resolved New spotting is great / New spotting is overcooked feedback thread

[James DeSouza]

13 hours ago, PeevishMonkey said:

Look at long wires far from you, angular resolution, calculated by their diameter and distance, may be 10 times smaller then "standart" angular resolution of the human eye. But it still visible.

They aren't.  If you can see the mountings for the wires (such as pylons) your mind fills in the gap and you might believe you can see them but you can't actually see them.  It's why chopper pilots kill themselves on power lines all the time, or why tank commanders get scissored between power cabled and their hatch.

 

It's the reason why you are supposed to fly over the pylon and not the line, you can actually see the pylon far enough away to be sure to overfly it.

[Tippis]

…hence why simulating perception cannot rely on naive trigonometry: because it's not just what the hardware in the eyes are capable of, but also what the wetware processing behind it can do. It's also why you'll achieve better and more realistic results with proxy objects and non-linear scaling curves than just making a simplistic mapping of “size X at distance Y = Z pixels”.

[Simultaneous]

Hi BN & team,

Loving 2.9 and working well....except for this spotting dot renderer.

Whereas I could understand some folks wanted it, I didn't, and furthermore it is a real immersion breaker for me. Huge dark squares all over the sky for aircraft 40 miles away, which disappear when you get closer. As a spotting tool I can't even imagine it works well as when you get closer they disappear and the aircraft become much harder to spot at closer distances. You are distracted by aircraft much further away sticking out very unrealistically.

I wouldn't mind if it could be disabled, but the autoexec toggle doesn't work for me, so I am stuck with this - I assume minority driven - cheat code.

Please remove or make it so I can disable it. My specs are below and spotting was very realistic before this for me.

 

Thanks and kind regards,

 

Simultaneous.

[Davee]

I think it is fantastic. For years I have wanted improvement in spotting both distant but primarily against the ground. I fly WWII primarily. This upgrade is super for spotting now both in 2d @ 4k and in VR.

Thanks ED.

 

[Tippis]

37 minutes ago, Simultaneous said:

Whereas I could understand some folks wanted it, I didn't, and furthermore it is a real immersion breaker for me. Huge dark squares all over the sky for aircraft 40 miles away, which disappear when you get closer. As a spotting tool I can't even imagine it works well as when you get closer they disappear and the aircraft become much harder to spot at closer distances. You are distracted by aircraft much further away sticking out very unrealistically.

What are your graphics setting and at what point do they disappear and reappaer?

Are you running any kind of AA or resolution scaling?

This is what it looks like at 10nm — nothing huge, nothing particularly dark (they appear a bit sooner but very faint and at 10nm is where they start to stick out against the background properly). I'm not sure exactly what FoV the F12 camera is supposed to have, but if it's in the 70° range, the 1.5–2.5px width of the spotting dots is pretty much spot on, and the range is about right as far as how faded in they should be.

[PeevishMonkey]

7 часов назад, Simultaneous сказал:

Huge dark squares all over the sky for aircraft 40 miles away

What kind of aircraft do you see 40 miles away?

6 часов назад, Tippis сказал:

Are you running any kind of AA or resolution scaling?

+

there are 3 types of AA

DLAA+DLSS (quality):

DLAA only:

MSAA x4:

[James DeSouza]

17 hours ago, Simultaneous said:

I wouldn't mind if it could be disabled, but the autoexec toggle doesn't work for me, so I am stuck with this - I assume minority driven - cheat code.

"Hey I am going to spend several hundred dollars on modules and several thousand dollars on a computer and flight sim hardware for this realistic flight sim, but being able to spot planes at realistic distances is bad so I want them to be black boxes the size of small buildings."

(Though I say small buildings, you can still see them 30-40 miles out pretty clearly while the aircraft is nose on to you, might be getting in to large building territory at that point).

If you find some way to disable it that works please let me know.

[Tippis]

49 minutes ago, James DeSouza said:

"Hey I am going to spend several hundred dollars on modules and several thousand dollars on a computer and flight sim hardware for this realistic flight sim, but being able to spot planes at realistic distances is bad so I want them to be black boxes the size of small buildings."

(Though I say small buildings, you can still see them 30-40 miles out pretty clearly while the aircraft is nose on to you, might be getting in to large building territory at that point).

If you find some way to disable it that works please let me know.

What are your graphics settings?

As shown above, they're not black boxes, nor the size of small buildings, nor visible out to 40 miles. This sounds more like you're running some kind of overlay.

e: In fact, I changed my test setup to really fill the sky with them out to 50nm, and the result was… not that.

Edited November 2, 2023 by Tippis

[MarkP]

1 hour ago, Tippis said:

What are your graphics settings?

As shown above, they're not black boxes, nor the size of small buildings, nor visible out to 40 miles. This sounds more like you're running some kind of overlay.

e: In fact, I changed my test setup to really fill the sky with them out to 50nm, and the result was… not that.

 

i dont know man, youre the only person in this 10 page thread that is defending the new spotting dots. Most people dont like it (including myself). Why still bothering?

[Tippis]

5 minutes ago, MarkP said:

i dont know man, youre the only person in this 10 page thread that is defending the new spotting dots.

Nah.

And since the purpose of the thread is providing feedback, I'm doing just that. Why bother? Because some tweaks are still needed – visibility is still a bit high and starts a bit too far out, for instance.

Also out of morbid curiosity about where some of the seeming exaggerations (and outright misinformed claims) are coming from.

[Why485]

2 hours ago, James DeSouza said:

"Hey I am going to spend several hundred dollars on modules and several thousand dollars on a computer and flight sim hardware for this realistic flight sim, but being able to spot planes at realistic distances is bad so I want them to be black boxes the size of small buildings."

(Though I say small buildings, you can still see them 30-40 miles out pretty clearly while the aircraft is nose on to you, might be getting in to large building territory at that point).

If you find some way to disable it that works please let me know.

Please stop with this strawmanning. It's incredibly counter-productive to this entire conversation. 2.9 has issues, and it's important to point them out, but characterizing anybody who doesn't see the same thing you do as a cartoon character of your creation is not helping.

Everybody involved wants what's most realistic. This is a complex problem that is only made more difficult by the wide variety of hardware involved. Not everybody is seeing the same things, and that's part of the problem here. It's important to show screenshots of the problems and what hardware you're viewing it on because the biggest issues right now are these discrepancies, especially when it comes to VR.

Edited November 2, 2023 by Why485

[Buzz313th]

Seems to me that this is simply a VR vs Flat screen issue.  I run the sim both in VR and on a 1440 flat screen.  IMHO, In VR, the spotting dots are too large, have too much contrast with the background and when the object gets close enough that the model renders, object visibility drops significantly as relative size drops and contrast drops.   On a 1440 flat screen, I have too search for the spotting dots, will sometimes lose sight and have to visibly search again.  I do believe the contrast is too high.  When I get close enough to the object, the transition between the spot dot and model/textures is less apparent.

My humble opinion on realistic aircraft recognition..   Highly dependent on  environmental and lighting conditions.  I have spotted small GA planes from a traffic advisory out to ~ 5-7 miles with 20 miles vis, and sometimes couldn't see them within 1 mile in the same visibility but different lighting conditions.   Large aircraft on a crisp clear day, ~20 miles without a reflection hit, 30-50 with a reflection hit.  With Trails, whatever current visibility is.  Looking for a small aircraft below and against a city, good luck.  Looking high against a blue sky, you can see everything.  Spotting aircraft at night with Pos and AntiCol lights on, forever, just form lights, within a mile.  For DCS to nail this variability and then make it right for flat screen AND VR users would be pretty impressive. I would expect a FF F15C before we get 90% realistic and accurate visibility model for spotting distant objects. 

Spotting a small GA aircraft beyond 3-4 miles close to the relative horizon (Where it gets neutral) is not easy and if you take your eyes off of them, then it's just as hard to find them again.  Spotting a medium size fighter in average lighting and contrast conditions with appropriate camo on an average skyscape is just as hard if not harder as finding that small GA aircraft at 3-4 miles.  There is a reason why the airforces require flight candidates to have perfect vision.... and if you want realistic spotting, be prepared for lots of squinting at your monitors. 

Edited November 2, 2023 by Buzz313th

[Tippis]

8 minutes ago, Buzz313th said:

Seems to me that this is simply a VR vs Flat screen issue.  I run the sim both in VR and on a 1440 flat screen.  IMHO, In VR, the spotting dots are too large, have too much contrast with the background and when the object gets close enough that the model renders, object visibility drops significantly as relative size drops and contrast drops.   On a 1440 flat screen, I have too search for the spotting dots, will sometimes lose sight and have to visibly search again.  I do believe the contrast is too high.  When I get close enough to the object, the transition between the rendered model and textures is less apparent.

I've seen conflicting reports on it, and my headset is just too old and jank to really give a good indication, but do you get the effect in VR where the parallax and distance is off? Where the dot appears to be at infinity when it's just about to switch over to full 3D model, or vice versa, where it appears closer than the 3D model is when the switch happens?

It certainly seems like they're going to have to tweak this separate for the pancake and the VR render paths, not just to get the size and colour right, but also to try to hunt down that issue. Of course, since I've heard different people say different things about the issue — which way the error goes, or whether it even exists to begin with — it's going to be… a fun one… to hunt down.

[Buzz313th]

5 minutes ago, Tippis said:

do you get the effect in VR where the parallax and distance is off?

Not at all.  I mean, the spotting dot's should be at a focal distance of infinity with the rest of the background in VR and it feels that way.  Best example you could use to check this would be to have a friend take off from an airport at "Spotting dot distance" and if the airport and spotting dot feel like they are at the same distance and at infinity, then we are good.  

BTW, I'm on a Quest 3 for hardware comparison and a Sony 52" TV for flat screen.

[Tzigy]

My 2 cents; I fly WW2 95% of the time, PvP, PvE. 4k LG OLED48. 

I like the new dots/render. Its not perfect but in my case (large 4k monitor), definitively going in the right direction

Edited November 3, 2023 by Tzigy

[SharpeXB]

2 hours ago, Buzz313th said:

For DCS to nail this variability and then make it right for flat screen AND VR users would be pretty impressive.

Easy. Just stop having these dots added to the game. What you would see in the sim then would pretty well nail what you’re describing above. 
 

2 hours ago, Buzz313th said:

Spotting a small GA aircraft beyond 3-4 miles close to the relative horizon (Where it gets neutral) is not easy and if you take your eyes off of them, then it's just as hard to find them again.

If DCS was realistic like this, players would just howl for their dots back ED can solve this if they stop trying to tailor “realism” to what players think it should be vs what it is. 
And if people don’t like it there’s always dot labels. 

[Tippis]

10 minutes ago, SharpeXB said:

Easy. Just stop having these dots added to the game

That would not achieve the objective, no. Nor is it easy. That's why this thread exists.

10 minutes ago, SharpeXB said:

What you would see in the sim then would pretty well nail what you’re describing above.

No. If they stopped having these dots, we'd be back to how things were before and that was nowhere near what he's describing. The lighting aspects in particular were ridiculously off.

11 minutes ago, SharpeXB said:

ED can solve this if they stop trying to tailor “realism” to what players think it should be vs what it is. 

That is pretty much what they're doing now. Hence why super-visibility at long ranges and sub-visibility at shorter ranges are both going away. The problem is that what you think it should be is not where this will end up, and you've been arguing against any change against your preferences since forever, using all kinds of contradictory pretzel logic to try to make a claim that your preference was more realistic. Now they're seemingly going for something that's actually realistic, which will rob you of the advantages you had, and that's too bad… well for you. It's very good or the game.

[Buzz313th]

The problem is this...

IRL at rough maximum spotting distance in optimum atmospheric and lighting conditions, lets say a small fighter at ~8-10 miles.  The size of the object you would see is relatively smaller than one pixel on a device running a native resolution of 1440 and maybe roughly the correct size of one pixel on a 4k screen.  But then you have the variable distance that players sit from their monitors.     VR poses another problem..  You may be using a per eye resolution of 2k x 2k, but you are also looking at a much smaller portion of the total raster so you are in essence much closer to the screen than a 2d player so your POV is to see larger pixels.  The best ED can do with the size, is to draw one pixel at what they believe is their best maximum distance for a particular size object in particular atmospheric and lighting conditions.  But, where I believe the most progress can be made to getting a realistic representation of distant object visibility is by putting  more work in to the color contrast that these single pixels represent when seen by the player against a particular background, at a particular altitude, in particular atmospheric and lighting conditions.  At greater distances color contrast would be less, then increase as distance decreases.  This same method can also be used to represent greater visibility during more optimal atmospheric and or lighting conditions.

Edited November 2, 2023 by Buzz313th

[SharpeXB]

What should have been done:

• Improved “spotting dots” added (WIP):
  • Increased decreased dot size at high low resolutions
  • Zoom is not affecting dots size

 Solved

[rob10]

3 hours ago, MarkP said:

i dont know man, youre the only person in this 10 page thread that is defending the new spotting dots. Most people dont like it (including myself). Why still bothering?

I'm not sure if they're all in this thread or mostly in other threads, but there are lots of people (including myself) defending the new dots.  The main difference seems to be that most of them realize that they aren't working for everyone and would like to see them tweaked to be better for those who don't like the new ones, versus a large number of those who don't like them just wanting it to go back to the way it was and not acknowledging that it is a significant improvement for many.

Edited November 2, 2023 by rob10

[SharpeXB]

3 minutes ago, Buzz313th said:

IRL at rough maximum spotting distance in optimum atmospheric and lighting conditions, lets say a small fighter at ~8-10 miles.

You can see aircraft this far away in 2.8 but it’s very hard. Kinda like it would be in reality. In good conditions and if it’s moving you can pick up a fighter maybe 14 miles away in 2.8. 

[Buzz313th]

24 minutes ago, SharpeXB said:

You can see aircraft this far away in 2.8 but it’s very hard. Kinda like it would be in reality. In good conditions and if it’s moving you can pick up a fighter maybe 14 miles away in 2.8. 

14 miles near the horizon, no way.  Against a clear and crisp blue sky, no contrails, no engine smoke at least 40 degrees above the horizon, maybe.  But once you stop staring at them, you will lose em.  Above ~15 miles in perfect conditions it needs to be a large aircraft to be consistently spotted and then it's easy to lose sight.  Against the ground, never.  Looking horizontally anywhere below ~FL18 and your looking through lots of atmospheric haze, this causes a condition of low contrast.

One last thing..  Optimal atmospheric conditions for maximum visibility doesn't happen often.  In certain parts of the globe and during certain seasons, more often and in others it's rare.  So for the most part, don't use maximum visibility distance in optimal conditions as the status quo for an example to follow.

Edited November 2, 2023 by Buzz313th

[Tippis]

32 minutes ago, Buzz313th said:

The problem is that...

IRL at rough maximum spotting distance in optimum atmospheric and lighting conditions, lets say a small fighter at ~8-10 miles.  The size of the object you would see is relatively smaller than one pixel on a device running a native resolution of 1440 and maybe roughly the correct size of one pixel on a 4k screen.

Weeeeeeell…

Let's take the F-16 as a “small fighter:” a wingspan of 9.5m, at 8–10nmi (14,816–18,520m) comes out as a 0.51–0.64 mils at a head-on aspect. So about twice the lower limit of a good eye's angular resolution.

If this was drawn as a single pixel on a 2560×1440 display, your FoV would be in the 74–95° region. So even a pretty modest zoom level would increase them to more than 1px.
On a 4k display with the same FoV range, they'd obviously come out 1.5× larger.

And remember, the default zoom in many cockpits is at the lower end of that range: 75–85° or so. So For the most part, without even touching the zoom axis, that small fighter would already be more than 1px wide in many cases. But that's just the rendering part. As you point out, that then runs face first into the issue of how close you sit to your monitor in relation to its size and pixel density, and whether or not you can even see individual pixels at that distance (and we haven't even dug into how the dots are shaded over range to remove contrast, and what happens when you slap AA on top of that). It doesn't seem like much of a stretch of the imagination that your average flight-simmer has enough junk in front of them to require the monitor to sit even farther away than every-day desktop use, so the single pixel that represents a just-above-visible target then becomes invisible to player's eye.

tl;dr: the issue these days probably isn't so much the monitors — it's more a factor of the geometry of the desktop setup.

 

30 minutes ago, SharpeXB said:

Solved

• Fails to solve the problem that aircraft are seen too early and far away, and that this is hardware- and settings-dependent.
• Fails to solve the problem that are at the same time both too visible and not visible enough at upper WVR range, and that this contradiction is also hardware- and settings-dependent.
• Fails to solve the problem of creating an equitable transition from unseeable to minimal model rendering size.
• Fails to offer any functionality for simulating perception.

The only thing your change “solves” is that some people that aren't you lose their advantage that you don't have whereas you get to keep an advantage that they don't have.

 

Edited November 2, 2023 by Tippis

[Buzz313th]

12 minutes ago, Tippis said:

Let's take the F-16 as a “small fighter:” a wingspan of 9.5m, at 8–10nmi (14,816–18,520m) comes out as a 0.51–0.64 mils at a head-on aspect. So about twice the lower limit of a good eye's angular resolution.

If this was drawn as a single pixel on a 2560×1440 display, your FoV would be in the 74–95° region. So even a pretty modest zoom level would increase them to more than 1px.
On a 4k display with the same FoV range, they'd obviously come out 1.5× larger.

I'm impressed you figured that all out...  Kudos..  But at head on, the aircraft, your F16 example might be rendered at 1.5 pixels wide due to the wingspan, the height would be a fraction of a pixel at the fuselage and much smaller for the wing height.  How do you represent that when all you can render is one pixel, or two pixels?  If you use one pixel, then the aircraft is more spottable because the dimensions are square 1x1 pixel, versus 1.5 x 0.2 pixels.  There is more surface area to the players POV.   

 

Its a complex problem to solve especially with so many user variables.

 

The best solution IMHO is to er on the side of being "Fair" instead of ultra realistic.    Use one pixel at max visible distance under perfect conditions.  Then use color contrast to make the pixel scale in user visibility based on all the other variables discussed.  Regarding color contrast, I wonder if scaling the spot dot transparency would work, or if it would look weird.

Edited November 2, 2023 by Buzz313th

[SharpeXB]

26 minutes ago, Buzz313th said:

14 miles near the horizon, no way.

Oh I agree. The point being even in 2.8 it’s possible to stretch the realistic detection range. It doesn’t need to be exaggerated like it is in 2.9.