better than average games for the platform. However, as far as languages goes there is C++, C#, Python, Perl, Ruby, and of course Java so it would be fairly flexible in terms of programming languages goes.

My biggest concern is attracting new developers to your console. Most audio game developers as you know are Windows-only users and developers. Far too many only program using Windows specific languages like Visual Basic 6 or BGT which clearly won't work on a Raspberry Pie system. They have to rewrite their games from scratch in a language they probably don't already know using APIs they will have to learn. So attracting developers could be problematic.

Then, there is the subject of research and development. Obviously, you will have to buy the hardware, develop any modifications you want to the OS, and probably develop some sort of game engine or tools to help ease development on your console. I just don't see any one person or even a group of people really following through with the time it will take to get this conceptual project off the ground. Even if you succeed then you are still going to have to charge enough money to make the research and development worthwhile as well as sell the hardware to the end users. So I'm not sure it will be very affordable for either the developers or the end users. It seems to me like it would be a costly project with little reward for the people who create this console.

I could certainly give it a go, though. And if it fails? That will jus encourage me to try harder to make even better products. I intend to use a raspberry pie to actually begin product development, though.

Hi,
Unfortunately, the raspberry pie does not have Windows, and I can't install it on the SD Card by default. I'll have to purchase one capable of holding windows, plus a way to restrict users from, say, browsing the internet to only playing games. This is a gaming console, after all. Not a full-fledged computer.