Your thoughts about "standard" vs. "premium"

  • Thread starter LP670-4 SV
  • 10,183 comments
  • 862,621 views

What would you have rather had PD do about "premium" vs. "standard" cars

  • Keep everthing the same

    Votes: 324 19.1%
  • Release the game later with all the cars "premium"

    Votes: 213 12.6%
  • Not do "premium" cars at all but focus on other features i.e. dynamic weather

    Votes: 134 7.9%
  • DLC packs after the release

    Votes: 844 49.8%
  • Wished PD didn't get are hopes up, lol

    Votes: 180 10.6%

  • Total voters
    1,695
Some staindard look bad especially in close-ups but it is fine while playing and looking at in replay as long there is no zoom. But sadly some gfx glitches in GT5 is main problem as far looks are concerned. It distract otherwise a photo realistic game

Exactly..

What bothers me more than the graphical quality of the standard cars is the fact we can't even customize the wheels on them! :banghead: Its just ridiculous! I really wanted to see a Veyron with Samba-Bus wheels :dopey:
 
Indeed I can confirm that a large number of GT4 cars still used Alpha maps for things, It really rather depended on the modeler who built the car and where he decided to use the polys, this is why some cars have it and some cars don't. Simply because one modeler chose to do it that way... and the other guy who was working on a different car chose not to.

I noticed the track repeating as well when I looked at that pic.

Interestingly I believe that pic was "screen capped" from a high res video capture rather then an actual Photo-Mode output. So maybe its an artifact of the motion blur effect used when your actually watching the replay rather then the more high-tech motion blur when used in photo mode.

If it is the mo-blur from in-game then I doubt you would notice anything when watching it.

I'll take your estimate of 2hours over my 30-mins, just for sake of argument, since I'm pretty sure 2 hours is more than enough time to tidy up those bitmaps to at least a presentable level and, hell, maybe subdivide the mesh a couple of times while you're at it. One guy, working 8hour shifts, monday to friday for 5 years could do 5200 cars by my reckoning. Given that the modellers and texture artists weren't doing any engine coding or sound modelling or any of that malarkey, they should have been on the job since day 1. So once again - what the hell have they been doing for the last five years?

Looking at some of the screenshots and taking it as read that they are genuine GT5 caps and the jaggies on the car bodywork are not caused by some video capture effect, the whole thing smacks of - those cars were added in a real hurry at the last minute just to try and cram in as much content as they could.

I've always said KY is a visionary and a real inspirational figure in the gaming industry but he's out of touch with the real world. This whole game is starting to look like a cautionary tale of aiming too high. PD needs a realist put in charge for any subsequent releases. Someone who will say "Okay that sounds great but come back next month with evidence that it's achievable or we'll have to write it in for the next release"

I'm pretty sure if they'd had someone like that in charge then GT5 would have been a launch title, albeit minus a couple of cars that only the PS4 or 5 could show off in their full glory. Personally I could have lived with that.
 
I'm mildly disappointed with some of the screenshots of premiums on Taxigamer :ill: I can't even read the dash on some of the interiors they've shown.

http://www.taxigamer.com/gameplay/sony-playstation-3/gt5-gameplay

5194962672_772f5b55a3_o.jpg

5194363017_e1f78c5fee_o.jpg

5193085685_0ff21b00dd_o.jpg


I hope its just artifacting from screen capture and the actual dashes are clear and readable. Has anyone mentioned what resolution GT5 will be running at?
 
i think that updates should either introduce new cars to the game or upgrade some from standard to premium that way we can all be excited when new updates arrive.:yuck:
 
What the hell have they been doing for the last five years?

I couldn't agree more. If this was a PC title I would be the first person in the mod community to get to work on updating the standard cars but alas that can't happen here. :yuck:

My take on it is that for the vast majority of the production (say 3-4 years) they were intending on building all the cars to premium quality, so there would be no point in having staff working on updating old cars that were just going to be thrown away. Once it came to the crunch to get GT5 ready and the decision was made to port over the GT4 cars maybe there simply wasn't the manpower available to put them to work on them... But they could have taught the office intern how to do it and let him work his way through them lol.

Who knows what really happened. It is annoying to see something that I could rectify myself with my average 3D modeling/texturing skills though.
 
I couldn't agree more. If this was a PC title I would be the first person in the mod community to get to work on updating the standard cars but alas that can't happen here. :yuck:

My take on it is that for the vast majority of the production (say 3-4 years) they were intending on building all the cars to premium quality, so there would be no point in having staff working on updating old cars that were just going to be thrown away. Once it came to the crunch to get GT5 ready and the decision was made to port over the GT4 cars maybe there simply wasn't the manpower available to put them to work on them... But they could have taught the office intern how to do it and let him work his way through them lol.

Who knows what really happened. It is annoying to see something that I could rectify myself with my average 3D modeling/texturing skills though.

Exactly! And someone must have known, after a couple of weeks or maybe months, just how long the car modelling was going to take. All you need is a calculator. 200 hand modelled (what the hell is with that lunacy anyway?) premium cars = 5 years work. Any project leader with the barest semblence of a functional central nervous system would have known that pretty much from the outset. So 1000 cars is going to take you a quarter of a century.

At that point you either order your team to quit pissing about hand modelling light fittings and get on with something productive or you hand the teaboy a copy of photoshop and a drive full of lowres UV maps and tell him to get on with it. Either way you aint trying to fob off 800-odd zx81-quality gfx assets as "next-gen gaming"
 
I couldn't agree more. If this was a PC title I would be the first person in the mod community to get to work on updating the standard cars but alas that can't happen here. :yuck:

My take on it is that for the vast majority of the production (say 3-4 years) they were intending on building all the cars to premium quality, so there would be no point in having staff working on updating old cars that were just going to be thrown away. Once it came to the crunch to get GT5 ready and the decision was made to port over the GT4 cars maybe there simply wasn't the manpower available to put them to work on them... But they could have taught the office intern how to do it and let him work his way through them lol.

Who knows what really happened. It is annoying to see something that I could rectify myself with my average 3D modeling/texturing skills though.

No it was redcided from the beginning. At TGS some 3-4 years back there is evidence of it also if I remember correctly.
 
No it was redcided from the beginning. At TGS some 3-4 years back there is evidence of it also if I remember correctly.

That's what I thought too but then I seem to recall that what they said was a little nebulous something like there would be multiple classes of cars or something but that they hadn't decided on the divide yet... so could have been this, could have been they were shooting for all newly modeled/premiums at the time but were going to give special care to some...

I was pretty certain at one point we were going to see race cars with damage and production cars without based on what PD had said.
 
Just watch the GTHD videos from 4-5 years ago. They already included standard and premium cars. GTHD was dropped but the concept remained.
 
I hate to pollute a GT5 forum with anything Forza related but here is an interesting video which shows how I personally think PD should have modeled their cars.

http://www.youtube.com/watch?v=F1a5ipjAHNQ

I don't believe PD did it this way, instead choosing to laser scan the cars, and then build a model to that laser scanned template. Which takes allot longer! The tape and Position Sensing arm used by Forza seems like a much faster way to be honest, I mean your inputting the polygons right there! a day later you have a perfect car model!
 
I hate to pollute a GT5 forum with anything Forza related but here is an interesting video which shows how I personally think PD should have modeled their cars.

http://www.youtube.com/watch?v=F1a5ipjAHNQ

I don't believe PD did it this way, instead choosing to laser scan the cars, and then build a model to that laser scanned template. Which takes allot longer! The tape and Position Sensing arm used by Forza seems like a much faster way to be honest, I mean your inputting the polygons right there! a day later you have a perfect car model!




Forza has bad gfx but photomode is good
 
I hate to pollute a GT5 forum with anything Forza related but here is an interesting video which shows how I personally think PD should have modeled their cars.

http://www.youtube.com/watch?v=F1a5ipjAHNQ

I don't believe PD did it this way, instead choosing to laser scan the cars, and then build a model to that laser scanned template. Which takes allot longer! The tape and Position Sensing arm used by Forza seems like a much faster way to be honest, I mean your inputting the polygons right there! a day later you have a perfect car model!

I'll tell you an even quicker way - just ask the nice guys at the car companies for the bleeding CAD files. 1000 cars - 10 mins work, 10,000 cars - pretty much the same. Why spend all that time and effort reinventing the wheel when the cars were already modelled in 3D, years before they were even built?
 
Exactly! 200 hand modelled (what the hell is with that lunacy anyway?) premium cars = 5 years work. Any project leader with the barest semblence of a functional central nervous system would have known that pretty much from the outset. So 1000 cars is going to take you a quarter of a century.

It only took them 3 years model 200 cars.
 
I hate to pollute a GT5 forum with anything Forza related but here is an interesting video which shows how I personally think PD should have modeled their cars.

http://www.youtube.com/watch?v=F1a5ipjAHNQ

I don't believe PD did it this way, instead choosing to laser scan the cars, and then build a model to that laser scanned template. Which takes allot longer! The tape and Position Sensing arm used by Forza seems like a much faster way to be honest, I mean your inputting the polygons right there! a day later you have a perfect car model!

PD laser scanned cars? From how long it took them I assumed they hand drew them from pictures :P

Laser scanning will give you a bunch of points... a LOT of points. You can turn that into a model and then throw a polygon mesh over it. There's programs which can then automatically throw an optimised mesh over a 3D model to get the minimum number of polygons required to represent the model in the detail you require.

According to this video it took them 35 minutes to scan the part to get a mesh from it. The mesh looks a few thousand polygons large, so even though a car is a lot bigger, PD wouldn't need that level of detail.
http://www.youtube.com/watch?v=gvvw2DEfqEw&feature=related

I'll tell you an even quicker way - just ask the nice guys at the car companies for the bleeding CAD files. 1000 cars - 10 mins work, 10,000 cars - pretty much the same. Why spend all that time and effort reinventing the wheel when the cars were already modelled in 3D, years before they were even built?

That would be nice, though manufacturers really dont like giving out their CAD files. Maybe they might give them to PD, but I was working for a race team and we were sponsored by Honda, they gave us free parts and rebuilds of engines... but they wouldn't give us a CAD model so we could get all the mounting points and check clearances... we had to get it 3D scanned ourselves. Didn't take too long, expensive though. We were planning on swapping to another engine manufacturer (to save weight) and only one or two (dont remember who) were willing to give us any form of CAD... they'd give us free/heavily discounted parts and engines in exchange for us advertising them on our car and trailer, but no CAD :P
 
Last edited:
Forza has bad gfx but photomode is good

You are confusing intentional reduction in detail during in game with a flaw in a modeling system.

If the model exists in good quality ANYWHERE then the modeling method was good.

When you see a low LoD model used (I believe that jaggy headlight turned out to be an AI car way in the background somewhere that got zoomed in on like the current GT5 ugly miata shot probably is) it does not reflect on the modeling process, only on the choice made to manage resources at that point in game.
 



Forza has bad gfx but photomode is good

I wasn't talking specifically about Forza's graphics... I was talking more about how they create the 3D models for the cars.

I'll tell you an even quicker way - just ask the nice guys at the car companies for the bleeding CAD files. 1000 cars - 10 mins work, 10,000 cars - pretty much the same. Why spend all that time and effort reinventing the wheel when the cars were already modelled in 3D, years before they were even built?

Unfortunatly its not quite that simple. CAD files cannot easily be translated into low-ish polly meshes that aren't horribly inefficient. You end up with a very high polly mesh which is pretty much the same as what you end up with when you laser scan a car.. And as you know its ALLOT harder to make a polly mesh less complex than it is to make it more complex (meshsmoothing etc) so the car surfacing technique seems to be the optimal technique for creating cars.. you get pollys exactly where they should be IRL, you end up with a mesh that isn't ridiculously dense but also retains good efficiency and at if the mesh you create is too "low-poly" you just whack a mesh-smooth iteration on it and bam you have a damn fine looking car model. I have done a bit of that type of surfacing in my past as well as laser scanning and I have always found it allot easier with the manual points model.

If you really want to cover both bases.. Laser scan the car at the same time!
 
For fear of being banned, I won't post the image, but I can back-up Scotracer's point that it is an in-replay image of an AI Cooper on the lowest LoD, rather than the image a player sees during a race intro.

I had posted it just before Jordan warned everyone and some got banned, so I won't be posting it again.
 
No. It was a shorter, up and down affair either side of Times Square in FM2 and 3. It went in and along the side of Central Park in FM1 if I remember, similar to how it was in GT4.
 
I wasn't talking specifically about Forza's graphics... I was talking more about how they create the 3D models for the cars.



Unfortunatly its not quite that simple. CAD files cannot easily be translated into low-ish polly meshes that aren't horribly inefficient. You end up with a very high polly mesh which is pretty much the same as what you end up with when you laser scan a car.. And as you know its ALLOT harder to make a polly mesh less complex than it is to make it more complex (meshsmoothing etc) so the car surfacing technique seems to be the optimal technique for creating cars.. you get pollys exactly where they should be IRL, you end up with a mesh that isn't ridiculously dense but also retains good efficiency and at if the mesh you create is too "low-poly" you just whack a mesh-smooth iteration on it and bam you have a damn fine looking car model. I have done a bit of that type of surfacing in my past as well as laser scanning and I have always found it allot easier with the manual points model.

If you really want to cover both bases.. Laser scan the car at the same time!

It aint hard to make a high poly model into a low poly one at all. Just retopo the bloody thing. This is industry standard procedure now for any game model - sculpt it in high poly with umpteen million polys then bake your normals and wrap it onto a your low poly. With gran turismo you wouldn't even need to do the texture baking - it'd actually be quicker.

The dude that did the enemies and guards for the new Assasins Creed just posted a bunch of test renders at zbrush forums - took him and one other guy 3 weeks, Hipoly model and tex, plus retopo to 4-6k polys for in game. It's a standard workflow these days.

With Cars it's a bit different cos they already exist in real world so the choice is scan, remesh and hand model all the bits the scan missed or import the autocad files, remesh and the job's done.
 
That is NOT Forza 3 in-game. It was done by taking an image of a AI car in photo-mode (i.e. low LOD).

I took this picture of Forza 3 photomode today. It's ****ing gorgeous:

What about in moving replay's? In game to him I think he means real time or real world, not photo mode.
 

Latest Posts

Back