- 720
- Austria
- ron1986
No, he means that PCars damage is better, but nowhere near to realism. Realism in general.
Gran Turismo 6 vs Gran Turismo 5: Visual Damage Comparison
- Thread starter ToyGTone
- 248 comments
- 20,218 views
- 952
- Jackson Heights, NY
- jamesman22
- jagenigma
To be fair, no game that has come out for the current generation of consoles has been complete or has been a full game without the need to pay for expansions, dlc, (Im looking at you Destiny, what a letdown). The gaming world is becoming crap because companies are trying to please everyone left and right but getting nowhere at the same time. Nothing is good enough for everyone so gaming companies arent gonna try as much as they used to.
- 28,470
- Windsor, Ontario, Canada
- Johnnypenso
Why does there have to be a "winner"? You are free to prefer and use any level of damage your games allow, that's your winner. The rest of us are having a discussion about different games having different damage models. No two games are identical in any regard, damage included. PCars damage model is clearly well ahead of GT's at the moment, no question about it and they may well be up against their contractual limitations. Whether GT7 steps up to the plate or not is up to them but the bar has definitely been raised in driving games on console. You seem to be emotionally invested in the GT series. I am not. I want every game to get better and I'm not afraid to point out what I like and what I don't like. If you can't clearly see that the PCars damage model, even 150 builds ago is superior to GT I don't know what else to tell you. It's as obvious as night and day.
- 13,890
- Adelaide
- Neomone
DLC is not the same, though.
When there's paid DLC, it's understood by the customer that paying the price for the disc gets them X, Y and Z. If they want A and B as well then they have to pay extra. It's simply a modular way of pricing content/features. I don't particularly agree with how many companies do it, but it's obvious up front what you get for your money.
The "incomplete" games are ones where the price you pay for the disc gets you X and Y, and Z at some point in the future. You've already paid for Z with the disc, but until it's actually delivered you don't have all that you paid for in the "base package" of the game.
It's different in the sense that with DLC you haven't paid for the content, although you might feel stiffed that the company didn't include that in the base package. That's just modern marketing bollocks trying to squeeze out every last dollar though. With incomplete games you have paid for the content, you just haven't received it yet.
I disagree.
Smart developers know what they want to make and do so. It's not so different to movies in this respect: one wants their movie to have a wide appeal so that they make as much money as possible, but realistically it's usually better done by making a really good film than by trying to cater to all different types of viewers. A dramatic comedy horror love story kids film pleases no one. A really good kids film that is well written enough that older folks can enjoy it too, like say Toy Story, will sell by the ton.
Good developers pick what they want to do and do it really well, knowing that most gamers will enjoy a really good game even if it's not quite in their genre of interest.
Bad developers try to please everyone, and fail. This is the point where having a strong vision is really important for a game. The Last of Us knows exactly what it's trying to do. Red Dead Redemption knows exactly what it's trying to do. Wipeout knows exactly what it's trying to do.
Is Gran Turismo the Toy Story of console racing games? It once was, it was really common for people with a Playstation to have Gran Turismo 1, even if they weren't really car people. GT6...not so much, unfortunately.
Well most (if not all) of you are comparing some Last gen Video game that already gone to it's PS3 limits to games that's on Ps4, XB1 and PC with different number of developers and a lot less cars...
I have to agree thought because Gran Turismo started to fall down sadly especially with their Broken promises, Better Racing games stepping all and the fact today's gamer's favorite type's games are ether FPS or Some Zombie Survival horror game, No one is really interested in games like GT anymore.
I have to agree thought because Gran Turismo started to fall down sadly especially with their Broken promises, Better Racing games stepping all and the fact today's gamer's favorite type's games are ether FPS or Some Zombie Survival horror game, No one is really interested in games like GT anymore.
Last edited:
- 6,937
- United Kingdom
- 14,436
- Los Angeles, CA
- Lma_robot
- lImaRobotl
What exactly is taking a step back, compared to whats in GT right now?
You're right. If PD tried, they can probably make something that most would be content with.
- 6,937
- United Kingdom
- 14,436
- Los Angeles, CA
- Lma_robot
- lImaRobotl
The thing is, the amount of "impact" needed is just a bit ridiculous. So what you dont like about FM4's damage model is that scratches appear on the sides of the car, on a head on collision, yet GT6 does this too it seems. So I guess its just the wing flying off from a full speed crash, that is bothering you?
Still, I'm glad that FM seems to progress with each iteration. Hopefully GT/PD follow suit.
- 6,937
- United Kingdom
GT approach seems more iterative and do think that GT5 was better regarding visual damage while FM4 seems to have fixed damage models it swaps in. On that video you linked, it does not look as bad as the FM4 video IMO.
PS4 has some serious compute performance so do think that is one area they can have big leaps in if they wanted to.
- 14,436
- Los Angeles, CA
- Lma_robot
- lImaRobotl
The fact is, it does the exact same effect which was your original arguement as to why Fm4's was a step back. It's not exactly a step back anymore is it? It's also showing that GT's model is as "fixed" as FM4's model. Seeing as the Honda Fit will not even come close to achieving the same speeds as that Zonda, I'm wondering if that has to do with it not "looking as bad." Also, hitting a stationary object would sure be less of an impact, compared to two supercars hitting each other head on at full speed.
Even when using a far less dramatic situation, it still causes the same effect. I'm wondering what would happen if we can replicate the situation completely.
There's alot of things that could have played out differently, "if they wanted to."
Last edited:
- 6,937
- United Kingdom
- 14,436
- Los Angeles, CA
- Lma_robot
- lImaRobotl
Going off that quote, it sure does sound nice. Look what that process achieved though. The cars looked like they drove through a lava pit and managed to drive out of it. I'm guessing thats why the eliminated it with GT6.
However, I am excited to see what they produce as well.
- 6,937
- United Kingdom
I think it likely takes quite a bit power to run and GT6 has stuff like Tessellation and MLAA being done on the CPU, maybe it didn't allow them to keep / improve it. Soft-body physics takes a lot of power to run it seems so if they wanted to get it to that level then only PS4 they got some chance of achieving it.
- 14,436
- Los Angeles, CA
- Lma_robot
- lImaRobotl
Thats the problem, they need to find a nice balance instead of pushing something to max and leaving little for everything else. I didnt pick up GT6(or another ps3 for that matter) because it didn't look like much changed, in my opinion. I'm waiting anxiously to see how much change they can bring to the table with this jump in hardware. Maybe it'll be enough to have me foolishly purchase a PS4, when I probably shouldn't.
- 6,937
- United Kingdom
Do think they used the things such as Tessellation and MLAA to use hardware more efficiently but they have obviously quite limited resources especially with such a poor GPU on the PS3.
You can see for example on this regarding cloth simulation how much more power they have regarding compute on PS4:
- 1,148
- United States
Quakebass
Premium
- 1,899
- Washington
- quakebass
Actual deformation of the 3D mesh, for one. Also a greater number of parts that can be knocked loose. Right now GT only uses texture modifications to display damage, and while that's incredibly clever and looks very good, it can only go so far in terms of display of visible damage. PD already did reasonably decent deformation in GT5, it just had too many odd and often glitchy parameters/outcomes, so I'd bet they can likely improve upon that. Considering most premium cars are made out several separate parts, it would seem very possible for premium cars to have the dislodgeable bits. Combining all of these together can look fantastic.
It won't be a perfect simulation of reality, but nothing in GT is. They need to start somewhere.
Last edited:
- 8,722
GT5's method is definitely where PD are going. GT6 is clearly a stopgap on the visual damage front. With a few more cycles allowed for collision handling, different materials can be better represented and extents and self-intersection (clipping, weird floating bits etc.) better handled at the point of impact. The GT5 E3 demo is probably a good representation of what GT7 will bring.
What's weird is that the GT6 model appears to need more memory. All GT5 needed was the mesh to deform in-place and some procedurally applied dirt / scratches (low-res, so kind of dark dirt). GT6 has apparently higher resolution dirt / damage maps, plus they have a tiling per-vertex normal perturbation map, to give the impression of dents in the reflection mapping and other shaders that utilise the surface normals.
(I'm assuming that car models aren't instanced in GT5, because you can't guarantee that every car won't be a different Premium, so you have to budget your memory for that one worst case. That means each car gets its own copy of vertices to displace, and may also explain why replays "duplicated" damage when rewound etc. The "virgin" geometry must be re-streamed back in after the race and at the re-starting of the replay, or at least only the part that describes the vertices' locations).
Both approaches are innovative, possibly unique in their specific combination of implementations, and reflective of PD's willingness to try bespoke solutions.
I didn't see it mentioned, but I've skipped over the bickering a bit, so I may have missed it, that a major advantage of procedural damage modeling is the fact that separate "damaged" states don't have to be modeled in, nor textured up. That's great in a game with so many cars - its also great for subtle variety and hence an impression of detail. If all cars end up with a high-resolution external mesh in the next game, then we could be in business.
I do wonder how on earth PD are going to get that real-time deformation working with the sub-division ("tessellation"), and even the progressive meshes (the pre-computed splitting and merging of vertices for gradual LoD reduction was done based on their original locations, not their displaced locations...) That's two non-trivial problems, as far as I can tell, and both will probably require yet another unique solution.
What's weird is that the GT6 model appears to need more memory. All GT5 needed was the mesh to deform in-place and some procedurally applied dirt / scratches (low-res, so kind of dark dirt). GT6 has apparently higher resolution dirt / damage maps, plus they have a tiling per-vertex normal perturbation map, to give the impression of dents in the reflection mapping and other shaders that utilise the surface normals.
(I'm assuming that car models aren't instanced in GT5, because you can't guarantee that every car won't be a different Premium, so you have to budget your memory for that one worst case. That means each car gets its own copy of vertices to displace, and may also explain why replays "duplicated" damage when rewound etc. The "virgin" geometry must be re-streamed back in after the race and at the re-starting of the replay, or at least only the part that describes the vertices' locations).
Both approaches are innovative, possibly unique in their specific combination of implementations, and reflective of PD's willingness to try bespoke solutions.
I didn't see it mentioned, but I've skipped over the bickering a bit, so I may have missed it, that a major advantage of procedural damage modeling is the fact that separate "damaged" states don't have to be modeled in, nor textured up. That's great in a game with so many cars - its also great for subtle variety and hence an impression of detail. If all cars end up with a high-resolution external mesh in the next game, then we could be in business.
I do wonder how on earth PD are going to get that real-time deformation working with the sub-division ("tessellation"), and even the progressive meshes (the pre-computed splitting and merging of vertices for gradual LoD reduction was done based on their original locations, not their displaced locations...) That's two non-trivial problems, as far as I can tell, and both will probably require yet another unique solution.
- 8,434
- Rockay City
Another non-trivial problem I didn't hear yet (or maybe I didn't look too hard
) is that to have realistic visual damage you also want different bits and pieces of a car to fall off (and interact with the rest of the environment), on top of morphing the meshes/models. Which would require those parts not only to be modelled separately, but also have correct physics properties applied to them to make them fly off (perhaps also break/shatter) and interact with the environment in the correct way. Because of this, you can't really look at visual damage apart from the physical thing IMO.Not sure we would want PD to do that for 1200+ cars BTW.
- 28,470
- Windsor, Ontario, Canada
- Johnnypenso
Aren't the premium cars already modelled with separate body panels and other parts?Another non-trivial problem I didn't hear yet (or maybe I didn't look too hard
Not sure we would want PD to do that for 1200+ cars BTW.
- 8,434
- Rockay City
Yes, and it's a good start, but I think the current parts are limited and only the visible portion is modeled. And how far would we like to see this go? See pistons, rod, cam belts, etc. fly off in a realistic fashion in a crash? But I guess for the first iteration it would be enough to improve the current parts.
- 8,722
The physics properties can be handled procedurally. Assign a material, precompute centre of mass and some 3D representation of MMOI, bingo. The constraints are a little harder, but can realistically be generic in most cases.
The modeling issue is problematic. Hitting an invisible body part because only the backfaces were facing you is a bit rubbish.
Maybe they can super Premiumificate the Premiums at some point. I'd say don't expect bits to fall off consistently.
It's easy enough to spawn masses of random debris, though, like most games do.
The modeling issue is problematic. Hitting an invisible body part because only the backfaces were facing you is a bit rubbish.
Maybe they can super Premiumificate the Premiums at some point. I'd say don't expect bits to fall off consistently.
It's easy enough to spawn masses of random debris, though, like most games do.
