PDA

View Full Version : Field of view (FoW) Some thoughts to contemplate.



F19_Ob
04-22-2007, 12:45 AM
<span class="ev_code_YELLOW">I posted this far down in another thread but thought I should give it a go here, so the insights are not lost in the noise, and when I now anyway took the time to write it. </span> http://forums.ubi.com/groupee_common/emoticons/icon_rolleyes.gif


I've tried to understand and read about how the eye works and sees things, but also how the brain interprets the information.

The general problem with simulations is that the views are already processed before we see them, thus we can't have the widest field of view ( wich is about 50% too narrow) and the moost inzoomed view (with most correct perspective) at the same time, as we do in RL.

As you may have noticed screenshots taken with widest FoW looks warped and the plane appears to have too long wings.
However while ingame, and rotating the view, wide field of view appears almost natural. This is because of the motion and the fact that we process info about a 3-D object in realtime and even in a simulation the 3-D image is compensated for in the brain (to some extent).
Although the smallest FoW gives best perspective one may instinctively feel this FoW is to narrow.

Ingame this state of fact has some consequenses.
In a hunting- and attacking- situation it doesn't matter so much because the prey is centered in our wiew already.
The unsuspecting prey however has the visibility cut by about 50% wich makes it harder to be vigilant and to look around properly.
This causes us to fool ourselves to some extent because we have expectations of what we should see when we turn our head to the side or to the rear and when when we have performed these tasks we 'think' we have looked around properly, not realising the cut visibility.

Next time you have recorded a track watch it from cockpit and study how you track and look around and you will see what I mean.
One really see less than what one would expect in the real world but usually we are content with Few and small headmovements ingame and belive we really looked well around.

So really one more really wide FOW should be added with more 'wide-angle' effect to try to mimmick the natural FOW with periphery.
This would give a 'boubble-canopy' feeling.
(Actually in the 10 year old CombatflightSimulator'1' this type of FoW was present although a bit rudimentary.)
As expected it was so zoomed out that the instruments was too small to read (no problem really)but it also gave a bit better Fow upward wich ment that no excessive tracking was needed to keep track of an enemy.

Anyway I explained these effects in detail to several simulatorcompanies at the time but the lack of feedback made me loose interest and I keep most ideas to myself these days (silly I know but my stamina is gone).

F19_Ob
04-22-2007, 12:45 AM
<span class="ev_code_YELLOW">I posted this far down in another thread but thought I should give it a go here, so the insights are not lost in the noise, and when I now anyway took the time to write it. </span> http://forums.ubi.com/groupee_common/emoticons/icon_rolleyes.gif


I've tried to understand and read about how the eye works and sees things, but also how the brain interprets the information.

The general problem with simulations is that the views are already processed before we see them, thus we can't have the widest field of view ( wich is about 50% too narrow) and the moost inzoomed view (with most correct perspective) at the same time, as we do in RL.

As you may have noticed screenshots taken with widest FoW looks warped and the plane appears to have too long wings.
However while ingame, and rotating the view, wide field of view appears almost natural. This is because of the motion and the fact that we process info about a 3-D object in realtime and even in a simulation the 3-D image is compensated for in the brain (to some extent).
Although the smallest FoW gives best perspective one may instinctively feel this FoW is to narrow.

Ingame this state of fact has some consequenses.
In a hunting- and attacking- situation it doesn't matter so much because the prey is centered in our wiew already.
The unsuspecting prey however has the visibility cut by about 50% wich makes it harder to be vigilant and to look around properly.
This causes us to fool ourselves to some extent because we have expectations of what we should see when we turn our head to the side or to the rear and when when we have performed these tasks we 'think' we have looked around properly, not realising the cut visibility.

Next time you have recorded a track watch it from cockpit and study how you track and look around and you will see what I mean.
One really see less than what one would expect in the real world but usually we are content with Few and small headmovements ingame and belive we really looked well around.

So really one more really wide FOW should be added with more 'wide-angle' effect to try to mimmick the natural FOW with periphery.
This would give a 'boubble-canopy' feeling.
(Actually in the 10 year old CombatflightSimulator'1' this type of FoW was present although a bit rudimentary.)
As expected it was so zoomed out that the instruments was too small to read (no problem really)but it also gave a bit better Fow upward wich ment that no excessive tracking was needed to keep track of an enemy.

Anyway I explained these effects in detail to several simulatorcompanies at the time but the lack of feedback made me loose interest and I keep most ideas to myself these days (silly I know but my stamina is gone).

Codex1971
04-23-2007, 08:30 AM
I have thought about many times and I think the limitations are high-lighted in your post, size of the object and the warping look of that object.

The only "real" solution for a sim would be the VR head set, something that covers the eye's natural FOV, about 160 degrees on the horizontal and 80 degrees in the vertical.

Manos1
04-24-2007, 04:02 PM
I understand the philosophy

but I saw the light!

Actually, this simulation can cope a lot better with the "wide field of view" than most of us think.

The question is... can your graphic cards cope with the wide field of view? (3072x768)...

Photo with the side monitors aligned in order to better take the picture...
http://www.e-335thgr.com/Athos/TripleHead2Go_01.JPG

http://forums.ubi.com/images/smilies/16x16_smiley-wink.gif

Codex1971
04-24-2007, 04:20 PM
Yeah I saw this in your "bug" thread. That is a more feasable solution, certinally cheaper than a VR headset. I'd say todays 8800 series card should handle the 3 monitors no problem.

Lurch1962
04-24-2007, 09:09 PM
F19_Ob,
You indicated that the original CFS offered an even wider (&gt;90 degrees) FOV than does IL-2. It's been a little while since I fired up CFS, but I seem to recall that its widest FOV was around 90 degrees, too.

But if indeed CFS's widest FOV was greater than 90 degrees, the effect of object size reduction due to "de-magnification" was mitigated by a kind of cheat. In CFS, all objects (planes, trees, buildings, boats, etc.) as seen from inside the cockpit were rendered *3 times larger* than their proper size. (As seen from an outside-the-plane viewpoint, they were magically restored to their true size!!!) So just think... all that time you were chasing fighters around the sky which were rendered as being the size of 4-engine bombers! http://forums.ubi.com/images/smilies/16x16_smiley-very-happy.gif

To get back to IL-2...

AS you correctly point out, the widest FOV settings result in stretching distortion. As I've revealed in previous posts, this is the result of the so-called Gnomonic projection used by all flight simulators. If you think this distortion is prominent now, it would be very much worse with any significant increase in the view angle.

Personally, I'd be loath to use any FOV wider that about 100 degrees just because of the marked variation in scale across the screen.

In the game, the greatest impediment to good scanning is not as much the mechanics of controlling the view direction as it is the fact that the virtual pilot can't physically move his head about the cockpit so as to see around the framing/look over the side/peer behind the headrest. In other words, our virtual heads are forced to do a Linda Blair-like swiveling on a set-in-concrete torso. Not to mention the usual blind spots such as the under the fuselage/wings and often times dead astern, which necessitates weaving if the full 4*pi steradians of the external environment is to be searched.

I've made the task of scanning less onerous than it otherwise would be by making the mini-joystick on my CH throttle act as a mouse for panning the view. And I've given it an aggressive responsiveness so that I can whip the view at a fairly brisk rate.

Naturally, the ultimate cure-all (given the aforementioned limitations) would be the acquisition of a TrackIR unit.

--Lurch--

F19_Ob
04-25-2007, 03:44 AM
Thanks for your input guys.

Well Lurch, I'm not 100% sure exactly what differed in the CFS1 FoW (been a long time for me to) or if there was some other component to it that made the view more wide angled in the 3-D cockpit. (I however compared the view's with CFS2 and other sims at the time. One was Sturmovik first edition.)
However The instruments became too small to read but there was more air upwards and to the sides.
This ment one could track an enemy in a turn with less need to use the tophat. In any case only little tracking was needed to follow the enemy in dogfights .
The penalty was a boubble canopy feeling, wich one get used to though (like present wide-view).

Note this viewtype would just be an addition to the other FoWs we already have and would not ruin anything for anyone since it would be optional like the other FoW's.
The gain would be a more realistic perifery view.
(I don't know of anyother way doing it on one screen.)

I think I understand the resoning though, that the views we have now is same for all.
The major flaw is that it's a bit shortsighted idea and infact result the present bad view-balance for attacker and attacked as I described earlier.
A high price to pay indeed just because one did not add one more 'optional' and wider FoW.
Wouldn't you say?

It's the tactical technical mistakes I want to illuminate here. Although I understand that lack of time and pressing production is a factor I still think that issues like this can (should) be fixed in following patches.

Please note that this is my opinion about these misses in the sim and there may be explanations to why it is like it is in the sim now, but I've heard no other logic explanation to why during all these years.

To continue with what I think was mistakes, one big was implementing different view-variables and positions in similar planes, wich result in advantages in some and disadvantages in others (for no reason).
It would be logical if for example the Fw190 got the "shift+F1"-view of the similar canopy me262 to avoid some the view-problems that has been complained about since the fw190 from day 1 of the sim.

Some planes have the head placed so close to the headrest that they have no rear-view to speak of, although there would be room to move the head a bit forward and thus aquire the sought after rear-view.


Trying to be as construcive as I can but my small English vocabulary limits me a bit.

regards

Lurch1962
04-25-2007, 04:40 PM
F19_Ob,
Don't apologise for your use of the english language... it's better than that of some for whom english is their *only* tongue! http://forums.ubi.com/images/smilies/25.gif

I still have CFS on my 'puter... Next time I'm gaming I'll fire it up and measure its actual widest FOV.

Indeed, to have as an option for those who wish to use it, an even wider FOV should pose no obstacle. However, there might be a fly in the ointment, so-to-speak. Hop in a Bf-109, switch to the widest FOV, and pan your view either left or right (it's best to use the pan view mode, and not the "default" look left right/left views in order to better see what I'm pointing out here). You'll see "see-through" portions of the wing's top surface texture near the wing root. This might be a hint to underlying limitations to even wider FOVs.

Your comments regarding the pilot head position are noted and have been addressed in your more recent post (with which I'm in full agreement).

--Lurch--

F19_Ob
04-27-2007, 03:44 AM
Thank you m8. http://forums.ubi.com/images/smilies/blush.gif

I also thought of loading it again for fun but I must have given it away when I gave away a bunch of old games and other stuff to a church basar some time ago.
If u decide to load it could u please check the p51, wich I think had the best unobstructed 3D view.
Remember to zoom out fully. http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif


Concerning the 109's texturefault on the wingroot I think it is not nessesarily because of the wideview, since the majority of planes are unaffected.
On the 109 a wider view may or may not make the texturefault more visible but on the contrary it seems likely that a slight tuning of the current View positions would fix it or improve a bit. Sometimes a few degrees is enough. One have to test.

In Ms flightsims one can freely test custom headpositions, so anyone can examine and understand the nature of the problem with textures.
I though think the overall gain and solution of the old view issues may be worth a try with a wider view.
<span class="ev_code_YELLOW">If it doesn't work it's just easy not using it or disabling the key for it.</span>In that respect the possible gain is greater than the loss.

I think that we who like different theatres are stuck with this sim a long time yet, before all planes and theatres are replicated in BoB (if at all).
BoB will be a good complement when flying BoB theatre though. http://forums.ubi.com/images/smilies/halo.gif