FOV discussion thread

My best bet/hypothesis is that the issue is related to car/world rendering in cockpit view, since the car seems to pivot around the driver’s seat (the ‘center wheel sensation,’ as some call it). But cars aren’t obligated to rotate around us (egocentric simracers) .
External views show no evidence of disconnection, so I conclude that the problem is not related to car physics but to the graphics engine itself.

 

In reply to these videos (that are advertisements of sorts, designed to impress, right??), I am watching one at a time and commenting immediately after each:

AMS 2:

• Ridiculously unrealistic FOV/seat position, as if I am sitting in the middle of the hood/bonnet while driving
• I may be perceptually impaired, but I get zero sense of any floatiness from this video--it looks quite plausible to me

LMU:

• Uses a much closer to realistic FOV/seat position, but still not quite correct
• Car response seems dull and slow, with lots of understeer (not an optimal race setup), which to me could be described as floating or sliding across the surface...though I think it is really just understeer

iRacing:

• Even closer to realistic FOV/seat position
• Handling looks plausible and quite frankly almost identical to AMS 2

The three videos are using completely different FOV and horizon lock settings so are effectively useless for comparison purposes IMO. Of course, we also have zero ability to determine any FFB sensations, good or bad, from a video.

Real life:

• Not a driver-view perspective, so....not sure how this can be used?
• Sawing at the wheel in corners, which almost every real racer does regularly, is something I need to do in AMS 2 more than in any other sim (see also the second video from @Ace's post above for that)

I know you are trying to be helpful and contribute, but these videos provide an example of how videos are almost never definitive, or even useful, to diagnose physics issues.

I would like to explore your statement about the cornering handling in AMS 2 GT3s. That you can rotate the car through a corner and get on the throttle very early. Is this with all aids off and TC=0?

And, we do all realize that the holy grail of race car set-up is to achieve exactly what you have described--a car that can be rotated by the driver/throttle in a controlled manner through a corner, with the earliest possible exit acceleration? Teams all spend as much time and money as they can (or are allowed) to get to this goal. Are you saying the setups in AMS 2 are too good?

 

FOV is 100% depending on your rig setup anyway, not sure how you want to judge if it is a "ridiculous" or "realistic" setting when the monitor size/setup and distance to the monitor is unknown, these settings might be either and it is impossible to tell from looking at a video.

 

I don't drive with my nose pressed against the wind screen and the steering wheel inside my abdomen. It's not surprising that the car appears to swing too wildly in corners when using such a view. It's only a video--many ways to edit or distort it. But if the original view of the person sitting in their rig is the same as in that video, it is not suitable for any physics-related commentary.

 

You have absolutely no clue about the setup the person recording these videos is using. Maybe they drive on a 20" monitor or a 60" TV screen, maybe they sit 40cm away and maybe they sit 90cm away or anything in between. All of it influences the correct settings of the FOV and you simply can not tell from watching the video if the setting is realistic or ridiculous.

 

No, if your chin is on top of the steering wheel rim, or pressed up against the wind screen, your view settings are not correct and the turn pivot point will be out of whack. An unrealistic turn pivot point is the easiest way to distort your perceptions in a sim. We know how far in front of us the steering wheel should be, and it is probably the only consistent element across most vehicles, and how much of the wheel we see through our peripheral (upper and lower) vision (since we have a physical one in front of us)...try setting the view so it matches reality instead of an FOV formula and you might be surprised how well you can control vehicles in a sim. Your brain responds to what it sees on the screen, including memory of sitting in and driving real vehicles. Adjust accordingly....unless you have a super-elaborate surround cockpit that can actually simulate your entire range of peripheral vision and place all the controls close enough to the screen that formal FOV calculations will equal reality. Very rare that any of us have that sort of configuration.

Like lowering centre of gravity, immense amounts of time and money are spent to put the driver as close to the turn pivot point as possible--because that's where natural control of the vehicle by a human is optimized (See McLaren F1 history of the central seat position as an example). Drive a Caterham or old Corvette where you are almost sitting on top of the rear axle and realize how different that feels to a mid-engined single-seater versus a bus where you are sitting in front of the front axle. The video of AMS 2 above we are referring to reminded me of one of those buses. It would be impossible for me to play the game like that--regardless of screen size, screen proportions, triples, curved, flat, etc. It's just wrong to the point that you cannot expect brains to process it as realistic.

 

You are confusing FOV with seat position apparently, because FOV does exactly what you say you want and it has to be adjusted to your setup to deliver that, but according to your logic there is exactly one correct FOV setting for everyone regardless of his setup.
Because FOV does not affect whether "your chin is placed on top of the steering wheel rim"(that is what seat position does), it takes into account what your screen shows with how you are actually sitting in front of it to provide the most natural way to display possible.

 

I used both terms in the original message. You can destroy any sense of realism with either or both settings. That's the only point I am making.

 

But in-game tips talk about the importance of using a FOV calculator...
I do agree @br1x92 about the FOV (there's only one correct setting) and agree with you (@Marc Collins), that specific AMS2 setup is... hum... sub-optimal at best (makes some kind of tunnel view effect). But that's what many people have .

 

Not the best pick, some may not have this car to test.

I've tried it (23.AGO.2025, real weather, 15h00, Default+, 4 laps, default setup).
Pace: 1'34" (perhaps 0.4-0.8s to gain on these conditions, some missed apexes, some overdriving)

Changing the preload to 200Nm:

• car handling: as expect, less prone to rotate (but still can spin it)
• changes nothing "float" related to me

For example, I would expect the car to shake badly through Brittens and the Hislops–Knickerbrook complex, but unless I hit the sausage curb, it just “floats” over those sections. I cannot feel the nuances of adhesion or the car’s composure (mainly roll, which feels too smooth — pitch and yaw are ok, but also on the smooth side).

That’s my definition of “float,” and it’s mainly visual, as FullFFB and buttkickers show there’s actually more going on (physics engine).

Using Default+, I can’t gauge the rear end until it’s too late — usable for hot laps, but not for race consistency.





Lap 2:

Lap 4:

 

Now I am very curious about your ingame cockpit camera settings (head movement etc)? Could you share those? What do you like and why?
EDIT: sorry I already replied a couple of pages back but should have read more. I will try these settings, but I am still curious why you set them at these levels and why new and not legacy settings. THanks!

 

@Marc Collins What is the difference between your FOV and the calculated one? (my guess you are between +8º to +12º).

Looking at @GFoyle posted videos lead me to do this comparision (AMS2 vs LMU vs iRacing FOV).
Here's my calculated FOV (58º):


And here my FOV+10ºC (68º), compare the most distant objects:


And here an irrealistic situation. I moved my seat way forward, but it makes a huge improvement on car rotation feeling:


AMS2 seems to have the camera on the rear of the correct position and then closes the shoot (zooms in). This does compress the background quite a lot compared with other sims.
The reason is technical, and for those interested is such just search "dolly zoom" or "vertigo effect" to see what it does in full glory.
I don't accept the excuse this is a creative choice (Reiza, I'm looking at you... ).

Many moons ago, I did, and reported my conclusions here. The chase view seems to be more connected to FFB.

I used the legacy camera to ground the camera to the car. That did not work on very bumpy tracks (like Sebring). Perhaps I'll try it again and allow for some movement.
The other sliders just make things worst as they make movements too damped - more floaty feeling.

I don’t use Default+… but at the time I set this reference, it felt okay to me.

 

Thx, but why in moving camera forward with a correct FOV irrealistic? I find having two steering wheels in my view (one real one and one in the game) strange and usually move the camera forward to only show the dash or at least get rid of the digital steering wheel also to get it in a more realistic camera position. Very curious to learn from you

 

Not FOV. Moving the seat moves your POV (Point Of View).
The unrealistic part is that you end up too close to the dashboard (not just apparently close).
Here's the full picture from triples. Check the central LCD mirror viewing angle - we are almos side-by-side. Also, the right mirror is gone. Everything look weird as they should in those circunstances!



So the solution would require someone at Reiza to feel the same issue and test different seat position/FOV setups to see if they work well. If so, an option should be added so that currently happy simracers don’t lose what they like, while others can use a different approach (something like a fancy “PERSPECTIVE” slider/switch).
It hurts my soul to feel AMS2 just inches away from greatness.

I vouch for that! AMS2 requires firm but controlled inputs, which doesn’t seem far from reality.
We can brake hard for T1 at Monza, but to make the chicane, a careful lift-off is required before turning in.
Many turns won’t even allow hard braking (although I usually set max braking force to 100%) and that's correct from my point of view.

 

Thx for sharing camera settings. I think I am in minority not liking the horizon/track to bump like crazy. Ideally both horizon and car are not shaking too much, but if I lock to car it becomes too much horizon shaking for me.

 

OK, I am glad you felt the significant difference in rotation from a single adjustment (and not even necessarily the best or first thing to adjust).

Next, I have no idea how you could drive with 50% Damping on Default+. You either have steering that feels like the rack is stuck in a thick viscous goo, or, you have some wheelbase settings off-kilter/not neutral. You can't feel the back of the car or half the other forces if the damping is so over-the-top.

I am using cockpit FOV of 98 on my single ultrawide monitor. I almost always have to move the seat position back 8, 16 or 24 steps (the Reiza seat position is not consistent from one series to the next) to get the virtual steering wheel placed where it should be.

As a contrast, I am using Head Movement of 80, No Legacy, but with Lateral and Vertical G of 10 and Longitudinal G of 25. I noticed early on that the sensation when braking hard was "underplayed" compared to the lateral G so increased it. Again, no idea how someone could use the sim with everything zeroed-out as that would remove so much of the information I need to "feel" what the car is doing. Fascinating that you think it adds to the floaty feeling. Proof that we need all these adjustments to tailor to different brains and there is no right answer.

Thanks for trying the experiment!

 

There are so many variations in hardware and settings, I shouldn't have stated it so definitively, but just be sure you have a very dynamic level of feedback, not sluggish or slow. AMS 2 has the "loudest" FFB of am sim I have used.

 

When I align the in-game steering wheel using the seat position, I assume that I should have it so that my real steering wheel (if it is the same size, which i do match with the wheels I have made) maps over the on-screen wheel.
I saw a Niels Heusinkveld video where he appeared to have done this. It seems to me like the on -screen wheel should essentially be the projection of my own wheel onto the screen in such a way that looking at my wheel or the screen wheel should have the same scale, modulo depth (for perspective). This is given by my wheel perfectly obscuring the screen wheel (again, if they are the same size, so in say the Caterham/90s F1 i have a matching size wheel).

Is this correct?

 

I spend most of my life with a dashboard and wheel in front of me , having the foreground of dash and windscreen frame move at all is the most unnatural feeling for me .
Looking to the distance ahead irl removes most sense of vertical movement.

I find almost locking , but not completely , to the horizon represents this better . putting head movement at zero will overstate the vertical movements too, so that’s why I recommend having 5-10 movement when using legacy as it mimics your own eye and neck dampening the sensations of vertical movement.

I can feel the pressure releasing through the steering rack . to give an example, I turn into turn one at Laguna , turning left , most of the weight is on the front , and is almost even across the steering rack with a little more pressure on the more loaded right .

I make the mistake of lifting off at the same time of downshifting too early and the rear drive axle under rotates ( or possibly momentarily locks the rear wheel, but not always) and the car pitches and sends most of the weight to the outside ( right in this case ) front wheel, and I feel that as the left side load releasing . my wheels resistance to left hand down reduces if you can imagine that .

 

I experimented with this a long time ago.First you need the correct FOV then you can adjust your seat position.But if everything is right in sense of feeling like "sitting in a real car" by projecting your wheel size to the size of the wheel on your screen,i doubt this.I'm much faster and have better car control when i sit nearer to the windshield.So what the sim gives as default FOV is in my opinion correct.But the virtual wheel is much to big.why ? i don't know.
If someone could explain this it would intrest me too.