Your thoughts about "standard" vs. "premium"

[Akira AC]

I didn't say that at all... they could drive very differently between GT4 and 5... but a standard in GT4 might still not have as much accuracy as a premium if it's missing some data points.

Super simplified exmple:

In GT4 they collect weight of car and square inches of tires touching the ground. That's it.

They use these to figure frictional coefficient and when the car will break loose and the forumlas they use to do it are only recorded as integer values.

Great.

Now we get GT5. They sample weight of car, tire contact pad, downforce from spoiller and weight of fuel tank and the formulas used are more refined and their values stored as floats.

Now they can figure out a more accurate result, not only is the basic accounted for, but if you are going faster you get extra traction from downforce and when your tank is emptier your car is more squirly.

Now we bring in that car from GT4, feed it's 2 data points into this new physics engine which can calculate 4 data points worth of info but only is given two.

Now maybe this new physics engine is better written than the previous one, so even with only 2 data points the experience is better than it was in GT4 after all the results of the physics calculations are now recorded to a much higher level of accuracy (float vs int) however you still get no value for increased downforce at speed or squirlyness as your tank runs dry... but without going back and updating the cars data, you are not on par with premium GT5 cars.

Scale that up many hundreds of data points and advanced calculations etc and you see where I am going with GT5.

Again, only one of many possibilities, certainly not saying this is in any way a fact, just that it's possible and we have no accsess to standard cars to rule it out.

That is another way of seeing it and very accurate by you,in that case I'm complete agree with you there is some values like advance fuel consumption based on weight calculations for extra 3d elements,spoilers,¿rain?¿snow? accumulation and so on.

On a personal note I'll say that these values seem to be more defined in constant variables based on pure linear calculations and vectorial calculations with scalar numbers,as far as I know integers calculations tents to fail in huge amounts of data flow,but that's each developer story.

 

[JDMKING13]

Kaz said himself that all the cars run on one physics engine. Its the data they allocate to each car that makes them handle differently using the same physics model.

Yes sir. 👍

 

[SlipZtrEm]

I didn't say that at all... they could drive very differently between GT4 and 5... but a standard in GT4 might still not have as much accuracy as a premium if it's missing some data points.

Super simplified exmple:

In GT4 they collect weight of car and square inches of tires touching the ground. That's it.

They use these to figure frictional coefficient and when the car will break loose and the forumlas they use to do it are only recorded as integer values.

Great.

Now we get GT5. They sample weight of car, tire contact pad, downforce from spoiller and weight of fuel tank and the formulas used are more refined and their values stored as floats.

Now they can figure out a more accurate result, not only is the basic accounted for, but if you are going faster you get extra traction from downforce and when your tank is emptier your car is more squirly.

Now we bring in that car from GT4, feed it's 2 data points into this new physics engine which can calculate 4 data points worth of info but only is given two.

Now maybe this new physics engine is better written than the previous one, so even with only 2 data points the experience is better than it was in GT4 after all the results of the physics calculations are now recorded to a much higher level of accuracy (float vs int) however you still get no value for increased downforce at speed or squirlyness as your tank runs dry... but without going back and updating the cars data, you are not on par with premium GT5 cars.

Scale that up many hundreds of data points and advanced calculations etc and you see where I am going with GT5.

Again, only one of many possibilities, certainly not saying this is in any way a fact, just that it's possible and we have no accsess to standard cars to rule it out.

I couldn't explain it any more clearly than this. It is a super-simplified example, yes, but it covers the gist.

Taking it a step further; surely people understand that PD has probably collected more detailed calculations and values for cars nowadays than they first did in GT1 times, right? That is the same thing (I think) Deve is trying to get at. As a physics engine becomes more complicated, it requires more in-depth info to be plugged in. We don't know if these new values were recorded by PD back when they first created these assets, or if everything has been re-done, from scratch, specifically for GT5, re-collecting all the cars to get these more in-depth figures.

It's simply an interesting thing to think about, since we don't know one way or the other. An example could be the incredibly complex 4WD system in a modern Evo X. How accurately can GT5 simulate the constant shuffling of power from front to back, and left to right? And now, hypothetically, if we had that same 4WD system existing in real life back during GT1, how accurately could it pull it off the simulation?

Red - The tires, as an example; the GT5 engine probably has a lot more figures tied into tires and their physics (well, we assume it's more in-depth, anyways). If these new values were never recorded for the Standards during their creation leading up to GT4, they would need to get the figures now, however they might come by them. I think that's all Deve is trying to get at.

I can play a decade-old-created mp3 on the newest version of iTunes, or an old one. When I check out the file info though, there are places to store album artwork, lyrics, and assorted other features in the new program. The older one would just have the basic slots for Artist, Title, and Album. It still works fine in the newer program, but I'm going to have to fill in those blanks to get the same experience as other, newer mp3's. Not a great analogy, but closest I can come up with after a weekend of being sick .

 

[Akira AC]

Also adding to what slipstream said before we could get some examples of the improvement of the new engine based on the TT build and somethng that cross in my mind earlier Tourist Trophy which makes me think about what PD learn about weight distribution and tire contact,here is an clear example of the tourist throphy build,apologies for not getting a better quality but judge based on that.

(thus far we will know everything in less than 10 days perhaps when first game reviewers get the early copy,kotaku jp says October 21 but is yet to confirm).

 

[Nish]

I love GT5 and the sound doesn't bother me but the motorcycle in the video sounds just like the racecars on GT4.

 

[Devedander]

Well seeing as how they are using the tire models from GT5 and not GT4 I don't see how that is relevant.

They measured COG in GT4, once again, your point?

Minus the COG business the wind resistance point could be very true. It will be interesting to see what some basic tests will come back with.

Once again wheel and tires are from GT5, not GT4.

This is all a general assumption for an example, I am not saying they did or did not do things this way:

Let's say in GT4 they measured the COG. The car is consisdered one big chunk of matter with a COG somewhere around the stick shift. Now when you put this car on a sloped road where the left wheels are 2 feet lower than the right wheels (let's say a very banked mountain side road) where does the COG go?

If you are going off GT4 measurements, the weight is distributed evenly towards the left.

But what if in GT5 premiums they have actually measured the weight of the frame vs the weight of the windows vs the weight of the roof... ie the car is no longer one solid thing that weighs x amount but it's weight distribution can be figured out?

Well if it's a top heavy car like a Scion TC with a solid glass roof, the car is going to tip a lot more than if it's a bottom heavy car with a light roof material and heavy frame it will shift much less. This might effect how the car drives in some circumstances, but without that data, the car will be less accurate.

This is just one example of how things might be different with more data and just saying "they measured it" isn't enough.

The problem I see is that a lot of people around here (no offense intended) don't know what they don't know and assume that if they don't know it, it must not be true.

There are any number of facets of measuring the car that could be improved and or missing from GT4 measurements. That was just one such made up example.

D I just don't get what your saying majority of cars in GT where in GT4 the premiums are just recreated physically ,however they do not use GT4 physics Its a new Physics system. Again those Premiums was once GT4 cars when they collected all the data. So know your saying its a possibility that the standards wont drive like the Premium i don't know man I highly doubt that .

I don't think you understand the difference between a physics engine and the variables associate with each model that are fed to the engine. I get the feeling a lot of people think a physics engine is like one huge formula that contains instructions for what to do with every car in every circumstance, when in reality, it's just a formula waiting for data to be fed into it. That data is stored in each cars data file and loaded into the physics engine when the car is loaded and run.

Here is a very simple pseudo code physics model:

Let W=weight of car
Let A=angle of front wheels
Let S=Speed of car
Let F=Whether car has grip or no

If W+A+S > 1000 then F = 0
Else F=1

Now when you run your car, if F=0 you have broken grip and you slide all around, if F=1 you have full grip and drive normally.

So lets say you have a 900 lb car and you are driving at 80MPH, by this simple formula anything over 20 degrees of front wheel angle will cause the car to loose grip as the total combined values will be over 1000 which sets friction (F) to 0.

If the car is lighter, you can go faster and or turn the wheel more before loosing grip and heavier cars you cannot go as fast or turn the wheel as far before loosing grip (obvious upper limit on car weight is 1000 before you can't go or turn the wheels at all and anything under about 700 could go top speed while making 90 degree turns )

For this physics engine, all you need to collect is 1 data point, (W) weight of car for each car. The other variables are passed from other modules which are figuring those out from your controller/wheel input and throttle input. So every car is an asset with 3 main data modules: A wireframe and texture package for the graphics, some sound files for the sound, and a data file with 1 value in it with the weight of the car.

This is a pretty **** physics model becuaes you only have two states, full girp and no grip. Also it's a linear measurement that weights speed and angle of wheels and weight of car all equally.

Probably this physics model isn't even good enough for RC Pro am, but lets jsut go with it for now and call it the GT4 model for the purpose of this example.

Now let's say I want to improve this model, I want cars with wide tires to have more grip and I want to give cars with a low center of gravity better grip.

I now measure 2 more data points: T for tires (2 means wide tires and 1 means normal tires) and C for center of gravity (2 for low and 1 for high).

Now I change my formula:

If (W+A+S)/C > 1000*T then F=0
Else F=1

So what just happened? If the car has a low center of gravity it's weight+speed+angle can be TWICE as big without throwing off it's friction/grip. Also if it has wide tires it has also double the potential as the W+A+S must now be greater than 2000 to force it to loose grip.

But now I must capture values for T and C so my GT5 assets now include for each car: wireframe and textures, sound files, data module with 3 values... but I didn't capture those values for GT4! Oh no!! What do I do?

Well I either go get those values, or I just fake those values and fill them in with dummy data. So I dont ahve time to go back and get the data, so for GT4 cars I just set T and C both =1 and the net effect is you have the same results you did back in GT4 despite using my new physics engine (the net effect of setting both to 1 is that you divide or multiply by 1 on both sides of the equation, dividing or multiplying by 1 does not change anyhting so it's as if it wasn't there).

Now:

If T=(null) then T=1
If C=(null) then C=1
If (W+A+S)/C > 1000*T then F=0
Else F=1

Voila! New physics engine that can use either new comprehensive GT5 data or old GT4 data... I toss out my old GT4 physics engine that doesn't use T and C, use only the new GT5 engine, but guess what? Those old GT4 cars that don't have T or C measured? They won't benefit from it.

That is how you toss out the old physics engine, build a new one from scratch, but still suffer from lack of data in your assets.

Hopefully now there is one less thing you don't know you don't know

BTW I know in GT4 they would have measured whether a car had wide tires or low COG (well I hope they did) the example just uses very basic things for simplicity of writing it out. Obviously if this were to be a real issue with GT5 the missing data would be much more complex than just whether a car has wide tires. And to be precise this is not even an entire physics engine, it's just the chunk that would determine whether you have grip or not.

Kaz said himself that all the cars run on one physics engine. Its the data they allocate to each car that makes them handle differently using the same physics model.

Please see above. If I read right, we are on the same page.

That is another way of seeing it and very accurate by you,in that case I'm complete agree with you there is some values like advance fuel consumption based on weight calculations for extra 3d elements,spoilers,¿rain?¿snow? accumulation and so on.

On a personal note I'll say that these values seem to be more defined in constant variables based on pure linear calculations and vectorial calculations with scalar numbers,as far as I know integers calculations tents to fail in huge amounts of data flow,but that's each developer story.

Yes these are all super simplified examples that wouldn't be used in anything but the most simplistic games or demos. I am talking old DOS game with ascii art for cars.

And again I say we have nothing to make us think this is what will happen, but we also have nothing to assure us that it won't happen. I am just putting it out there to remind people that just like we all once assumed GT5 would have all premium cars, we shouldn't assume that all cars will be equally equipped unless it's officially stated or we have some hands on evidence.

 

[JDMKING13]

This is all a general assumption for an example, I am not saying they did or did not do things this way:

Let's say in GT4 they measured the COG. The car is consisdered one big chunk of matter with a COG somewhere around the stick shift. Now when you put this car on a sloped road where the left wheels are 2 feet lower than the right wheels (let's say a very banked mountain side road) where does the COG go?

If you are going off GT4 measurements, the weight is distributed evenly towards the left.

But what if in GT5 premiums they have actually measured the weight of the frame vs the weight of the windows vs the weight of the roof... ie the car is no longer one solid thing that weighs x amount but it's weight distribution can be figured out?

Well if it's a top heavy car like a Scion TC with a solid glass roof, the car is going to tip a lot more than if it's a bottom heavy car with a light roof material and heavy frame it will shift much less. This might effect how the car drives in some circumstances, but without that data, the car will be less accurate.

This is just one example of how things might be different with more data and just saying "they measured it" isn't enough.

The problem I see is that a lot of people around here (no offense intended) don't know what they don't know and assume that if they don't know it, it must not be true.

There are any number of facets of measuring the car that could be improved and or missing from GT4 measurements. That was just one such made up example.



I don't think you understand the difference between a physics engine and the variables associate with each model that are fed to the engine.

Here is a very simple pseudo code physics model:

Let W=weight of car
Let A=angle of front wheels
Let S=Speed of car
Let F=Whether car has grip or no

If W+A+S > 1000 then F = 0
Else F=1

Now when you run your car, if F=0 you have broken grip and you slide all around, if F=1 you have full grip and drive normally.

So lets say you have a 900 lb car and you are driving at 80MPH, by this simple formula anything over 20 degrees of front wheel angle will cause the car to loose grip as the total combined values will be over 1000 which sets friction (F) to 0.

For this physics engine, all you need to collect is 1 data point, (W) weight of car for each car. The other variables are passed from other modules which are figuring those out from your controller/wheel input and throttle input. So every car is an asset with 3 main data modules: A wireframe and texture package for the graphics, some sound files for the sound, and a data file with 1 value in it with the weight of the car.

This is a pretty **** physics model becuaes you only have two states, full girp and no grip. Also it's a linear measurement that weights speed and angle of wheels and weight of car all equally.

Probably this physics model isn't even good enough for RC Pro am, but lets jsut go with it for now and call it the GT4 model for the purpose of this example.

Now let's say I want to improve this model, I want cars with wide tires to have more grip and I want to give cars with a low center of gravity better grip.

I now measure 2 more data points: T for tires (2 means wide tires and 1 means normal tires) and C for center of gravity (2 for low and 1 for high).

Now I change my formula:

If (W+A+S)/C > 1000*T then F=0
Else F=1

So what just happened? If the car has a low center of gravity it's weight+speed+angle can be TWICE as big without throwing off it's friction/grip. Also if it has wide tires it has also double the potential as the W+A+S must now be greater than 2000 to force it to loose grip.

But now I must capture values for T and C... but I didn't capture those values for GT4! Oh no!! What do I do?

Well I either go get those values, or I just fake those values and fill them in with dummy data. So I dont ahve time to go back and get the data, so for GT4 cars I just set T and C both =1 and the net effect is you have the same results you did back in GT4 despite using my new physics engine (the net effect of setting both to 1 is that you divide by 1 on both sides of the equation, dividing by 1 does not change anyhting so it's as if it wasn't there).

Now:

If T=(null) then T=1
If C=(null) then C=1
If (W+A+S)/C > 1000*T then F=0
Else F=1

Voila! New physics engine that can use either new comprehensive GT5 data or old GT4 data... I toss out my old GT4 physics engine that doesn't use T and C, use only the new GT5 engine, but guess what? Those old GT4 cars that don't have T or C measured? They won't benefit from it.

That is how you toss out the old physics engine, build a new one from scratch, but still suffer from lack of data in your assets.

Hopefully now there is one less thing you don't know you don't know

BTW I know in GT4 they would have measured whether a car had wide tires or low COG (well I hope they did) the example just uses very basic things for simplicity of writing it out. Obviously if this were to be a real issue with GT5 the missing data would be much more complex than just whether a car has wide tires.



Please see above. If I read right, we are on the same page.



Yes these are all super simplified examples that wouldn't be used in anything but the most simplistic games or demos. I am talking old DOS game with ascii art for cars.

And again I say we have nothing to make us think this is what will happen, but we also have nothing to assure us that it won't happen. I am just putting it out there to remind people that just like we all once assumed GT5 would have all premium cars, we shouldn't assume that all cars will be equally equipped unless it's officially stated or we have some hands on evidence.

Thank you for that explanation, but when you say old GT4 data, for example The Honda Integra Type R was in GT4. It was made premium, but a car like a old vette which is left standard your saying might not have the same data collected as the Type R which was made premium but was also in GT4. I don't know man This is to technical for me. I drove the Audi R8 in the GT5 demo and it handles differently than in GT5P can you explain the difference between GT5P and GT4 switching to a new physics engine and data collected? Thank you.

Edit: So wouldn't even premium cars suffer from different physics that where from GT4? Only new cars such as the SLS would have more detail data which would give it a different driving feeling compared to your theory? I don't know man i think all cars will feel great. 👍

 

[Devedander]

Thank you for that explanation, But when you say old GT4 data, for example The Honda Integra Type R was in GT4. It was made premium, but a car like a old vette which is left standard your saying might not have the same data collected as the Type R which was made premium but was also in GT4. I don't know man This is to technical for me.

Yes that is exactly what I am saying. Becuase just like they went and upgraded the graphics for the Teg, they would have upgraded the data for it but simply imported what they had (graphics, data and sound files) for the Vette.

Why they chose what they chose to upgrade to Premium I don't know, but they did, and those cars got more attention than the ones they didn't choose to upgrade. All I am saying is it's quite possible those cars that didn't get graphics upgrade attention, also didn't get data upgrade attention.

It's simple, let's say all 800 standard cars had tire width measured but not height.

They say "Let's make the teg premium, go measure it from scratch again, not only do we need more data, our physics engine now can accept data accurate down to 1/64 of an inch while the GT4 engine only needed data down to 1/4 inch, so our old data is rough anyway. "

So they go, remodel the teg from scratch, remeasure everything they already had measured down to a much more detailed level and in the process measure the few hundred things they now want to use in GT5 that they weren't thinking about in GT4.

The vette they say "we aren't upgrading it to premium, just use the graphics and data from before, it's not as accurate but we don't have time".

Again, clarify here, not saying they did, just showing how they could have.

Edit: So wouldn't even premium cars suffer from different physics that where from GT4? Only new cars such as the SLS would have more detail data which would give it a different driving feeling compared to your theory? I don't know man i think all cars will feel great. 👍

You put this edit in after I started typing, this is explained above.

Really try to understand what I am saying here... you seem like a guy who should be able to understand it, and as long winded as I am, I didn't just type that all up for my own joy, I would really appreciate if you tried to understand what I am saying... if we are all more educated on the subject, the conversation will be a lot less frustrating

I drove the Audi R8 in the GT5 demo and it handles differently than in GT5P can you explain the difference between GT5P and GT4 switching to a new physics engine and data collected? Thank you.

Look back and try to seperate a physics model from the data, really understand what I wrote. I simplified it as much as possible so it really isn't that technical and you shoudl be able to figure it out even if you aren't familiar with coding.

The way a car with the same data can drive differently between two different physics engines is that the engine changes... what it does with the numbers you feed it changes.

Example:

I take W+A+S > 1000 then F=0

This means when the weight of yoru car, plus speed plus wheel angle is greater than 1000 you loose grip. A 900lb car going 80mphs can not turn the wheel any more than 20 degrees without loosing grip.

Now let's say I tweak that physics engine and say:

W+A+(.8 * S) > 1000 then F=0

What just happened? I reduced the weight of the speed portion by 20%. So a 900lb car going 80 mph can now turn it's wheel 25 degrees before loosing grip.

.8 * 25 = 20

Thus 900 + 80 +(.8 *25) is still not greather than 1000, therefore you still have grip. The result in feeling to you is this car has more grip, you can go through turns faster and sharper than you could before with the old physics engine.

I didn't change the data of the car, my data still says it weighs 900 lbs... what I changed was hot he game handles what's going on with that 900 lb car.

That is how a car with the same data can handle differently from GT5P and Gt5.

Again, you need to realize a physics engine is not something that includes the rules for every car in every situation, it is a formula that each car feeds it's data into and the result is what the car does. Change the formula, you change what the car does without having to change the information about the car.

 

[JDMKING13]

Yes that is exactly what I am saying. Becuase just like they went and upgraded the graphics for the Teg, they would have upgraded the data for it but simply imported what they had (graphics, data and sound files) for the Vette.

Again thank you for your explanation It does make sense D, but you said what I was going to ask you next. How can you say they wouldn't want to get more information on the vette (standard car), for the fact of GT5 is going to have a better physics Engine than GT4? If I was a game developer creating a new physics engine, I would want to collect more data on the cars to make the game feel better driving wise. Now this is not a fact, But reading from the official website the difference between the premium cars and standard cars say nothing about handling characteristic difference. It does explain the difference between, detail, features (wipers), damage. Nothing about handling. As I said befor thank you for your explanation If they follow your format I can understand where your coming from 👍 I will let you do your thing I will let you know when its in the PS3 👍

Edit: The way a car with the same data can drive differently between two different physics engines is that the engine changes... what it does with the numbers you feed it changes.

Example:

I take W+A+S > 1000 then F=0

This means when the weight of yoru car, plus speed plus wheel angle is greater than 1000 you loose grip. A 900lb car going 80mphs can not turn the wheel any more than 20 degrees without loosing grip.

Now let's say I tweak that physics engine and say:

W+A+(.8 * S) > 1000 then F=0

What just happened? I reduced the weight of the speed portion by 20%. So a 900lb car going 80 mph can now turn it's wheel 25 degrees before loosing grip.

.8 * 25 = 20

Thus 900 + 80 +(.8 *25) is still not greather than 1000, therefore you still have grip. The result in feeling to you is this car has more grip, you can go through turns faster and sharper than you could before with the old physics engine.

I didn't change the data of the car, my data still says it weighs 900 lbs... what I changed was hot he game handles what's going on with that 900 lb car.

That is how a car with the same data can handle differently from GT5P and Gt5.

Again, you need to realize a physics engine is not something that includes the rules for every car in every situation, it is a formula that each car feeds it's data into and the result is what the car does. Change the formula, you change what the car does without having to change the information about the car.[/QUOTE]

Now i don't think you got me here I was not clear enough. Compare that GT5P car to a New car thats going to be used in GT5 EX Integra Type R to SLS when PD got Information on the Integra type R that was before the new GT5 physics engine was created. The SLS data was gathered closer to the GT5 physics. Going of your theory about obtaining information cant we say that PD gather more minute information than the Type R which information was gathered years befor the SLS?

 

[Devedander]

Again thank you for your explanation It does make sense D, but you said what I was going to ask you next. How can you say they wouldn't want to get more information on the vette (standard car), for the fact of GT5 is going to have a better physics Engine than GT4? If I was a game developer creating a new physics engine, I would want to collect more data on the cars to make the game feel better driving wise. Now this is not a fact, But reading from the official website the difference between the premium cars and standard cars say nothing about handling characteristic difference. It does explain the difference between, detail, features (wipers), damage. Nothing about handling.

It doesn't say that there will be any handling difference. It does say all the cars will get to run on the new GT5 physics model.

However it also doesn't say "Standard models will be lower poly count and less detailed than premium" but it does say "they have been carefully recreated in GT5" (paraphrase quote) yet truth is they pretty much look to be just what we had before.

So when they say they get to use the new gt5 physics model, that doesn't in my mind assure me at all that they will have been updated in any way to take advantage of it in terms of capturing more data points. It doesn't mean they will be at a disadvantage either for sure, but based on PDs recent history, I am not cutting any slack.

As for why wouldn't you get more info for the vette and other standards? Why wouldn't you want to upgrade them graphically to premium standard?

Exactly... you would want to of course, but if you are working one way you are working that way.

Here's another example of how it might have happened (and again I am not saying this is how it happened, just this is a way it might have to answer your question infering it's not reasonable it would happen this way):

PD has some teams dedicated to working on cars. Each team has a few guys who measure and model, some guys who record sound and some guys who gather data for the physics engine. These teams work together and generally crank through a car together start to finish (ie the modelers model while the sound guys record sound and the data guys gather data and they all get done about the same day).

KY says "We could break off some of the data guys to go scavenge missing data for the standard but that screws up the team dynamic for the other cars and my main goal is to get as many premiums done as possible, so I am not sacrificing parts of teams taht will hamper how many premiums we get done. Thus either a car gets the full premium treatment (ie it's built fromm scratch on all fronts) or it gets imported as a standard and that's it".

That's one example of how it could happen. There are many other ways I am sure.

As I said before, many people were quick to point out that if we were disspointed in standard cars it's only becuase we expected what we weren't explicitly promised... well I am just taking that advice here...

The way you say it sounds like you know that they are going to follow that format 100%. As I said befor thank you for your explanation If they follow your format I can understand where your coming from 👍 I will let you do your thing I will let you know when its in the PS3 👍

Actually you will see I have been explicity stating at every turn that this is not something that I think will necessarily be true or even has been hinted at directly. And in fact I am careful to state it's one of many possiblities... which is the exact opposite of knowing that this is the format they are going to follow... they might differ in any number of ways... I am specifically trying to NOT put out a false dichotomy of "it will be this way or that" I am specifically saying there are some things that might be a surprise down the road, and this might be one, and here is one example of how that might be possible.

It might be possible any number of other ways, it might not be an issue at all down the road.

I am only illustrating how we cannot say now that it's impossible and clarifying for those who say "it's not reasonable" or "explain how that could possibly happen" so you can see that yes, it could possibly happen.

Could possibly being the keywords there... again I am in no way saying it's certain to happen this way.

What I am doing is similar to someone say how a football team might go all the way this year by starting this player and trading that... it might happen that way, it might not, it might be slightly different, it might be drastically different...

I see a lot of people saying "how would that happen" or "why would that happen" infering that they can't come up with a reason that it would thus sinde there is no reason that it would, its safe to say it would not.

I am just showing you reasons that it could happen and thus illustratinng why it's not safe to say that it won't happen that way since there is no reason for it to.

Now i don't think you got me here I was not clear enough. Compare that GT5P car to a New car thats going to be used in GT5 EX Integra Type R to SLS when PD got Information on the Integra type R that was before the new GT5 physics engine was created. The SLS data was gathered closer to the GT5 physics. Going of your theory about obtaining information cant we say that PD gather more minute information than the Type R which information was gathered years befor the SLS?

Again, difference between data points and phyics engine formulas.

They may have standardized data points long ago (like before GT5P) but have since tweaked the engine.

Look at my example... I tweaked the physics engine, the game plays differently now, but I didn't have to get any new data points to do it...

So lets' say in Gt4 they got 800 data points.

When they made GT5p they collected 1000 data points. Those are the standard for GT5 and all premium cars have 1000 data points.

The Integra that was made premium has 1000 data points that were collected for GT5P. However the Vette might still be using it's 800 data points from GT4.

GT5P comes along, you get 1000 data points with the GT5P physics engine.

Then GT5 comes along, they use the same 1000 data points, but they calculate what happens differently. So in GT5P they say "when weight plus speed plus angle is greater than 1000 loose grip" in GT5 they say "when weight plus speed plus angle is greater than 1200 then loose grip." No new data points are needed for this change and the car feels different now.

I think you are still not getting the relationship of how a physics engine works with the physics information.

Think of it like a BMI scale. A BMI chart does not have information about every person in the whole world, but you plug in your weight and your height and it tells you how obese you are. If you had 1000 people and you had their height and weight, you could tell how obese they were and plot a graph of how obese the average person was.

Let's say they change that calculation for the BMI. This is the same as tweaking the physics engine. Now the BMI is more leniant and you have to be fatter/shorter to be obese. Suddenly the same 1000 people, the same weights and heights yield a different result....

I don't know how else to explain it... the physics engine is the formulas, the data from the cars is the data... changing EITHER changes how the car feels in the end.

But simple answer to your question: No, I would assume that the standard for the amount and type of data is set early on. GT4 they decided what data to get, they went and got it. GT5 they decided what data to get and went and got it.

In both cases hopefully the standard was set before the first car was measured. It's really bad buisiness to change the standard of what needs to be collected constantly throught the process so every few weeks you decide "let's collect 3 more data points from now on!"

It really sucks to go changing standards half way through the process. It would be like building 200 hundred apartments, but not deciding on the floor plans before you start building and half way through you decide to do things a little different... you now have to go back and fix all the ones you did already.

So no, I think data captured for GT5P would have already been standardized and any car captured and created since work started on GT5 (including GT5P) would be the same set and same accuracy. From a project management perspective, you want the very first car measured and the very last car measured to be done to the same standard.

 

[JDMKING13]

It doesn't say that there will be any handling difference. It does say all the cars will get to run on the new GT5 physics model.

However it also doesn't say "Standard models will be lower poly count and less detailed than premium" but it does say "they have been carefully recreated in GT5" (paraphrase quote) yet truth is they pretty much look to be just what we had before.

So when they say they get to use the new gt5 physics model, that doesn't in my mind assure me at all that they will have been updated in any way to take advantage of it in terms of capturing more data points. It doesn't mean they will be at a disadvantage either for sure, but based on PDs recent history, I am not cutting any slack.

We saw a small standard car clip and your saying they looked liked what we had before, please give me your copy of GT4 Even though that mean nothing because thats your opinion and some think different It just makes me laugh. Come on D If you cant understand the differences between the the type of cars thats pretty sad. It clearly puts the premium cars on a high standard than the standard cars. It the same thing you where telling me about the cockpit view for the standards( Yes you were right) It clearly states the difference between the two type of cars on the official web page.

"Exquisitely Detailed, Premium Cars Are Stunningly Recreated Both Inside And Out

From the seductive curves of exotic sports cars, to the inviting sight of the road ahead when you slide into the driver’s seat and peer past the glass, down to the intricately detailed instrument panel and dashboard... Every one of the over 200 premium cars recreate every last detail of the car, inside and out."

Gran Turismo 5 contains over 200 of these "Premium" level cars.

Updated Cars From Past Gran Turismo Games Present An Unrivaled Lineup

"The massive lineup of cars from past Gran Turismo games has been beautifully recreated through the latest technology and the Playstation 3’s cutting-edge graphics."

Gran Turismo 5 feature an astounding 800+ cars, a vast collection covering a wide range of eras and categories."

*Standard cars do not support vehicle interior camera views.

It also has pictures so if you cant understand the difference between the to cars I'm sorry.

As for why wouldn't you get more info for the vette and other standards? Why wouldn't you want to upgrade them graphically to premium standard?

Exactly... you would want to of course, but if you are working one way you are working that way.

D the man said it would of took a long time to make 1000 cars to that level and he felt good about the 200 car roster. Im pretty sure its faster to collect data than build 1000 cars with that quality.

Here's another example of how it might have happened (and again I am not saying this is how it happened, just this is a way it might have to answer your question infering it's not reasonable it would happen this way):

PD has some teams dedicated to working on cars. Each team has a few guys who measure and model, some guys who record sound and some guys who gather data for the physics engine. These teams work together and generally crank through a car together start to finish (ie the modelers model while the sound guys record sound and the data guys gather data and they all get done about the same day).

KY says "We could break off some of the data guys to go scavenge missing data for the standard but that screws up the team dynamic for the other cars and my main goal is to get as many premiums done as possible, so I am not sacrificing parts of teams taht will hamper how many premiums we get done. Thus either a car gets the full premium treatment (ie it's built fromm scratch on all fronts) or it gets imported as a standard and that's it".

That's one example of how it could happen. There are many other ways I am sure.

As I said before, many people were quick to point out that if we were disspointed in standard cars it's only becuase we expected what we weren't explicitly promised... well I am just taking that advice here...



Actually you will see I have been explicity stating at every turn that this is not something that I think will necessarily be true or even has been hinted at directly. And in fact I am careful to state it's one of many possiblities... which is the exact opposite of knowing that this is the format they are going to follow... they might differ in any number of ways... I am specifically trying to NOT put out a false dichotomy of "it will be this way or that" I am specifically saying there are some things that might be a surprise down the road, and this might be one, and here is one example of how that might be possible.

It might be possible any number of other ways, it might not be an issue at all down the road.

I am only illustrating how we cannot say now that it's impossible and clarifying for those who say "it's not reasonable" or "explain how that could possibly happen" so you can see that yes, it could possibly happen.

Could possibly being the keywords there... again I am in no way saying it's certain to happen this way.

What I am doing is similar to someone say how a football team might go all the way this year by starting this player and trading that... it might happen that way, it might not, it might be slightly different, it might be drastically different...


I see a lot of people saying "how would that happen" or "why would that happen" infering that they can't come up with a reason that it would thus sinde there is no reason that it would, its safe to say it would not.

I am just showing you reasons that it could happen and thus illustratinng why it's not safe to say that it won't happen that way since there is no reason for it to.

No prob man I know your style as I said before I'm really excited GT5 is coming out so for these next weeks I will be happy and positive. I will let you come up with all the negative speculation or you being a realist, and do your critique thing which is cool. I respect everything your saying, however I will think on the bright side I have been waiting for this game for a long time and it is almost here. 👍

 

[iLex]

This is a very interesting read guys, thanks.

The thing I'd like to see your opinion on too is the practical side of this.

The way I see it we will never notice the difference in physics between premium and standard.

1) None of the cars will handle 100% like their real life counterpart, so it's hard to really compare. In other words: we wouldn't know if they implemented all the data correctly, as long as the bulk of the data is right.

2) PD recorded more data than needed for the GT4 engine. I don't know this ofcourse, but it seems very reasonable for a company that wants to continue creating great racing games that they want as much data that can be possibly used in the future. My guess is that all data they thought they could use in the future that doesn't take a very long time to collect has been collected.

3) Between the 200+ premium cars they should have a generic counterpart for pretty much each of the 800 standard cars. So if they haven't collected a certain datapoint for a certain standard car, they can take the datapoint of a similar premium car and use that. For example, they can take the way one light hatchback's fuel consumption reacts to a spoiler and use the same formula on another light hatchback, without anyone ever noticing or caring.
Ofcourse this wouldn't be accurate, but it would be accurate enough (considering point 1 too) for the game.
Note that the data they haven't collected in my theory is data that most likely doesn't have a huge influence on characteristics as I think the most important data have been collected for all cars (partly in GT4 and partly through point 2).

4) The labour of updating the data of the cars is nothing compared to the modelling, sound, weather engine etc. You can probably have an intern or two do the job

Saying the cars will handle differently sounds like the standards will behave in a very different way.
I think the standards will at the most behave a bit less accurately like the original car than the premiums will.
This might even be insignificant because of the precision of the physics engine.

All this thinking is ofcourse based on the physics engine theory/discussion in the above posts.

 

[Organ-Donor]

I didn't say that at all... they could drive very differently between GT4 and 5... but a standard in GT4 might still not have as much accuracy as a premium if it's missing some data points.

Super simplified exmple:

In GT4 they collect weight of car and square inches of tires touching the ground. That's it.

They use these to figure frictional coefficient and when the car will break loose and the forumlas they use to do it are only recorded as integer values.

Great.

Now we get GT5. They sample weight of car, tire contact pad, downforce from spoiller and weight of fuel tank and the formulas used are more refined and their values stored as floats.

Now they can figure out a more accurate result, not only is the basic accounted for, but if you are going faster you get extra traction from downforce and when your tank is emptier your car is more squirly.

Now we bring in that car from GT4, feed it's 2 data points into this new physics engine which can calculate 4 data points worth of info but only is given two.

Now maybe this new physics engine is better written than the previous one, so even with only 2 data points the experience is better than it was in GT4 after all the results of the physics calculations are now recorded to a much higher level of accuracy (float vs int) however you still get no value for increased downforce at speed or squirlyness as your tank runs dry... but without going back and updating the cars data, you are not on par with premium GT5 cars.

Scale that up many hundreds of data points and advanced calculations etc and you see where I am going with GT5.

Again, only one of many possibilities, certainly not saying this is in any way a fact, just that it's possible and we have no accsess to standard cars to rule it out.

I would obviously have to see what they are actually recording to make an edcuated statment, but a guess could be ventured:

tire material and viscocity relative to temperature
engine mount location and resultant change in cog
weight of door panels and body panels and glass and roof (to better calculate cog and effects of resistive forces like wind)
wheel weight and resultant momentum and frictional losses

Who knows... remember back when GT4 was made they futureproofed the cars and we couldn't imagine what more they would need to record about them to get better models... then look what happened.

You can always record more data.



Money... lots and lots of money.

I honestly don't know, that guy is just super lucky.

But that's not what I am tlaking about when I say standards of care (no matter what you do for people, some cancer patients will die fast and some will outlive the odds by a lot or even go into remision) what I mean is stuff like:

Must have IV antibiotics for 3 days post surgery
Must scrub hands before surgery for 2 minutes
Must maintain sterile surgery suite

At some points none of those were standards of care, but once they became established, you didn't revert. However you also didn't go back and consider previous surgeries botched because they didn't do those before they were the standard.

This is all a general assumption for an example, I am not saying they did or did not do things this way:

Let's say in GT4 they measured the COG. The car is consisdered one big chunk of matter with a COG somewhere around the stick shift. Now when you put this car on a sloped road where the left wheels are 2 feet lower than the right wheels (let's say a very banked mountain side road) where does the COG go?

If you are going off GT4 measurements, the weight is distributed evenly towards the left.

But what if in GT5 premiums they have actually measured the weight of the frame vs the weight of the windows vs the weight of the roof... ie the car is no longer one solid thing that weighs x amount but it's weight distribution can be figured out?

Well if it's a top heavy car like a Scion TC with a solid glass roof, the car is going to tip a lot more than if it's a bottom heavy car with a light roof material and heavy frame it will shift much less. This might effect how the car drives in some circumstances, but without that data, the car will be less accurate.

This is just one example of how things might be different with more data and just saying "they measured it" isn't enough.

The problem I see is that a lot of people around here (no offense intended) don't know what they don't know and assume that if they don't know it, it must not be true.

There are any number of facets of measuring the car that could be improved and or missing from GT4 measurements. That was just one such made up example.



I don't think you understand the difference between a physics engine and the variables associate with each model that are fed to the engine. I get the feeling a lot of people think a physics engine is like one huge formula that contains instructions for what to do with every car in every circumstance, when in reality, it's just a formula waiting for data to be fed into it. That data is stored in each cars data file and loaded into the physics engine when the car is loaded and run.

Here is a very simple pseudo code physics model:

Let W=weight of car
Let A=angle of front wheels
Let S=Speed of car
Let F=Whether car has grip or no

If W+A+S > 1000 then F = 0
Else F=1

Now when you run your car, if F=0 you have broken grip and you slide all around, if F=1 you have full grip and drive normally.

So lets say you have a 900 lb car and you are driving at 80MPH, by this simple formula anything over 20 degrees of front wheel angle will cause the car to loose grip as the total combined values will be over 1000 which sets friction (F) to 0.

If the car is lighter, you can go faster and or turn the wheel more before loosing grip and heavier cars you cannot go as fast or turn the wheel as far before loosing grip (obvious upper limit on car weight is 1000 before you can't go or turn the wheels at all and anything under about 700 could go top speed while making 90 degree turns )

For this physics engine, all you need to collect is 1 data point, (W) weight of car for each car. The other variables are passed from other modules which are figuring those out from your controller/wheel input and throttle input. So every car is an asset with 3 main data modules: A wireframe and texture package for the graphics, some sound files for the sound, and a data file with 1 value in it with the weight of the car.

This is a pretty **** physics model becuaes you only have two states, full girp and no grip. Also it's a linear measurement that weights speed and angle of wheels and weight of car all equally.

Probably this physics model isn't even good enough for RC Pro am, but lets jsut go with it for now and call it the GT4 model for the purpose of this example.

Now let's say I want to improve this model, I want cars with wide tires to have more grip and I want to give cars with a low center of gravity better grip.

I now measure 2 more data points: T for tires (2 means wide tires and 1 means normal tires) and C for center of gravity (2 for low and 1 for high).

Now I change my formula:

If (W+A+S)/C > 1000*T then F=0
Else F=1

So what just happened? If the car has a low center of gravity it's weight+speed+angle can be TWICE as big without throwing off it's friction/grip. Also if it has wide tires it has also double the potential as the W+A+S must now be greater than 2000 to force it to loose grip.

But now I must capture values for T and C so my GT5 assets now include for each car: wireframe and textures, sound files, data module with 3 values... but I didn't capture those values for GT4! Oh no!! What do I do?

Well I either go get those values, or I just fake those values and fill them in with dummy data. So I dont ahve time to go back and get the data, so for GT4 cars I just set T and C both =1 and the net effect is you have the same results you did back in GT4 despite using my new physics engine (the net effect of setting both to 1 is that you divide or multiply by 1 on both sides of the equation, dividing or multiplying by 1 does not change anyhting so it's as if it wasn't there).

Now:

If T=(null) then T=1
If C=(null) then C=1
If (W+A+S)/C > 1000*T then F=0
Else F=1

Voila! New physics engine that can use either new comprehensive GT5 data or old GT4 data... I toss out my old GT4 physics engine that doesn't use T and C, use only the new GT5 engine, but guess what? Those old GT4 cars that don't have T or C measured? They won't benefit from it.

That is how you toss out the old physics engine, build a new one from scratch, but still suffer from lack of data in your assets.

Hopefully now there is one less thing you don't know you don't know

BTW I know in GT4 they would have measured whether a car had wide tires or low COG (well I hope they did) the example just uses very basic things for simplicity of writing it out. Obviously if this were to be a real issue with GT5 the missing data would be much more complex than just whether a car has wide tires. And to be precise this is not even an entire physics engine, it's just the chunk that would determine whether you have grip or not.



Please see above. If I read right, we are on the same page.



Yes these are all super simplified examples that wouldn't be used in anything but the most simplistic games or demos. I am talking old DOS game with ascii art for cars.

And again I say we have nothing to make us think this is what will happen, but we also have nothing to assure us that it won't happen. I am just putting it out there to remind people that just like we all once assumed GT5 would have all premium cars, we shouldn't assume that all cars will be equally equipped unless it's officially stated or we have some hands on evidence.

Yes that is exactly what I am saying. Becuase just like they went and upgraded the graphics for the Teg, they would have upgraded the data for it but simply imported what they had (graphics, data and sound files) for the Vette.

Why they chose what they chose to upgrade to Premium I don't know, but they did, and those cars got more attention than the ones they didn't choose to upgrade. All I am saying is it's quite possible those cars that didn't get graphics upgrade attention, also didn't get data upgrade attention.

It's simple, let's say all 800 standard cars had tire width measured but not height.

They say "Let's make the teg premium, go measure it from scratch again, not only do we need more data, our physics engine now can accept data accurate down to 1/64 of an inch while the GT4 engine only needed data down to 1/4 inch, so our old data is rough anyway. "

So they go, remodel the teg from scratch, remeasure everything they already had measured down to a much more detailed level and in the process measure the few hundred things they now want to use in GT5 that they weren't thinking about in GT4.

The vette they say "we aren't upgrading it to premium, just use the graphics and data from before, it's not as accurate but we don't have time".

Again, clarify here, not saying they did, just showing how they could have.



You put this edit in after I started typing, this is explained above.

Really try to understand what I am saying here... you seem like a guy who should be able to understand it, and as long winded as I am, I didn't just type that all up for my own joy, I would really appreciate if you tried to understand what I am saying... if we are all more educated on the subject, the conversation will be a lot less frustrating



Look back and try to seperate a physics model from the data, really understand what I wrote. I simplified it as much as possible so it really isn't that technical and you shoudl be able to figure it out even if you aren't familiar with coding.

The way a car with the same data can drive differently between two different physics engines is that the engine changes... what it does with the numbers you feed it changes.

Example:

I take W+A+S > 1000 then F=0

This means when the weight of yoru car, plus speed plus wheel angle is greater than 1000 you loose grip. A 900lb car going 80mphs can not turn the wheel any more than 20 degrees without loosing grip.

Now let's say I tweak that physics engine and say:

W+A+(.8 * S) > 1000 then F=0

What just happened? I reduced the weight of the speed portion by 20%. So a 900lb car going 80 mph can now turn it's wheel 25 degrees before loosing grip.

.8 * 25 = 20

Thus 900 + 80 +(.8 *25) is still not greather than 1000, therefore you still have grip. The result in feeling to you is this car has more grip, you can go through turns faster and sharper than you could before with the old physics engine.

I didn't change the data of the car, my data still says it weighs 900 lbs... what I changed was hot he game handles what's going on with that 900 lb car.

That is how a car with the same data can handle differently from GT5P and Gt5.

Again, you need to realize a physics engine is not something that includes the rules for every car in every situation, it is a formula that each car feeds it's data into and the result is what the car does. Change the formula, you change what the car does without having to change the information about the car.

It doesn't say that there will be any handling difference. It does say all the cars will get to run on the new GT5 physics model.

However it also doesn't say "Standard models will be lower poly count and less detailed than premium" but it does say "they have been carefully recreated in GT5" (paraphrase quote) yet truth is they pretty much look to be just what we had before.

So when they say they get to use the new gt5 physics model, that doesn't in my mind assure me at all that they will have been updated in any way to take advantage of it in terms of capturing more data points. It doesn't mean they will be at a disadvantage either for sure, but based on PDs recent history, I am not cutting any slack.

As for why wouldn't you get more info for the vette and other standards? Why wouldn't you want to upgrade them graphically to premium standard?

Exactly... you would want to of course, but if you are working one way you are working that way.

Here's another example of how it might have happened (and again I am not saying this is how it happened, just this is a way it might have to answer your question infering it's not reasonable it would happen this way):

PD has some teams dedicated to working on cars. Each team has a few guys who measure and model, some guys who record sound and some guys who gather data for the physics engine. These teams work together and generally crank through a car together start to finish (ie the modelers model while the sound guys record sound and the data guys gather data and they all get done about the same day).

KY says "We could break off some of the data guys to go scavenge missing data for the standard but that screws up the team dynamic for the other cars and my main goal is to get as many premiums done as possible, so I am not sacrificing parts of teams taht will hamper how many premiums we get done. Thus either a car gets the full premium treatment (ie it's built fromm scratch on all fronts) or it gets imported as a standard and that's it".

That's one example of how it could happen. There are many other ways I am sure.

As I said before, many people were quick to point out that if we were disspointed in standard cars it's only becuase we expected what we weren't explicitly promised... well I am just taking that advice here...



Actually you will see I have been explicity stating at every turn that this is not something that I think will necessarily be true or even has been hinted at directly. And in fact I am careful to state it's one of many possiblities... which is the exact opposite of knowing that this is the format they are going to follow... they might differ in any number of ways... I am specifically trying to NOT put out a false dichotomy of "it will be this way or that" I am specifically saying there are some things that might be a surprise down the road, and this might be one, and here is one example of how that might be possible.

It might be possible any number of other ways, it might not be an issue at all down the road.

I am only illustrating how we cannot say now that it's impossible and clarifying for those who say "it's not reasonable" or "explain how that could possibly happen" so you can see that yes, it could possibly happen.

Could possibly being the keywords there... again I am in no way saying it's certain to happen this way.

What I am doing is similar to someone say how a football team might go all the way this year by starting this player and trading that... it might happen that way, it might not, it might be slightly different, it might be drastically different...

I see a lot of people saying "how would that happen" or "why would that happen" infering that they can't come up with a reason that it would thus sinde there is no reason that it would, its safe to say it would not.

I am just showing you reasons that it could happen and thus illustratinng why it's not safe to say that it won't happen that way since there is no reason for it to.



Again, difference between data points and phyics engine formulas.

They may have standardized data points long ago (like before GT5P) but have since tweaked the engine.

Look at my example... I tweaked the physics engine, the game plays differently now, but I didn't have to get any new data points to do it...

So lets' say in Gt4 they got 800 data points.

When they made GT5p they collected 1000 data points. Those are the standard for GT5 and all premium cars have 1000 data points.

The Integra that was made premium has 1000 data points that were collected for GT5P. However the Vette might still be using it's 800 data points from GT4.

GT5P comes along, you get 1000 data points with the GT5P physics engine.

Then GT5 comes along, they use the same 1000 data points, but they calculate what happens differently. So in GT5P they say "when weight plus speed plus angle is greater than 1000 loose grip" in GT5 they say "when weight plus speed plus angle is greater than 1200 then loose grip." No new data points are needed for this change and the car feels different now.

I think you are still not getting the relationship of how a physics engine works with the physics information.

Think of it like a BMI scale. A BMI chart does not have information about every person in the whole world, but you plug in your weight and your height and it tells you how obese you are. If you had 1000 people and you had their height and weight, you could tell how obese they were and plot a graph of how obese the average person was.

Let's say they change that calculation for the BMI. This is the same as tweaking the physics engine. Now the BMI is more leniant and you have to be fatter/shorter to be obese. Suddenly the same 1000 people, the same weights and heights yield a different result....

I don't know how else to explain it... the physics engine is the formulas, the data from the cars is the data... changing EITHER changes how the car feels in the end.

But simple answer to your question: No, I would assume that the standard for the amount and type of data is set early on. GT4 they decided what data to get, they went and got it. GT5 they decided what data to get and went and got it.

In both cases hopefully the standard was set before the first car was measured. It's really bad buisiness to change the standard of what needs to be collected constantly throught the process so every few weeks you decide "let's collect 3 more data points from now on!"

It really sucks to go changing standards half way through the process. It would be like building 200 hundred apartments, but not deciding on the floor plans before you start building and half way through you decide to do things a little different... you now have to go back and fix all the ones you did already.

So no, I think data captured for GT5P would have already been standardized and any car captured and created since work started on GT5 (including GT5P) would be the same set and same accuracy. From a project management perspective, you want the very first car measured and the very last car measured to be done to the same standard.

 

[CoolColJ]

Gees people are typing novels in here!
They should bundle all the posts in this thread, and publish a book.

GT5:Stands vs Premium, the Saga continues
An epic and riveting tale of emo rage

 

[Devedander]

we saw a small standard car clip and your saying they looked liked what we had before

I said they look to be. That means they appear to be based on what we have so far. Could that be considerably wrong come release? Sure. Likely? I don't think so for many reasons...

Come on D If you cant understand the differences between the the type of cars thats pretty sad.

I am pretty sure I understand the differences quite thoroughly... and on top of that I recognize where there might be differences we yet don't know about or have confirmed...

"The massive lineup of cars from past Gran Turismo games has been beautifully recreated through the latest technology and the Playstation 3’s cutting-edge graphics."

Now if I was to just read that, I would think these have been torn down and recreated somehow... yet according to Kaz's own statements in interviews and what we have seen, they are most likely to be just GT4 models shown in higher resolution (which is a given considering the platform they are ported to) and with a new lighting engine.

That's why I am skeptical of when they say they will also use GT5's physics engine. I am not at all certain they don't mean that in the loosest possible way which is that sure they run on the formulas in GT5's new physics engine, but they use only the data from before. PD has shown many times they are willing to be crafty with what they say...

It also has pictures so if you cant understand the difference between the to cars I'm sorry.

Wait... the few tiny car pictures on GT's website are supposed to be something that spells it all out whereas that direct feed video clip we looked at that shows low poly cars is to be dismissed? I am not so sure that logic follows... and pictures and videos do not tell you anything about how the cars feel to drive so again, they don't rule anything out.

D the man said it would of took a long time to make 1000 cars to that level and he felt good about the 200 car roster. Im pretty sure its faster to collect data than build 1000 cars with that quality.

I explained one way in which this could have panned out with the result being no going back and collecting data... this is exactly what I am talking about, people say "I don't think that could happen" because they haven't thought of a way it could or a way it makes sense, so I put forth a way it could make sense.

I used to work with a guy who was pissed that when he would get a package shipped via Fedex, Fedex wouldn't pull it off the line and hold it for customer pickup until they had tried to deliver it once. His reasoning was it surely saves them money to just put it aside and not try to deliver it since they don't waste money and gas on the attempted delivery. The fedex guy tried to explain how that sounds reasonable but in reality the way they load the trucks down from the big trucks to the smaller ones for delivery was such that it's actually slows down the process a lot to stop the belts and pull the occasional box and put it to the side. Basically they hava streamline process that really doesn't allow for them to do this even though it sounds like it should be a good idea. Once it's been out for delivery once, the next nights unpacking process is designed to allow for packages to be pulled and so it works quite well then.

This is just one case of how what you are sure must be reasonsble, isn't always.

Like I said, perhaps they have teams that work together and breaking apart a team throws off the rest of the process slowing everything down.

Or perhaps gathering some of this data is time consuming and actually involves getting the car and running it (like heat disspation numbers, tire flex under load things that you actually have to measure or record happening and can't just pull from a spec sheet) so getting the data is really almomst as time consuming as the whole process.

If it involves getting the car and running it to capture data, well then it makes sense to capture the audio data too since you have gone through the trouble to get a real car and run it... and heck if those guys are all there and you have the car, might as well re scan it for high res modeling right?

What I am getting at is sometimes one part of the process can be so intensive it becomems a case if do it all or don't do it at all because each part takes 80% of the time it takes to do the whole. That sounds impossible at first (wouldn't that make the whole take 240% of the time if each of 3 parts takes 80%) but again, if you think about it, if finding and getting your hands on a car in the right conditions to record something takes 2 months of a 3 month process, then there is really no reason to do it unless you can do the whole thing.

And that brings us full circle to what cars get attention and don't, those that get attention get the whole shebang and those that don't get nothing.

All I am saying is you can't just say "I am sure it takes less time to get data than remodel a whole car". Sometimes what takes time and how long can really be surprising!

Quite often it's one thing to say "surely it would be faster to do things this way" but the reality is more complex.

No prob man I know your style as I said before I'm really excited GT5 is coming out so for these next weeks I will be happy and positive. I will let you come up with all the negative speculation or you being a realist and do your critique thing which is cool. I respect everything your saying, however will think on the bright side 👍

Hey be excited! I am excited! I have been waitinng for years, I bought a friggin PS3 and a G27 pretty much for this game! I am sure not going to hide my dissapointments in some facets or pretend there aren't still issues I am worried or concerned about, but I am definitely excited, if for nothing else so this ordeal can finally be over and the waiting and suspence behind what GT5 will really be like can be done with...

Of course I am sure it will all be quickly replaced by the same game with GT6

This is a very interesting read guys, thanks.

The thing I'd like to see your opinion on too is the practical side of this.

The way I see it we will never notice the difference in physics between premium and standard.

1) None of the cars will handle 100% like their real life counterpart, so it's hard to really compare. In other words: we wouldn't know if they implemented all the data correctly, as long as the bulk of the data is right.

Sure, but as you go back through the GT series, none of them handled like their real life counterparts, but you can tell the difference in the experience between each iteration... there are more options than "just like real life" and "not like real life" in terms of how accurate something is.

2) PD recorded more data than needed for the GT4 engine. I don't know this ofcourse, but it seems very reasonable for a company that wants to continue creating great racing games that they want as much data that can be possibly used in the future. My guess is that all data they thought they could use in the future that doesn't take a very long time to collect has been collected.

We would assume they did, but that's an assumption, and as I have said before, futureproofing is a fools game. Recording more info is one thing, but knowing what to record is another.

You might record a gagillion extra things, but it all depends on what you want down the road that matters. GT4 cars are a perfect example... they recorded more info than they used, but ultimately it was different and yet more info that they needed.

3) Between the 200+ premium cars they should have a generic counterpart for pretty much each of the 800 standard cars. So if they haven't collected a certain datapoint for a certain standard car, they can take the datapoint of a similar premium car and use that. For example, they can take the way one light hatchback's fuel consumption reacts to a spoiler and use the same formula on another light hatchback, without anyone ever noticing or caring.
Ofcourse this wouldn't be accurate, but it would be accurate enough (considering point 1 too) for the game.
Note that the data they haven't collected in my theory is data that most likely doesn't have a huge influence on characteristics as I think the most important data have been collected for all cars (partly in GT4 and partly through point 2).

Sure, that could be a case... but there could still be a lot of data that doesn't fit that bill.

4) The labour of updating the data of the cars is nothing compared to the modelling, sound, weather engine etc. You can probably have an intern or two do the job

You would assume so, but then again, maybe not... see my reposne above.

]qote]Saying the cars will handle differently sounds like the standards will behave in a very different way.
I think the standards will at the most behave a bit less accurately like the original car than the premiums will.
This might even be insignificant because of the precision of the physics engine.[/QUOTE]

Could very well be. Could be the standards and premiums have exactly the same data points and they haven't increased the data points at all since GT4 only impmroved the physics engine. It's just one of many possibliities.

Gees people are typing novels in here!
They should bundle all the posts in this thread, and publish a book.

GT5:Stands vs Premium, the Saga continues
An epic and riveting tale of emo rage

And we could publish as shorter coffee table companion book:

GTP Forums: The members who love to contribute little to the conversation, but complain a lot about those who do.

And maybe even include some keychain charms:

Posts that hope to appear witty in order to cover up their lack of actual value:

 

[Organ-Donor]

Hey be excited! I am excited! I have been waitinng for years, I bought a friggin PS3 and a G27 pretty much for this game! I am sure not going to hide my dissapointments in some facets or pretend there aren't still issues I am worried or concerned about, but I am definitely excited...

Deve there is a big difference between voicing your concerns with certain issues of the game vs. stating the same things over and over and over again. We are at 361 pages and you have been pretty much the only consistent person left in here from the start, saying the same exact things, crying about the same stuff, educating people on how this is unacceptable. Looking through your post history gives me the same feeling as leaving a funeral. I love cockpit view, it's the only view I like to use. When I found out we will have standard cars not supporting cockpit view I was shocked and pissed. I have accepted it and moved on like an adult (not swept it under the rug, but used adult logic and understand we won't have them and don't piss and moan about it 6 months later).

 

[UN_OC]

You're either a pesimist or an idiot.


Gt5 isn't going to create 2 seperate physics engines for everything, each has to reflect teh different roads (gravel/tarmac/snow etc...), each has to show rain each has to be able to do standard and pro levels of physics ecah has to do etc...

Let me explain how an 'engine' in a game works.

An 'engine' is created so that each aspect of the game doesn't have to be crated individually for each alternative (each part on each car on each track under each weather setting under each etc...). This adds continuity to all as some cars aren't reacting differently to others. This engine controls all the parts of its aspect. (i.e. a weather engine controls the weather, and probably lighting of day and night), a physics engine controls the physics.

It makes sense as every game made commercially in the last forever.


The weather, road and choices are all plugged into the engine, the car is then plugged in via its parameters and then you play it.

You can't be missing parameters. If some aer missing, then because the code relies on each paramter to interpret what should be happening, cars would glitch out everywhere, if that worked at all. Coming from someone who has written a few small things, I can tell you you can't just block off loose ends - the whole things works or it doesn't. The engine can't do only part of it because it all works with each other.

And secondly, they aren't going to have two seperate engines, one for standard cars and one for prem cars. You can't run them at once because of once again glitches when they occasionally meet and because it would be way way way way too hard for even a PS3 to run to complete engines at once. Hence standard cars can only be run with standard cars. This sounds even more limiting and unlike PD. Not only that, but then you have to have the code for teh engine for both. That takes even more space and would waste even more time.


I believe that the engine has been remodeled and hence is acting differently, but that data is still teh same for all of it. Yes the weight of the car makes the car drive differently under the new physics engine, but they don't have to go and reweigh the car to find out its the same.


THe main difference I believe between them is jsut the drivers cam, and the level of graphics. They need to be done seperately for each and every car. Hence you can have them both running together and each one has to be done up alone allowing PD to release the game with different standards.

 

[amar212]

You're either a pesimist or an idiot.

You can call him ignorant or missleading or subjective or whatever ike that, but please sustain for calling anybody an idiot. Especially the member who has wasted a severe portion of his life in deep explanation of his view on some matters.

I also do not agree with him in may fields, but I would never cal him an idiot nor would think that.

As far as "measuring" process for obtating the numbers and values for each car in the game is concerned, I think all of us who monitor GT series development can agree that PD probably have enourmous data for each car ever appeared inthe game.

You can just wath an DVD that came with GT4: Prologue and witness vast amaount of data being collected during just one day shown there.

So, presuming that Standard cars will have "lower" amount of data than Premium is just fishing in the deep waters of assumption.

 

[alucard0712]

Have you guys ever think that maybe new GT5 engine has much more parameters to define physics of a car?If it's like this then PD need to re-create physics of all 800 cars AGAIN.I don't think they have time to do so...I this case standard cars MAY have different physics engine, for example STANDARD (or do not have PROFESSIONAL).But again it can lead to new difficulties such as problems with mixing two types of a cars in one race.There is ONE answer: wait until release.

 

[WolfRacer543]

As for the "standards "may" handle differently" I really would like to know what kind of data PD might have collected on the premiums that is different that the standards? I mean what possible difference in data could there be? Please, enlighten me. As far as I have ever seen, right down to the hp and torque curves, almost every detail for every car is pretty darn close to accurate (I'm talking about GT4 here). So please, enlighten me.

There's a huuuuuuge amount of data you can collect about a car to determine its performance. Very few cars are understood well enough (even by their designers) to generate accurate simulations of how they'll perform.

Without going into huge detail and ignoring what goes into tyres alone, you have...

Roll resistance
Pitch resistance
Heave resistance
Roll damping
Pitch damping
Heave damping

All of which change with both position and speed, meaning each one isn't simply a constant, but rather a huge spreadsheet of data.

You then have what components of those variables above come from

Slop in joints
Suspension
Chassis

Then you talk about centre of gravity, but you also need to take into account rotational inertia in pitch, roll and yaw. You need to know what of those inertia's and masses are sprung and unsprung, and in the case of suspension components which are semi-sprung, you need to know how sprung they are.

Then a brief look at aerodynamics. You need to know how much downforce and drag is produced and the centres of pressure, but all of those variables will change with pitch and rideheight, as well as left/right aero balance changes with roll. This would require huge spreadsheets of data which would take days in a high quality wind tunnel to produce.

I haven't even talked about engines and drivetrain there's heaps of other stuff.

No idea what of those variables are used in GT, but you'd hope as the physics engine evolves, it requires more data about the car. Though a lot of the data is probably just made up.

 

[JDMKING13]

I also do not agree with him in may fields, but I would never cal him an idiot nor would think that.

As far as "measuring" process for obtating the numbers and values for each car in the game is concerned, I think all of us who monitor GT series development can agree that PD probably have enourmous data for each car ever appeared inthe game.

You can just wath an DVD that came with GT4: Prologue and witness vast amaount of data being collected during just one day shown there.

So, presuming that Standard cars will have "lower" amount of data than Premium is just fishing in the deep waters of assumption.

Yes Sir. 👍

 

[CoolColJ]

From memory, I remeber reading that the engineer behind one of the cars that were scanned by PD was asked to come up with 70 or so parameter values for the physics engine. The Mustang at SEMA

This one. Pretty sure there was a thread here about that, and they recorded it's sounds as well
https://www.gtplanet.net/closer-look-at-gt-award-winning-1970-mustang/

 

[DG_Silva]

Has anyone considered the possibility that the data of the Standard models regarding physics is just the same as the Premium models? I haven't seen anywhere that the simulation data representing the car is reduced, just the aesthetic modelling of the car.

If the standard models have less data recorded or interpreted into a physics model, I'd like to know where that has been said or written, because I've never seen anything saying that.

 

[SlipZtrEm]

Has anyone considered the possibility that the data of the Standard models regarding physics is just the same as the Premium models? I haven't seen anywhere that the simulation data representing the car is reduced, just the aesthetic modelling of the car.

If the standard models have less data recorded or interpreted into a physics model, I'd like to know where that has been said or written, because I've never seen anything saying that.

Well, ideally, given the years of development time, all of the Standard cars present in GT4 have been revisited more recently to record these new parameters, as well as get the necessary images and measurements down to start the Premiumization (yes, Premiumization ) process. The values related to the physics though could be implemented in fairly short order once the date has been collected though; the modeling can happen later.

I kind of look at it like a spreadsheet document; say, just throwing a number out there, the GT4 physics engine had 200 variables, values for each car, like front tire width, rear tire width, tire aspect ratio, things like that. GT5 has 400 variables. Unless the values for these 400 were recorded back in GT4 days (future-proofing/guessing), there would be blank cells in the new game.

Strictly speaking, depending on how in-depth things like aerodynamic drag are modeled in relationship with the new simulated damage, the Standards could already be missing out on certain parts of the physics engine. If my 458's hood is bent at an odd angle, and the game takes into account how bent body panels can wreak havoc on aero properties, then there's an area Standards just can't take advantage of.

I'm in the camp that believes it's far more likely that Kaz and crew have indeed grabbed most of these new-to-GT5 values for the Standard cars, these new variables that add to more nuanced physics calculations, but playing devil's advocate, we just don't know for certain. I'd think they've looked at far more than 200+ cars in real life between GT4 and GT5 though .

(and on a good note; despite me still seeing it as more of a random track generator and not a "course creator" in the expected sense, the newest video of it in action has me incredibly, incredibly excited about hopping on a road that's looking more and more like just a random, perfect back road. I see myself spending lots of time with that!)

 

[VXR]

I suppose there could be concern that in between porting the data from GT4 and making new models that fit within the GT5 physics, that does leave the modern cars with much more up-to-date data for the new game's physics, which could be quite different from how GT4 was.

Hopefully for the guys not bothered about standards, they collected more data than was used in GT4's physics engine, otherwise, I can't see them getting details on all 800 cars again to add extra detail within GT5's physics calculations. I guess in that situation, it could leave the majority of standard cars feeling less individual between themselves through lack of extra data, but the 'physics' would still be the same, regardless.

 

[Anes]

Holy crap Devedander !! 👍

 

[Griffith500]

@ SlipZtrEm: actually, it's Premiumification

My thoughts on this are that PD have likely always had more data points than they need for the "simulation", I'd bet that's still true even for GT5.

Even if they can't find an accurate value to slap in a given position, they can always tweak a value from a similar car or a car with a similar physical attribute that this number is governed by. People do this all the time, with things that peoples' lives depend on (I'm looking at you, Mr. Civil Engineer! )

The only way to know if there is a difference would be to test a car that is both Standard and Premium; then again, they would probably use the same data points.
I'm sure the standards will feel just as "rich" in the game, to drive, as the Premiums will. Otherwise, they needn't have bothered with them.

 

[Gejost]

Some people are really addicted to this forum or have way too much free time on their hands.

 

[5Liter]

Some people are really addicted to this forum or have way too much free time on their hands.

LOL i was thinking the same thing....at the end of the day when you're finally playing GT5, its just a bunch of pixels on a screen.

 

[yellowclone]

There is a saying about what happens when you assume... your whole post is based on a few assumptions about what I meant and not what I actually said.

See how that assuming works out? Try actually reading what you are responding too instead of tossing out a knee jerk gut response:

I didn't assume that's going to happen, I simply said it's a possibility that can't be ruled out and if it turns out to be true will be yet another way standards are broken.

This is all a general assumption for an example, I am not saying they did or did not do things this way:

Again, you said I assumed, which I might have, but you also said not to assume. You said that you didn't assume anything, and then a few post later, you said it was a general assumption

Deve is just mentioning a possibility; the people screaming "NO, ASSUMPTIONS!" are being equally presumptuous to think Standards have all the same values as Premiums. We simply don't know one way or the other.

I kept saying assumption because he told me not to assume, but yet he was doing the same thing in a sense, but not admitting to assuming. To me, the word unlikely, alone, is an assumption, so he kind of brought it on himself. Just because he gives a LONG educated guess does not make it true, and therefore, it is an assumption/speculation.

Deve there is a big difference between voicing your concerns with certain issues of the game vs. stating the same things over and over and over again. We are at 361 pages and you have been pretty much the only consistent person left in here from the start, saying the same exact things, crying about the same stuff, educating people on how this is unacceptable. Looking through your post history gives me the same feeling as leaving a funeral. I love cockpit view, it's the only view I like to use. When I found out we will have standard cars not supporting cockpit view I was shocked and pissed. I have accepted it and moved on like an adult (not swept it under the rug, but used adult logic and understand we won't have them and don't piss and moan about it 6 months later).

👍👍