New onboard view

[HornetMaX]

Thx !

As it seems the new view mode is appreciated, it may be interesting to understand if the improvement comes from the pitch change, from the yaw change or from both.

Two things I was wondering about:

1. If MXB already has the same pitch view behavior (sight aligned to the ground), how does it cope with whoops and similar stuff ?
Is the link "ground slope --> view pitch angle" rigid ? Some sort of filtering / deadzone / averaging of the ground slope over a given distance ?

2. Maybe to limit the weirdness in case of wheeling (looking at the tank), you could make it so that the sight is kept aligned to the ground only if the pitch angle of the bike is below a certain threshold. Example: flat track surface, small wheeling (let's say 10deg), the sight stays horizontal, but big wheeling the sight goes up. Of course you could make this more sophisticated (perfect correction below some threshold, filtered outside that). Or maybe just some filtering will work fine: will compensate bike pitch movements (relative to "ground pitch") but only within a limited control bandwidth, i.e. it will react with limited speed: hitting a ramp too fast will cause the sight to stay horizontal in the first moments and become parallel to the ramp only after.

[tchemi]

@MAx,
Yes, it is smart and if we take the actual hypothesis, this should be achievable.

Standard view : pitch (of the view) aligned to the bike
Implyes that we can align a camera axis with bike elevation
Mode=1 view : yaw (of the view) aligned to the bike direction of motion
Implyes that we can align camera with bike direction of motion

If we can align a camera angle with ground/ actual bike elevation / bike actual direction, we can have a smooth and high end cam behaviour on wheelies (sorry, when I say 'we can', you know what I mean)

Then, your last sentence is very interesting !!
It made me think that there is something you seamed to have miss. There is a longitudinal movement from front to back / back to front. If you grab the brake, you will move forward like if your balls where smashing on the tank.
is this movement from back to front related to G force / actual deceleration of the bike ?? or static/constant variable ?
Your example with the ramp made me thing about it. The fact that in reality you will still look in front of you for a few moment is due to G force. -decided to skip explanations-

If whe have -in real time- bike speed, acceleration / gforce, direction of the bike etc... We can implement a logic on the pitch alignment.
Let's find a good formula !!

[TFC]

MaX, good question with the dead zone. I'd also be interested to know. Its a shame you can't load up mxb tracks in gpb to find out..

[HornetMaX]

@MAx,
Yes, it is smart and if we take the actual hypothesis, this should be achievable.

Standard view : pitch (of the view) aligned to the bike
Implyes that we can align a camera axis with bike elevation
Mode=1 view : yaw (of the view) aligned to the bike direction of motion
Implyes that we can align camera with bike direction of motion

If we can align a camera angle with ground/ actual bike elevation / bike actual direction, we can have a smooth and high end cam behaviour on wheelies (sorry, when I say 'we can', you know what I mean)

Not sure I understand what you're asking here. Mode 0 (default) and 1 (new) set both pitch and yaw behaviour

Then, your last sentence is very interesting !!
It made me think that there is something you seamed to have miss. There is a longitudinal movement from front to back / back to front. If you grab the brake, you will move forward like if your balls where smashing on the tank.
is this movement from back to front related to G force / actual deceleration of the bike ?? or static/constant variable ?
Your example with the ramp made me thing about it. The fact that in reality you will still look in front of you for a few moment is due to G force. -decided to skip explanations-

My understanding is that, at least in MXB (don't know in GPB), the rider arms/legs(and maybe torso too) are some spring/damper links to the bike, hence the dynamics of the rider (or its torso) is driven by physics. Don't know if it moves fwd (as in translate fwd), but for sure it "bends" (pitch) forward.

But that "only" controls the camera position (assuming the camera is bolted in the rider head), not the direction the camera points to. or, if it controls the direction also (camera attitude = head attitude) then there's some extra sight adjustment on top. That last part (camera attitude adj) is what has been changed with "mode = 1", not the rest (camera pos). That's my understanding, at least.

[HornetMaX]

MaX, good question with the dead zone. I'd also be interested to know. Its a shame you can't load up mxb tracks in gpb to find out..

On the GPB Mega tracks page there's a "track" named funring that has 4 ramps.
They are extremely basic (just flat planes, with different angles).

The "track" is just a huge (and mostly empty) space so if you want to figure out where the ramps are, be sure to raise thew drawdistance in your profile.ini (from 1000, to 100000m), so that you can see where they are.

[tchemi]

But that "only" controls the camera position (assuming the camera is bolted in the rider head), not the direction the camera points to. or, if it controls the direction also (camera attitude = head attitude) then there's some extra sight adjustment on top. That last part (camera attitude adj) is what has been changed with "mode = 1", not the rest (camera pos). That's my understanding, at least.

Ho, you may be right !! I didn't saw that.

Also I was thinking about one thing. Why are we talking about tuning the view for wheelies and stopies ? Do we need more realism, more immersion or just more feedback.
In my opinion, the actual mode=1 is very realist and a tuning could be just a matter of feedback. It is actually hard to know, Is the angle of the bike that dangerous or can I continue full throttle ?

But in any case, as I remember my wheelies IRL, even if we talk about little ones, big ones, slow ones, surprises ones... I always had my tank in front of me and I was guessing the road in front of me. Also, I'm not a stunter and I never made wheelies for kilometers.

[Vini]

IMO it's already realistic, i don't see a problem with wheelies.

the only thing that is missing would be the "dampened" rider/cam from MXB to simulate the G-forces acting on the rider (particularly, pulling the head/cam down a bit under braking).

..but just to make sure, having the option of controlling yaw and pitch independently would be good.

[HornetMaX]

It was just loud thinking: the first wheelies with the new view looked very weird, but maybe it's just because I (and others) were too accustomed to the old view ...

At any rate, personally I'd make the new view (mode = 1) the default for GPB.

I'm still curious how exactly the "ground" direction is computed (especially in MXB where the track may be very bumpy).
Toying around on Funring (that has some sort of pipe section) I've seen some weirdness with the new view: sometimes the view varies in a weird manner (but that was when trying to "climb" on the pipe wall, a hardly reasonable situation for GPB) and sometimes there was some sort of fast vibration of the view. But maybe it's just due to that specific "track".

[tchemi]

I don't know how the direction is computed but I guess it is something like  "actual pos - last post" and this is computed every T (period). the more little T is the more fluid the camera will be. The greater T is, the more fluid the camera is.

Is it how direction is computed ? What is the size of T and could it be changed ? Is it static, computed ?
Maybe this is a matter of camera actuation.


=>  TL;DR how the bike direction is computed ?

[HornetMaX]

So I did a few more testing.

TL/DR version: I.LOVE.THE.NEW.1ST.PERSON.VIEW !!

Details: so I did another short session, putting some more thinking in it.

First, I disabled auto rider lean (both l/r and f/b). This has allowed me to confirm that the GPB rider behaves like the MXB one: if you brake, the torso pitches forward. Good ! Actually you can see it even when shifting gears up (little fwd/back shake).

Second I was thinking on which track we could really stress test this and .. bam: Ledenon ! It's not a track, it's a rollercoaster (I've been there for a round on french SBK, amazing).

Result: I find the new view lovely. Seriously lovely. Fuckin' lovely in fact. I absolutely love the way it behaves when you have an head-shake: if it's moderate (wobble), you only see the bike shaking, but if it becomes more serious (weave), the view shakes too. Just perfect.

I'm finding the behaviour in case of wheeling more tolerable, so yeah, i't may be just a matter of getting accustomed to this.

I still think there's some filtering to apply (wouldn't know where though, either on the ground reference or on the sight pitch movement): in some low-speed turns in which the bike is seriously leaning there seems to be some sort of high-freq vibration in the sight. There's a chance this may be due to poor track surface as the track is pretty old (don't know, maybe Hawk can check this out), but I'd tend to say that the onboard view should work reasonable even with a poor surface. If somebody can try the same I'd like to hear your feedback.

Overall, I'd put the mode=1 as the default one or, if for whichever reason this can't happen, then it would be nice to have an in-game option to switch between mode 0 and mode 1. With mode 1 by default, of course

Well done PiBoSo !

P.S.
I'm pondering ditching the 3rd person view. And before somebody comes up saying "3rd person is for sissies", let me tell you: the old 1st person was for masochist sissies

=>  TL;DR how the bike direction is computed ?

This comes more or less straight out of the physics, should be smooth enough.

[Vini]

First, I disabled auto rider lean (both l/r and f/b). This has allowed me to confirm that the GPB rider behaves like the MXB one: if you brake, the torso pitches forward. Good ! Actually you can see it even when shifting gears up (little fwd/back shake).

Hmm, I don't see it. The bike just moves but the G-forces are not pulling the rider (whose muscles are apparently simulated as dampers in MXB).

[h106frp]

What 3rd person view? Thought that was just for replays   

[HornetMaX]

First, I disabled auto rider lean (both l/r and f/b). This has allowed me to confirm that the GPB rider behaves like the MXB one: if you brake, the torso pitches forward. Good ! Actually you can see it even when shifting gears up (little fwd/back shake).

Hmm, I don't see it. The bike just moves but the G-forces are not pulling the rider (whose muscles are apparently simulated as dampers in MXB).

it's not the bike that moves, it's the rider that moves relative to the bike.

[Vini]

i just looked very closely at the horizon/ground under braking and it doesn't move so the cam is not affected by g-forces. only the bike moves forward/"down" (dash drops out of the FoV).
the g-forces should pull on the neck of the rider and consequently rotate the cam towards the bike (with some dampening in between), so that the relative movement between cam and bike is not as big.
perhaps it should even be the other way around: the dash should not "drop away from you" but rotate towards you to give better feedback of how hard you're braking.

[HornetMaX]

i just looked very closely at the horizon/ground under braking and it doesn't move so the cam is not affected by g-forces.

I'll take another look tonight but then why when you're accelerating an you shift up a gear you have a small jerk of the view then ?
To me camera pos is affected by acceleration (but notice that camera direction may be unaffected by that as it is "stabilized" to the horizon by default or to the local ground with the new mode).