*spinning* Next-Gen Discs

Well... I am in a dicotomy here, too... I usually don't mind video per se... But I do HATE CGI trailers... it's like a movie trailer, which only shows B-Roll footage, not in the movie. It's great for "PR", but nothing else. The Star Wars 1313 guys touched upon this topic, too, recently. Either make the in-game videos look like the game (or vice versa), or don't bother. The same should (imho) apply to trailers, too.
 
I think a good rule of thumb is that assets will scale proportionally to memory if we assume everything else in the system is perfectly balanced (the large chunk of memory reserved for the OS doesn't count). So my guess is that 128GB, which we are likely to have, should be fine for next gen. FMV doesn't scale up at all, in fact it should be halved once they use H265. We also should need less and less baked light because finally we're moving toward real light engines.

Here's what I found on the cost of producing BDROM: An additional layer causes a 5% drop in yield, a few hundred dollars for the additional master glass (replaced every 10000 disks), a few cents more per disk, time is raised to 4.5 seconds versus 4 seconds for a single layer. So with this, I expect 4 layers BDXL to be practical and economical at launch, and I also expect the console's drive to be compatible with much more layers, so that 250GB+ games can still be done later on if needed.

I wish they would do at least the MLR upgrade because we really need the bandwidth increase, and it's a very simple upgrade with very little cost to the drive. But going beyond 128GB, there's a significant problem with both HoloDisks and the proposed upgrades to bluray (any of them). The metallized glass stamper needs a big increase in precision and would degrade very fast. They might use the same "technique", but unlike BDXL, they need new machines and new stamping and mastering technologies, it's basically proportional to to increase in data density, so they are all stuck with the same problem. The holographic principle requires a resolution of 10 times the wavelength, MLR for bluray would have the same pitch but it still requires 3 times the "pit length" precision, and it would degrade fast. All of that only for the very few games that would require more than 128GB? Why take the risk? What justifies the expense? If Microsoft wants to get rid of optical drives, they sure won't spend a lot on the drive next gen, they'd be in an even worse situation than PS3 was with bluray.
 
trailers based on in-game cutscenes that use the engine can be misleading too, you get dialogue, animations, viewing angles and actions you will find nowhere else in the game ; you can stage explosions and carefully load your scene with lots of lights and stuff.
so if the cutscene looks like a movie, and then you're dropped in a mediocre 3rd person beat'em'up that feels like rehashed PS2 shovelware, you've been ripped off by the trailer.
 
No, not really - the max is to work at 2x texture resolution, but they don't paint in any extra detail that wouldn't be visible after the downsize.

The geometry assets they're using for the normal map detail are also completely unfit for in-engine use, and any other use, really - we had to rebuild everything on our current project. Impossible to UV map, rig, sometimes even to render.

For nextgen, they'll have to build more detailed game models that'll still be 'low poly' and paint more details into the textures. It's not 'removing quality removal' at all. Also, expectations will increase in stuff like quality of facial animation and secondary deformations like muscles and cloth/hair dynamics, so assets will have to be higher quality in technical terms as well.

Completely unique virtual texturing is another level of additional work on top of that.

Thanks for the lesson Laa-Yosh


I still think nand is the way to go next gen. Prices have dropped like a rock with 128 gig drives going for as low as $40 after rebates (yes they are older generations) and even newer gen drives going pretty cheap , i just got a 256 gig for $150 and its putting up scores with the fastest drives out there.

Obviously you wouldn't use such fast nand or as expensive of a controller but i think nand is certianly becoming something thats possible in the upcoming gen as a replacement for discs
 
I was re-reading info about Mega Meshes from Lionhead.

So, in summary. Mega meshes give a more artist centric approach to modelling Detail only where you need it. It’s pretty tolerant to late changes. The representation we have here maps really well onto surface parameterisations,and therefore current console hardware. Ideally, I would have liked to have used displacement maps, but I never got it running quick enough on the xbox to be viable. Recently, however, I’ve been doing some tests using displacement maps in direct X 11 and got some very exciting results here, so I think the tech has good prospects for future hardware. Even for voxel octrees, the pipeline we’ve developed is one way of addressing the modelling requirements

A Microsoft owned first party studio seems to be on the similar wavelength as John Carmack and trying to let artists be more free with content creation.
 
one big question is, can a holo disc be "pressed"?
In theory injection molding can go to 10s of nm, so you could press an embossed hologram and read it out with a ccd for increased data density ... this however would be fundamentally different from the holographic discs such as GE's, which use analog recorded volume holograms.
 
OT - question I had a dvd a few montsh ago from the video shop, now it wouldnt play, I tried it in my old pc/dvd player & I was able to rip it. but it was like ~13gb IIRC

from wikipedia
4.7 GB (single-sided, single-layer – common)
8.5–8.7 GB (single-sided, double-layer)
9.4 GB (double-sided, single-layer)
17.08 GB (double-sided, double-layer – rare)

now this wasnt doublesided i.e. I didnt have to turn it over, so how is this possible?
 
Back
Top