Is there a "best" FOV for Matching Via Viewspeed from 2D?

[Excalibur]

@DPI Wizard also my method with counting pixels gives me slightly different results when I change default warframe fov from 65 to 75 and I will keep default csgo fov 90. Idk if somehow my method is wrong when I change default FOV in game or mouse sensitivity calculator is not accurate. Could you please take a look on this when you have time? 

[potato psoas]

On 29/04/2017 at 2:14 AM, KandiVan said:

Yea i think it depends a lot on the game. H1 requires a lot of headshots, so I think a little bit lower than 106.26 hfov is ideal.  Arena shooters like quake with large hitboxes can have much higher fovs (upwards of 120-130 hfov). I think for most games, hovering within 10 hfov of 106.26 is the "ideal" range.

If you picked a suitable sensitivity you wouldn't need to worry about using a lower FOV for games that require better aim. I can shoot targets one pixel small with my sensitivity. Honestly, you can do it with the fastest sensitivity. But what is suitable depends on what you can comfortably use according to the best grip style for you.

Edited May 8, 2017 by potato psoas

[CaptaPraelium]

On 04/05/2017 at 8:36 AM, Drimzi said:

It is 90 FOV 4:3 base.

Since monitors are mostly 16:9 now, the hor+ scaling that games use keep the vertical FOV fixed (73.74) and then expand the horizontal FOV to fill the aspect ratio, which ends up being 106.26 horizontal FOV on 16:9 monitors. This was the FOV used by the Quake, Unreal Tournament, and Counter-Strike back in the day and is still being used by Source games.

Generally increasing the FOV further is just causing too much distortion and makes targets in the center of your screen too small, causing a disadvantage. If a game has a zoom / aim down sight function, you can increase the FOV without any downsides but it will still depend on your tolerance for pincushion distortion.

90 FOV 4:3 base may not be the most optimal but it is pretty close. I think it would be somewhere between 90 (106.26) and 100 (115.63). The most optimal way to increase FOV further would not be to increase your 4:3 base FOV to 120+ but to increase your monitor's width and aspect ratio and keep the 4:3 base FOV fairly close to 90-100. For this to happen though, you are going to have to wait until wider monitors are made. 21:9 aspect ratio is the next step, will be years before there is anything bigger.

 

Thanks mate I understood all that already, I was just wondering why you chose that FOV in particular. It seems it's just for historical reasons (as in, that's how it was done before). I was hoping that it was actually optimal but it turns out that it's not - although as you say, it is close.

For rectilinear projection,
x=tan λ
y=tan λ /cos Φ

So optimal FOV will be 90 (45 degrees either side of zero) and anything beyond this becomes hyperbolically more lossy.

This is fine if we have a circular monitor but we don't.... And obviously there is a balance between taking a wider FOV which may contain lossy parts, and a more narrow FOV which avoids lossy projection but at the cost of losing (by not displaying) optimal projection around the edges of the monitor.  I'm guessing that on this forum, there is a math nerd who knows how to find the volume of two intersecting objects, and if there is, they would be able to make a formula to find the actual optimal FOV given a particular aspect ratio.

Edit while discussing this with a friend, he has pointed out that my source for the rectilinear projection formula (this PDF: http://www.tawbaware.com/ptassembler_mod_rect_projections.pdf) may be wrong, as he has found this more reliable source: http://mathworld.wolfram.com/GnomonicProjection.html

Edited July 3, 2017 by CaptaPraelium

[CaptaPraelium]

Just as a follow-up to the above, the formula from tawbaware is correct. The one from wolfram is specifically for projecting the sphere onto the plane, as in, a map of the earth. It doesn't apply here, because we aren't mapping a sphere (and that's why future diagrams from me regarding viewspeed have NO circles)
@potato psoas Thought you may find this interesting.

[potato psoas]

50 minutes ago, CaptaPraelium said:

Just as a follow-up to the above, the formula from tawbaware is correct. The one from wolfram is specifically for projecting the sphere onto the plane, as in, a map of the earth. It doesn't apply here, because we aren't mapping a sphere (and that's why future diagrams from me regarding viewspeed have NO circles)
@potato psoas Thought you may find this interesting.

Yeah all the stuff you've shown me has definitely changed what I think about everything. But I don't understand why you don't like circles... it's just representing 360 distance.

[CaptaPraelium]

I'm avoiding circles because they are misleading. It seems like a good way to represent what we're doing and seeing, but actually represents something rather different... So we are kinda thinking about it wrong because we are visualising something that never happened.
Sorry I'm taking such a long time to get back to you with the drawings bro. IRL has a funny way of changing plans :\ I expect that the xmas period should offer me some time off to do fun stuff like this Hope you're having a good break.

[potato psoas]

On 12/23/2017 at 06:32, CaptaPraelium said:

I'm avoiding circles because they are misleading. It seems like a good way to represent what we're doing and seeing, but actually represents something rather different... So we are kinda thinking about it wrong because we are visualising something that never happened.
Sorry I'm taking such a long time to get back to you with the drawings bro. IRL has a funny way of changing plans :\ I expect that the xmas period should offer me some time off to do fun stuff like this Hope you're having a good break.

Yah, merry christmas. Will await your explanation.