Upgraded 360 with Natal (Rumor)

[ninelven]

I can believe in a new MS console for Fall 2010. It makes sense for MS to leverage DX11 as early as possible. With adequate RAM, PC ports/co-development shouldn't be too much trouble.

Six cores seems a bit much for the CPU, but I suppose it's possible. Another IBM based CPU makes sense, but I wouldn't be too surprised to see a quad core Athlon II 32nm (it would be about the same die size as the current CPU in Jasper and make cross platform development with the PC almost trivial).

On the other hand, 1GB of ram would be a bit of a letdown IMO; 1.5-2GB would be more appropriate. The GPU could easily offer an 8x performance increase, which is really where it's at for the next gen consoles.

I'm guessing Natal would continue to be an add-on for the 720/whatever.

 

[AzBat]

SeattlePI Microsoft blog debunks rumor...

http://blog.seattlepi.com/microsoft/archives/171274.asp?source=rss

TeamXbox goes further...

Today, I can tell you that I have obtained the first details on the next Xbox (info that we'll reveal in the coming weeks) and that's why we can confirm that the rumor reported by 1UP is false. Microsoft is NOT launching a new, upgraded Xbox 360 disguised as a new Xbox in 2010 (although built-in Natal technology is a possibility for a new edition of the Xbox 360, similar to the Elite or Arcade flavors).

It's not our intention to tease you, but for now the only thing I'm going to say about everything I heard regarding the next Xbox is that it won't launch until a certain type of television becomes more widespread because... (be sure to check out our Twitter account to learn more)

http://news.teamxbox.com/xbox/19960/Rumor-New-Xbox-Console-in-2010-We-Tell-You-the-Truth/

Looks like we might be getting some news sooner rather than later.

Tommy McClain

 

[manux]

http://news.teamxbox.com/xbox/19960/Rumor-New-Xbox-Console-in-2010-We-Tell-You-the-Truth/

Looks like we might be getting some news sooner rather than later.

Tommy McClain

Based on teamxbox it sounds much like microsoft is betting on 3d display devices and hdmi 1.4 for next xbox. If blu-ray 3d is on track that should mean big push for 3d in late 2010. Ofcourse that would be too early for microsoft as they need bigger install base of tv's to actually have people to sell the console to.

I would much rather see a solution where single console renders 2d image and user could buy a second console to enable 3d mode... Though I guess smart hardware and programming can go to great lengths in reducing performance hit from rendering scene twice.

the only thing I'm going to say about everything I heard regarding the next Xbox is that it won't launch until a certain type of television becomes more widespread because...

edit. Sony displayed earlier this year 3d gran turismo 5 which was rendered using several ps3's. If that mode was enabled for users owning multiple ps3's and 3d display device that would be super tempting(assuming displays would be available in decent price range)

 

[liolio]

Ms would be crazy to limit themselves to "3D TV" owners. By the time their next system launch the only certainty is that at last all owners play at least on "hd ready" display.

 

[manux]

Ms would be crazy to limit themselves to "3D TV" owners. By the time their next system launch the only certainty is that at last all owners play at least on "hd ready" display.

Why would it be 3d displays only? Isn't it pretty straightforward to support both regular and 3d display... though 3d could be the big gimmick why to upgrade from existing console or buy nextbox and not pc or ps4(assuming they wouldn't have same capability).

 

[Malibu]

Yeah, it will be similar to this generation. HDTV's weren't widely adopted (and still aren't), but one isn't required nor needed to enjoy the PS3 and 360.

 

[liolio]

3D would be great that for sure but it's likely that each frame would have to be rendered twice so whatever the improvments in IQ would be for the vast majority of the users an half would be lost.
(not exactely an half but a lot anyway and for a user base likely way tinier than what were HD tv users at the begining of this gen).

 

[manux]

3D would be great that for sure but it's likely that each frame would have to be rendered twice so whatever the improvments in IQ would be for the vast majority of the users an half would be lost.
(not exactely an half but a lot anyway and for a user base likely way tinier than what were HD tv users at the begining of this gen).

Most likely smart tricks, techniques and hardware can be used to make the performance hit a lot less. And memory hit isn't that bad either if we assume 720p or maximum 1080p framebuffer. Though performance was the reason why I proposed using a second console for generating the second image. That way even current gen consoles could output 3d image(as proven by the gran turismo earlier this year).

I think the additional rendering power used for for second image could be used on 2d use case to do ssaa/temporary aa/ to improve image quality for traditional display devices.

 

[upnorthsox]

we can confirm that the rumor reported by 1UP is false. Microsoft is NOT launching a new, upgraded Xbox 360 disguised as a new Xbox in 2010 (although built-in Natal technology is a possibility for a new edition of the Xbox 360, similar to the Elite or Arcade flavors).

Well, that's just saying that MS won't Wii-ify their next gen console. I thought that was a given. They do hint at a builtin natal box which is the real updated 360 in disguise.

 

[Nesh]

Most likely smart tricks, techniques and hardware can be used to make the performance hit a lot less. And memory hit isn't that bad either if we assume 720p or maximum 1080p framebuffer. Though performance was the reason why I proposed using a second console for generating the second image. That way even current gen consoles could output 3d image(as proven by the gran turismo earlier this year).

I think the additional rendering power used for for second image could be used on 2d use case to do ssaa/temporary aa/ to improve image quality for traditional display devices.

The stereoscopic 3D Sony demonstrated with GT and Wipeout actuallu run on two consoles? Or was it one console outputing the stereoscopic image?

 

[manux]

The stereoscopic 3D Sony demonstrated with GT and Wipeout actuallu run on two consoles? Or was it one console outputing the stereoscopic image?

I think it was running on two consoles, both rendering separate images. Polyphony also demoed super resolution in which the image was rendered using 4 playstation3 consoles(each rendered 1/4th of screen). And I think they also had the super fps demo where 4 playstation3's rendered separate image to achieve 240fps.

 

[liolio]

Most likely smart tricks, techniques and hardware can be used to make the performance hit a lot less. And memory hit isn't that bad either if we assume 720p or maximum 1080p framebuffer. Though performance was the reason why I proposed using a second console for generating the second image. That way even current gen consoles could output 3d image(as proven by the gran turismo earlier this year).

I think the additional rendering power used for for second image could be used on 2d use case to do ssaa/temporary aa/ to improve image quality for traditional display devices.

In regard to rendering cost, my understanding is that is basically twice as costly. On the cpu side it's less.
Two consoles? Well actually if the system launchs pretty late (2013) them MS could pack quiet some power for cheap in their next project actually I wouldn't be shocked if they ask users having enough money to have top of the class 3D tv to pay twice 250/300 $/euros. Thus they would have to design the thing properly I mean in a way that the combination of to system would be elegant.

 

[Crossbar]

Well, that's just saying that MS won't Wii-ify their next gen console. I thought that was a given. They do hint at a builtin natal box which is the real updated 360 in disguise.

From where is your quote?

 

[manux]

In regard to rendering cost, my understanding is that is basically twice as costly. On the cpu side it's less.
Two consoles? Well actually if the system launchs pretty late (2013) them MS could pack quiet some power for cheap in their next project actually I wouldn't be shocked if they ask users having enough money to have top of the class 3D tv to pay twice 250/300 $/euros. Thus they would have to design the thing properly I mean in a way that the combination of to system would be elegant.

My experience with 3d movies(imax 3d and some local movie theater with barco 3d-projectors) leads me to believe that 3d effect works best with gigantic screen. Regular size displays such as mine(2.5m viewing distance, 50") would be too small for proper immersion. I believe buying a decent 3d display actually means buying a projector or insanely expensive BIG tv, bigger than what one would consider comfortable for 2d movie viewing. From usual or small screen the 3d effect seems really stupid as the 3d just pops from inside a postcard. In my opinion the 3d display has to cover whole field of vision, even the edges where eye usually doesn't focus. This also means content has to be authored so that main thing is on the center and there is stuff on the edges that you usually don't need to care about, it is just there to keep up the immersion.

Hence I would propose give a normal console system to regular people. Those who are willing to invest proper big 3d display device would be willing to buy a second console as it's not such a big cost in the grand scheme of things. Last I checked proper 3d projector was around 30ke. Once big 3d displays and projectors are cheaper there would be time for new revision of console that has dual display chips inside.

 

[RudeCurve]

Thus they would have to design the thing properly I mean in a way that the combination of to system would be elegant.

That would be easy, design them like 1U blades and just stack them.

 

[Shifty Geezer]

In regard to rendering cost, my understanding is that is basically twice as costly.

Yes, it's the same scene viewed from two different positions, exactly the same cost as rendering two different frames with the player moving slightly between them.

 

[AzBat]

Microsoft, I don't want to have to buy a new tv again, just give me this...

http://www.youtube.com/watch?v=Jd3-eiid-Uw

Thanks.

Tommy McClain

 

[manux]

Yes, it's the same scene viewed from two different positions, exactly the same cost as rendering two different frames with the player moving slightly between them.

Prequisite, I assume unified shader architecture inside box.

The question becomes, how much of the work done for the first eye can be reused when rendering what second eye sees. Camera position is only slightly different for each eye. Ofcourse the same z-buffer doesn't work so visibility needs to be somewhat calculated again but perhaps parts of the visibility information can also be reused between frames and hence more saved work. We might for example be able to determine while rendering first image that some object is definitely occluded for both eyes and not even try to render it for second eye. I would expect that performance wouldn't drop to half except in some kind of silly worst case scenario and no optimizations.

I would assume also part of the gpu time is spent calculating stuff like physics and animations and that part is calculated only once per frame and hence hit for gpu is less than 50% of the time gpu time spent calculating the whole scene.

In theory if we would say that 20% of gpu time is used to calculate physics and vertex work that can be reused between frames that would then mean 80% of time is used for rendering stuff to screen, hence the performance drop would be 40% when adding stereoscopic rendering. If we would assume we can reuse some visibility information and so on the hit to frame could go quite a bit lower... Even to something like only 20-30% performance hit for stereoscopic rendering.

edit. It might also be possible to do something like rendering both images at the same time tile by tile and use caches optimally(perhaps). Dunno, haven't really thought about this enough.

 

[Shifty Geezer]

Prequisite, I assume unified shader architecture inside box.

The question becomes, how much of the work done for the first eye can be reused when rendering what second eye sees. Camera position is only slightly different for each eye.

I suppose yes, as the scene isn't changing at all, the models, textures etc. can be taken as exactly the same. Scene-wide optimisations could be reused in principle, and you'd just have the grunt work of rendering the triangles. I dunno what percentage of rendering a scene that is, but I imagine it's a lot more than 20-30% that you'd need to double up on.

 

[AzBat]

Here's a funny, but interesting idea that G4TV had in response to TeamXbox's new rumor...

But as long as we’re tossing rumors around, here’s one for you: In addition to hands-free motion-control support, the next Xbox will also support Microsoft’s Surface technology with your TV being the main screen and the Surface display as your secondary screen, sort of like a living-room sized version of the Nintendo DS. Now we play the waiting game to see whose rumor is right.

http://g4tv.com/thefeed/blog/post/6...w-Xbox-360-Have-Been-Refuted-And-Revised.html

Don't know how plausible, but it's interesting. Probably be way too expensive to get traction though.

Tommy McClain