Dear ED - visibility

[Tippis]

So a factor of 2x at 2.65 miles as actually quite huge.

No. Because at 2.65 miles, the target is actually quite tiny.

 

At 14,000', a 40' target (eg. a Hornet) will cover 2.86 mils — double that and it becomes 5.7 mils. Your bog-standard 1440p display will sport roughly 110ppi. At arms-length distance (± weird arm lengths), 5.7 mils equates to 0.17", or all of 19 pixels. This is the scaled-up size of a unit that “really” should just be 9½ pixels.

 

It would be this (now replaced with an actual image of the respective sizes):

 

At the distances where the scale factor is high, the plane is not big enough to look odd, even at low altitude. It is big enough to give you a good sense of its orientation, which was the goal all along.

 

As the target came closer, you'd see it grow bigger — as things tend to do when they get close — only at a imperceptibly slower pace, which would look entirely natural because it's a smooth curve that doesn't vary between particularly extreme values.

 

Also what if that target aircraft was on a carrier deck?

Then it would be in size with the carrier. Chances are you wouldn't even notice anything unless you zoomed in, in which case that zoom needs to be discounted from the scaling since you're no longer using the 1:1 vision that the base scale factor applies to.

 

Players today have large 4K displays where this oddity would be much more noticeable than it was on 2003 era CRT monitors

The 2003 era monitors had roughly the same pixel densities as we have now.
Edited July 28, 2019 by Tippis

[Ziptie]

I use VR, and can regularly spot bogeys from 15-20 nm. Oculus Rift CV1. Not sure why the difference is so much different on a monitor, which unless I am incredibly wrong in my very basic knowledge of computer stuff - I'd think having higher resolution should enable you to see things from farther away......

 

Long winded way to say that it seems like there are settings that might be adjusted incorrectly on the OP's machine..

 

Happy days.

[3WA]

Personally, for now, I just turn on labels. I can't see crap in this game. I have an 1080p 23" monitor.

[KTFBGB]

I use VR, and can regularly spot bogeys from 15-20 nm. Oculus Rift CV1. Not sure why the difference is so much different on a monitor, which unless I am incredibly wrong in my very basic knowledge of computer stuff - I'd think having higher resolution should enable you to see things from farther away......

 

Long winded way to say that it seems like there are settings that might be adjusted incorrectly on the OP's machine..

 

Happy days.

 

That and most people sit too far from their monitors and tv’s. You can even find charts that show optimal viewing distance for different size screens. But basically you need to sit close enough that that the edges of the screen are all the way out to your peripheral vision. You should be able to see the whole screen barely, and need to move your eyeballs some to focus on the edges.

[SharpeXB]

Then it would be in size with the carrier. Chances are you wouldn't even notice anything unless you zoomed in, in which case that zoom needs to be discounted from the scaling since you're no longer using the 1:1 vision that the base scale factor applies to.

The problem with upscaling objects without scaling their surroundings is that it makes them effectively huge. This is the problem that Model Enlargement had. It makes distant targets too visible because they stand out huge compared to their surroundings. This Serfloss study seems all about air to air. What about ground targets? You can’t apply scaling to those without it looking odd because they’re always seen next to unscaled context. And when you change the scale factor with zoom view, again that’s going to look odd as well. Scaling is assuming that the wide angle FOV from your monitor is somehow “correct” but of course it’s not. The real 1:1 size you would see from the cockpit is equivalent to zoomed in. Scaling is only applied to the zoomed out view and making the targets remain the same size while everything else shrinks with perspective. That’s going to look odd.

 

Targets in DCS are easily seen at 2-3 miles using the zoom view. Why use scaling which only enlarges the target when zoom enlarges everything to life sized and doesn’t look odd?

 

The 2003 era monitors had roughly the same pixel densities as we have now.

Pixel density isn’t relevant. You can make a 2003 CRT the same pixel density as a 60” 4K TV by just sitting farther away from it. But there were no 60” UHDTVs in 2003 so all the math in that study is outdated.

Edited July 28, 2019 by SharpeXB

[Tippis]

The problem with upscaling objects without scaling their surroundings is that it makes them effectively huge.

Again: you won't see that at the distances involved.

 

What about ground targets? You can’t apply scaling to those without it looking odd because they’re always seen next to unscaled context.

Sure you can, for much the same reason: at the distances involved, the supposed oddity won't actually be all that noticeable. Also, you could scale surrounding assets as well, or just use a different scaling curve (Serfoss was interested in A2A perception — how well that translates to ground targets would be a different study).

 

And when you change the scale factor with zoom view, again that’s going to look odd as well.

It won't look all that much more odd than the zoom already does. You'd get a slight dolly zoom/vertigo effect going as the scaling counter-compensated for the altered FoV but zoom is already an unrealistic affordance so it's not particularly silly if it actually created fancy zoom effects when you used it. ;)

 

Scaling is assuming that the wide angle FOV from your monitor is somehow “correct” but of course it’s not. The real 1:1 size you would see from the cockpit is equivalent to zoomed in. Scaling is only applied to the zoomed out view and making the targets remain the same size while everything else shrinks with perspective.

You know how I've been throwing in “normalisation” in this discussion for the last couple of threads. Yeah, there's a reason for that. The only tricky bit (and I use that term very generously here) in all of this is deciding what the baseline FoV should be. As you move away from that using a FoV/zoom slider, the scaling adjusts to compensate — if you've given yourself 200% better vision, you no longer need a 100% increased target display size, after all.

 

Pixel density isn’t relevant.

Eh, yes it is. It's actually what determines the final size on your screen because it's all just trigonometry. Size / distance = angular coverage -> angular coverage × PPI = pixels on screen. All of that can be calculated, normalised, and compensated for. Pixel density does not change with distance. That's kind of the whole point of the exercise — the user's decision as to how far away they want to be from the screen is the only uncontrollable variable, and it leaves open the ultimate smug answer to all spotting complains: just sit closer. :D

 

Coincidentally, a 60" UHDTV has much much worse density than those 2003 CRTs — you'd have to go back to the early 1990s to get monitors with such low density.

Edited July 28, 2019 by Tippis

[Mars Exulte]

@Sharpe

That's not how pixel density works. Diameter divided by resolution. It's extremely straightforward and has nothing to do with where you're sitting.

 

Say 1080 on 24'' diameter, that means a 4k TV over 48'' will have lesser density than it, under 48'' will have more.

[Rikus]

ED don´t do Smart Scaling not because they don´t like, they don´t do because Waggs said the actual engine don´t support

[lucky-hendrix]

Then they should modify the engine, it is their own engine after all.

 

There needs to be a solution!

 

Sent from my VTR-L09 using Tapatalk

[SharpeXB]

Again: you won't see that at the distances involved.

You would perceive it on the large 4K screens and TVs people have today. That wouldn’t have been the case in 2003

 

You know how I've been throwing in “normalisation” in this discussion for the last couple of threads. Yeah, there's a reason for that. The only tricky bit (and I use that term very generously here) in all of this is deciding what the baseline FoV should be. As you move away from that using a FoV/zoom slider, the scaling adjusts to compensate — if you've given yourself 200% better vision, you no longer need a 100% increased target display size, after all.

That’s what is basically flawed with scaling. It’s trying to generate the correct angular size of distant objects. But it’s not taking into account the size of the screen or the viewers distance. It’s also not taking the FOV into account which by necessity is variable ie zoom view. There is no “baseline FOV”. So the whole concept is a bust.

 

the user's decision as to how far away they want to be from the screen is the only uncontrollable variable

Exactly. So this whole scaling study calculation is flawed. It’s not taking this into account. Nor is it considering the size of the screen and the FOV.

 

@Sharpe

That's not how pixel density works. Diameter divided by resolution. It's extremely straightforward and has nothing to do with where you're sitting.

 

Say 1080 on 24'' diameter, that means a 4k TV over 48'' will have lesser density than it, under 48'' will have more.

Ok so what that definition means is that today you can have a screen twice the size with the same apparent resolution. So the conclusion that you won’t see scaling oddities isn’t the same today as it was in 2003.

 

When I say pixel density I’m talking about your perceived pixels per degree.

Edited July 28, 2019 by SharpeXB

[Magic Zach]

ED don´t do Smart Scaling not because they don´t like, they don´t do because Waggs said the actual engine don´t support

The 2.5 engine? Nine said it was causing more problems than it was worth, but scaling was supported before, for a little time.

[SharpeXB]

but scaling was supported before, for a little time.

That was Model Enlargement. Different than scaling. It was only introduced briefly in the stable version and was mostly present in the Open Beta. It didn’t work well and so it was removed.

[nighthawk2174]

Even so it was better than nothing imo

[SharpeXB]

Even so it was better than nothing imo

Tons of threads about it here to search. It was actually worse than nothing. Primarily because it made distant aircraft so visible that using radar was pointless.

[Fri13]

https://why485.itch.io/smart-scaling-demonstration

 

Didn't know there are so many different ways.

 

https://apps.dtic.mil/docs/citations/ADA414893

[Bogey Jammer]

https://why485.itch.io/smart-scaling-demonstration

 

Didn't know there are so many different ways.

 

https://apps.dtic.mil/docs/citations/ADA414893

 

 

Extremely interesting :book:

[SharpeXB]

https://why485.itch.io/smart-scaling-demonstration

 

Didn't know there are so many different ways.

 

https://apps.dtic.mil/docs/citations/ADA414893

 

As that link points out, this is why Serfoss scaling won’t product accurate results for a PC sim game like DCS:

“The main drawback with this method is that it was finely tuned for a specific projector system and dome. It does not take into account either variable resolution, or a variable field of view”

Serfoss isn’t taking into account the varied FOV, display sizes and resolution that PC players use. So all the scaling calculations are flawed for that use. It would only be valid on a pro simulator with fixed hardware.

[Ignition]

Tons of threads about it here to search. It was actually worse than nothing. Primarily because it made distant aircraft so visible that using radar was pointless.

 

 

Because it was badly implemented.

[SharpeXB]

Because it was badly implemented.

There is no good way to implement it. It’s all been discussed to death if you search here.

[nighthawk2174]

Yes it has issues but still though... the science in is paper is sound and it makes spoting far more realistic. The "issues" are largely due to the original scope of the paper, but the issues of changing FOV and different resolutions is fixed/improved upon with the different versions that you so conveniently forgot to mention.

Edited July 28, 2019 by nighthawk2174

[SharpeXB]

Yes it has issues but still though... the science in is paper is sound and it makes spoting far more realistic. The "issues" are largely due to the original scope of the paper, but the issues of changing FOV and different resolutions is fixed/improved upon with the different versions that you so conveniently forgot to mention.

The idea that you can enlarge a target to its correct angular size on a monitor is flawed because the size of the screen, it’s distance from the viewer and FOV can’t be accounted for.

 

Those other methods don't account for the size of the screen or viewing distance. Which is impossible to account for yet has the greatest effect on the size of objects you see on the screen.

 

Also it’s only scaling up the target and not it’s surrounding environment which makes it artificially stand out.

 

The video is only showing air to air targets where the awkward relative scale problem won’t reveal itself. Let’s see that with ground targets or aircraft parked on a carrier.

Edited July 28, 2019 by SharpeXB

[nighthawk2174]

Ok, the point of smart scaling is to allow for a realistic level of detail discretion of target cues (including actually seeing the target in the first place) within WVR ranges. Simulators tend not to do this well at all due to the way our eye works. A screen ≠ a human eye and what you see in person cannot be simulated on a screen without adjustments being made. It is here that the idea of scaling comes in. Either the way the targets are displayed needs to be changed or our screens need to be improved greatly to account for the level of detail the human eye sees. The former solution is the far more cost effective solution and is what the paper focused on.

 

The whole purpose of the paper was how to enlarge the aircraft so it did not "... appear unrealistic or give the pilot a false sense of how the target would look in real combat". To do this he first looked at quite a few investigations into spotting ranges, both for max detection and max range that target orientation can be determined. (sources in the doc). This included looking at hundreds of A/A engagements and looking at the ranges and what cued in the pilot to the target.

 

From these he continues on to point out how vital being able to see these cues are to target detection. Additionally he also goes on to explain why just a constant or even linear increase in target size is not viable. "...without success – the target seemed much too large at close ranges and not large enough at a distance." Hence why DCS's original attempt was poor, it was just a simple model enlargement, and as stated by the author would be unsatisfactory. "It seems possible that if the target could be magnified only minimally at close ranges yet magnified significantly (perhaps 2X or more) at a 3-mile distance, the desired cue visibility could be achieved."

 

Serfoss

From here he goes on to discuss how the curve he made was made at some length, resulting in:

 

R = range(ft)/1000

Factor = 1.0 + 0.09226 * R - 0.00148 * R2

 

This is the equation he made to produce the same spotting and target orientation results in a sim as was achieved in the 6+ studies referenced and 100's of A/A engagements.

Essentially the original non-linear equation developed by Sefross had a set of parameters it had which did not account for varying FOV (via zooming), and resoltuion. To account for this a magnification factor needs to be applied to account for different FOV's along with resolution.

Chihirobelmo:

M = Magnification Factor;

D = Distance(ft)/1000;

R = Resolution Length(ex:FHD will have 1920);

F = Horizontal FOV;

M = 1 + (0.09226*D-0.00148*D^2) * {1.22e-7*R^2-1.17e-3*R+2.8} * (0.02*F-0.4)

 

From here this was further developed:

 

Sanpat v1

 

S = Reference Serfoss factor

F = Horizontal FOV

H = Horizontal Resolution

Factor = ((S - 1) * (((F * 60) / H) - 1)) + 1

 

This method seeks to maintain the desired spotting ranges and ability to discern visual cues even as FOV and Resolution may change.

 

My equation is not about the size on the monitor of the objects. It is to maintain the pixel number require to convey information perceptible IRL by pilots as per Serfoss’ research. So if it 5px on 1080p it will be 5px on 720p. Of course 5px in 720p will be a lot bigger than in 1080p. 

It will make things bigger so the size will not accurate as it is not the point of smart scaling. The different in current BMS implementation and my suggestion is at higher resolution than 1600x1200, mine will be smaller, but it is still magnified if needed and will be bigger than IRL.

Serfoss’ curve is not linear and the size in arcmin is not the focus. The focus is for pilots to be able to tell object information such as aspect, attitude, etc, at the same level as IRL.

 

That's kind of the point though to make it stand out as much as it should irl and allow for the pilot to determine stuff like basic shape and aspect .

 

Next we get:

 

Sanpat v1 (Clamped)

 

This simple "clamps" the updated equation preventing large changes in sizing in some situations, which if detected will force the program to revert to the original Serfoss value instead of the larger value.

 

Finally we get:

 

Sanpat v2

 

S = Reference Serfoss factor

F = Horizontal FOV

H = Horizontal Resolution

R = Range where 6 means 6000ft

Factor = S * ((F * 60) / (2 * H))

If (F*60) / H > 2 and range < 6 then

dampeningFactor = range/6

F = 1 + ((F - 1) * dampeningFactor)

 

 

This way the new proposed formula can be used with the res and fov setting worse than 20/40, by dampening the curve within 1 nm. The curve will now start from 1 at 0 nm as is desired. From 1 nm, the object will still maintain the correct pixel number from Serfoss’ research. May be this a better compromise for the case of worse than 20/40 setup. 

The trade off and downside of this method is, while maintains constant detail level at a given range as per the purpose of Serfoss’ research, the size will varies. People with low resolution monitors will see bigger objects, etc. But for its purpose, this is correct as the better screen resolution the lesser the need for scaling.

 

 

 

The idea that you can enlarge a target to its correct angular size on a monitor is flawed because the size of the screen, it’s distance from the viewer and FOV can’t be accounted for.

 

Latter updates to Serfoss take this into account. See above

 

Also it’s only scaling up the target and not it’s surrounding environment which makes it artificially stand out.

This is kind of the point, to increase how much it "stands out" to increase detection range and the range you can get directional cues from the bandit to that of rlf vision.

 

The video is only showing air to air targets where the awkward relative scale problem won’t reveal itself. Let’s see that with ground targets or aircraft parked on a carrier.

 

https://why485.itch.io/smart-scaling-demonstration

 

Then go download the program for yourself and see the effects... Additionally have you considered that it may be possible to turn off the scaling for aircraft on the ground or ground targets?!?! Additionally the latter version of the original equation take zooming into effect, zoomed out you won't notice much of a difference and zoomed in it takes this into account and adjusts the scaling.

Edited July 29, 2019 by BIGNEWY
1.15 Discussions of other game companies products are forbidden.

[SharpeXB]

Ok, the point of smart scaling is ... snip TLDR

Scaling can’t take the size of the screen into account so the whole concept is invalid

[nighthawk2174]

wow just wow........ you asked me to explain so the least you could do is read it. as I explain why that statement is false.

[SharpeXB]

wow just wow........ you asked me to explain so the least you could do is read it. as I explain why that statement is false.

I can’t follow that whole text wall and it doesn’t matter because the whole concept is flawed.

Serfoss was using some specific simulator setup. A dome projector with a specific resolution and viewing distance. That’s the only way you would get consistent results.

A PC game like this has such a wide range of monitor sizes and viewing distances that scaling can’t account for. So the whole concept just won’t work here.