Dear ED - visibility

[nighthawk2174]

Again, if you'd actually read the paper and what I wrote you'd not come to that conclusion. How can it not be adjusted to account for differences in FOV and Resolution? As was done in the updated versions.

[KTFBGB]

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.

 

I find this thread interesting, but have no dog in this fight. I am utterly uneducated on this stuff. But wouldn’t the scaling not care what size monitor or resolution you have? It seems to me that it works on individual pixels like everything else. So the change would be proportional to your pixel size?

[lucky-hendrix]

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.

"Blah blah, I am right and you are wrong, and I can't be bothered to read the evidence or to think. "

 

That's basically what you said.

 

Please refrain to post

 

Sent from my VTR-L09 using Tapatalk

[lucky-hendrix]

Nighthawk and the paper are super clear, there is nothing complicated to understand, just simple math. Nothing prevents this from being implemented in-game, at least partially, which would at least be better than the current method.

 

Sent from my VTR-L09 using Tapatalk

[SharpeXB]

Again, if you'd actually read the paper and what I wrote you'd not come to that conclusion. How can it not be adjusted to account for differences in FOV and Resolution? As was done in the updated versions.

The website says this about Serfoss which is correct

“The main drawback with this method is that it was finely tuned for a specific projector system and dome.”

He’s using a specific resolution, size and distance from the screen in a pro/military simulator. That’s how he arrives with these scale factors. Without the same hardware those factors are invalid.

The same thing can’t be done for a PC game where every player has a different sized screen.

How do you determine the scaling factor?

You can’t arrive at any realistic factor unless you account for the screen size and viewing distance.

So all those calculations are invalid and the concept doesn’t make sense outside that specific simulator

 

Nighthawk and the paper are super clear, there is nothing complicated to understand, just simple math.

It’s not simple math, frankly I have no clue nor does anyone else, how those factors are arrived at. And without considering the screen size they’re meaningless.

Edited July 29, 2019 by SharpeXB

[nighthawk2174]

Yes, and this is a drawback to that original equation, but did you not read below that how the following people had made it so that the equation was adjusted (a "skew" added) to account for that?

Edited July 29, 2019 by nighthawk2174

[SharpeXB]

Yes, and this is a drawback to that original equation, but did you not read below that how the following people had made it so that the equation was adjusted (a "skew" added) to account for that?

How would these following people know what screen size anyone is using? Obviously they can’t.

So how can they determine the scale factor?

For example Serfoss used 2.0 at 14,000’. But that’s calibrated to a particular dome projector.

Chihirobelmo and Sanpat appear to be using Serfoss factor as a base but without the same projector it’s invalid.

How would you determine a suitable factor when you can’t know the screen size?

So the whole thing is just nonsense.

[nighthawk2174]

Why are you so focused on screen size, resolution matters far more here and it can be inputed into the equations as a variable. Additionally why would Chihirobelmo and Sanpat's math be invalid what is your proof that they did not take this into account? That the adjustments to the equations they made to ACCOUNT for different resolutions is invalid?

[Magic Zach]

I think DCS would benefit greatly from very slight scaling (not making aircraft look terrible) and increased contrast. Perhaps that increase of contrast will make spotting aircraft against the ground much more feasible, or at the very least, not make an aircraft seemingly vanish when it flies over any noise in the terrain.

[wilbur81]

Just my opinion/2 cents worth on spotting:

 

I don't have a big problem with spotting in DCS per se (it is pretty hard since I play in 4K on a 55 inch screen). My problem is that spotting isn't the same across resolutions. If everyone I played with/against in MP was also playing in 4K, we'd all be on the same page mostly (vision acuity aside :-)). But as it is now, the person playing at a higher resolution pays for it against opponents using 1080p or lower res. Before I upgraded to 4K, I personally thought that spotting was way too easy. At 4K, I was like, "Hey, Hollywood...where'd he go? And Hollywood said, 'Where'd WHO GOOOOO!?'" :-)

[Tippis]

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

No you wouldn't because, as I keep explaining to you, the difference you're imagining between now and then is almost completely in your head. Again, the detail level they offered back then is pretty much exactly the same as it is now or in many cases actually worse — hell, in the case of the TV, it's worse than what we had in 1993, never mind 2003.

 

To wit:

• 3840×2160 @ 60" — 73ppi (this is your precious TV)

• 3440×1440 @ 35" — 107ppi

• 2560×1440 @ 27" — 109ppi

• 1920×1080 @ 24" — 92ppi

• 1920×1440 @ 21" — 114ppi (this is 2003)

• 1024×720 @ 15" — 84ppi (this is 1993)

 

Now, give this argument a rest because if you continue to make this false claim in spite of everything that has been told you, you will only manage to demonstrate a deep-seated ignorance and absolute unwillingness to understand how reality works.

 

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.

Incorrect. Taking the FoV into account is actually a core part of it, as the calculations and previous explanations have shown. Once you have the correct size of the object, then yes, the user can choose how far away from their screen they want to sit but… well… that's their choice. If they want to make things worse for themselves, then why should the game stop them?

 

Coincidentally, since you're a fan of modern technology, you'd be interested to know that there are monitors that can actually report back how far away the user is…

 

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.

Wrong has has been proven conclusively many times now, and the study did indeed take this into account, but you refuse to read it so of course you'd claim that it didn't and hope that this was the case.

 

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.

No. The conclusion is that the difference between then and now that you hinge your entire argument does not exist, and is in fact not even relevant to the issue at hand do begin with since it can be trivially accounted for anyway. Perceived pixels is not the same thing as pixel density, so don't confuse the two.

 

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.

Good news: they can be accounted for, and it's very trivial to do so as has been shown every time you make this false claim.

 

How would these following people know what screen size anyone is using? Obviously they can’t.

It's a single windows API call. Not only can it be done — it's a single line of code to do so. I've been using “trivial” a lot here; I've done so for a reason.

 

So how can they determine the scale factor?

The science has already done so, and it and subsequent studies have fully explained how to extend those results for other cases. They arrived at their scale factor by… you know… scientifically testing it. Your refusal to read up on the topic does not make the results nonsense. At no part have you been able to actually demonstrate any flaws in either the methodology or the results, and until you actually start referring to the content studies (fair warning: this means you have to read them), you don't have a leg to stand on.

 

I find this thread interesting, but have no dog in this fight. I am utterly uneducated on this stuff. But wouldn’t the scaling not care what size monitor or resolution you have? It seems to me that it works on individual pixels like everything else. So the change would be proportional to your pixel size?

Depends on the end goal. If all you want to do is apply raw Serfoss (because tbh, it's good enough for most uses), then yes, monitor parameters are wholly irrelevant. Apply multiplier; be happy. :D

 

If you want to apply it in an environment where FoV adjustments are available, then it matters a bit more because you need to determine the “inherent FoV” of the screen — the point at which you have a 1:1 parity between a displayed object and its real size. Only at this particular FoV should the full multiplier apply, and as you zoom in or out, it should be adjusted to compensate.

 

Now, this could be done a couple of ways. You could just grandly declare a specific FoV factor for all monitors — say, one inch = 50 mils. On a 1920px wide, 92ppi monitor (meaning it's 20.8" wide), this would mean that the width of the screen would cover 1043 mils, or just under 60°. So on that monitor, that would be considered 1x zoom, and all zoom-compensation scaling factors would be calculated form there.

 

Alternatively, you could figure out a likely actual FoV for the screen: you know its size, and you hope that people sit at arm's-length, and again grandly declare what that arm's length is. Now, that same 20.8"-wide monitor is assumed to sit at 30" (because that's how long the programmer's arms are for some reason) away from the user, and thus the screen takes up 693 mils (~40°) so that your 1x zoom level.

 

Or just could just grandly declare that all users should bring out their protractors and ensure that at all times, the monitor takes up 40° of their field of view, or else!… and proceed from there. Or just have a settings box where the user can input their distance to the screen. The options are (not quite) endless.

[nighthawk2174]

inb4 he says TLDR :music_whistling:

[nighthawk2174]

So? what precludes a game from having a good spoting system after all a sim is still a game.

[KTFBGB]

No you wouldn't because, as I keep explaining to you, the difference you're imagining between now and then is almost completely in your head. Again, the detail level they offered back then is pretty much exactly the same as it is now or in many cases actually worse — hell, in the case of the TV, it's worse than what we had in 1993, never mind 2003.

 

To wit:

• 3840×2160 @ 60" — 73ppi (this is your precious TV)

• 3440×1440 @ 35" — 107ppi

• 2560×1440 @ 27" — 109ppi

• 1920×1080 @ 24" — 92ppi

• 1920×1440 @ 21" — 114ppi (this is 2003)

• 1024×720 @ 15" — 84ppi (this is 1993)

 

Now, give this argument a rest because if you continue to make this false claim in spite of everything that has been told you, you will only manage to demonstrate a deep-seated ignorance and absolute unwillingness to understand how reality works.

 

Incorrect. Taking the FoV into account is actually a core part of it, as the calculations and previous explanations have shown. Once you have the correct size of the object, then yes, the user can choose how far away from their screen they want to sit but… well… that's their choice. If they want to make things worse for themselves, then why should the game stop them?

 

Coincidentally, since you're a fan of modern technology, you'd be interested to know that there are monitors that can actually report back how far away the user is…

 

Wrong has has been proven conclusively many times now, and the study did indeed take this into account, but you refuse to read it so of course you'd claim that it didn't and hope that this was the case.

 

No. The conclusion is that the difference between then and now that you hinge your entire argument does not exist, and is in fact not even relevant to the issue at hand do begin with since it can be trivially accounted for anyway. Perceived pixels is not the same thing as pixel density, so don't confuse the two.

 

Good news: they can be accounted for, and it's very trivial to do so as has been shown every time you make this false claim.

 

It's a single windows API call. Not only can it be done — it's a single line of code to do so. I've been using “trivial” a lot here; I've done so for a reason.

 

The science has already done so, and it and subsequent studies have fully explained how to extend those results for other cases. They arrived at their scale factor by… you know… scientifically testing it. Your refusal to read up on the topic does not make the results nonsense. At no part have you been able to actually demonstrate any flaws in either the methodology or the results, and until you actually start referring to the content studies (fair warning: this means you have to read them), you don't have a leg to stand on.

 

 

Depends on the end goal. If all you want to do is apply raw Serfoss (because tbh, it's good enough for most uses), then yes, monitor parameters are wholly irrelevant. Apply multiplier; be happy. :D

 

If you want to apply it in an environment where FoV adjustments are available, then it matters a bit more because you need to determine the “inherent FoV” of the screen — the point at which you have a 1:1 parity between a displayed object and its real size. Only at this particular FoV should the full multiplier apply, and as you zoom in or out, it should be adjusted to compensate.

 

Now, this could be done a couple of ways. You could just grandly declare a specific FoV factor for all monitors — say, one inch = 50 mils. On a 1920px wide, 92ppi monitor (meaning it's 20.8" wide), this would mean that the width of the screen would cover 1043 mils, or just under 60°. So on that monitor, that would be considered 1x zoom, and all zoom-compensation scaling factors would be calculated form there.

 

Alternatively, you could figure out a likely actual FoV for the screen: you know its size, and you hope that people sit at arm's-length, and again grandly declare what that arm's length is. Now, that same 20.8"-wide monitor is assumed to sit at 30" (because that's how long the programmer's arms are for some reason) away from the user, and thus the screen takes up 693 mils (~40°) so that your 1x zoom level.

 

Or just could just grandly declare that all users should bring out their protractors and ensure that at all times, the monitor takes up 40° of their field of view, or else!… and proceed from there. Or just have a settings box where the user can input their distance to the screen. The options are (not quite) endless.

 

That’s a lot of complexity there. Couldn’t the program just detect the native screen size of the monitor and apply it that way?

[Tippis]

That’s a straight up game. Not a sim. I personally hope that DCS never, ever happens to even remotely resemble that GAME. There are plenty of games out there. If This smart scaling makes DCS look that game does for spotting, which I have played plenty in the past, then my vote is absolutely no. Keep the arcade stuff out of DCS.

It might comfort you to know that the whole scaling study, and its subsequent improvements, were intended to improve the realism of perception in actual military simulation settings. That they're also applicable to games is a happy by-product.

 

 

That’s a lot of complexity there. Couldn’t the program just detect the native screen size of the monitor and apply it that way?

Oh, sure, that part is trivial, but you still need to decide what you want to do with that information. The crux of the matter is figuring out FoV, and that depends on a third distance that can't easily be determined (short of making distance measuring USB dongles a hardware requirement to run the game). They either have to be manually input by the user, or assumed by the developer based on what logic they prefer to lean on. But it's not all that complex — it's a lot of options to choose from, but none of them are difficult in and of themselves.

Edited July 29, 2019 by Tippis

[Magic Zach]

That’s a straight up game. Not a sim. I personally hope that DCS never, ever happens to even remotely resemble that GAME. There are plenty of games out there. If This smart scaling makes DCS look that game does for spotting, which I have played plenty in the past, then my vote is absolutely no. Keep the arcade stuff out of DCS.

 

We are talking about just the spotting systems, not graphics, GUI, flight models and performance, or gameplay elements.

[Ignition]

That’s a straight up game. Not a sim. I personally hope that DCS never, ever happens to even remotely resemble that GAME. There are plenty of games out there. If This smart scaling makes DCS look that game does for spotting, which I have played plenty in the past, then my vote is absolutely no. Keep the arcade stuff out of DCS.

 

 

I hope you're not using the zoom then.

[Fynek]

For having sat often in a real ULM, I can say for sure that the spotting implementation in DCS could be improved because IRL spotting is definitely easier than what it is in the game in certain conditions.

 

For instance when zooming at max in DCS, I can see contacts better than how i can see them in real life against the sky from a distance ( but i see better IRL a white aircraft against the green grass fields) , however if i unzoom a bit, at some point they suddenly completely disappear in DCS instead of this being more progressive.

 

This should not happen because in real, when you focus on something and clearly spot it, once you stop focusing on it, you still somewhat see it as your eyes/brain "know" it is there so you can still see that small detail or dot moving in the background.

 

But I 've never been in a fast jet, I can imagine that trying to spot planes that have camouflage flying super fast or against the terrain must be very, very difficult even for a trained eye.

 

ED had to balance things but if i can wish for an improvement , it would be the progressive aspect of spotting, the contacts should not totally disappear if you reduce your zoom once you've already spotted them, unless they suddenly go against the terrain and have camouflage or you get blinded by the sun or any other cause that would make sense for suddenly losing sight IMHO like light reflections or being disoriented or suddenly taking too many G's.

Edited July 29, 2019 by Fynek

[KTFBGB]

I hope you're not using the zoom then.

 

It’s not actually zoom. It’s a narrowed field of view to mimic focusing on a small area. Lots of posts and videos explaining this.

[Mars Exulte]

I'm all for improvements in spotting, particularly via contrast adjustment, etc. I disagree with scaling persay, as higher resolutions and larger screens become common place (the average size has nearly doubled in recent years) the need for artificial enlargement to see non pixelated detail is passing away.

 

As I observed last night with my own experience, it's not that the objects are not rendering, they tend to ''vanish'' into the background (they're still there, just blending in). AA and AF probably (almost certainly) aggravate that tendency. Point is, they're there to be seen, just the contrast, depending on aspect and blah blah makes them difficult to actually pick out.

 

And as has been noted before, another issue is the average dogfight in DCS takes place at several times the speed and over several times the area. Meaning a target can go from visible to distant or drastically change position far more quickly than in your average WWII sim (this affects DCS props, too, but to a lesser extent than it does jets).

 

I will also say again regarding 'zoom and fov' you are not truly ''zooming''. The 'default fov' most people use is very zoomed OUT in comparison to true scale. So, during combat to reduce your FoV a bit, effectively ''focusing your attention'' on a narrower area, is NOT a ''mersion breaking cheat'' it is a fix for a self inflicted ''unreality''.

 

@Sharpe

You're a bit willfully obtuse sometimes regarding this topic. As far as your res and monitor width, res is easily pulled ingame, and most the time Windows recognises monitor model numbers, which presumably also contain such information, or easily could. You could also simply tell it.

 

99.9% of people are sitting 24-36'' from their 24-27'' screens, so ''distance from the screen'' is not a legitimate concern in most cases. The odd person may game on a 75'' and sit a bit further back, but they're def the exception.

[SharpeXB]

Why are you so focused on screen size,

Because screen size is exactly what determines how big anything you see is. The same object on a 24” screen will appear 2x the size on a 50” one!

It’s not rocket science.

If scaling claims to be so accurate then in order to show you the correct size for a target, it would have to consider the screen size. You could have scaling otherwise but it wouldn’t be accurate at all.

 

I'm

You're a bit willfully obtuse sometimes regarding this topic. As far as your res and monitor width, res is easily pulled ingame, and most the time Windows recognises monitor model numbers, which presumably also contain such information, or easily could. You could also simply tell it.

 

99.9% of people are sitting 24-36'' from their 24-27'' screens, so ''distance from the screen'' is not a legitimate concern in most cases. The odd person may game on a 75'' and sit a bit further back, but they're def the exception.

Yeah but the Serfoss values everyone seems to worship are only valid for a specific size display.

Who says a 2x factor at 14,000’ is accurate? Why not 2.5x? Who knows?

So if you had scaling in DCS it would be just arcade game graphics nonsense and not accurate.

Edited July 29, 2019 by SharpeXB

[nighthawk2174]

Ok dude, did you read anything me or Tipis have said because I don't think you have... if you had you'd not ask that question. Or if you'd actually read the original paper either :doh:

 

May I refer you back too:

 

https://forums.eagle.ru/showpost.php?p=3992128&postcount=97

 

https://forums.eagle.ru/showpost.php?p=3992196&postcount=115

 

 

or better yet try out this app someone made to show the effects of smart scalling:

 

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

Edited July 29, 2019 by nighthawk2174

[SharpeXB]

No you wouldn't because, as I keep explaining to you,

And I’m not reading it because it’s all gibberish.

 

There’s no way any of these scaling factors know how big your screen is or where you’re viewing it from. So they’re all junk. And the conclusions are junk. The Serfoss study was only valid on specific display hardware. Citing it for DCS is pointless

[nighthawk2174]

If you actually read our stuff you'd realize all of what you said is utter nonsene. + what about the other scaling options such as sanpat? Additionally as Tillias said and Zuckov said you can make basic assumptions about where people are sitting and get all the hardware info via windows API's or via settings.

Edited July 29, 2019 by nighthawk2174

[David OC]

and as usual we come full circle about all this.:thumbup:

 

You think ED's engineers haven't banged their own heads together over this?;)

 

Even they had problems with trying to make it better and across the board in that demonstration.

 

Quote

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

 

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.

 

at higher resolution than 1600x1200, mine will be smaller, but it is still magnified if needed and will be bigger than IRL.

End Quote

 

The rest is just tac on and tac on in different situations to get it better. Chasing your own tail is how I see it.

 

Still not real fair across the board in multiplayer. Smaller screens could be still much better in SOME situations etc.

 

For IRL training this is would be fine and can be fully adjusted to suit the D level trainer res. For 1000's of users...bit harder I guess.

 

The sweet spot for me was going 1440P @ 27-inch. Good visuals at max graphics, great smooth performance and can still spot things. Seems close to real.

Also many real competitive players will even adjust the screen color for a visual advantage.

 

Just my opinion/2 cents worth on spotting:

 

I don't have a big problem with spotting in DCS per se (it is pretty hard since I play in 4K on a 55 inch screen). My problem is that spotting isn't the same across resolutions. If everyone I played with/against in MP was also playing in 4K, we'd all be on the same page mostly (vision acuity aside :-)). But as it is now, the person playing at a higher resolution pays for it against opponents using 1080p or lower res. Before I upgraded to 4K, I personally thought that spotting was way too easy. At 4K, I was like, "Hey, Hollywood...where'd he go? And Hollywood said, 'Where'd WHO GOOOOO!?'" :-)

Edited July 29, 2019 by David OC