Revving before a race...

[ilovethescudo]

I know it's frankly "customary" for cars to rev their engines before a race... my question is why exactly? I noticed in the License test with the Viper GTS, that revving it gives you slightly better power, but why is that? And what exactly is the best RPM range to rev the engine through? From idle to the redline? From halfway to the redline? Flooring it the entire time until the race starts?...

 

[R35 SLIDER]

it is usually done to get oil flowing through the motor so that there is oil in the motor while the car is racing. the optimum rev range is from right down low to just before redline.

 

[ItsHim]

It depends on both engine characteristic and tyres. You want to put as much torque on the tyres so they are slightly above their grip threshold. Slight squeal is faster than a silent start. The amount of torque is dependent how much the engine revs and how much torque the engine puts out at a certain rev range.

I'm afraid you'll have to try to figure it out yourself.

Edit: Maybe I didn't get the question. I used to rev my scooter because it took some time to get the engine up to optimum launch revs. With short bursts of throttle I held it in the optimum range and that was both quicker and gave me more cotroll than starting from an idling engine.

 

[paddycars1]

Sometimes second gear gives you some help on the start line

 

[Parnelli Bone]

Real-life: race cars often run lean, which means they take more air than usual into their intake. This causes them oftentimes to stumble at idle. This is how they manage to get the best air-fuel mixture at higher revs, tho, hence it's better to rev them higher at the start of a race, rather than possibly have a stall happen. The best acceleration can often be garnered at higher revs, since the clutch can be manipulated alot more effectively in real-life than it can in GT4.

The game: Some cars do best when accelerating from near-redline, while others do better at lower revs. Since we don't have to worry about air-fuel mixture, stalling, burnt clutches or any of that stuff, it's possible to let a car start at very low revs (especially FF or FR drivetrains) so that the drive wheels don't slip. It's also possible to simply drop the gas-hammer at very high revs without worrying about anything breaking or burning up.

Virtually every car in GT has what I call a "sweet spot". This is the best area on the tachometer to nail when accelerating from a dead-stop. It varies per car, mostly because some can't handle lots of revs right away without wheels slipping, and in some other cases, GT4 does simulate a bit of clutch-slipping, too, which I've noticed mostly with cars that won't spin wheels when you nail the throttle.

In other words, a good for-instance would be an all-wheel drive Skyline GT-R. Try and search for the best area on the tach to nail. License test A13 (Skyline at Trial Mountain) is a great example. You might think it would be best to simply floor the engine, so that it's pinging its RPM limiter. But I've found that it's best to start the test at 6,000 rpms, and once the test starts (BEEP!) to quickly nail the gas a little harder so that the tach needle lands at 7,000. I've gained precious tenths of seconds this way during this test. Simply overreving the GT-R will cause a loss of speed, which is supposed to be GT4 simulating clutch slip in my opinion.

Since I've done A13 over and over and over and over and over and over and over again, I've definitely noticed that the ghost from a test-run in which I overreved the engine always falls behind if I nail the tach more carefully. 💡

HTH

 

[Austin343]

I don't believe that it's clutch slip, Parnelli. That Skyline suffers a huge drop in power when you stray into the redline, with the power peak being much closer to your optimal launch speed. So the car accelerates slowly from the start when you're bouncing off of the limiter because there's little power available in this particular car when the revs are still falling; it only begins to accelerate quickly when the revs drop into the powerband. Also, when you're hitting the limiter, approximately half of the launches you make will have the timer start when the rev limiter still has the throttle released (that is, it's the half of the "bounce" off of the limiter during which the revs are falling). So you will only start to move a fraction of a second after the timer starts.

 

[Parnelli Bone]

I don't believe that it's clutch slip, Parnelli. That Skyline suffers a huge drop in power when you stray into the redline, with the power peak being much closer to your optimal launch speed. So the car accelerates slowly from the start when you're bouncing off of the limiter because there's little power available in this particular car when the revs are still falling; it only begins to accelerate quickly when the revs drop into the powerband.

Mm hmm. Could be.

Also, when you're hitting the limiter, approximately half of the launches you make will have the timer start when the rev limiter still has the throttle released (that is, it's the half of the "bounce" off of the limiter during which the revs are falling). So you will only start to move a fraction of a second after the timer starts.

Agreed. That's a rather n00bish action there--letting the tach needle bounce off the limiter, thinking it's the best way to accelerate. I've seen some posts in the past where someone has suggested it doesn't matter if you bounce off the limiter or not at the start of a race. But it's a bad practice.

 

[ilovethescudo]

I'm not exactly talking about launching at a certain RPM range... I'm talking about running the engine through a certain RPM range several times before the race starts... In virtually all standing starts with opponents, the constantly run their engines through a certain RPM range while waiting for the race to start. In driving mission 34, even the last car to start before you do is constantly running his engine up and down through an RPM range before he starts... my question is why do this? I know all about launching at a certain RPM, but why rev continuously up and down... does it serve to warm the engine somehow in the game?

 

[Austin343]

In the game, it does nothing (I don't know about its utility in real life). The AI cars do it to add a feeling of excitement for new players; having all of the cars sitting silently on the grid until just before the start would be quite boring in comparison.

 

[ilovethescudo]

But like I've said before, with the Viper, doing it seems to give the engine a tad bit more power than if I just either idle, redline it, or keep it at a certain RPM the entire time before launching. I'll have to do some tests on the test course after I get done with the GTWC...

 

[kalmjaws]

Hi All!
If I remember correctly, the slalom license tests have a split timer right after launch. This is where I noticed the best time at that split came from a launch at lower than peak RPM's. I think in this particular example for my auto tranny the best launch was between 5,500 and 6,000 rpm's

 

[Parnelli Bone]

I'm not exactly talking about launching at a certain RPM range... I'm talking about running the engine through a certain RPM range several times before the race starts... In virtually all standing starts with opponents, the constantly run their engines through a certain RPM range while waiting for the race to start. In driving mission 34, even the last car to start before you do is constantly running his engine up and down through an RPM range before he starts... my question is why do this? I know all about launching at a certain RPM, but why rev continuously up and down... does it serve to warm the engine somehow in the game?

Well, it wouldn't keep the engine warm. PD didn't program engine temperature, after all, like ToCA has. The way they rev doesn't do anything in GT except keep the Ai's engines up in a higher RPM range, which could help them accelerate the fastest. The sim cars don't have TCS on (my opinion) but in some cases, they are still able to get off-the-line without major wheelspin.

But this is all theory, I have no idea why PD programmed it this way. Perhaps it's to simply sound realistic. After all, if the entire field of cars just sat there idling, it wouldn't sound as exciting, know what I mean? Then everyone playing the game would start to complain.

 

[Austin343]

It's been pretty well demonstrated that opponent cars use the default driving aid settings (10/10/5 or 10/10/7, depending on the car). The only exceptions are at Motegi Super Speedway and George V Paris, where they retain the ASM but turn off the TCS.

 

[Giygas]

To make things even more complicated(maybe), some cars don't even reach full power before they top out at the redline! The spoon s2000 racer does this, also i think some other Japanese cars do as well.

 

[Parnelli Bone]

For what races? I totally disagree with whoever started that line of thought. Some examples why it's wrong.

1. When flying over large bumps (like at Seattle down the staircase-street area), TCS totally cuts power for any car we drive. Not true of the Ai. The power/gasoline meter still shows for them.

2. In general, ASM and TCS would interfere a heck of a lot more than it does while the Ai races around, really killing their spontaneity while cornering. It would slow them to a crawl, especially with default settings. The only reason the Ai typically drives worse than we do is due to the way they brake (too early/too often) and some of the bizarre decisions they make while turning-in and accelerating out.

I know what a GT4 car behaves like with TCS and ASM on, I highly doubt most of the cars in our game use it. Anyways, let's hear your side of the discussion.

 

[Austin343]

Parnelli, I think that it's plainly obvious that the AI cars use TCS (with the exceptions that I mentioned), and at the default setting. Try driving a car on which its effects are easy to see, such as the TVR Cerbera Speed 12: use full throttle from a standing start in first gear with TCS first on, and then off. Then compare the pattern of revs and the sounds that you hear with the AI Speed 12 in an Arcade Mode race or in, say, Blackpool Racers (the TVR manufacturer's race), ideally on a track with a standing start. You will find that the results of your driving with TCS exactly match those of the AI's driving.

The effects of ASM are harder to see with the limited information that you get from the display of the opponents in a replay, but I firmly believe that it is activated for them. Try taking a car with a lot of oversteer, such as the RUF Yellow Bird, and driving it first with the ASM off and then with it on. In particular, try carrying too much speed into a corner. Then watch an AI opponent do the same in the same car, and you'll find that its slides look and sound very similar: a slow drop in speed as the car seems to proceed very "inevitably" into the sand, grass, or barrier. And an even clearer demonstration can be found by watching the blue throttle bar of an AI car going through a long and fast corner in a replay. It continues to show full throttle even though the speed drops (the effect can only be seen in some cars in some corners, however). If you take the same corner in the same car without aids, you will find that you can easily continue to accelerate through it, but with ASM on, the loss of stability that it perceives will cause you to slow while turning, just as the AI do.

I'm sorry for hijacking the thread, but I really don't believe that there's any basis for the hypothesis that cycling through a range of RPMs before a race assists with the launch. GT4 just doesn't model such complexity.

 

[Parnelli Bone]

Sorry austin343. I see what you're saying. Guess we'll have to agree to disagree on this one.

I have an experiment I'd like to try out, but I'm at work now so probly won't get to it 'til later tonite or tomorrow.

Edit start: It's always been my theory that GT4 uses the same computer programming for B-spec and Ai cars. Within this programming, the game allows for either Bob or the Ai to 'learn' a track, getting slightly better (or trying to) lap after lap. This is how Battle Points, Machine Points, and Track Points go up.

I've noticed that when Bob is completely green (no points yet), he drives like the Ai does in Beginner-level races. Too many brakings/lots of confusion in those corners. He eventually gets better (as the Ai sometimes does in races lasting longer than 3 laps), yet still can't touch our skillz.

It's also always been my theory that driving aids aren't used in either B-spec, or for the Ai.

Okay, so what I've done is taken an ordinary PT Cruiser to the Opera/Paris track, put it in B-spec mode, and compared lap times in various ways. The PT was completely stock. No tuning, other than various manipulations with the Driving Aids. I should add that I used my 'top Bob'...the one that has the highest amount of points, for this experiment. Despite this, Bob drove really badly...occasionally slamming or rubbing walls and whatnot. His best times were the ones that had no slams, but perhaps a minor wall-rub or two.
-------------------------------------------------------
First I let Bob drive with Aids completely off. Speed was left on '3', and was not changed for the duration of the test.

His best lap time (after 5 laps): 1:52.994

Next, he drove with the Driving Aids at default (10/10/5).

Best lap time (after 5 laps) 1:52.994. Exactly the same.

I should note there were absolutely no differences in the way Bob drove with driving aids supposedly OFF or ON, so far as how Bob managed to weave the car in and out of Opera's 18 turns. There were occasionally small chirps of wheespin from those front wheels (very small chirps) if they were ON or OFF, as though Bob cuts the throttle at the very slightest mention of wheelspin.

````````````````````````````````````````````````````````````````````````````
Next, I tried the experiment myself using an AUTOMATIC transmission in the same car, so I would be confined to the same gear limitations as Bob or an Ai car. I did not hold gears, nor did I use E-brake or any other tricks I'd normally use. To avoid wheelspin with the Aids OFF, I tamed the throttle 'til smoking tires stopped, which is (I believe) the way the Ai and Bob avoid wheelspin. With the Aids ON, I could keep throttle plastered almost the whole time (95% of the time), unless I was braking.

Driving Aids OFF Best lap after five? 1:45.234

Now the same five laps with Driving Aids on default (10/10/5). 1:50.007


So to summarize. Bob drives exactly the same no matter how we set those Driving Aids. But I can't manage the same times (nearly) with them on or off.

Conclusion? I believe B-spec and the Ai are run from the same set of computer codings, and use the same techniques for dealing with wheelspin and body sway issues. In effect, both the Ai and Bob cut throttle at the slightest bit of wheelspin and/or recalcitrance, and do not use Driving Aids. If Driving Aids were used, there would definitely be some sort of major difference in lap times and the PT's behavior around the track.

Thanks for your time.

 

[Austin343]

See, Parnelli, you're missing a major point. When we change the driving aids on a car, it does not change the aids that the AI (B-Spec) actually uses. It would just be way too much trouble for PD to implement AI that control wheelspin and slides on their own, so they choose to program them with unchangeable, default driving aid settings. I mean, given the identical lap times with aids off or on, do you really think that the AI cars somehow manipulate their controls "manually" in exactly the same way that the driving aids do? And it really is important to watch the red and blue brake and throttle bars that you can watch in the replay. They very clearly show the use of full throttle for the majority of the time, even when the tires seem to be straining to break free; there is no explanation for that other than TCS. I'm sorry to say it, but your test proves nothing at all. I urge you to try more dramatic tests like the one that I mentioned above that uses the Speed 12.

 

[Parnelli Bone]

See, Parnelli, you're missing a major point. When we change the driving aids on a car, it does not change the aids that the AI (B-Spec) actually uses. It would just be way too much trouble for PD to implement AI that control wheelspin and slides on their own, so they choose to program them with unchangeable, default driving aid settings. I mean, given the identical lap times with aids off or on, do you really think that the AI cars somehow manipulate their controls "manually" in exactly the same way that the driving aids do?

Actually, I wasn't sure if setting the Aids would make any difference, hence I was experimenting up above. I used a setting of "3" for speed because at this level, Bob is taking some risks, yet is still driving conservatively overall. This is the constant. Then I switched Driving Aids on and off (variables) to see if there would be any difference. Either way, Bob's not crashing into every wall, but he still makes some mistakes here and there.

If we make a change to transmission, limited-slip, or suspension, for instance, such changes do appear in B-spec mode, and can make a difference in lap times. So I just wanted to see if PD implemented the same set of possibilities for Bob (and henceforth the Ai) since both are computer-driven. 💡

It would be counter-productive if PD coded the same TCS/ASM that we use for the Ai, and then used a completely different set of rules for Bob. I was trying to see if it's even possible that changes in TCS/ASM make a diff for computer-driven vehicles.

And it really is important to watch the red and blue brake and throttle bars that you can watch in the replay. They very clearly show the use of full throttle for the majority of the time, even when the tires seem to be straining to break free; there is no explanation for that other than TCS. I'm sorry to say it, but your test proves nothing at all. I urge you to try more dramatic tests like the one that I mentioned above that uses the Speed 12.

The blue (gas) meter doesn't really prove anything either. It's possible to be using less-than-full throttle, for instance, killing wheelspin, yet the meter is still registering that full throttle is being used.

So far as the Speed 12 test goes...nah, I think I'll pass. Does the Speed 12 even appear as an Ai vehicle? I don't remember ever seeing it, but you'd probly know better than I. I was using the PT Curser as a lab rat because it ap`pears more frequently as an Ai car; hence we can observe the way it drives via computer, compare such driving to Bob's driving, and then take a spin ourselves behind the wheel. It is just as sloppy to drive as a Speed 12, and should cause the Driving Aids to go overboard, assuming the Ai uses the same system we do.

So when the computer drives it, it does so in a similar fashion (lots of leaning, lots of mistakes, etc) no matter if Bob or the Ai is doing the driving.

Now, it may be possible that you're right to some degree, and PD did implement a rudimentary Driving Aid system for the Ai & Bob. Such a system would be simpler than the one we use, and would allow a certain amount of parameters into the game's physics when the computer drives around. Get outside these parameters, and the computer does something about it; taming wheelspin and extraneous movement before the human eye/ear can even detect it.

But it ain't default TCS/ASM settings like we use (10/10/7). Unconvinced. No way would the PT Curser up above be able to even take those wide, wacky paths around Paris/Opera without some sort of vast interference mid-corner. Kaz/PD have admitted during interviews that the first four Gran Turismo games use the same driving physics model, which gets more detailed with each game. Since wheelspin is extremely rare in Ai cars from GT1 thru GT3, I'd say PD is simply using a modified version of the same rules they've been using for computer cars since the mid-'90s, which has nothing to do with our TCS/ASM system.

 

[Austin343]

Evidently, we'll have to agree to disagree. I just see no logical reason why PD would go to the trouble of developing an elaborate system for the AI to control their cars, when the default driving aid settings are very easy to implement and match the behavior of the AI cars exactly. And I have no idea how you figure that the throttle meter could be full when the "driver" of the car isn't inputting full throttle. Yes, less than full throttle can be used when the bar is full, but only because TCS is limiting it. It's exactly what happens when we drive a car with TCS: We may floor the throttle, which is represented with a full blue bar, but the system lets the car receive significantly less because wheelspin would develop otherwise.

 

[Parnelli Bone]

Evidently, we'll have to agree to disagree. I just see no logical reason why PD would go to the trouble of developing an elaborate system for the AI to control their cars, when the default driving aid settings are very easy to implement and match the behavior of the AI cars exactly. And I have no idea how you figure that the throttle meter could be full when the "driver" of the car isn't inputting full throttle. Yes, less than full throttle can be used when the bar is full, but only because TCS is limiting it. It's exactly what happens when we drive a car with TCS: We may floor the throttle, which is represented with a full blue bar, but the system lets the car receive significantly less because wheelspin would develop otherwise.

Yup, we'll agree to disagree. One last thing though. Actually two.

Firstly, just wanna say it's awesome to finally have a mature discussion on a topic such as this which hasn't downgraded into Radio GT-style mudslinging. I've known you a long time, Austin343, and consider you an online-friend. So this isn't a flame-war, it's merely a comparing of notes. 👍

So far as the blue meter goes, I drew that conclusion as I drove the PT Curser the other day with TCS OFF, but I've noticed it with plenty of other cars I've driven/raced. There were times when I've driven around that wheelspin starts to happen, I cut the throttle ever-so-slightly, and the wheelspin goes away, but the blue meter is still registering that full-throttle is being used. It's a minor flaw of the relationship between the dual-shock and computer, I believe.

Why didn't PD just use the same system of default Driving Aids for the Ai that we're given as default? Because it would make the entire game feel incredibly sluggish, way more sluggish than it already is. The Ai would constantly be getting penalized in and out of corners.


Anyways, this final thought is conclusive. You're saying the Ai uses default Driving Aid settings in GT4, I'm saying it doesn't.

1> Take a car (any car) and drive it round Seattle with TCS ON. Doesn't matter if the setting is 1 or 10, or anywhere in between.

2> When you get to the so-called "staircase" area (up hill jumps), notice how TCS interferes immediately, cutting off all power. The blue meter completely vanishes, no matter if we've got the throttle plastered or partially on.

3> I just did the PD Cup race at Seattle yesterday, but this experiment works for any Seattle race: FR Challenge, Pro-level Compact Car races, etc. When viewing the replay, if the Ai had TCS ON as we do, power should be cut as the Ai goes up the staircase. Or down it.

4> However, during the PD Cup replay, I noticed that not only does the blue meter stay fully on, but the Ai manages to keep its throttle floored the entire time it's doing those jumps. Result? Tach needles start bouncing past redline, and onto their RPM meters.

This proves it. If the Ai were using the same system of Driving Aids that we are, they'd be handicapped massively, and we'd be able to notice it, especially at Seattle. We'd all be complaining about how dreadful GT4 is even more than we already do. None of us would even be playing GT4 anymore; it would be so easy to us (even more easy than it already is) to blow the Ai away, and we'd all be discussing ToCA or some other game instead.

But like I said in the previous post, I believe in what Kaz admitted during that interview I read. The Ai is using the same driving physics (albeit modified) that PD had implemented for the first Gran Turismo. Any hint of wheelspin, it gets consevative with throttle. Any hint of body-sway recalcitrance and/or slipping tires, and it stops pushing as hard. This is why tire squeal is kept to a minimum in GT4; the ai is so focused on trying to make perfect arcs thru turns, with squeal-free cornering, that it doesn't truely push the way a human driver would. They're trying to solve these dillemas with GT5.

This can be called a form of Driving Aid if you like (since it's a computer calling the shots) but it ain't the same system of TCS/ASM taht we use.

 

[Austin343]

Parnelli, I just got around to watching AI cars at Seattle, and you're absolutely right: they bounce off of the rev limiter on the jumps while TCS on your own car keeps the RPMs more or less constant. (However, the blue meter does stay full, of course - I don't know what you were doing that made it disappear over the jumps. And by the way, when you perceive that you have partially released the throttle while the blue bar stays full, that's a function of the analog sticks; they are programmed with the "full throttle" position being slightly below the top of the stick's range, so that you don't register less than full throttle when you push the stick fully up but also slightly to one side.) I don't know how to explain the "Seattle jump phenomenon," but I'm not the type to discount a large pile of evidence with a single counterexample. My best guess is that there are certain exceptions programmed into the AI's driving aids, such as cutting them when airborne. Everything else that I have observed indicates that the AI cars use TCS and ASM, but this is puzzling, if not entirely inexplicable.

 

[Parnelli Bone]

How could driving aids be cut all the sudden while they're airborne, tho? That sounds way too complicated. I made a mistake, tho. You're right...the blue meter does stay full-on down when airborne, even tho TCS is cutting power to the drive wheels.

GTSail has sent me a PM regarding this discussion, and referred me to several races I should check out (which I haven't ever done yet) like the Blackpool TVR ones. So I'll view these and tell you what I see. TCS in particular definitely has a look to is when we use it. It'll be fairly obvious to view under these circumstances.

 

[Kenuwine]

To clear this up (because this is funny) they rev up in the game for a good launch...Have u ever raced an Evo Sti skyline or any 4wd car, they shoot out the gate..its for a better start in the game. In LIFE, u actually qaulify and dnt start at a standstill. imagine if everyone took off the same time on a real track, people are going to be dying to dive to the number one spot...It's real simple..They launch in the game, thats all the reving is for.