Bioshock catch-all thread: Art-Deco style

Pages

The Cult of Rapture just put up a video showing off how water's modeled, which looks much like one can see in the latest Silent Hunter game. Now if you'll excuse me, I have to go drain the lizard.

It's pretty, but I'm sort of surprised the splashing from the pipe (maybe 2/5 of the way in) isn't better done.

I'm nit-picking, though.

URL for the video

I can't wait for this game. I think my favourite water effect showcased there is the rippling and flowing of the water. I like the effect of a swathe of water flowing across a slope or down steps.

I think 2K Games is getting hammered from this video. They are timing out or failing to respond. Good to know there is a ton of interest in this game.

I like the audio for it, as well. I have a feeling I'm going to need to go pee a LOT while playing this game.

The coolest water feature is that when you set Splicers on fire they run for it to put themselves out, and then get up PISSED.

wordsmythe wrote:

It's pretty, but I'm sort of surprised the splashing from the pipe (maybe 2/5 of the way in) isn't better done.

True, though keep in mind the words "WORK IN PROGRESS" that are visible during the entire demo.

It's pretty, but I'm sort of surprised the splashing from the pipe (maybe 2/5 of the way in) isn't better done.

Let's look at the bright side of things.

If this game was developed by a tech wizard like John Carmack, it would show off his next-gen, spellbinding particle effects using DX10 shaders.

The game would take place in a mysterious abandoned city, buried deep in a sea of kerosene. All your weapons would be replaced with a flamethrower, a lighter, and a box of matches soaked in kerosene.

shihonage wrote:
It's pretty, but I'm sort of surprised the splashing from the pipe (maybe 2/5 of the way in) isn't better done.

Let's look at the bright side of things.

If this game was developed by a tech wizard like John Carmack, it would show off his next-gen, spellbinding particle effects using DX10 shaders.

The game would take place in a mysterious abandoned city, buried deep in a sea of kerosene. All your weapons would be replaced with a flamethrower, a lighter, and a box of matches soaked in kerosene.

Don't forget the Big Daddy closets. Boo!

Ya, there's something "not right" with the splashing from the pipe, but I'm sure they'll have it sorted out. And even if that's what it's going to look like, it's still some fantastic looking water so what do I care. The first few minutes will probably be Oooing and Ahhing over the water effects and then hopefully I'll be sucked into the story and gameplay and it will just be a part of the world like everything else.

Even the high-end guys have trouble rendering water properly. Liquids are very difficult to model, requiring unbelievable horsepower to do. It's not just flexing polygons, it's arbitrarily complex outlines that need a brand-new set of properly-textured and lit polygons to represent them every single frame.

That pipe splashing does look pretty crappy. They're just using particles to roughly approximate water. But if they were actually trying to render a real stream with anything resembling fidelity, your machine would grind to a halt.

Most of the stuff they're doing here is really clever work with textures, I suspect some form of procedural bump mapping/vertex shaders. Damn, does it look good. But as soon as you get into actual droplets of water, it breaks down, because home computers just don't have the oomph to do it right.

I'm wondering how good the framerates are gonna be. We may need the next generation of cards after the 8800/x2900 to render this stuff properly.

Malor wrote:

Most of the stuff they're doing here is really clever work with textures, I suspect some form of procedural bump mapping/vertex shaders. Damn, does it look good. But as soon as you get into actual droplets of water, it breaks down, because home computers just don't have the oomph to do it right.

Agreed; I'd be bold and say that the standing/"ambient" water sections are simply animated using displacement maps and such like, to give the illussion of a shifting surface. The water flowing down the steps, too, makes impressive use of specialist shaders along with more traditional texture offsets.

Fluid dynamics will always be really hard to replicate with (effectively) just a bunch of sticks, but I think the programmers and artists working on BioShock are definitely making some advances on it. Personally, I think it looks fantastic.

Wrestlevania wrote:

Agreed; I'd be bold and say that the standing/"ambient" water sections are simply animated using displacement maps and such like, to give the illussion of a shifting surface. The water flowing down the steps, too, makes impressive use of specialist shaders along with more traditional texture offsets.

Yeah, it's really not much different than FEAR using normal-mapped decals to have the illusion of semi-destructable environments...

Fluid dynamics will always be really hard to replicate with (effectively) just a bunch of sticks, but I think the programmers and artists working on BioShock are definitely making some advances on it. Personally, I think it looks fantastic.

Personally i can't wait until people start taking advantage of advanced liquid and cloth physics simulations like those on the physX card. I'd buy one of the new PCI-e versions if i thought it would be worth it in the next 6 months, but i give it until the end of 2008 before a game comes along that really uses these things in game. Just before anyone points out the physics engines on graphics cards, i think they're overrated and aren't much good beyond Havok and particle effects.

Well, that's my future prediction anyway...

Here's some REAL water simulation:

Scanline Flowline VFX.

Writeup on how some of it is done.

(both these links via Metafilter.)

Duoae wrote:

Personally i can't wait until people start taking advantage of advanced liquid and cloth physics simulations like those on the physX card. I'd buy one of the new PCI-e versions if i thought it would be worth it in the next 6 months, but i give it until the end of 2008 before a game comes along that really uses these things in game. Just before anyone points out the physics engines on graphics cards, i think they're overrated and aren't much good beyond Havok and particle effects.

IMHO, PhysX are just kicking around waiting to be bought by someone like Nvidia or ATI. And once their free "games" (large demo's) start turning heads, I think interest will begin to peak quite quickly.

So I'll take your prediction and run on with it; I think every games-orientated computer system will contain a dedicated physics chip by the next round of consoles (2009-11?).

Wrestlevania wrote:

So I'll take your prediction and run on with it; I think every games-orientated computer system will contain a dedicated physics chip by the next round of consoles (2009-11?).

I'd agree with that, but the current generation has shown us that their direction in this blooming multi-core era is more of a "get lots of processing cores, and have each one handle a different task" approach. The technology didn't quite bear out, but recall that the PS3's Cell setup was originally meant to handle everything, with no additional graphics chip necessary. The idea being that some cores handle graphics, others handle number-crunching, others handle sound, etc. The 360 takes a similar hybrid approach in that it still has a graphics chip, but six cores dedicated to different tasks. One (or at least part of one) is reserved for OS-level (blade, music player, etc) regardless of game, movie, etc. being played at a given time. By the next generation of consoles, I foresee at least an order of magnitude more cores available for processing; a few can handle physics, a few more handle graphics, and so on. Once multi-core programming really hits its stride, throwing more cores (rather than more gigahertz) at a task will result in the desired performance.

I agree Wrestlevania.

Baggachipz wrote:

The 360 takes a similar hybrid approach in that it still has a graphics chip, but six cores dedicated to different tasks.

The 360? You mean the PS3? but that has one main core and 6 SPU's.... the 360 has 3 cores. I'm confused.

By the next generation of consoles, I foresee at least an order of magnitude more cores available for processing; a few can handle physics, a few more handle graphics, and so on. Once multi-core programming really hits its stride, throwing more cores (rather than more gigahertz) at a task will result in the desired performance.

I disagree. We've barely scratched the surface of multicore programming (parallel stuff) and you think that in 6 years time we'll be able to double the number of cores and therefore the headache that programmers are already having to deal with? No way.

Also, as Cell has proven. You can't have a mutlicore CPU able to do different jobs. Cell is good for high throughput number crunching only (as the SPU's are very limited), the 360's PPC is also limited in the sense that it's still a general computing architecture.

My prediction for the next gen of consoles? AMD/Nvidia are working on the CPU/GPU hybrids right now. In the strictest sense they are not multicore with respect to their separate CPU/GPU architectures but the processing of information will be much quicker due to the proximity and available bandwidth of the respective components. These will be the mainstay of the Xbox 720 and PS4 if not specifically developed by those two companies.

Not to finally mention chip yields and wafer size. Currently, Cell is an expensive chip because of its complexity and size. I don't see how wasting even more chips around the edge of the wafer is going to make the business any better. From a cost point of view, it doesn't work.

What's killing Sony, financially, is the Blu-Ray drive. Cell alone might almost make sense, if they could push enough units, but with the Blu-Ray too, it's breaking them.

Worse, they're investing hundreds of millions in building things that are just sitting in warehouses.

Yeah, that's also true. But the xbox 360 isn't burdened with a next gen drive and, even with the cheap parts, they're still losing money on each box... I expect that the GPU's are a better cost effective part than their 3-core PPC chip. Though that is my completely unfounded guess

Actually, early cost estimates considered the GPU to be the most expensive piece of hardware in the 360.

Fair enough, though i was talking about cost effectiveness per wafer ie, how many chips are wasted in the process...

Do you know what the estimated cost of Nvidia's part is in the PS3?

One report, released in late 2006, saw RSX at around $130 (Cell was $90), which, if true, would sound severly overpriced given that the chipset is based on technology Nvidia was using two years ago (GF6800 series). As far as we know, Microsoft was never really happy with NVs pricing for the Xbox GPU either. One of the reasons that made them switch to ATI, I guess. Of course, Sony was in dire need of some GPU.

The RSX is comparable with the 7800 not the 6800 - if memory serves.
*Checks*
In fact it does:

http://www.theinquirer.net/default.a...

http://www.ps3forums.com/archive/ind...

Yeah, Nvidia have done well out of their deals with MS and SONY. Both times because Nvidia was in a position of having the other company over a barrel because they needed a component... quickly.

Whoops, I meant to type 7800, of course. After all, it was the G70 to be launched roughly two years ago.

Yeah, it's amazing how fast time goes...it feels like only last year.

Part of me wants to upgrade when i can afford it so i can play bioshock with really nice settings... but the other part of me (the sensible part) tells me that my 3GHz P4, 6800 will run the game fine - if not at a great resolution and with all the bells and whistles...

Perhaps that money would be better spent on living... or a 360

It looks like this game will mean goodbye to my trusty old 6600GT. Nevermind, 'tis but a small sacrifice to gods of Bioshock.

Duoae wrote:

I agree Wrestlevania.

Baggachipz wrote:

The 360 takes a similar hybrid approach in that it still has a graphics chip, but six cores dedicated to different tasks.

The 360? You mean the PS3? but that has one main core and 6 SPU's.... the 360 has 3 cores. I'm confused.

No, I meant the 360. It has 3 cores, running 2 threads each. Think of the "hyperthreading technology" intel used before their multi-core line. The OS actually sees two distinct processors, even though it's only one chip running two separate threads. Therefore, to the game programmer, the 360 has essentially six cores to work with.

I wrote:

By the next generation of consoles, I foresee at least an order of magnitude more cores available for processing; a few can handle physics, a few more handle graphics, and so on. Once multi-core programming really hits its stride, throwing more cores (rather than more gigahertz) at a task will result in the desired performance.

I disagree. We've barely scratched the surface of multicore programming (parallel stuff) and you think that in 6 years time we'll be able to double the number of cores and therefore the headache that programmers are already having to deal with? No way.

Yes, we have *barely* scratched the surface of multi-core programming. That's because the "old" way of programming involved throwing more cycles of a single cpu at a task. Those cycles are in one big bucket and the more you dedicate to one task, the fewer cycles you can dedicate to any other task in a given time window. The "new" way of programming games is clearly multi-core. Now, there are several buckets of cycles available, and the dedication of one or more cores to a task doesn't necessarily impact another task, if things are done right. This is a huge shift in thinking and methodology for programmers. Once the shift is made, however, it's a much more scalable approach. Like I said, throw more cores at tasks as they become available, without detriment to other tasks. To your point, look at how the power of processors has exponentially increased over time. Apply that trend instead to the number of cores available, and that's where I get my prediction. Simply doubling the number of cores in six years in comparison to what's available today is actually well below what I predict. Yes, each individual core will get a speed boost as well, but as cores can shrink and be packed into the same space available, and production becomes cheaper, expect the number of hardware threads to double more than once before the next consoles hit.

Also, as Cell has proven. You can't have a mutlicore CPU able to do different jobs. Cell is good for high throughput number crunching only (as the SPU's are very limited), the 360's PPC is also limited in the sense that it's still a general computing architecture.

The original stated goal was to have a collection of general-purpose cores that could be tasked with anything. That didn't bear out, due to yield issues and other wrinkles. If you ask me, this is simply because the cell idea is in its first incarnation, and perhaps a bit before its time. The shift in methodology can't be easily made, and the resulting hybrid designs used by both the 360 and the PS3 are there to bridge the gap.

My prediction for the next gen of consoles? AMD/Nvidia are working on the CPU/GPU hybrids right now. In the strictest sense they are not multicore with respect to their separate CPU/GPU architectures but the processing of information will be much quicker due to the proximity and available bandwidth of the respective components. These will be the mainstay of the Xbox 720 and PS4 if not specifically developed by those two companies.

Not to finally mention chip yields and wafer size. Currently, Cell is an expensive chip because of its complexity and size. I don't see how wasting even more chips around the edge of the wafer is going to make the business any better. From a cost point of view, it doesn't work.

Regardless of who's making the chips, the trend is clearly multi-core. From the 10000-foot view, doesn't it make more sense to have one chip with lots of cores, doing everything, than one chip doing some tasks and one or more specialty processors doing other things? From a production standpoint, it's much cheaper and easier to have a vast army of identical clones able to do everything, than it is to have a group of specialists that have to be clumsily integrated and shoehorned into a space.

I've derailed enough, we'll see where we end up soon enough. At least my predictions are now on record.

Here's a 7 min gameplay montage video from GT.

Looks interesting. I love the visual style, but hate the character design (except for the big daddies.) Did they hire American McGee when I wasn't looking?

Besides that, looks great. I was REALLY hoping that the game would have a bit more of a Myst-like serene vibe to it, at least now and again, but it looks as though it's going for the non-stop power-scare thing. Not a criticism, but the first demo they gave really got my hopes up for something a little more unique in atmosphere.

I feel the same way, but this could also be just a gameplay video of the action inside the game. So not all is lost yet.

The game needs some crosshairs.

Quintin_Stone wrote:

The game needs some crosshairs.

NO.

Okay, maybe an option.

Pages