Eagle Dynamics: Request a rethink on spotting mechanics.

[Waxer]

Thank you for the information / discussion folks. Okay, so basically Eagle Dynamics are aware but are doing nothing about the spotting mechanic because it is realistic. Seems reasonable. But what is therefore unreasonable - as pointed out by the last guy - is someone on 1080p getting a spotting buff due to unrealistic sprite blocks appearing early on their low resolution screen.

 

And answering the OP asking about me running in a lower resolution, yes that does make spotting easier. My 5K monitor is 5120x2880 and scaling that to 1080p would look horrible. But scaling to 1440p does not look too bad as 2560x1440 is an exact 1/2 fraction of my monitor scale on each axis. Also there are some other settings that help: turn off anti aliasing totally for example. On such a high resolutions screen you don't need it anyways. And set draw distance from High to Ultra.

 

Much as I love my 5K, 60Hz 27" monitor, I was looking at a new 1440p, 120Hz 49" monitor as a complement. I'd keep the 5K for another productivity / photography focused PC.

Edited May 16, 2019 by Milou

[Schmidtfire]

Thousands of DCS players cannot be wrong. Seems like new threads about spotting pops up every week here or on Hoggit.

 

In theory ED’s approach might be correct. But It does not factor in current hardware limitations.

Other issues connected to this are color/contrast were units blend with the sky, strange behaviour with units popping in/out during zoom etc.

 

Hope to hear a response from ED soon on this issue. If they want WWII to grow, proper spotting is essential.

[philstyle]

It is by no means a solution, but have you tried lowering the resolution? Would that improve spotting? I am genuinely interested if you would see the difference... Saying that, I could try it myself. I always just assumed it's my fault when I can't spot bandits, not DCS' (3440x1440 user here).

 

 

Lowering resolution (assuming physical monitor size and other graphcis options remain the same) has been empirically shown to increase the range at which contacts become first visible in DCS. The tests have been done and published.

[Xilon_x]

for example in this case I can't see the helicopter but instead I can see its shadow on the ground and it is only 5km away

 

 

https://forums.eagle.ru/showthread.php?t=221291

 

 

try also to position yourself at 5km of distance and tell me if you can see a helicopter or plane flying at low altitude.

Edited May 16, 2019 by Xilon_x

[veenee]

Indeed by no means a solution.....

 

lol, aren't you a darling? ;-)

 

 

I think I will try running DCS in 2560x1080, last time I tried it didn't look too bad.

 

If it helps spotting other planes, it's worth it.

Edited May 16, 2019 by veenee

[Abburo]

I do subscribe to this complain as well. Spotting needs to be worked more.

I don't understand either what's not technically compliant to ensure this. The spotting dots that ED tried are only a kind of workaround.

 

However, this issue lead me learn new tactics, applicable only for dogfights which helps me predict better the foe trajectory. Even so it happens to loose the track... but that happens in reality too... right?

[GGTharos]

Yes it does. Sometimes for different reasons, but you're on the right track.

Edited May 16, 2019 by GGTharos

[The_Dan]

On 1080p i can't spot fighters (or bigger planes like AWACS) at 30nm. I usually can spot fighters at 10nm. This might be realistic. But i have to use the FOV zoom on it's highest level (like in the pictures and videos posted regarding this topic) and that's unrealistic as hell.

A RL fighter pilot doesn't have a switch in his cockipt to zoom his view back and forth and the imagination of such a thing is hilarious.

For me the problem is less the spotting distance per se, but the spotting with a "realistic" FOV. This does make dog fights much closer than 10nm unrealisticly difficult. Spotting is hard in real life but doesn't have this FOV constraints. So it is more difficult in DCS compared to RL because you have a tiny fast moving FOV that gets blocked on the whole monitor from a cockpit bar and nearly no chance to spot something if you don't know exactly where to look (i know, in RL it is difficult too without knowing where to look, but you can orientate by o'clock positions, that's nearly impossible, at least much harder, in DCS).

Edited May 16, 2019 by The_Dan

[Mars Exulte]

for example in this case I can't see the helicopter but instead I can see its shadow on the ground and it is only 5km away

 

 

https://forums.eagle.ru/showthread.php?t=221291

 

 

try also to position yourself at 5km of distance and tell me if you can see a helicopter or plane flying at low altitude.

 

Oh yeah, there's lots of flakey stuff like that. That's an issue with the 3d model itself, specifically the LoDs (what causes much of this, imo) rather than a ''mechanic'' it's a missing, incorrect, or outdated LoD. I believe newer stuff suffers from this less than older, outdated models, but it's not uncommon. Hopefully, as they make their beauty pass over the old units, this sort of stuff will become less and less common.

[GGTharos]

On 1080p i can't spot fighters (or bigger planes like AWACS) at 30nm. I usually can spot fighters at 10nm. This might be realistic. But i have to use the FOV zoom on it's highest level (like in the pictures and videos posted regarding this topic) and that's unrealistic as hell.

 

It is and it isn't ...

 

A RL fighter pilot doesn't have a switch in his cockipt to zoom his view back and forth and the imagination of such a thing is hilarious.

 

Because to spot that aircraft you'll need the tiny little center of the eye where things are highly focused. There's no useful way to simulate this on a monitor. You're going to lose peripheral OR spotting distance, you can't have really them both in a realistic manner.

 

This does make dog fights much closer than 10nm unrealisticly difficult.

 

Not really, they're fine. Could be better, yes, there are some things that could be fixed possibly right now to help (mainly contrast/LOD change issues IMHO), but overall it's not the end of the world. Also, like real pilots, you anticipate the target's trajectory so you know where to look for him.

 

(i know, in RL it is difficult too without knowing where to look, but you can orientate by o'clock positions, that's nearly impossible, at least much harder, in DCS).

 

Actually it works fine in DCS. And BTW, regardless of what you might have read about spotting fighters, according to real combat testimonies for the most modern of conflicts, an eagle pilot didn't spot a pair of MiGs until out 7-8nm, and ONLY because the TD box in the HuD showed him exactly where to look. It's an anecdote, yes, but there you have it.

 

Spotting with a practiced and planned visual scan is considered to be 7nm on average for a large, head-on fighter (think F-14/15/22, Su-27 etc). I can spot like that without zooming in.

[Tippis]

I do subscribe to this complain as well. Spotting needs to be worked more.

I don't understand either what's not technically compliant to ensure this. The spotting dots that ED tried are only a kind of workaround.

 

There are a couple of issues at play here.

 

The first is that a tangle of legacy code is putting stopping blocks in the way of realistic simulation of perception. Right now, the rendering pipeline that creates the on-screen visuals you're using to spot things is also used to determine what your sensors see — adjust the on-screen appearance to make things easier to see and sensors will also have an easier time seeing them.

 

This is why sensor effectiveness is inversely proportional to computer hardware capabilities: lower resolutions and less fancy post-processing makes it easier for radars and EO/IR tracking to find and separate targets from the environment if you play in VR than if you're on a 4k screen. This is obviously bonkers and needs fixing, but it would require an absolutely massive rewrite of some of the core components of the game to sort out.

 

The second is that DCS simply does not simulate any of the spotting stuff. It uses a laughably naive, purely mathematical model of determining perspective. That's accurate enough if you want to do trigonometry; it is wholly inaccurate and unrealistic if you want to simulate perception. And you do. Not only that, you want to simulate the perception of people with exceptional capabilities of perception, rather than just your average wonky-eye-haver.

 

As someone pointed out, zoom the one concession to how vision works since it ever so slightly manages to let the game account for the differences in acuity in different parts of the eye, but as also pointed out, it's ridiculously over-modelled and not actually representative of the limits of that acuity. As much as people like to ignorantly throw mud on BMS' smart scaling, it actually offers a level of simulation that DCS simply cannot achieve at the moment by modelling some of the cognitive processes going on behind a pilot's eyes. Not all of them, granted, because it does not account for background all that much, but some simulation is still more than none at all. The brain has had a looong time to learn to latch on to glints, movements, discolourations and other unexpected contrasts, all of which means that the trigonometrical calculation of perspective is almost entirely besides the point. The “right” size of a rendered object is simply not the “realistic” size of that object, because the two have very little to do with each other — a realistic size would account for all the brain processing going on.

 

Funnily enough, a realistic simulation would make spotting a whole lot easier in many cases, but it would also make it a whole lot harder in many others. Yet the “realism” crew think that only the latter is an acceptable end result. A realistic approach would also remove some of the crutches the “realism” crew rely on to argue in favour of keeping things unrealistic. Above all, a realistic simulation would make it even across all kinds of hardware, as opposed to the current wide section of different outcomes where hardware is a hugely important (often illogically so) factor in what can be seen and what can't. For some reason, some players argue in favour of keeping that inequality intact as well…

 

But in the end, the crux of the matter is this: spotting in DCS can never be realistic unless it actually tries to simulate the cognitive process of perception. The cognitive processes of perception cannot be simulated without a significant change in the rendering pipeline. The rendering pipeline cannot be changed without a huge rewrite of the sensor code.

Edited May 19, 2019 by Tippis

[SharpeXB]

lower resolutions and less fancy post-processing makes it easier for radars and EO/IR tracking to find and separate targets from the environment

Do you have any evidence or test from the sim to support this?

[Tippis]

Do you have any evidence or test from the sim to support this?

The devs say so. You already know this.

[SharpeXB]

As someone pointed out, zoom the one concession to how vision works since it ever so slightly manages to let the game account for the differences in acuity in different parts of the eye, but as also pointed out, it's ridiculously over-modelled and not actually representative of the limits of that acuity.

The zoom view feature is not unique to DCS but universally used by all other flight sims and sim games in general. For the reasons stated previously. DCS level of zoom is comparable to other sims. Furthermore the degree really doesn’t matter as its self limiting, you’re simply trading one advantage for another.

As much as people like to ignorantly throw mud on BMS' smart scaling,

Smart scaling isn’t a magic fix. It introduces more problems than it solves mainly in that it looks really awful in game to see objects rendered out of scale and making distant objects too visible.

 

The devs say so. You already know this.

Wags said the LOD model sizes are related to RCS, which is why smart scaling won’t work in DCS, since it scales up the LODs. But not that RCS is related to your screen resolution.

Neither of us knows exactly how this all works in the game engine but the bottom line is Wags stated years ago that smart scaling isn’t smart.

Edited May 19, 2019 by SharpeXB

[ApoNOOB]

Not flying much A2A combat for years because of this. I recommend looking through the other (very very long) debates, there is a bunch of issues that meet in spotting complaints and they often get mixxed together. A very complicated topic but also very important.

 

 

When you look on twitch etc. you will also see most of the community working around DCS spotting issues with very much unrealistic zoom. I think it also plays into this a bit, people have been playing DCS a certain way, got used to engine quirks etc. and don't want to try anything else.

[SharpeXB]

When you look on twitch etc. you will also see most of the community working around DCS spotting issues with very much unrealistic zoom. I think it also plays into this a bit, people have been playing DCS a certain way, got used to engine quirks etc. and don't want to try anything else.

Again... zoom view is a universal feature for every flight sim. It’s the only way to simulate 20/20 vision and a realistic fov on a PC screen. It’s not unique to DCS

[nighthawk2174]

https://apps.dtic.mil/dtic/tr/fulltext/u2/a414893.pdf

[Gladman]

Very few people I talk to don't want an implementation of some kind to fix this issue. It did exist at one time but was removed. I thought it less than ideal but good enough at the time.

There always will be those that believe their way is the only way everyone should do it but it is my recommendation that the system previously used be re implemented or a new one brought in. This is a game and we need to enjoy it as much as it needs to be realistic.

[Tippis]

The zoom view feature is not unique to DCS but universally used by all other flight sims and sim games in general.

…and it is still hugely unrealistic and a rather poor substitute for what it tries to imitate and compensate for. It's kind of funny how people how many people claim they want “realism” but also want to keep the zoom and indeed use it as an argument against actually increasing the realism. :D

 

Smart scaling isn’t a magic fix. It introduces more problems than it solves mainly in that it looks really awful in game to see objects rendered out of scale and making distant objects too visible.

No it does not. As we clarified the last time, this is just you making assumptions about something you've never actually seen or experienced yourself. In actuality, it does none of those things.

 

It's not a magic fix, no, but it is by far the most realistic way of simulating perception in any consumer flight sim available today. With a few additions, it can be made even more realistic.

 

Wags said the LOD model sizes are related to RCS, which is why smart scaling won’t work in DCS, since it scales up the LODs. But not that RCS is related to your screen resolution.

As clarified the last time, the latter inherently follows from the former. If smart scaling is impossible because it screws up RCS, then the RCS data must by necessity be picked from a part of the rendering pipeline where it will also be affected by resolution.

 

Either Wags was saying something that is incorrect, or sensor effectiveness is an inverse function of display resolution.

Edited May 19, 2019 by Tippis

[SharpeXB]

https://apps.dtic.mil/dtic/tr/fulltext/u2/a414893.pdf

That’s been posted 2,000 times. It’s irrelevant because it’s using 2003 display technology.

 

…and it is still hugely unrealistic and a rather poor substitute for what it tries to imitate and compensate for.

What’s “realistic” about looking at the world on a PC monitor? It’s the only solution to the problem though.

 

It's not a magic fix, no, but it is by far the most realistic way of simulating perception in any consumer flight sim available today. With a few additions, it can be made even more realistic.

Smart scaling was a solution from a different era. When game graphics were so poor that nobody cared how bad anything looked. It would look very awkward in a sim like DCS with such beautiful graphics on today’s large high res displays. Smart scaling applied an egregious level of enlargement. IIRC something like 2x to an aircraft only 3 miles away. So you’d see double sized F-18s hanging off the deck of a carrier. Not something most players want.

 

And the point is fully moot because it’s been stated clearly by Wags that DCS can’t do smart scaling.

Edited May 20, 2019 by SharpeXB

[Nodak]

I'd like something new, take an area of the view near what the Hornets HMD sight circle is, and any contact that's located in the circle gets magnified by a factor, except if penetrates a small predetermined distance bubble. Nothing else gets zoomed. You'll still be required to use your eyes with a good scan. And the zoom will never interfere with gunnery or in close engagements.

[Tippis]

That’s been posted 2,000 times. It’s irrelevant because it’s using 2003 display technology.

No, because it has nothing to do with display technology, and is as relevant as ever. The paper explains why. You should read it. You should also read up on “2003 display technology” — it's not actually all that different from what we have today.

 

We've been through this: you were proven wrong, so don't drag that nonsense out again.

 

What’s “realistic” about looking at the world on a PC monitor? It’s the only solution to the problem though.

No, there are other solutions — ones that actually simulate perception.

 

Smart scaling was a solution from a different era. When game graphics were so poor that nobody cared how bad anything looked.

No, it's a solution from an era of 1440p displays, and has nothing to do with good or bad graphics — it has to do with simulation of perception, which is as valid now as it was back then. Indeed, as DCS' implementation highlights, its even more relevant now due to a much wider arrange of display resolutions and rendering hardware, where you get these inverse relationships between hardware capability and the ease of spotting.

 

You can keep making these deeply ignorant comments but since you already admitted that you have not actually read the paper or seen it in action, and since you've increasingly try to buttress them with outright false statements about historical facts, everything you say trivially be dismissed by pointing out that you are only making uneducated guesses with no connection to reality.

 

All because, for some reason, you want DCS to not realistically simulate perception, evenly and fairly. It really makes no sense.

 

Smart scaling applied an egregious level of enlargement. IIRC something like 2x to an aircraft only 3 miles away. So you’d see double sized F-18s hanging off the deck of a carrier. Not something most players want.

Again, no, that's not how it works — you're just guessing because you've never actually seen it and you have not bothered to read up on how it works. At no point in any of the threads you've tried this in have you ever been able to make a true statement about how scaling works, how it looks, and what it does, because by your own admission, you simply don't know.

 

Stop making arguments from ignorance, please, and before you try to refute a claim, please look up and investigate what you're talking about.

 

And the point is fully moot because it’s been stated clearly by Wags that DCS can’t do smart scaling.

…and again, no, that's not what he actually said. The point remains valid because they've clearly stated that they're trying to figure out a way to solve the problem — and what Wags said meant that there is an immense problem that must be solved.
Edited May 20, 2019 by Tippis

[S-GERAT]

Right now the spotting has improved a lot over time, at least in VR. I remember than in the past that putting a low resolution or a low supersampling value allows to spotting a lot better than with high res or ss which was weird.

 

Now I can see the other planes from miles of distance and the reflections allow to see them a lot easily during dogfights. Of course in some cases is hard to find the other planes in the sky, but it happen in the real world too.

 

There are some problems with how the models look at some distances and is very annoying because you are seeing the contact and it disappear suddenly while you are are closing.

[CallsignFrosty]

Tippis thank you for the high IQ posts and responses - we need to get this fixed.

 

I frequent these threads about resolution vs scaling every time they are made and it's always the usual suspects attempting to 'shut it down'.

Until this is fixed I cannot really compete in any meaningful way in WVR combat (without sensors to compensate) as I am on a 4K 28" monitor.

4K is not something I will be going back from since I have been flying in 4K with DCS since at least 2014 or so and my eyes would never forgive me for willingly giving up on the most beautiful simulator experience you can achieve.

Not to mention that (VR users aside perhaps - no experience of DCS VR myself) DCS has excellent performance at high resolutions - even with sensors exported to another 1080p monitor.

 

The old scaling experiments that ED released, while comical is some cases, were vastly superior to the current scaling (or lack thereof) issues experienced but high resolution users.

[SharpeXB]

…and again, no, that's not what he actually said.

Just to be clear, go to 1:40 for exactly what Wags said about smart scaling and why they didn’t use it.

 

 

it's a solution from an era of 1440p displays,

You mean circa 2003 1440p 17” 4:3 CRT monitors. Not exactly relevant today.

 

Again, no, that's not how it works —

You mean smart scaling doesn’t scale up models? Sure it does. The problem is that it doesn’t scale up their surroundings and so would look really awkward and terrible. You’re glossing over that fact.

Edited May 20, 2019 by SharpeXB