Jump to content
Sign in to follow this  
Bies

Visual Search in Air Combat

Recommended Posts

From:

Center of Strategic and Budgetary Assessments

 

TRENDS IN AIR-TO-AIR COMBAT
IMPLICATIONS FOR
FUTURE AIR SUPERIORITY

 

https://csbaonline.org/uploads/documents/Air-to-Air-Report-.pdf

 

Quote

For about fifty years, pilots relied on the human eye as the primary air-to-air sensor and
machine guns and automatic cannon as their primary weapons. The physical limitations of
human vision give it a relatively short effective range as an air-to-air sensor of about 2 nautical
miles (nm). Aircraft can be seen farther away if the highly sensitive central vision is focused
on them, but with central vision limited to a cone roughly 2 degrees wide, pilots searching for
opposing aircraft without some sort of cue to limit their search are unlikely to detect them
until the less acute peripheral vision is able to resolve them at about 2 nm.

Quote

The region surrounding an aircraft where a pilot can reliably expect to detect approaching
enemy aircraft extends to about 1.5 to 2.5 nm. Under conditions of good visibility, favorable
lighting, minimal clutter, etc., it is possible to see modern fighter-size aircraft at ranges of 10
nm or more if they fall within the highly focused central vision. Aircraft are sometimes seen
at these longer ranges, especially if the observer is cued and able to limit the search area to a
few degrees, but uncued observers are extremely unlikely to detect enemy aircraft at anything
approaching maximum theoretical range.  13

13.   See S. Schallhorn et al., Visual Search in Air Combat (Pensacola, FL: Naval Aerospace Medical Research Laboratory,
1990), particularly 5–11, for the discussion on visual search and the practical use of the human eye as an air-to-air sensor.

 

FIGURE 5. VISUAL SEARCH LIMITS

image.png.a487acbfca135472c0b1270ea70cee06.png

Quote

 

Systematically searching an area of sky requires the observer to focus on a distant object such
as the horizon to ensure proper focus.15 The shaded area in the illustration on the left of Figure
5 represents the visual “lobe” thus formed where an opposing aircraft could physically be
detected by the human eye in one “fixation.” At extreme ranges, the lobe is only about 2
degrees wide, so aircraft A would only become visible on the third fixation, or deliberate
shifting of the visual lobe. During fixation 3, aircraft B would not be detected, even though it is
closer to the observer than aircraft A, because it lies outside the observer’s central vision.
Aircraft C would be detected on fixation 3, even though it is at the same angle to the observer
as aircraft B, because it is close enough to be detected by the less sensitive peripheral vision.
This explains why even when aircrew use disciplined search patterns and fly in formations
where members are assigned different search sectors, the likelihood of detecting enemy
aircraft beyond about 2 to 3 nm is low.16

For example, a pilot searching a relatively small sector
90 degrees wide by 20 degrees high might be physically able to see a target at 7 nm range, but
the probability it would fall within his 2 degree central vision on any given fixation is just
1/450 (0.002). This per-fixation probability increases to only about 1/110 (0.009) at 3 nm and
is still only about 1/5 at 2 nm. The illustration on the right of Figure 5 shows the cumulative

probability a pilot searching each 90-degree sector with 20 fixations per minute would detect
an aircraft approaching from various directions by range.17 The cumulative probability of
detecting the approaching aircraft remains below 0.50 until it is between 1.9 and 2.8 nm. For
simplicity, the series of figures that follow will use a circular 2 nm area to illustrate the region
where visual search is likely to detect an approaching enemy aircraft.

 

 

15.   When not focused on a distant object, the eye muscles tend to relax, resulting in a focus distance of 10 feet or less.

16.   Disciplined search patterns consist of eight to twelve relatively widely spaced fixations per 90-degree horizontal sector.
Aircrew in a formation are generally assigned to search a specific sector and do so in a systematic manner usually
proceeding from the front to the rear of the sector slightly above the horizon, then from the rear to the front of the sector
slightly below the horizon, or vice versa. The goal is to cover the sector in a reasonable amount of time with fixations
spaced about 15 degrees apart. This maximizes the chances of detecting a target with peripheral vision while sacrificing
little in terms of relatively low-probability long-range detections.

Edited by Bies
  • Thanks 1
  • Upvote 3

Share this post


Link to post
Share on other sites

Very informative, thanks.

 

What is particularily interesting is "pilots searching for opposing aircraft without some sort of cue to limit their search are unlikely to detect them until the less acute peripheral vision is able to resolve them at about 2 nm." It is something that can be easily checked in real life. From an airliner you can spot other airliners quite far away but once you have "lost" them it is rather difficult to locate them again unless you look exactly at the right place.

 

My feeling is that long range detection when you know where to look is just fine in game. The plane appearing as one pixel is appearing far enough. However, medium range, peripheral view detection (which according to this document should be possible under 2nm), typically locating a plane under you in a glimpse during a dogfight, is harder in game (VR) than in reality. In reality in normal condition you instantly and effortlessly locate a plane at medium range (1 km?) in just a glimpse, even if you don't know it is there. In game, when the background is the ground, you clearly need to scan the background to locate the plane, which requires concentration and takes more time. At least in VR.

 

This is easy to understand and very difficult to solve. In reality the sharpness of the contour of the plane contrasts with the blurred background, and allows stereoscopic vision making the plane looking closer that infinity at moderate distances. In game the lower resolution does not make that possible.

 

 

 

 

Edited by haltux
  • Upvote 2

Share this post


Link to post
Share on other sites

What IRL is very easy to see, is the moving in front of a not moving background. This is also very important in ground war. You can have very good camouflage, as soon as you make fast movements, you can be detected. Peripheral view works especially with movements.

  • Upvote 1

Share this post


Link to post
Share on other sites

Something that I have noticed is that when you fly with an Instrutor or another Pilot, the one in Command will often Spot Traffic much, much later if at all compared to the one who simply acts as Passenger/Observer/Second Pilot.

 

Even if you as PIC actively look around to find Traffic while your Observer/Second isn't, he will still have an easier Time spotting Traffic.

Share this post


Link to post
Share on other sites
27 minutes ago, 6./ZG26_Klaus_Mann said:

Something that I have noticed is that when you fly with an Instrutor or another Pilot, the one in Command will often Spot Traffic much, much later if at all compared to the one who simply acts as Passenger/Observer/Second Pilot.

 

Even if you as PIC actively look around to find Traffic while your Observer/Second isn't, he will still have an easier Time spotting Traffic.

It might have to do with, even if you are not fully concentrated on flying, your brain still needs some capacity for the flying, so you are, even if looking around actively, not as receptive as someone, who is just passenger.

Share this post


Link to post
Share on other sites
3 hours ago, 6./ZG26_Klaus_Mann said:

Something that I have noticed is that when you fly with an Instrutor or another Pilot, the one in Command will often Spot Traffic much, much later if at all compared to the one who simply acts as Passenger/Observer/Second Pilot.

 

Even if you as PIC actively look around to find Traffic while your Observer/Second isn't, he will still have an easier Time spotting Traffic.

This is something I have noticed while driving in wildlife-prone areas. Lots of deer on the highways around where I am, and my wife will spot them on the side of the road before I do. And that's with a relatively focused scanning area.

Share this post


Link to post
Share on other sites
13 minutes ago, RedKestrel said:

This is something I have noticed while driving in wildlife-prone areas. Lots of deer on the highways around where I am, and my wife will spot them on the side of the road before I do. And that's with a relatively focused scanning area.

 

put a wife as a back gunner to spot then

  • Upvote 1

Share this post


Link to post
Share on other sites
38 minutes ago, Cpt_Siddy said:

 

put a wife as a back gunner to spot then

She missed her calling as a Pe-2 gunner

Share this post


Link to post
Share on other sites

This very interesting pilot view capability analysis does answer some of the requests here for high zooming quality feature to detect planes over 10 km. It seems to me not realistic.

I am building  a simple mission  where i fly a fighter plane  at slow speed straight towards another fighter plane circling at 15km . The system calculates the distances in 500 mt steps and displays the distance as I approach. Then I can test at what distance I can see it or not. Because he circles I do not know where he will appear when I will first see it. Then one can change the position of the Sun and have cloud tops in various densities and see how it affects the vision. You can also change the altitudes and spot the plane against the sky or the ground (winter/summer). I will make a summary of that in a near future when done and see how it looks.

  • Like 2
  • Upvote 2

Share this post


Link to post
Share on other sites
7 hours ago, IckyATLAS said:

This very interesting pilot view capability analysis does answer some of the requests here for high zooming quality feature to detect planes over 10 km. It seems to me not realistic.

 

It's worth remembering that a "modern fighter-size aircraft" is roughly the size of a B-25.

Share this post


Link to post
Share on other sites
2 hours ago, RobCarter3 said:

 

It's worth remembering that a "modern fighter-size aircraft" is roughly the size of a B-25.

This brings even a stronger argument, meaning that the distances in the study have to be even smaller.

Share this post


Link to post
Share on other sites

Ok I did my tests. I use my 4K, 43" screen, well calibrated. This means that the cockpit size (in default FOV, no Zoom) is not real size but is not too far from it. When I fly my eyes are at about 80cm from this huge screen. The smaller the pixels or the higher the screen resolution and the shorter the viewing distance. This means that like in real life I have to scan the screen.

I fly a P39 because it has one of the best cockpit in terms of visibility. I did my tests in "real conditions" I mean with the cockpit and its structure visible, which means you have obstructions here and there. There are fighters with much worse visibility, but lets not do it too difficult.

I made tests with the sun high above and behind me hoping this would have a better contrast to spot the plane. The sun behind makes things worse because you get reflexion, shades, glittering glass scratches, and above all your propeller disc becomes much more visible and shiny. So the cockpit glass the propeller create a kind of additional blurry filter that does not help to spot a few pixels. Just for the record, I have an excellent near and far view (I am sixty  and appreciate that) still near to 10/10, so no glasses "yet".

There is also the HUD visor glass that adds a another layer that disturbs the contrast. 

 

The P39 flies at 3000 mt straight line same altitude towards the enemy plane which stands st 3400 mt. This is to have it a little above the P39 flight line to ensure it can be spotted. The P39 flies at 250 km/hr starting from distance of 15 kilometers. The enemy plane circles at 180 km/hr, again this is not to make things too difficult.

Blue sky lapino summer map. Scattered clouds at 1'000 mt. Time of the day 13:00 Sun high in the sky. It is at that time that the propeller disc interferes less. As we can see all this is optimal conditions, and I know that I fly towards the enemy plane.

 

Planes tested for visibility are U2vs, A20B, B25, HE111, BF109, P47, Tempest, FW190, Pe2, P51, Yak1

 

To make things as objective as possible, during the test there is absolutely nothing displayed (I mean distances, icons etc.) and the moment I have spotted the plane I hit the pause button, then I display the distance. here the results which are averages and I will explain why they can differ.

 

U2vs        5km

A20B       9km

BF109     4/5 km 

HE111    10/11 km

B25         10/11 km

P47          8/3 km

P51          8/9 km

Tempest  8 km

FW190     4 km

Yak1         3/4 km

 

Using the max Zoom FOV improves at best these values by 1 km. So its effect is negligible. But pilots didn't have zooming eyes.

 

Some comments here. When an aircraft with metallic finish glitters the pixels are white and against the light blue sky the are seen better hence at a higher distance. The P51 is typical. As I said the plane is circling so it has the wings turned and so with a higher surface reflecting the sun. It is at that moment that you detect it best. In other angles you do not see it. The problem is that at 8 or 9 kilometers it is a few white very small pixels and if you are not just looking there at that moment no chance. As I said I have to scan my screen so I miss here and there.

 

The P47 is similar but if you missed the glitter and he is higher so less reflections towards you then detection falls drastically. The german fighters are difficult to detect, same for the Yak.

The big planes as it can be expected you can detect them much further consistently. The He111 at that distance is still about 10 pixels. So you could detect it a little further, but then you need to concentrate, and on my 4K I cannot see just one grey pixel against the light blue. Maybe one white pixel but it is the eye limit here and it is too tiring.

The U2vs has two wings so during the turn there is a chance you have a larger wing surface and then a better aspect ratio.

 

On way to improve things could be to push color contrast on the Monitor, but this becomes an artificial solution.

 

I did recordings to check if the plane was displayed before my detection. And yes it is displayed but detecting it is another story.

 

Conclusion: In battle with fighters flying at 400 km/hr, crossing speed at 800 km/hr head on (this is 222 mt/sec) and with your plane not flying straight but with turns that means, variable lighting conditions in the cockpit, and those specks that are enemy fighters against ground and they can be anywhere, no way, you detect at best at 2 kilometers if you are good. 

This seems to me pretty realistic. 

 

Those who request special zooms to see planes well at 15 km, sorry but that's the arcade version of the game.

 

 

 

 

 

 

 

 

 

 

  • Thanks 2
  • Upvote 3

Share this post


Link to post
Share on other sites
3 hours ago, IckyATLAS said:

 

Some comments here. When an aircraft with metallic finish glitters the pixels are white and against the light blue sky the are seen better hence at a higher distance. The P51 is typical. As I said the plane is circling so it has the wings turned and so with a higher surface reflecting the sun. It is at that moment that you detect it best. In other angles you do not see it. The problem is that at 8 or 9 kilometers it is a few white very small pixels and if you are not just looking there at that moment no chance. As I said I have to scan my screen so I miss here and there.

 

The P47 is similar but if you missed the glitter and he is higher so less reflections towards you then detection falls drastically. The german fighters are difficult to detect, same for the Yak.

 

 

Problem is, the metallic finish is currently wrongly applied to Allied aircraft that don't have metallic finish.  

 

Allied aircraft that have camouflage paint finish are being unfairly highlighted as metallic when they are not; for example, Spitfire, Tempest and P51 Mustang with British camouflage. 

 

Consequently, Allied fighters stand out and are easier to follow and German fighters are harder to detect and harder to follow and maintain in visual contact.

 

It is very uncomfortable to fly Allied aircraft, even though camouflaged, in the knowledge that one is highlighted in this way and the opposition are not; this is a clear unfair advantage to the Axis side.  It is one of the reasons I have not been flying much lately and have gone off IL-2 somewhat.  I hope the developers fix this soon; I would hope it is easy to fix by correcting aircraft LOD codes.

 

https://forum.il2sturmovik.com/topic/12255-graphics-models-and-maps/page/12/

 

 3df4a8b29c688e43cbd623592b12b7c9.png

 

https://forum.il2sturmovik.com/topic/56416-aircraft-visibility-bug-bopb/?tab=comments#comment-866877

 

Happy landings,

 

56RAF_Talisman

 

 

 

  • Upvote 1

Share this post


Link to post
Share on other sites
3 minutes ago, 56RAF_Talisman said:

 

Problem is, the metallic finish is currently wrongly applied to Allied aircraft that don't have metallic finish.  

 

Allied aircraft that have camouflage paint finish are being unfairly highlighted as metallic when they are not; for example, Spitfire, Tempest and P51 Mustang with British camouflage. 

 

Consequently, Allied fighters stand out and are easier to follow and German fighters are harder to detect and harder to follow and maintain in visual contact.

 

It is very uncomfortable to fly Allied aircraft, even though camouflaged, in the knowledge that one is highlighted in this way and the opposition are not; this is a clear unfair advantage to the Axis side.  It is one of the reasons I have not been flying much lately and have gone off IL-2 somewhat.  I hope the developers fix this soon; I would hope it is easy to fix by correcting aircraft LOD codes.

 

https://forum.il2sturmovik.com/topic/12255-graphics-models-and-maps/page/12/

 

 3df4a8b29c688e43cbd623592b12b7c9.png

 

https://forum.il2sturmovik.com/topic/56416-aircraft-visibility-bug-bopb/?tab=comments#comment-866877

 

Happy landings,

 

56RAF_Talisman

 

 

 

This^^^

 

I'd rather they bring back old visibility system, now spotting plane 5-10km away or even further but when getting close (1-4km) it dissapears in certain conditions (which happens a lot) even keeping eye on him all the time.

Like they have cloaking device in their planes.

It feels totally broken and even makes lottery out of the dogfight despite being on enemies six o'clock.

I know devs used RL tests data but they didn't took in consideration pc hw limitations and VR users.

Going online i'm not sure are others cheating or not.

In this area game totally degraded!

 

 

VR user!

Share this post


Link to post
Share on other sites
1 hour ago, 56RAF_Talisman said:

 

Problem is, the metallic finish is currently wrongly applied to Allied aircraft that don't have metallic finish.  

 

Allied aircraft that have camouflage paint finish are being unfairly highlighted as metallic when they are not; for example, Spitfire, Tempest and P51 Mustang with British camouflage. 

 

Consequently, Allied fighters stand out and are easier to follow and German fighters are harder to detect and harder to follow and maintain in visual contact.

 

It is very uncomfortable to fly Allied aircraft, even though camouflaged, in the knowledge that one is highlighted in this way and the opposition are not; this is a clear unfair advantage to the Axis side.  It is one of the reasons I have not been flying much lately and have gone off IL-2 somewhat.  I hope the developers fix this soon; I would hope it is easy to fix by correcting aircraft LOD codes.

 

https://forum.il2sturmovik.com/topic/12255-graphics-models-and-maps/page/12/

 

 3df4a8b29c688e43cbd623592b12b7c9.png

 

https://forum.il2sturmovik.com/topic/56416-aircraft-visibility-bug-bopb/?tab=comments#comment-866877

 

Happy landings,

 

56RAF_Talisman

 

 

 

This perfectly true. I have also tested the P47 in camouflage and it glitters the same if it has the right orientation according to sun and player.

  • Thanks 1

Share this post


Link to post
Share on other sites

Don't confuse nautical miles with kilometers either, the very high spot ability cited between 1.9 to 2.8 nm's equates to 3.5 to 5.2 Km's.  Pilots have a greater than 50% chance of immediately detecting if they scan, but we can't even track them in this region.

Edited by [CPT]Crunch
  • Upvote 1

Share this post


Link to post
Share on other sites

I think the current vis is almost perfect, based on real life experience and the reasons above.

 

The loss of contacts at short/med range against the ground (the moving target as described above) is the only problem I have, and even that has improved in the short time I have been playing.

 

 

 

 

Share this post


Link to post
Share on other sites
On 12/18/2019 at 4:25 AM, Bies said:

From:

Center of Strategic and Budgetary Assessments

 

TRENDS IN AIR-TO-AIR COMBAT
IMPLICATIONS FOR
FUTURE AIR SUPERIORITY

 

https://csbaonline.org/uploads/documents/Air-to-Air-Report-.pdf

 

13.   See S. Schallhorn et al., Visual Search in Air Combat (Pensacola, FL: Naval Aerospace Medical Research Laboratory,
1990), particularly 5–11, for the discussion on visual search and the practical use of the human eye as an air-to-air sensor.

 

FIGURE 5. VISUAL SEARCH LIMITS

image.png.a487acbfca135472c0b1270ea70cee06.png

 

15.   When not focused on a distant object, the eye muscles tend to relax, resulting in a focus distance of 10 feet or less.

16.   Disciplined search patterns consist of eight to twelve relatively widely spaced fixations per 90-degree horizontal sector.
Aircrew in a formation are generally assigned to search a specific sector and do so in a systematic manner usually
proceeding from the front to the rear of the sector slightly above the horizon, then from the rear to the front of the sector
slightly below the horizon, or vice versa. The goal is to cover the sector in a reasonable amount of time with fixations
spaced about 15 degrees apart. This maximizes the chances of detecting a target with peripheral vision while sacrificing
little in terms of relatively low-probability long-range detections.

It should also be noted that the rate of fixations they give, and the following probability is completely arbitrary and incorrect. It is not in the original paper either. Your eyes make more fixations faster when you read this post.

 

http://www.scholarpedia.org/article/Human_saccadic_eye_movements

 

https://www.britannica.com/science/saccade

 

"It is during these periods, which last on average about 190 milliseconds, that the eyes take in visual information. " This would equate to over 300 fixations per minute. Your eyes also make two fixations at the same time, and see everything from a slightly different angle, which increases the minute of arc that the eyes can see.

 

 

It is correct however that contacts at or under 2.5 miles should be easy to locate, and that contacts further out should require more careful scanning to find. However, given a rendering system on the screen that represents contacts in some proportional manner to real life contacts (contrast, size, etc) this problem more or less solves itself since these principles apply to searching for small numbers of pixels as well. However it does not work if the contacts are unreasonably difficult to see like we currently have in game.

  • Upvote 1

Share this post


Link to post
Share on other sites

I tested the Spitfire with my test bench. And first even if camouflaged it glittered. This is not normal, it should not glitter. So two/three white pixels appeared just by luck in front of where I was looking and down just above my cockpit dark surface so the eye picked that contrast and this was a 12,2 kilometer contact, the maximum detection so far. The glitter lasted three seconds then faded and the contact disappeared. This is because the plane was in a turn just right for max glitter, but the orientation changed and the contact disappeared, even if I could see the direction. Then I lost it completely. 

This is exceptional because I knew the plane was in front of me, and I was flying straight and steady.

Share this post


Link to post
Share on other sites

Exaxtly, the rendering 'should' be equal to or even better what we see with our eyes in the 2° cone IRL, after all we still got to use our own eyes to spot the pixels on a monitor that far exceeds the 2° range also. 

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×
×
  • Create New...