Setting the FOV - Field of View

[tone12 0]

after reading the pdf on setting the FOV i was left wondering, isnt that what the ingame zoom keys are for?

 

or is there a bit more to it?

 

 

[Olham 164]

There is much more to it.

With zooming in or out, you just get everything closer, or more remote - while your FOV may still be set up wrong.

With setting the FOV, you are in control of setting the correct width-to-height ratio for your monitor resolution.

There are so many different resolutions available now, that this is an important bit.

 

If you take the default values of 23:16 for example: they may right for a 4:3 monitor, but absolutely wrong for a 16:9 or 16:10

And even for a 4:3 monitor, they might be wrong, as this example shows.

 

My old monitor had a res of 1280 x 1024.

If I divide the vertical resolution by 16, I get:

1024 : 16 = 64

If I now divide the horizontal resolution by this factor, I get:

1280 : 64 = 20

 

So the ideal FOV setting for this monitor was 20:16 - and not 23:16.

[+Lothar of the Hill People 6]

As discussed off-topic in this thread, Olham's simple calculations, while maybe a useful first approximation, aren't very accurate. Nor are they useful for people in different display situations. So I've created a simple Excel-based FOV Calculator for OFF. Just enter in your resolution and it suggests a range of the most spatially accurate FOV settings.

 

Here are a few common aspect ratios for those of you without Excel. Just divide your horizontal by vertical resolution to calculate your aspect ratio to find which example applies to you. The first example will provide a good contrast to Olham's method:

For a true 4:3 screen the perfect FOV would be: 20 x 15 (= 4:3)

 

4:3 aspect ratio (r = 4/3 = 1.333), such as 800x600, 1024x768, 1280x960, and 1600x1200:

 

 

As illustrated in the program output, 20:15 FOV renders a picture close to a 4:3 aspect ratio, but not perfectly. The narrow angle of this FOV setting (width 1) suggests it's good at some distance from a small CRT monitor, but 25:19 (width 4) is nearly as accurate in terms of aspect ratio and much more appropriate for sitting close to a large bank of monitors or on the floor in front of an old low-def big-screen television.

 

Here are some more common aspect ratios:

 

16:9 aspect ratio (r = 1.778), such as 1280×720, 1366×768, 1600×900, and 1920x1080:

 

The 35:20 suggestion isn't possible in OFF, 33:19 being the widest FOV that's both feasible and relatively accurate.

 

16:10 aspect ratio (r = 1.6), such as 1440x900, 1680×1050, and 1920x1200:

 

 

5:4 aspect ratio (r =1.25), such as 1280x1024:

 

 

And an uncommon one for Olham:

66:35 aspect ratio (r =1.886), such as 1980x1050:

 

Poor fellow, only two of his suggestions are feasible.

 

But I have no idea what weird resolution for which OFF's default of 23:16 would be appropriate... the rendered aspect ratio is r = 1.448.

Edited December 4, 2012 by Lothar of the Hill People

[Herr Prop-Wasche 7]

Thanks, Lothar. Glad to see that you are back and tweaking!

 

I've just downloaded your FoV calculator now. Will give it a look tonight.

[+Lothar of the Hill People 6]

Just updated the calculator to version 1.3, which compared to the original expands the search space and eliminates accurate FOV combinations not possible in OFF.

 

Having just 4 combinations of accurate settings to test beats going through all 88 by trial and error! But it's great we have the choice with OFF. Most graphics engines fix one or even both of the horizontal and vertical field of view, which is one reason cheap conversions of console games to PC look so weird.

 

I can no longer edit the above post, but here are the updated results for a few common and uncommon aspect ratios.

 

4:3 aspect ratio (r = 4/3 = 1.333), such as 800x600, 1024x768, 1280x960, and 1600x1200:

 

 

16:9 aspect ratio (r = 1.778), such as 1280×720, 1366×768, 1600×900, and 1920x1080:

 

 

16:10 aspect ratio (r = 1.6), such as 1440x900, 1680×1050, and 1920x1200:

 

Notice the 0.001 error on the ultra-wide 30:19 FOV. If you're going to go massive multi-monitor or huge projection screen that fills your peripheral vision, you definitely want to go for an overall 16:10 aspect ratio for almost no spatial distortion in OFF.

 

5:4 aspect ratio (r =1.25), such as 1280x1024:

 

 

66:35 aspect ratio (r =1.886), such as 1980x1050:

 

A couple more options for Olham!

Edited December 5, 2012 by Lothar of the Hill People

[Olham 164]

Sorry, Lothar, but wether I use your numbers or calculate them in my way - the result remains the same for my resolution.

(I had made a typo for my own monitor - my res is 1680 x 1050 - sorry for that)

 

Here is my simple formula for verticals 19, 17, 15 and 14:

 

1050 : 19 = 55,263

1680 : 55,263 = 30,4

Rounded result: 30 : 19

 

1050 : 17 = 61,765

1680 : 61,765 = 27,19

Rounded result: 27 : 17

 

1050 : 15 = 70

1680 : 70 = 24

Result: 24 : 15 (clean - not rounded)

 

1050 : 14 = 75

1680 : 75 = 22,4

Rounded result: 22 : 14

 

But it's great anyway, that you have made a calculator for all those who are unsure about maths.

(...and have EXCEL)

Edited December 5, 2012 by Olham

[+Lothar of the Hill People 6]

Hey Olham, the simple linear scaling gives an okay approximation, but the relationship between FOV and aspect ratio is nonlinear. So your method can't tell which of the combinations are better than the others, and if anything the intuition can be misleading.

 

For example, while it looks like 24:15 is the most accurate ("clean") because you don't have to round, really 30:19 is more accurate spatially despite the 30.4 you got. And what about VFOVs of 13, 16, 18, and 20? Doing 3D spatial correction shows that you're not going to get as good of results using any of these compared to the set I suggest. 26:16, which you recommended to Creaghorn and others, is downright awful compared to 27:17 FOV at a 16:10 aspect ratio. I'm curious which one you're actually using--what seemed to feel best, math or not.

 

Think of it like gamma correction, which is done to photographed and rendered images because of the nonlinearity in how we perceive luminosity. Non-gamma-corrected images look "alright", in that darks are darker than lights, but the degree of darkness and lightness is off. OFF would look a lot uglier if you disabled gamma correction in your graphics driver.

Edited December 5, 2012 by Lothar of the Hill People

[Olham 164]

Well, maybe it's beyond my understanding - after all I came to the same numbers as you, didn't I?

 

For my monitor (1680 x 1050) I use a FoV of 24:15.

[+Lothar of the Hill People 6]

I looked around a bit and not much of the stuff Google returned about field of view and gaming is correct or useful, so it's no wonder OFFers ended up on a slightly wrong track. But let's see if I can make sense of things a little better. First we need to understand the difference between aspect ratio and field of view.

 

Olham's calculations are perfectly fine for a 2D world. If we were trying to play Pac-Man on our modern high-resolution wide-screen monitors, linear scaling is exactly the approach to make sure the title character is rendered as a perfect circle and not stretched into another ellipse. In this flat universe, aspect ratio and field of view are the same thing, just in different units. Talking about the square footage of your house versus measuring it in square meters doesn't change the shape of your kitchen, and rendering scale models such as floor plans is pretty intuitive.

 

This 2D math, with two parallel planes, is a rough enough estimate for making sure a round cockpit gage that's flat and directly in front of you looks round. But all the other gages in your peripheral vision may look like coins run over by trains and seem ten feet away. What's missing?

 

For 3D flight sims like OFF, a more realistic presentation of the game world is critical for immersion. The CFS3 engine is basically maps a three-dimensional model onto a 2D plane. Imagine sitting front and centered before your monitor. Notice that the pixels in the middle are closer to your eyes than the pixels far in the corners. As we move out from the center of our vision, the number of pixels required to cover each degree of vision decreases. I modified the field of view example from the Wikipedia article to illustrate this:

With the camera at the center and looking right, the right side of the circle is divided into four 45° angles, so the 90° field of view (in green) is divided into upper and lower halves. You can see that the first 45° from the center of the view covers about 64 pixels, but the next 45° to the top of the circle is only covers 26 vertical vertical pixels (but more horizontal pixels). Angle degrees don't trade for pixels one for one--the relationship changes nonlinearly. This is why aspect ratio (pixels or inches) and field of view (angles) are not the same thing in 3D, and while Olham's linear scaling doesn't work.

 

But what does this mean? Objects in the game world don't just have width and height, but also depth. I've said this before: despite what a lot of stuff on the web says, Field of View has nothing to do with zoom. Zooming in and out, such as you can do in game, changes the size of things but not the depth--physically it's no different from zooming in and out of a screenshot from the game, doesn't change the perspective of how things are viewed.

 

Instead, Field of View has everything to do with perspective. A simple example might help, but instead of screens and lenses think mirrors--specifically the passenger side mirrors of automobiles. In the US and Canada at least they all have warning stickers saying "Objects in mirror are closer than they appear", emphasis added on closer. It's not a zoom effect--it's not that objects are larger than they appear. The convex shape of these mirrors changes the field of view, distorting perspective rather than scale. Zooming in on the mirror, or getting closer to it, does not "fix" the shape of the reflected image.

 

With our two eyes we mostly think about perspective in terms of binocular vision, but that's only useful for stuff really close. Much of how we judge distance is due to angles, processing that happens in our visual system of which our conscious minds aren't even aware. When the FOV is so wrong that your plane in OFF appears taller than it is wide, that's enough to consciously know something's amiss. But most FOV problems are far more subtle, and far more insidious in how they break immersion--distorting our sense of distance and depth. No wonder forced perspective is one of oldest and most effective special effects tricks in film and photography.

 

The true test of FOV settings is your perception of distance. Is the spot that looks halfway down your wing, actually halfway down your wing? Do your wings get too narrow farther away from you at the tip, or not narrow with distance enough? Does the wingspan of your E.III actually feel like nine and a half meters? Does that enemy craft you're firing at from thirty feet actually appear to be thirty feet away, and not 25 or 40?

 

With the wrong FOV settings, even if stuff in a flat plane doesn't appear distorted (cockpit gages are round), the sense of perspective can be so off as to make you feel out of scale. OFF's aircraft may feel more like toys than some of the most advanced killing machines of the day. It's a shame so many flyers have had the wrong settings for so long.

 

I just remembered a scene in this bad comedy Ski Patrol, where a character is pranked to wake up in a miniature version of his house. In the movie it's suppose to make a short character feel like a giant, but in reality his vision system would recognize that walls and doors are closer and not just relatively smaller, and that the ground is the same distance from head as always. But project that 3D world onto a flat screen, and we're much easier to confuse. This is why FOV is so important, and why it's so tragic that pretty much every gamer is looking at all these virtual worlds from the wrong angle.

 

Anyway, my calculator's only a starting point--just gets all the really bad settings out of the way. Which of the recommendations yields the proper sense of perspective depends on more than just your monitor's aspect ratio. 24:15 is a good place for you to start, Olham, but depending on how big your monitor is and how close you are too it you may want to go wider or narrower on the chart.

Edited December 5, 2012 by Lothar of the Hill People

[Olham 164]

Okay, I think I have halfways understood this in theory.

Basically, it seems to mean that some seemingly 'correct' width:height ratio settings would still distort

the perspective view more than other 'correct' width:height ratio setups?

 

Now I'd still like to see the 3 best (least distorting) choices for my 1680:1050 screen

(as I said, I don't have EXCEL and can't use your calculator).

[+Lothar of the Hill People 6]

Okay, I think I have halfways understood this in theory.

Basically, it seems to mean that some seemingly 'correct' width:height ratio settings would still distort

the perspective view more than other 'correct' width:height ratio setups?

Yep!

Now I'd still like to see the 3 best (least distorting) choices for my 1680:1050 screen

(as I said, I don't have EXCEL and can't use your calculator).

Exactly why I posted the tables in the post above... see the 16:10 aspect ratio. But this still doesn't account for your viewing distance and monitor size, which is why you want to move through and test the different widths to find the one that works best for your setup. Think I can make that easier as well; just replied to your PM.

[+Lothar of the Hill People 6]

Bad news for my FOV calculator. The math is right, but the Workshop settings are wrong. The values aren't actually horizontal and vertical FOV as they're labeled in the Workshops. If you look at the XDP files:

<Station Name="Pilot" Type="pilot_station" View="0" FovUp="23" FovDown="16">

Up and Down, not Horizontal and Vertical. If you experiment, you'll see that increasing the so-called "horizontal" doesn't widen the angle of view directly, but adds relatively more above you. In the default case as in this example, 23+16=39 degree total vertical FOV, which about fits a 59.1 degree horizontal FOV on a 16:10 screen. So that makes a lot more sense, even though it's not at all how the Workshops indicate things work.

 

Going to have to rethink how it's handing the projection (really, ironing out how it handles the "zoom" levels). It doesn't actually render the 59.1:39 FOV directly, but renders a wide-angle for which CFS3 was oft-criticized then "zooms in" to approximate the right FOV at the right scale.

 

Some back-of-the-post-it-note suggestions for optimal viewing:

• 16:9 aspect ratio: 20:16 Workshop FOV, sit 75% away from screen's diagonal
  
• 16:10 aspect ratio: 20:20 Workshop FOV, sit 73% away from screen's diagonal
  
• 4:3 aspect ratio: 30:20 Workshop FOV, sit 67% away from screen's diagonal
  

For example, if you run OFF at 1920x1080 (1080p) resolution, take 1920/1080 to find your aspect ratio. That's 1.77... which is the same as 16/9, so set your FOV to 20:16 in the OFF Workshops. If your monitor has a 20 inch screen, your eyes should be 15 inches away. It turns out we knew what we were doing as kids when our folks yelled at us for sitting too close to the TV--getting the perspective right. And OFF benefits from screen real estate like few other games!

 

Stick with the default "zoom" level for the normal scale of human vision and correct perspective. You're probably used to zooming out more, but if you're sitting close enough to the screen you should be in the best view for firing up close and spotting enemies (with TrackIR--set your x,y,z movements to 1cm:1cm and likely turn down pitch and yaw). At the recommended distance things should appear as close as the actually are in the game world. The bottom of your monitor should feel like the lip of your cockpit--which means you can't reach your machine guns from six feet away on the couch!

[Herr Prop-Wasche 7]

Thanks for the info, Lothar. I thought I noticed, using your settings, of a bit of a "fisheye" effect, particularly when pitching up or down. I have a 16:9 monitor at 1920 X 1080 and had been using a 28:16 FOV setting in workshop. I'll try the 20:16 and let you know what I think.

[HumanDrone 4]

For those who don't have Excel, Open Office should be able to open Lothar's sheet. It's an older format Excel file, so the spreadsheet app in Open Office ought to handle it easily.

[+Lothar of the Hill People 6]

For those who don't have Excel, Open Office should be able to open Lothar's sheet. It's an older format Excel file, so the spreadsheet app in Open Office ought to handle it easily.

Good advice, HumanDrone, but don't yet bother as the sheet's all wrong since things are mislabeled in the Workshops.

[HumanDrone 4]

That's all right. I'm using a 23" monitor, and hadn't fancied getting that close anyhow. And that concerned me about your settings anyway, as I would think that, as you say, if you are trying to simulate a 60° wide FOV (for instance) that you should be in such a position that the edges of your screen subtend an angle of 60° with your eye position at the angle's vertex. That's a lot closer than I can comfortable sit. Guess it's time for a new 40" monitor!

 

(the Human Drone wanders off, whistling "It's beginning to look a lot like Christmas"...)

[Herr Prop-Wasche 7]

So far, I'm getting a lot less distortion with 20:16 than I did with 28:16. I still need to experiment around a little more.

[+Lothar of the Hill People 6]

That's all right. I'm using a 23" monitor, and hadn't fancied getting that close anyhow. And that concerned me about your settings anyway, as I would think that, as you say, if you are trying to simulate a 60° wide FOV (for instance) that you should be in such a position that the edges of your screen subtend an angle of 60° with your eye position at the angle's vertex. That's a lot closer than I can comfortable sit. Guess it's time for a new 40" monitor!

That's exactly how it's suppose to be, HumanDrone. Though with multi-monitors you can focus on getting the vertical FOV right for distance and let the horizontal take care of itself with the added peripheral vision.

 

Does take a lot of screen real estate to get the most out of a simulator, but there's more to it. The correct field of view and distance are critical so objects in the gameworld are as close as they appear: enemy planes are actually where you think they are when you're shooting at them, you can judge the distance to that clearing as you're gliding down without fuel, etc. Mostly it's a subconscious effect of whether or not you feel like you're there in the cockpit or flying a radio-controlled plane. The goal is really to put your eyes and the 3D engine's virtual camera in about the same place.

 

Here are a couple of examples I found to illustrate the difference between a proper visual simulation, in this case the Honda driving simulator:

 

and a cheesy arcade game:

Notice here how the real and virtual steering wheel and front tires are seemingly meters apart. Perspective is bizzaro, and it's impossible to intuitively judge distances and angles to make turns on the racetrack. Even if only the sitting position is wrong, it hiccups your brain just a little to have to mentally switch perspectives--a subconscious effect that can destroy immersion. All together, the most garish kit has multiple ways of inducing nausea.

 

So it's not just about more, bigger screens, but how you use them. Getting the math right is key (and labeling the FOV settings incorrectly in the Workshops doesn't help!). My FOV calculator is moving to join the OFFice incomplete campaign editor suite in its next release, but oh it will do so much more than just calculate... stay tuned.

[HumanDrone 4]

Yessir, I see what you mean (and of course have experienced it...). And you seem perfectly capable of handling the math, so I'll leave it to you. What I take from what you said above is that the workshop settings of "up" and "down" would seem to set degrees above horizontal (for "up") and degrees below horizontal (for "down"; then the program uses the total angle to calculate the horizontal FOV angle from the total vertical angle (or something like that). If that's the case, it must be a holdover from pre-TrackIR days!

 

Since all that zoom should do is lower the FOV angles (proportionately, one would hope), it seems to me that if you get the no-zoom angles, distance to monitor, etc, correct then it's all good from there. But it seems to me that "distance to monitor" is a key part part of this.

 

Won't our wives/girlfirends (or in a few cases, husbands/boyfirends) love this! Us sitting there with a tape measure from our schnozzola to the screen! Anyway, good luck! If by any chance you need a hand, holler.

[Herr Prop-Wasche 7]

This topic is interesting not only in regard to OFF, but also for viewing anything on a computer monitor.

 

A few months ago, I upgraded from an old CRT monitor to a new LCD 22" Dell widescreen monitor. Although it is a 16:9 instead of a 16:10 monitor, its native resolution is 1980 X 1050. 1920 X 1080. At that resolution, text was too small for my old eyes to read on the computer unless I utilized the zoom feature. However, even with zoom enabled, I was experiencing a noticeable amount of eye-strain. Even worse, I was experiencing bouts and waves of dizziness and nausea that lasted for hours after using the computer for any lengthy period of time. It was only after I had reduced the resolution to 1650 X 1080 1680 X 1050. that these problems went away, so screen size and resolution DO seem to interact in some way to influence what feels right to your eyes and brain. Something to think about, anyway.

 

--Edited to correct resolution numbers.--

Edited December 13, 2012 by Herr Prop-Wasche

[+Lothar of the Hill People 6]

A few months ago, I upgraded from an old CRT monitor to a new LCD 22" Dell widescreen monitor. Although it is a 16:9 instead of a 16:10 monitor, its native resolution is 1980 X 1050. At that resolution, text was too small for my old eyes to read on the computer unless I utilized the zoom feature. However, even with zoom enabled, I was experiencing a noticeable amount of eye-strain. Even worse, I was experiencing bouts and waves of dizziness and nausea that lasted for hours after using the computer for any lengthy period of time. It was only after I had reduced the resolution to 1650 X 1080 that these problems went away, so screen size and resolution DO seem to interact in some way to influence what feels right to your eyes and brain. Something to think about, anyway.

You sure about those resolutions, HPW? Neither 1980x1050 nor 1650x1080 are 'proper' resolutions. 1920x1080 is 16:9, and 1680×1050 is 16:10. What's the model number of you Dell monitor?

[+Lothar of the Hill People 6]

What I take from what you said above is that the workshop settings of "up" and "down" would seem to set degrees above horizontal (for "up") and degrees below horizontal (for "down"; then the program uses the total angle to calculate the horizontal FOV angle from the total vertical angle (or something like that). If that's the case, it must be a holdover from pre-TrackIR days!

Yep that's what it looks like-good ol' CFS3, released in 2002. But it lets me replace the OFF Workshops settings with a simple slider that moves through the full vertical FOV allowable by OFF while preserve the correct aspect ratio for your monitor.

 

Since all that zoom should do is lower the FOV angles (proportionately, one would hope), it seems to me that if you get the no-zoom angles, distance to monitor, etc, correct then it's all good from there. But it seems to me that "distance to monitor" is a key part part of this.

Yep again. It's possible to calculate where in the room CFS3 thinks its virtual camera is, so you know where to put your eyeballs.

 

Won't our wives/girlfirends (or in a few cases, husbands/boyfirends) love this! Us sitting there with a tape measure from our schnozzola to the screen! Anyway, good luck! If by any chance you need a hand, holler.

Takes a special someone to put up with Barmy OFFers!

[Olham 164]

I guess, HPW has a 1680 x 1050 (16:10) monitor, like I have. It's easy to mix up the "6" with the "9", as I also had done.

 

HPW, I don't think you should lower the resolution of your screen, but rather zoom in another notch, or a third one?

 

This FoV in OFF is still a mystery to me - I experimented a bit after Lothar's discovery,

that both values are part of the vertical angle - but I couldn't say I found something out???

Edited December 13, 2012 by Olham

[Herr Prop-Wasche 7]

You sure about those resolutions, HPW? Neither 1980x1050 nor 1650x1080 are 'proper' resolutions. 1920x1080 is 16:9, and 1680×1050 is 16:10. What's the model number of you Dell monitor?

 

I guess, HPW has a 1680 x 1050 (16:10) monitor, like I have. It's easy to mix up the "6" with the "9", as I also had done.

 

Sorry about the resolutions being wrong. I wasn't at my home computer and didn't get the numbers right.

 

In fact, I do have a Dell Ultravision U2312HM 16:9 aspect ratio monitor. The native resolution is 1920 X 1080 but I had to back it down to 1680 X 1050 (confirmed this is the actual resolution on my 16:9 monitor). The next lowest resolution is 1600 X 1024.

 

However, I run OFF at the 1920 X 1080 resolution.

Edited December 13, 2012 by Herr Prop-Wasche

[+Lothar of the Hill People 6]

In fact, I do have a Dell Ultravision U2312HM 16:9 aspect ratio monitor. The native resolution is 1920 X 1080 but I had to back it down to 1680 X 1050 (confirmed this is the actual resolution on my 16:9 monitor). The next lowest resolution is 1600 X 1024.

 

However, I run OFF at the 1920 X 1080 resolution.

 

1680x1050 is 16:10 and doesn't work well at all on your 16:9 Dell. You're either getting a distorted (squished) or truncated (missing top&bottom) view. With LCD screens you pretty much always want to run them at native resolution.

 

If stuff on your desktop is too small, DON'T lower your resolution. Windows Vista/7/8 which were designed in the LCD era instead have options to scale up the render itself . Check Control Panel\Appearance and Personalization\Display, though you'll have to logoff and log back in for the changes to take effect. This should give a much better result than using non-native resolution in the wrong aspect ratio!