FOV discussion thread

That's the way I do it, though there is a bunch of theory about FOV calculations that suggest otherwise. As I stated above, I would only use the FOV calculator as the definitive setting if I had a full surround view rig (which I do not and neither do 99%+ of people).

It works better for VR?

 

Well FOV and wheel size on screen can, I suppose, be changed somewhat independently. Keeping the wheel location fixed and changing the FOV will change the shape of the car in the view - something like the 'terminator' shot in cinema (classic example is Robert De Niro and Ray Liotta in the Cafe at the end of Goodfellas - the camera is moved as the zoom is changed in such a way that leaves the subject the same size). So, keeping the steering wheel fixed in size, and then in theory adjust the FOV so that the car is 'correct' (and not elongated or squished) is the best way. To my eyes, changing the seating position in the game doesn't seem to really change how far you appear from objects in the road, only where the car looks to be in your view.

To handle this, I trust the triple screen setup in AMS2 to do that work for me, then move the seating position to fit the wheel. It works pretty well. However I'm not sure how to justify a given choice of FOV (which i read as car size). Ideally I could find a way to choose the car 'size' or 'shape' on the screen that was correct. I'm not sure how to do that - but do you have any ideas/suggestions?

I agree that a standard FOV number of 70 is a very poor way of solving this problem, since obviously it depends on how far my screen is from my eyes. This depends even on how much i might be leaning forward or back in my seat on that day. Obviously in a real car it doesn't matter, but since the screen is supposed to represent the correct perspective projection, moving my head back and forth changes what would be the correct angle of all the lines in the image. When i move my head forward in real life, objects grow bigger AND change relative shape to match, but this is not true of the screen (they only grow bigger since i'm closer to the screen, but the perspective is now incorrectly projected) since the screen doesn't know I'm moving my head (no matter how much I shout at it).

 

@wegreenall
Your explanation sounds interesting.Will try it tomorrow.When i speak about FOV i men the correct arithmetic one (FovCalculater).I meant AMS2 cockpit perspective is correct with the FOV.
Edit: Coudn't resist to try it .
Great - this is what i was missing.Now i know where my missing seconds are...
Sadly the sim can't save FOV per car.
Thanx for the tip and the explanation.

 

The FOV calculator only works if the graphics engine is accurate. Looking at various sims, we can conclude that’s not always the case (and AMS2 is in a class of its own, not in a good way).

In AMS2, by using a shallower FOV, the background is compressed, making objects appear closer together than they really are. The most obvious example is when taking a turn with a car in front of us: a shallower FOV makes the car ahead appear to remain at an almost constant distance, while a wider FOV shows it approaching under braking and pulling away as it accelerates out of the corner.

I took your observations and I’m going to spend the next few days with the FOV set between 68 and 72 (the FOV calculator says 58º in my case!!). It looks promising.

You just described the dolly zoom effect I referred to earlier. And no, you need to change both in the same proportions to keep your object/foreground (your car in this example) constant. Most recent movies do this very well because now then can robotize the camera movement (dolly movement).
Your car/wheel dimension depend on both zoom (FOV) and your position (seat postition).

 

I have used default+ with 50 damping in the past a lot, works just fine and it’s been also the recommend start value from Reiza for DD for a long time (default+ definitely benefitted from higher damping amount, I normally use as little damping as possible)

 

Yes - implicitly I mean that you counteract the change in FOV with a change in the seat position until it's correct. I had not seen the rest of this post so I apologise if I was redundant in my comment regarding the dolly effect. Marius moved the post here from the physics discussion thread, so the context for the post I made was the discussion there.

I was incorrect when I said they can be changed "independently". What I meant was, by holding the wheel size fixed, you now have essentially two 'unknowns' (fov and seat position) and one 'known' - you can make a change to the FOV and presumably always make a corresponding change to the seat position to return the wheel to its correct size. Perhaps the correct FOV is the solution with fixed "zero" seat position, and correct wheel size, so now you just have one variable. However, Marc Collins' point perhaps means this is not the case.

 

This is new to me, so if you change your seat camera position you also change your FoV with that? I thought the two were independent.

 

Field of view and point of view are two completely separate and independent things. You are correct.

 

OK so people are talking at cross purposes here. YES you can change them independently. But you can't change them independently without also messing up the relative size of the wheel, which is a way to make sure your view settings are correctly scaled. Joaquim said "And no, you need to change both in the same proportions to keep your object/foreground (your car in this example) constant."

That's what I meant that they are not independent - they won't be if you want to get the right values, because there is only a subset of the possible combinations of FOV and seat position that properly convert the 2d projection of the 3d space you're in to the right one for where your screen is.

 

Yap, I've tested with a value of 10 and it only affects very fast movements, like noise (as it should). Using 50 is OK, I don't loose anything car behaviour related. Perhaps less FX and less damping is the correct way to go, but I won't spend a lot time on that... (I wonder why )
I now tend to use more damping because it gives a better "rubber like" FFB behavior.