• Welcome to PiBoSo Official Forum. Please login or sign up.
 

It's a long way to the top ( if you wanna rock'n'roll )

Started by PiBoSo, October 22, 2016, 08:34:46 PM

Previous topic - Next topic

HornetMaX

Quote from: PiBoSo on October 23, 2016, 05:51:03 PM
Unfortunately there isn't much left to improve on the physics' side, so a new comparison video will not make such an impact.
Not that they would be visible in a video, but a couple of things that come to mind physics side (aside the already mentioned problems leaning uphill):

  • Gearbox sim
  • Brake thermal simulation (not sure where we stand on this: I thought there was none, but I may be wrong)
Very low speed handling is also sometime problematic (more a virtual rider issue than real physics).

Hawk

Quote from: HornetMaX on October 24, 2016, 07:08:54 AM
Quote from: PiBoSo on October 23, 2016, 05:51:03 PM
Unfortunately there isn't much left to improve on the physics' side, so a new comparison video will not make such an impact.
Not that they would be visible in a video, but a couple of things that come to mind physics side (aside the already mentioned problems leaning uphill):

  • Gearbox sim
  • Brake thermal simulation (not sure where we stand on this: I thought there was none, but I may be wrong)
Very low speed handling is also sometime problematic (more a virtual rider issue than real physics).

As far as I understand from what I've been reading about the "Pacejka Magic Tyre Formula", it doesn't work well at low speed. If I remember rightly it's because of a situation were it starts dividing by zero in a calculation? I didn't really understand it fully but that's the impression I got that the formula gets into a situation at very low speed were it becomes unstable and dev's have to artificially stabilize the formula until a certain speed is attained?

Hawk.

Stout Johnson

Quote from: HornetMaX on October 24, 2016, 07:08:54 AM
Quote from: PiBoSo on October 23, 2016, 05:51:03 PM
Unfortunately there isn't much left to improve on the physics' side, so a new comparison video will not make such an impact.
Not that they would be visible in a video, but a couple of things that come to mind physics side (aside the already mentioned problems leaning uphill):

  • Gearbox sim
  • Brake thermal simulation (not sure where we stand on this: I thought there was none, but I may be wrong)
Very low speed handling is also sometime problematic (more a virtual rider issue than real physics).
3. realistic high-sider physics
4. better simulation of powerslides and trail-braking (probably related to tyre model and virtual rider)
5. less rigid linkage between bike and rider
(6. rider ragdoll ?)

Then it would be perfect imo...
    -----------   WarStout Kawasaki Team   -----------

doNico

Quote from: HornetMaX on October 24, 2016, 07:08:54 AM
Quote from: PiBoSo on October 23, 2016, 05:51:03 PM
Unfortunately there isn't much left to improve on the physics' side, so a new comparison video will not make such an impact.
Not that they would be visible in a video, but a couple of things that come to mind physics side (aside the already mentioned problems leaning uphill):

  • Gearbox sim
  • Brake thermal simulation (not sure where we stand on this: I thought there was none, but I may be wrong)
Very low speed handling is also sometime problematic (more a virtual rider issue than real physics).

Thats one of the biggest issues. Less experienced riders (like me) always crash again after a reset... especially if you reset the bike and you face the wrong race direction or your on the grass ... Once the virtual rider has put his foot down on the ground, that shouldnt be possible. I used to put myself into similar positions in real life (Most, Sachsenring and Oschersleben aswell) and i NEVER felt while going low speed. Pib has to fix that.
In GPB, you sometimes  steer to the right and bike falls to the left while going slow ... sometimes it does a flip flop :S weired. IMHO thats one of the points, people dont buy the game because you need the feeling of success. The first answer i got on the forums was "GP Bikes is the top tier of demoralizing games ever" ... not sure who said that but its true. Of course its a sim ... but it would be easier for less experienced players or beginners to have some ingame helps. 

(Im still racing with clutch help ... it seems like racing with clutch help dissabled has some advanteges too as the engine brake is a bit better (best possible to try at Jerez) )

~doN

Stout Johnson

Quote from: doNico on October 24, 2016, 04:55:10 PM
Quote from: HornetMaX on October 24, 2016, 07:08:54 AM
Quote from: PiBoSo on October 23, 2016, 05:51:03 PM
Unfortunately there isn't much left to improve on the physics' side, so a new comparison video will not make such an impact.
Not that they would be visible in a video, but a couple of things that come to mind physics side (aside the already mentioned problems leaning uphill):

  • Gearbox sim
  • Brake thermal simulation (not sure where we stand on this: I thought there was none, but I may be wrong)
Very low speed handling is also sometime problematic (more a virtual rider issue than real physics).

Thats one of the biggest issues.  [...]  I used to put myself into similar positions in real life (Most, Sachsenring and Oschersleben aswell) and i NEVER felt while going low speed. Pib has to fix that.
In GPB, you sometimes  steer to the right and bike falls to the left while going slow ... sometimes it does a flip flop :S weired.
You are right, the low-speed crashes are a bit of a pain. Especially if you know, in real life you would just put your foot down to stabilize. Also I think the virtual rider should have some sort of filter to eliminate certain behaviour that would be causing a stupid crash.

But in general those low-speed crashes are owing to the fact, that Piboso has implemented a real counter-steer behaviour which only really takes place from a certain speed on (everybody that rides a motorcycle in real life knows what I mean). And on the edge of that speed or below, it is actually quite hard to simulate I guess. So yeah it needs some sort of improvement or VR-help, but I take the edgy behaviour over anything uncomplicated but also very unrealistic. Just look at how turns in Ride2 are initiated - there is no counter-steering whatsoever. Although a counter-steering to initiate a turn is only very nimble, it still looks odd when it is missing.
    -----------   WarStout Kawasaki Team   -----------

iVolution

Quote from: Stout Johnson on October 24, 2016, 07:55:05 AM
Quote from: HornetMaX on October 24, 2016, 07:08:54 AM
Quote from: PiBoSo on October 23, 2016, 05:51:03 PM
Unfortunately there isn't much left to improve on the physics' side, so a new comparison video will not make such an impact.
Not that they would be visible in a video, but a couple of things that come to mind physics side (aside the already mentioned problems leaning uphill):

  • Gearbox sim
  • Brake thermal simulation (not sure where we stand on this: I thought there was none, but I may be wrong)
Very low speed handling is also sometime problematic (more a virtual rider issue than real physics).
3. realistic high-sider physics
4. better simulation of powerslides and trail-braking (probably related to tyre model and virtual rider)
5. less rigid linkage between bike and rider
(6. rider ragdoll ?)

Then it would be perfect imo...
Guys don't start on physics improvements now... i am still thanking baby jesus and the holy spirit for the fact that mister P. is actually crushing some bugs now. Checking the daily development makes me happy nowadays.

Warlock

Quote from: iVolution on October 24, 2016, 11:03:02 PM
Guys don't start on physics improvements now... i am still thanking baby jesus and the holy spirit for the fact that mister P. is actually crushing some bugs now. Checking the daily development makes me happy nowadays.

Exactly. Those problems are minimal compared to the issues that are under investigation right now, so please have some patience.
Let Piboso solve the big issues, then i'm sure he will be happy to work again improving physics, once the unpleasant work is finished.

Stout Johnson

October 25, 2016, 05:51:10 AM #22 Last Edit: October 25, 2016, 05:54:09 AM by Stout Johnson
Quote from: Warlock on October 24, 2016, 11:59:00 PM
Quote from: iVolution on October 24, 2016, 11:03:02 PM
Guys don't start on physics improvements now... i am still thanking baby jesus and the holy spirit for the fact that mister P. is actually crushing some bugs now. Checking the daily development makes me happy nowadays.

Exactly. Those problems are minimal compared to the issues that are under investigation right now, so please have some patience.
Let Piboso solve the big issues, then i'm sure he will be happy to work again improving physics, once the unpleasant work is finished.

Exactly my thoughts - I did not state anything else ??? I also read everyday in the daily dev blog. Every time I read something related to netcode or dynamic track, I am all smiles  :)

But that does not mean, we cannot discuss the way further down the road. If we really should be afraid to talk about anything here, then we should close the whole forum down until netcode is fixed.
    -----------   WarStout Kawasaki Team   -----------

Warlock

But Piboso could be tempted to work on the physics !   ;D  don't distract him!  :)

Stout Johnson

October 25, 2016, 06:49:42 AM #24 Last Edit: October 25, 2016, 06:55:17 AM by Stout Johnson
Quote from: Warlock on October 25, 2016, 06:09:23 AM
But Piboso could be tempted to work on the physics !   ;D  don't distract him!  :)

Then shut down the forum.  ;) (Btw, we answered to something he wrote, not the other way around.)

But I really agree... Piboso should lock himself into fixing the netcode stuff.
    -----------   WarStout Kawasaki Team   -----------

HornetMaX

Quote from: Stout Johnson on October 24, 2016, 05:05:27 PM
You are right, the low-speed crashes are a bit of a pain. Especially if you know, in real life you would just put your foot down to stabilize. Also I think the virtual rider should have some sort of filter to eliminate certain behaviour that would be causing a stupid crash.

But in general those low-speed crashes are owing to the fact, that Piboso has implemented a real counter-steer behaviour which only really takes place from a certain speed on (everybody that rides a motorcycle in real life knows what I mean). And on the edge of that speed or below, it is actually quite hard to simulate I guess. So yeah it needs some sort of improvement or VR-help, but I take the edgy behaviour over anything uncomplicated but also very unrealistic. Just look at how turns in Ride2 are initiated - there is no counter-steering whatsoever. Although a counter-steering to initiate a turn is only very nimble, it still looks odd when it is missing.
As far as we can see, the VR is an adaptive controller, in the sense the value of the parameters depend on the bike speed. This in principle should allow to have "somehow different controllers at different speeds", but it's likely hard to find a tuning that gives the expected results at very low and low to very high speeds. It seems like the current tuning is OK for low to very high speeds (what you usually have while on track), but fairly bad for very low speeds: it's not even granted it is possible to find a tuning that satisfies us at all speeds.

As the bike behavior is very different at very low speeds, I'd try with the current VR for low to high speeds and a dedicated controller for very low speeds. But that requires handling the "crossover" between the two controllers, not exactly trivial.

Not a huge deal to me anyway, maybe something he can think about once he's back from "operation netcode" :)

P.S.
I'm really happy to read in the daily dev posts that stability issues (that may be independent from the netcode stuff) are taken care too !