Spotting aircraft issue

[Kling]

Today I had a flight to Berlin so decided to take some photos.

First is a Cathay Pacific 747-400 at 5 nautical miles (9.26km). A 747 is 70m long and a Heinkel 111 is 18 m long so a 747 is 3.8 times as big. Now imagine how small a He111 will be at this distance.
http://img17.imageshack.us/img17/6105/8leh.jpg (http://imageshack.us/photo/my-images/17/8leh.jpg/)

Next is a 737 at 2.5 nm (4.6km). Not sure what version it is but still its alot bigger than any bomber we have in game (maybe fw200 comes close)
http://img600.imageshack.us/img600/5129/4b7w.jpg (http://imageshack.us/photo/my-images/600/4b7w.jpg/)

And finally a picture of my onboard radar to prove that this 737 is 2.5nm away. The target shows as blue with a arrow pointing down, telling me that he is descending, and the -36 means that he is 3600 feet below me.
http://img19.imageshack.us/img19/7324/puyw.jpg (http://imageshack.us/photo/my-images/19/puyw.jpg/)

Remember these planes are NOT camouflaged so its alot easier to see these than a camouflaged bomber... Just to put things in perspective ;)

Regards
Kling

[92 Sqn. Philstyle (QJ-P)]

Nice one Kling.

Good to see such images!

[ATAG_Snapper]

Great shots and excellent perspective, Kling. :thumbsup: Thanks for sharing this. In my extremely limited experience doing touch 'n go's in a Grumman TR2 trainer decades ago, I remember how panicked I felt trying to spot other private aircraft in the circuit. And that was KNOWING where they were from the continual radio traffic with the tower! My hat is off to ALL real pilots like yourself that spot other aircraft as a matter of course.

As I mentioned elsewhere here, some factoring must be done simply for gameplay, especially since we don't use icons. Otherwise, we'd have to rename Cliffs of Dover to "Marco Polo"! LOL. I'm looking forward to this upcoming TF release for many reasons; not the least being the treatment of the LOD situation.

[DennydD]

Nice pictures. Dude you should save for a better camera. :devilish:

[Kwiatek]

Are you sure about 737 was at 2.5 nm (4.6km)? It looks for me really much further. As for 747 it looks ok to me.

[kopperdrake]

Yeah - but if he uses a lower res camera then the dots are easier to see!

[Kling]

Im using my Samsung Galaxy II phone and it has 8megapixel... Obviously the zoom is almost non existing... So the quality is what it is..

Kwiatek.. The Tcas (Traffic collision avoidance system) is very accurate with distance and altitude but no so much with location on the screen. Usually your Tcas will say that a target is at you 2oclock but he could very well irl be at your 3 etc etc... But for the distance its correct. I took the picture of the 737 first and then the photo of the Tcas so the 737 is most likely closer on Tcas 10 secs later when I took that photo.. Closing speed is something like 1500km/h.. Maybe the correct distance would be 4nm for the 737... ;)

[Kwiatek]

Yea 4nm sound more beliveable to me :)

[Dutch]

Thanks for this Kling. I really enjoy seeing your inflight pics. :thumbsup:
Now imagine what a mostly pale blue 109 would look like at these distances.........:-)

[Ohms]

I have had quite a few test flights to checkout TCAS systems and always you have a couple of contacts which are hard to see even when you know where they are. TCAS is a wonderful thing.:thumbsup:

[gavagai]

How big should they be on my monitor and how far away do I sit to have the correct field of view?

[14./JG5_Capt.Stubing]

Great Perspective... I'm a pilot and own my own Bonanza and fly out of Southern California. We are one of the busiest airspaces in the world. I too have a version of TCAS which is very helpful. There are times especially with haze that I don't see aircraft unless they are about 2 miles or less. That's with Radar calls and TCAS. Also ground clutter makes it hard to pick up targets as well. But with that said I think the difference is we are looking at the world through a much smaller aperture and in much lower res that the MK1 eyeball. There has to be some balance of making targets stand out vs. making them disappear like they do now.

[DennydD]

I coudn't resist....

Picture of a FW200, ingame Dot, 3.83 km distance:

3718

The 737 put into the original picture (without size manipulation or whatever)

3719


Spotting isn't easy. Fact.

But to make it harder than reality does not help.

[Kling]

fear not... in the next patch the issue is gone ;)

[DennydD]

fear not... in the next patch the issue is gone ;)

Thank god there is TF! :salute:

[AKA_Knutsac]

Check out this relevant paper from NAMRL (interesting place, I used to drink with some of the researchers and staff...a long time ago); good stuff starts on page 4. http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA241347

Also these bits (and consider the size of modern fighter a/c compared to BoB):

"A total of 759 training engagements at the Naval Air Station Oceana Tactical Air Combat Training System (TACTS) range revealed that in 624 of the
engagements the pilots first sighted the target as a dot against the background at an average distance of 5.67 nmi (Hamilton & Monaco, 1986; Monaco & Hamilton, 1985). In the remaining 135 engagements exhaust smoke, contrails and sun glint off the aircraft allowed the pilots to detect the aircraft at even greater distances."

"Two very closely matched groups of eyeglass and non-eyeglass wearers had average detection ranges of 4.52 and 5.64 nmi respectively when using all detection means including aircraft sighting, target glint, contrails and exhaust smoke. When limiting subjects to aircraft-only detections, the corresponding distances were 4.35 and 5.54 nmi respectively."

"In 1983, Kress & Brictson studied 87 air-to-air engagements at the Yuma TACTS range. Average unaided detection distances for the target F-5 and F-4 aircraft were 3.1 nmi."

"Another study by Hutchins in 1978 at the Air Combat Maneuvering Range (ACMR), which is the earlier name of the TACTS, involved 45 air combat training engagements. The mean detection distance of the A-4 targets was 3.09, with a range of 0.38 to 6.23 nmi."

From http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA414893

~S~

AKA Knutsac

[ATAG_Lewis]

Great Pics Kling.....Thanks for sharing...

[9./JG52 Ziegler]

Great Perspective... I'm a pilot and own my own Bonanza and fly out of Southern California. We are one of the busiest airspaces in the world. I too have a version of TCAS which is very helpful. There are times especially with haze that I don't see aircraft unless they are about 2 miles or less. That's with Radar calls and TCAS. Also ground clutter makes it hard to pick up targets as well. But with that said I think the difference is we are looking at the world through a much smaller aperture and in much lower res that the MK1 eyeball. There has to be some balance of making targets stand out vs. making them disappear like they do now.

Haze is the worst. I flew for several years out of BWI on the east coast and in summer, like now on a humid day, it was tough. Night flights were much easier.

[RAF74_Buzzsaw]

Check out this relevant paper from NAMRL (interesting place, I used to drink with some of the researchers and staff...a long time ago); good stuff starts on page 4. http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA241347

Also these bits (and consider the size of modern fighter a/c compared to BoB):

"A total of 759 training engagements at the Naval Air Station Oceana Tactical Air Combat Training System (TACTS) range revealed that in 624 of the
engagements the pilots first sighted the target as a dot against the background at an average distance of 5.67 nmi (Hamilton & Monaco, 1986; Monaco & Hamilton, 1985). In the remaining 135 engagements exhaust smoke, contrails and sun glint off the aircraft allowed the pilots to detect the aircraft at even greater distances."

"Two very closely matched groups of eyeglass and non-eyeglass wearers had average detection ranges of 4.52 and 5.64 nmi respectively when using all detection means including aircraft sighting, target glint, contrails and exhaust smoke. When limiting subjects to aircraft-only detections, the corresponding distances were 4.35 and 5.54 nmi respectively."

"In 1983, Kress & Brictson studied 87 air-to-air engagements at the Yuma TACTS range. Average unaided detection distances for the target F-5 and F-4 aircraft were 3.1 nmi."

"Another study by Hutchins in 1978 at the Air Combat Maneuvering Range (ACMR), which is the earlier name of the TACTS, involved 45 air combat training engagements. The mean detection distance of the A-4 targets was 3.09, with a range of 0.38 to 6.23 nmi."

From http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA414893

~S~

AKA Knutsac

It is important to understand that modern jet fighters are closing at a much greater speed than WWII types.

While a pilot is scanning a different area of the sky, they can cover a very large distance.

Also, headon, a modern Jet fighter does not have a significantly larger profile than a WWII fighter

For that reason, this survey cannot necessarily be taken as an accurate gauge to the distances spotting might occur in WWII situations.

There is no question that enabling spotting out to 25 km may be allowing the players too much visual acuity, but reducing the visibility range too much will not provide an accurate environment.

One thing which most RAF pilots reported, was an ability to spot a formation of bombers from a very long distance. Fighter escort over those bombers would only be noticed after the pilots had moved much closer. Individual bombers were also more difficult to see than a formation. Formations act as multipliers.

[Mattias]

I coudn't resist....

Picture of a FW200, ingame Dot, 3.83 km distance:

3718

The 737 put into the original picture (without size manipulation or whatever)

3719


Spotting isn't easy. Fact.

But to make it harder than reality does not help.

:salute:

Nice comparison shots, but please note that we don't know the FOV of Klings camera :D

Cheers/m

[ChiefRedCloud]

WOW! Nice ...... thanks

[Kling]

:salute:

Nice comparison shots, but please note that we don't know the FOV of Klings camera :D

Cheers/m

this is very true.
The only thing i did was to zoom in until the planes on my phone screen had the same size as the real object. of course its not a scientific way of doing it but was more for fun ;)

[AKA_Knutsac]

It is important to understand that modern jet fighters are closing at a much greater speed than WWII types.

While a pilot is scanning a different area of the sky, they can cover a very large distance.

Also, headon, a modern Jet fighter does not have a significantly larger profile than a WWII fighter

For that reason, this survey cannot necessarily be taken as an accurate gauge to the distances spotting might occur in WWII situations.

There is no question that enabling spotting out to 25 km may be allowing the players too much visual acuity, but reducing the visibility range too much will not provide an accurate environment.

One thing which most RAF pilots reported, was an ability to spot a formation of bombers from a very long distance. Fighter escort over those bombers would only be noticed after the pilots had moved much closer. Individual bombers were also more difficult to see than a formation. Formations act as multipliers.

Buzzsaw,

I was throwing these documents out there because they provide some real world empirical data on visual target acquisition ranges from professional fighter/attack pilots and because they describe the environmental and physiological reasons why target acquisition ranges may be less then the lay person, mathematical modeler, or armchair sim pilot, might imagine. I concur groups of a/c would be visible at greater distances, as would contrailing a/c, smoking a/c, etc, however, baring these factors, individual a/c should remain pretty difficult to spot, as in real life (2 to 5 mile draw distance?).

I also found the focus of the second reference (the dissertation) quite interesting in that the author is attempting to address one of the main shortcomings of flight simulators...don't know if a similar algorithm to improve target detail at range is used in CloD or not.

~S~

AKA Knutsac

[RAF74_Buzzsaw]

Buzzsaw,

I was throwing these documents out there because they provide some real world empirical data on visual target acquisition ranges from professional fighter/attack pilots and because they describe the environmental and physiological reasons why target acquisition ranges may be less then the lay person, mathematical modeler, or armchair sim pilot, might imagine. I concur groups of a/c would be visible at greater distances, as would contrailing a/c, smoking a/c, etc, however, baring these factors, individual a/c should remain pretty difficult to spot, as in real life (2 to 5 mile draw distance?).

I also found the focus of the second reference (the dissertation) quite interesting in that the author is attempting to address one of the main shortcomings of flight simulators...don't know if a similar algorithm to improve target detail at range is used in CloD or not.

~S~

AKA Knutsac

CoD uses 7 LOD's for the aircraft, which is as much or more than any combat aircraft simulation out there. That includes RoF and BoS.

Future sims may provide attempt to provide a closer replication of what the real human eye sees, but they won't really come close until we get massive improvements in resolution and the addition of real 3D. (not what is sold as 3D now)

As I mentioned before, 25k is likely too optimistic a range for the game to set for spotting, but your suggestion of 8k is low by most of the accounts I have read.

[AKA_Knutsac]

Draw distances beyond 8km are reasonable, maybe even out to 25km for large a/c, but at longer range the dots should be very small and very faint (low contrast). As in real life, we should really have to work hard to spot distant contacts (IMHO).

~S~

AKA Knutsac

[Mattias]

this is very true.
The only thing i did was to zoom in until the planes on my phone screen had the same size as the real object. of course its not a scientific way of doing it but was more for fun ;)

It would be interesting to know what equivalence of FOV you used -I bet it was way under FOV 70 :)

Cheers/m

[Broodwich]

One thing which most RAF pilots reported, was an ability to spot a formation of bombers from a very long distance. Fighter escort over those bombers would only be noticed after the pilots had moved much closer. Individual bombers were also more difficult to see than a formation. Formations act as multipliers.

I was gonna say, finding a bunch of contacts relatively closely grouped together is obviously going to be a lot easier to find than a single one.

[DennydD]

It would be interesting to know what equivalence of FOV you used -I bet it was way under FOV 70 :)

Cheers/m

The iPhone 4 has a FOV of ~60°.

Let's calculate it:

Klings picture is 1600 pixels wide. The 737 is in a distance of 4.6km. We assume it is a 737 800 then it is ~40m long and has nearly the same wingspan. We do not need to be too exact here because this is just a small factor compared to the distance. The 737 is 10x4 pixels wide in the picture. Means that we see 160 times more (horizontal).

Lets see:

FOV (horizontal) ~ 160 * 2 * arctan (20/4600) ~ 79.7°

If you look up the specs of these cameras you find that this fits very well.

The current dot I show in this little picture is 4x2 pixels wide. The FW200 has ~32m wingspan (but it is nearer in the pic :-P)
This means the 737 dot is 5 times bigger.

:thumbsup:

[Kling]

The iPhone 4 has a FOV of ~60°.

Let's calculate it:

Klings picture is 1600 pixels wide. The 737 is in a distance of 4.6km. We assume it is a 737 800 then it is ~40m long and has nearly the same wingspan. We do not need to be too exact here because this is just a small factor compared to the distance. The 737 is 10x4 pixels wide in the picture. Means that we see 160 times more (horizontal).

Lets see:

FOV (horizontal) ~ 160 * 2 * arctan (20/4600) ~ 79.7°

If you look up the specs of these cameras you find that this fits very well.

The current dot I show in this little picture is 4x2 pixels wide. The FW200 has ~32m wingspan (but it is nearer in the pic :-P)
This means the 737 dot is 5 times bigger.

:thumbsup:

Uhm what?!? Too early in the morning for me! :p

[gavagai]

The iPhone 4 has a FOV of ~60°.

Let's calculate it:

Klings picture is 1600 pixels wide. The 737 is in a distance of 4.6km. We assume it is a 737 800 then it is ~40m long and has nearly the same wingspan. We do not need to be too exact here because this is just a small factor compared to the distance. The 737 is 10x4 pixels wide in the picture. Means that we see 160 times more (horizontal).

Lets see:

FOV (horizontal) ~ 160 * 2 * arctan (20/4600) ~ 79.7°

If you look up the specs of these cameras you find that this fits very well.

The current dot I show in this little picture is 4x2 pixels wide. The FW200 has ~32m wingspan (but it is nearer in the pic :-P)
This means the 737 dot is 5 times bigger.

:thumbsup:

Well, now I can answer my earlier question:

"How big should they be on my monitor and how far away do I sit to have the correct field of view?"

Thanks!

[Mattias]

The iPhone 4 has a FOV of ~60°.

Not when zooming, which Kling did :)

[DennydD]

Not when zooming, which Kling did :)

I calculatet it for you. FOV~70-80°. Depends on if you take äqudistant objects or you use the intercept theorem. Does not realy matter cause the difference is not big. An hes got a Samsung. The Samsung has a much wider FOV in camera mode than the iPhone4. Just read the tech specs (focal length 4,28mm [iPhone] vs. 3,97mm [SGS2]).

:-)

[Mattias]

I calculatet it for you. FOV~70-80°. Depends on if you take äqudistant objects or you use the intercept theorem. Does not realy matter cause the difference is not big. An hes got a Samsung. The Samsung has a much wider FOV in camera mode than the iPhone4. Just read the tech specs (focal length 4,28mm [iPhone] vs. 3,97mm [SGS2]).

:-)

:salute:

I'm not picking a fight about this, but please reconsider how a FOV of 70-80 is possible on a camera with a FOV of 60 when fully zoomed out, given he zoomed in when taking the picture :-)

EDIT: I read too fast and missed it was a Samsung, not an iPhone :doh: Still he must have had a FOV way lower than 70-80 if he zoomed in until the plane he saw matched his screen in size :thumbsup:

[DennydD]

:salute:

I'm not picking a fight about this, but please reconsider how a FOV of 70-80 is possible on a camera with a FOV of 60 when fully zoomed out, given he zoomed in when taking the picture :-)

Kling has a Samsung that has a wider FOV (>60) than the iPhone. I just gave the figure of the iPhone to have a rough feeling of the FOVs and it was the 1st thing i found. :D

But lets just focus on CloD. We know the propagated FOV there.

3757

This is a picture of a FW200 at 4000m with a FOV of 30° and a wingspan of 32m. The contact is 19 pixels wide.

32m <=> 19 pixels

Horizontal border to border at a distance of 4000m (lets call this D):

D=2*4000*tan(15)=2140m

2140m <=> 1920 pixels

=>

32m <=> 29 pixel

Conclusion: The contact renders only 66% of the size it should have.

:thumbsup:

[AKA_Knutsac]

"Conclusion: The contact renders only 66% of the size it should have."

If correct, that would seem to be a fundamental problem with the code. I don't know if anybody took the time to read either of the references I linked earlier, but the second one (http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA414893) is a dissertation describing a proposed method for improving the military's combat sims. Basically an algorithm, premised on real-world data from combat pilots, that applies a magnification factor to contacts based on their range. The goal is to better represent target orientation on distant contacts without misrepresenting acquisition range. The paper is about ten years old, so I have no idea whether any sims have adapted such a technique, or whether it would really work.

~S~

AKA Knutsac

[Mattias]

Kling has a Samsung that has a wider FOV (>60) than the iPhone. I just gave the figure of the iPhone to have a rough feeling of the FOVs and it was the 1st thing i found. :D

But lets just focus on CloD. We know the propagated FOV there.

3757

This is a picture of a FW200 at 4000m with a FOV of 30° and a wingspan of 32m. The contact is 19 pixels wide.

32m <=> 19 pixels

Horizontal border to border at a distance of 4000m (lets call this D):

D=2*4000*tan(15)=2140m

2140m <=> 1920 pixels

=>

32m <=> 29 pixel

Conclusion: The contact renders only 66% of the size it should have.

:thumbsup:

:salute:

I'm not really following your calculations but I have to admit I'm very tired at the moment :D I fail to see how a plane that's zoomed in until it has the same size on the tiny camera display as the eye sees it can have a FOV even close to the in game screenshots you have :)

I'll leave this discussion with this; if the contact is indeed too small you should contact Microsoft as this means the perspective projection matrix in DirectX is wrong. The Clod engine feeds DirectX with a model of the correct size at the correct distance, so any inconsistences is not the fault of the Clod engine :thumbsup:

Cheers/m

[Mattias]

"Conclusion: The contact renders only 66% of the size it should have."

If correct, that would seem to be a fundamental problem with the code. I don't know if anybody took the time to read either of the references I linked earlier, but the second one (http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA414893) is a dissertation describing a proposed method for improving the military's combat sims. Basically an algorithm, premised on real-world data from combat pilots, that applies a magnification factor to contacts based on their range. The goal is to better represent target orientation on distant contacts without misrepresenting acquisition range. The paper is about ten years old, so I have no idea whether any sims have adapted such a technique, or whether it would really work.

~S~

AKA Knutsac

:salute:

Magnifying distant contacts is a different beast, but I don't disagree this could be a good solution :thumbsup: Not enough time for this patch though :D

[DennydD]

As far as I know the projection equations are writen by the guy who does the engine an do not come with MS.

Was an illuminating debate. :salute:

[Mattias]

As far as I know the projection equations are writen by the guy who does the engine an do not come with MS.

Was an illuminating debate. :salute:

:salute:

I always like a good debate :thumbsup:

The code's calculations going to DirectX are standard vertex graphic equations :thumbsup:

Cheers/m

[hnbdgr]

Interesting discussion. It should be noted that whilst FOV in degrees or MM (if you're coming from a photography environment) works fairly nicely as analogy to human eyesight, it shouldn't be used this way.

The Acuity of human vision is much more complex. Whilst the FOV of a human being is generally accepted around 47mm (based on 35mm systems) it will be much more capable to recover detail at greater distances. Hence the zoom function in game. It's there to simulate what the human eye does naturally. Taking a picture with a camera, even if at the same FOV will not represent an accurate picture of what the eye is actually capturing.

Also, I have no real life experience as a pilot so can't comment, but the distance of 8~10km seems fairly realistic as a "first contact" however very faint (assuming ww2 fighter silhouette). How much is it right now out of interest?

anyway just my 2 cents:)