New Caucasus Map (and tech needed)

[GEIST]

Hi! It's time...

I don't run the business at ED but if you offer a free flight weekend and with all modules and someone loads up caucasus, that potential client will probably fly off somewhere else. But that's not why I am posting - I own quite a lot of modules and all maps but it's just not quite there yet. The new maps are better and I'll post my statement here but some essential problems still remain with these - they all have a comical look to them.

Thank you whoever made this video as it shows it in perfect clarity. Comparing the two you can see a few major issues:

• The color mapping is off (it has always been and looks like it's the case with the new maps) which gives DCS it's distinct comical look. Saturation is too high. Maybe some RGB CYMK conversion issue that happened when the game was developed but there is clearly something off. And this only related to the maps - the planes do look great! Or maybe it's just how the color range is rendered... I know there have been many discussions about it always concluding it's correct but it isn't. Color Theory/Colorspaces/Rendering are hugely complex topics and if the rough RGB value is in the same realm does not mean that is reflects it accurately. There is some logical flaw in how this is rendered which gives this odd perception of the world.
• Trees do not blend into the scenery. Trees usually look like the scenery and tend to blend into it. In DCS it looks like they have 1 color and that's set everywhere so they stand out in the bring laser green with a hard edge compared to the ground which doesn't work very well. I think some tech is needed that blends the color of the trees based on the satellite data (like the grass) to blend them better into the environment
• 10m height levels with normal maps are not cutting it anymore unfortunately. I can see satellite data is being used but it seems so low-res. I know new maps are coming but I don't think that was resolved yet.
• Use photoscan workflows. The blockout building and instance them around the scene based on satellite maps isn't working very well. My suggestion would to scan 1 city with a drone flyover and then you can at least build your library of houses, streets, sights based on real datasets.
• I've tested creating ground 3D assets in the past and the engine can handle it (I went overboard with my tests and you need a 3080 to run it but I feel like there is a middle ground between performance and pushing the visual fidelity a bit more).

Thanks for reading!
 

[Rudel_chw]

23 minutes ago, GEIST said:

that potential client will probably fly off somewhere else.

please, where would that somewhere be? … I have actually looked, but have been unable to find another Combat sim that is better than this one

[sirrah]

In my opinion, Caucasus still looks quite amazing in the winter

Sure, its scenery could use some love, but then again, it's free..

 

Oh, and by the way:

34 minutes ago, GEIST said:

Thank you whoever made this video ...
 

What video are you referring to?

[GEIST]

@Sirrah: Sorry had to remove the link - just search dcs in youtube and you'll come across it. 
Also what I forgot to mention is that the issue with the maps goes beyond the look good factor.

• There is a much reduced sense of speed when flying at low altitude as everything looks very samey
• Callouts during a mission make limited sense. 'The big house next to the main road' usually could be any of the 20 standing around that area. Cities at least have roads, in open terrain there is just not much to go by. There are no unique features to them to be more specific. Just need to give GPS or laser coordinates.
• Sense of scale is also off. It's better in VR but somehow there is a bit of a minitiature world feeling in DCS

@Rudel. Noted.

[cfrag]

43 minutes ago, GEIST said:

If you offer a free flight weekend and with all modules and someone loads up caucasus, that potential client will probably fly off somewhere else.

Perhaps. Probably the free Syria map. IMHO, although I do understand the gist of what you are saying, I'm not convinced that the map in which you are flying will have the most impact on your decision if you like the simulation of a military jet that can also fly in other maps. If you understand the concept that better planes than the ones that come free with DCS (the Su-25T and T51), I believe that we can assume that the potential client also understands the concept that better maps are also not free. Wouldn't you agree?

 

43 minutes ago, GEIST said:

Thank you whoever made this video as it shows it in perfect clarity.

Please be advised that there is no video in evidence with your post. But I'm not sure that it matters much

43 minutes ago, GEIST said:

Comparing the two you can see a few major issues:

Unfortunately, since we do not which 'two' you are speaking of, we can't follow your reasoning.

43 minutes ago, GEIST said:

• The color mapping is off (it has always been and looks like it's the case with the new maps) which gives DCS it's distinct comical look. Saturation is too high. Maybe some RGB CYMK conversion issue that happened when the game was developed but there is clearly something off.

Are you aware that you can apply filters to enhance your own experience in 'Settings-->Color Grading'? And why do you assume some RGB-CYMK translation issue? DCS does not produce output for print, so I doubt that there will be any CYMK conversion involved. Most graphics cards work in RGB space.

Since I do not know which videos you talk about, you may also want to remind yourself that most videos, when done professionally, run through a post-production process that (sometimes significantly) changes the colors of the output material. That's because reality often looks too boring, and many video producers prefer a more colorful kick, or want to establish a color theme. That's art, though.

43 minutes ago, GEIST said:

Color Theory/Colorspaces/Rendering are hugely complex topics

They are not. They are straightforward linear transformations. How color is perceived and interpreted by the human eye, on the other hand is. 

43 minutes ago, GEIST said:

There is some logical flaw in how this is rendered which gives this odd perception of the world.

That appears to be your perception. I'd like some demonstrable fact rather than 'it feels wrong to me'. I'm not saying that I don't believe you or that there is nothing wrong - I simply would like to see examples, and some form of normative reference to establish what is 'correct'. We can go from there.

43 minutes ago, GEIST said:

Trees do not blend into the scenery.

Indeed. No discussion there. 

43 minutes ago, GEIST said:

10m height levels with normal maps are not cutting it anymore unfortunately.

While I agree that the density of the Caucasus map grid is a bit lacking, it does produce the best performance results, though. If I need a well-performing map, Caucasus is the map to go to. Probably because of low detail.

43 minutes ago, GEIST said:

Use photoscan workflows.

It sounds to me as if you a proposing a solution before you define the problem. I recommend that we agree on the issue before we leap to conclusions on how the solve it.

43 minutes ago, GEIST said:

I've tested creating ground 3D assets in the past and the engine can handle it (I went overboard with my tests and you need a 3080 to run it but I feel like there is a middle ground between performance and pushing the visual fidelity a bit more).

I'm sot sure what you are saying here except that you agree that high detail comes at the cost of performance.

What is it that you are looking for? More love for Caucasus? I don't think many would disagree here. Expect I'd like better visuals without the cost of performance. That may not be in the cards at this time.

Edited May 3 by cfrag

[Beirut]

2 minutes ago, sirrah said:

In my opinion, Caucasus still looks quite amazing in the winter

Sure, its scenery could use some love, but then again, it's free..

 

And it provides a safe space for performance that everyone has. If you want to try a heavy scenario, seems that's the map to do it on. If it could be further optimized, okay, but I wouldn't want anything done to it to make it harder to run. Good to have an easy fast map, so to speak.

[OldFlyer]

1 minute ago, Beirut said:

And it provides a safe space for performance that everyone has.

I think this is key. I agree that the map could do with an update but as a freebie it's a good entry-level map that isn't too demanding on the PC.

[cfrag]

4 minutes ago, GEIST said:

search dcs in youtube and you'll come across it. 

and 

4 minutes ago, GEIST said:

There is a much reduced sense of speed when flying at low altitude

and 

5 minutes ago, GEIST said:

Callouts during a mission make limited sense.

and

5 minutes ago, GEIST said:

Sense of scale is also off.

Are you trolling?

[GEIST]

15 minutes ago, cfrag said:

They are not. They are straightforward linear transformations. How color is perceived and interpreted by the human eye, on the other hand is. 

Thanks for your response. Good points made on your end. This one however I have to disagree. Colorspaces are hugely a complex topic and depend on lots of factors to get it right. Assuming everything is RGB and using linear transformations will get you where we are. CMYK was just an example if anything was based off photo reference there is a chance conversions went wrong - e.g. compare DCS to a photo of a magazine image printed some time ago and say the colors look correct. That's where your perception part comes in but that's not grounded in any data - perception of color is relatively uniform (unless a person has some health related issue) - scale on the other hand is very subjective.

6 minutes ago, cfrag said:

and 

and 

and

Are you trolling?

I am not trolling. I see we don't see eye to eye on this but that's all good.. different perceptions I suppose.

Edited May 3 by GEIST

[sirrah]

5 minutes ago, GEIST said:

Thanks for your response. Good points made on your end. This one however I have to disagree. Colorspaces are hugely a complex topic and depend on lots of factors to get it right. Assuming everything is RGB and using linear transformations will get you where we are. CMYK was just an example if anything was based off photo reference there is a chance conversions went wrong - e.g. compare DCS to a photo of a magazine image printed some time ago and say the colors look correct. That's where your perception part comes in - but perception of color is relatively uniform (unless you have some health related issue) - scale on the other hand is very subjective.

I am not trolling. I see we don't see eye to eye on this but that's all good.. different perceptions I suppose.

I think cfrag is referring to the already countless (sometimes pretty heated) discussions about these topics. Especially the "sense of speed" one. (you might want to do a quick forum search on that one, to understand cfrag's concern )

[cfrag]

Just now, GEIST said:

compare DCS to a photo of a magazine image printed some time ago and say the colors look correct.

Seriously? Do you think a magazine uses imagery that is not enhanced in post to look better? Never use magazine prints as reference. Get yourself a color reference card (any good photography shop has them), and have them present inside your establishing shots so you can color correct in post.

2 minutes ago, GEIST said:

Colorspaces are hugely a complex topic and depend on lots of factors to get it right.

Again, no. The getting it 'right' part is complex, but that has nothing to do with color space, and all with the way biology works.

4 minutes ago, GEIST said:

perception of color is relatively uniform

Again, seriously? It's not. Never was, never will. Different eyes have different perceptions (there are eyes that can't see some colors, you know?), and the same eye has different perception based on the length and intensity it looked at other colors a second before. Human eye's color perception cannot be relied on for anything. That's what we have machines that look at wave length for.

[GEIST]

Ah gotcha! Yeah definitely didn't intend to troll - just some observations I've made. 
Will try and find some topics related to that now

 

[GEIST]

9 minutes ago, cfrag said:

Again, no. The getting it 'right' part is complex, but that has nothing to do with color space, and all with the way biology works.

Jup! Super easy!
http://www2.mat.dtu.dk/people/J.Gravesen/pub/48-2015-colour.pdf
And of course are we talking about perception of color - that's what colorspaces are all about.

11 minutes ago, cfrag said:

Different eyes have different perceptions (there are eyes that can't see some colors, you know?)

As I've stated above I am excluding people with eye related issues (color blindness included). But that's not what we're talking about here, right?

But let's leave it at this hehe - I think I can say with quite certainty that the colors are off in DCS. I can't say for certain in any way if that's based on colorspace, rendering techniques or filters applied but something is not right.

Thanks for the suggestion of the filters btw - will give that a go and see if that helps!

[cfrag]

17 minutes ago, GEIST said:

Jup! Super easy!
http://www2.mat.dtu.dk/people/J.Gravesen/pub/48-2015-colour.pdf
And of course are we talking about perception of color - that's what colorspaces are all about.

I really do not want to be contrarian, so I think that you may simply want to read up a little bit more about color space. Color space is about representing the exact same color in multiple different 'spaces' -- by determining the source color's coordinates in that 'space'. There are multiple color spaces (for example CYM(K) and RGB) that are created for different purposes. Going from one color space to another is usually a linear transformation - meaning that the same color may have different coordinates in different spaces, but going back and forth between those spaces will always return the same color. Much like switching between Cartesian and polar coordinates. Using one representation can be beneficial for some applications (for example to simulate subtractive color mixing in CYM, and additive in RGB). But the color that is referenced in both is always the same color. The human eye, on the other hand, is very different, and so far we have not found a way transform a color (any color space) into that of the human eye - we are still looking for a color space that linearly can transform any color to 'human eye space". That's also pretty much what the paper that you kindly referenced to states:

Quote

no matter what inner product you put on the space the resulting Euclidean distance does not corresponds to human perception of difference between colours

So maybe you want to re-read that very nice paper.

Edited May 3 by cfrag

[Exorcet]

Caucuses looks great to me, I don't understand the complaints against it. It's also the only map with seasons. I don't use it much anymore because of the close proximity of all the airfields and the lack of any airbases in Turkey, but as is it seems fine.

[aceviper]

Try this texture replacement mod

Caucasus is still beautiful

[GEIST]

  

50 minutes ago, cfrag said:

I really do not want to be contrarian, so I think that you may simply want to read up a little bit more about color space. Color space is about representing the exact same color in multiple different 'spaces' -- by determining the source color's coordinates in that 'space'. There are multiple color spaces (for example CYM(K) and RGB) that are created for different purposes. Going from one color space to another is usually a linear transformation - meaning that the same color may have different coordinates in different spaces, but going back and forth between those spaces will always return the same color. Much like switching between Cartesian and polar coordinates. Using one representation can be beneficial for some applications (for example to simulate subtractive color mixing in CYM, and additive in RGB). But the color that is referenced in both is always the same color. The human eye, on the other hand, is very different, and so far we have not found a way transform a color (any color space) into that of the human eye - we are still looking for a color space that linearly can transform any color to 'human eye space". That's also pretty much what the paper that you kindly referenced to states:

Okay let's dive into this topic a bit more. I can see where you are coming from - you're referring to the mathematical part of converting one colorspace into another. Yes that's easily done - with 1 click of a button in any modern software package. Does that make the color 'correct' - no it doesn't. It simply converts one colorspace into another. This has absolutely nothing to do with the accuracy of the colors compared to real life. You were referring to a color chart earlier - that's a great idea but only works if that color chart was present when the photo was taken... in any other case it's completely useless, you can't do it after the fact - as the conditions when taking the image matter. 

Let's take a photo from caucasus (RGB, CMYK doesn't matter - let's assume it's correct) and you pick a color of the trees and apply that same color to the trees in DCS and put it in the RGB render engine and on our screens. Does that mean the color is correct - No. Human perception aside that green may or may not be the correct value unless that color chart was in the picture when it was taken and the photo color corrected. And let's say even that is all correct. There was a color chart when the picture was taken and the RGB values are correct there is the rendering part of it all - where simply using RGB is not enough to accurately represent the colors and other factors come into play on how it's all calculated.

From your replies I assume you're getting all that - what I am referring to is that by simply saying it's RGB, it's linear is oversimplifying the problem ... yes its the defined colorspace but it doesn't in any way mean that the colors are 'correct' and that goes beyond perception. In certain lighting conditions your colors might be completely off while you're in the correct colorspace doesn't mean your values are 'correct'. On top of that is perception which is a uniquely complex as well but overall we're talking about a median here.

btw. kudos for checking out the paper!

Edited May 3 by GEIST

[cfrag]

Just now, GEIST said:

This has absolutely nothing to do with the accuracy of the colors compared to real life.

Ah, I get it. Let's look at the process or 're-creating' color, where many applications have an issue. Let us assume that an object (say, uniformly colored cube) uniformly emits light of the color 'C' (let's say a definite shade of yellow). We then use a device (maybe a CCD camera) that 'samples' that color (takes a picture), and encodes (converts) the sensor information into a file. Let's say the format is now RGB. This file is accessed by a computer, and displayed on another device (a TV screen, a cathode ray tube with separate r, g, and b rays) as an image, and a human looks at the screen representation.

There are multiple conversions happening that each can introduce errors: when the ccd's sensors trigger and convert the light information into signals that get interpreted into RGB space,  when the computer's output processor converts the RGB signal to drive the CRT beams, when the electrons hit the screens surface to elicit a response (red, green, and blue reagents to emit photons), and finally the eye receptors that get triggered by the photons. Each time, there is error, and each error is device-specific and can vary greatly: the TV can be old, etc.). In the industry, there are means to try and compensate: there are very expensive 'calibration' tools designed to minimize the difference between what emanates from the source (the cube) and is displayed on the screen, so physically, the TV accurately (calibrated) outputs the color that was input. That's how claibrated screens work: they are guaranteed (within limits) to accurately reflect an incoming color independent on the color spaces that were used internally (from camera to TV screen) - calibration is meant to cut out the middle man.

The problem is now that we are still back to square one: two human eyes can (and usually do) perceive the color displayed on the same calibrated screen differently. Cut out the middle, and have both humans look at the cube. You will find that two humans looking at the same color perceive it different. Meaning: color is always subjective to the viewer. No color space can prevent that. The only thing that we can do is try and minimize the error when we sample, encode, store and reproduce color. Colors can be accurately reproduced, calibration can ensure that. The human eye simply can't reliably see color. So color space has nothing to do with the issue at hand. It's personal perception.

So what we can do (and what is regularly done) is use systems to tweak the color that is produced so that you personally like it or think that it looks 'just like the real thing'. So what we are doing is take a (calibrated) color that has an exactly defined position in a color space, and then intentionally shift that color's position - falsify the color, introduce an error - until it looks good to us. We intentionally change the output so it appears better to us individually. We do the opposite of objectively calibrating color to reality: we calibrate the output subjectively to out eye's and mind's feeling. So we are not trying to make colors to (objectively) exactly match the source, we are tweaking the emitted colors to match our individual expectation. 

So, with apologies to "the Matrix": "there is no color". If Caucasus feels wrong to you that is because you are absolutely right: it feels wrong to you. It says nothing about the "true" colors of Caucasus or how well the representation matches the source - it feels wrong, and therefore it is. Color, as perceived by an individual is uniquely individual. So, if you feel that caucasus colors feel wrong, you are right. There may even be people who feel Caucasus colors feel right. Neither is absolute right, both are individually correct.

So to circle back: the colors of Caucasus feel wrong to you - and they also often feel wrong to me. And, I believe, that the Caucasus colors are a mixture of artistic choice and industrial pragmatism: compress the color palette to save memory, increase scenic dynamism (difference between bright and dark), perhaps even add a tonal filter during the shader pass. 

So the question for me isn't so much how we can get a more correct color experience. The question is: can we get an experience for Caucasus that pleases more people, and if so, can we do it in a way that does not break memory and/or performance constraints? I surely hope so.

[GEIST]

I think we're on a similar page now. I stand by it when I say it's a hugely complex topic beyond human perception.

And I think there are 2 different ways to look at perception which might come into play here. There is the one what you're referring to and how your chemicals react to the light(colorblindness aside e.g.). I'd argue that probably isn't very much at all. Eyes need such precision that they are overall quite similar between every human.
The other part is how humans perceive the image they see. A lot of people will see something like the caucasus map - that's what they've always known, they are familiar with it and they can't imagine it being 'wrong' until they see it. So in that regard I would say you're right - perception matters hugely but not in terms of how color is displayed but how the brain interprets it. 

I'll rephrase my original statement.
I would like the maps to more accurately represent the true colors by default. And I think there are ways to do so by measurement - accounting for everything you listed above. There will still be errors it can't be helped but I bet 3 chocolate bars and an apple pie our current maps will not be close.

[Wait4It]

I agree with the original post on most accounts, but not just for
- The use of color in every DCS maps looks too uniform. Everything has the same brown or green hues. Look at some pictures of the real locations to see a lot more variation. It also looks dark in most cases where the tarmac in the sun still looks dark grey (and the taxi light manages to overpower the sun)
- Trees, buildings/objects need to blend a lot better in the background and local texture. Currently a forest from far away is not really visible and a forest from close by is a green texture with sharp tree objects on top.
- Cities look very uniform, probably due to the reuse of the same buildings with very clean textures.
- Terrain mesh needs to be better when close and have good working LOD for performance when far.
- There should be clutter at low level for feeling of speed and sense of height. Everything is so clean.

In my opinion the new maps are having these same issues, and its visible where map makers struggle with engine limitations (height mapping/texture/object variation)

As i see it the only real way forward to make it future-proof would be a large feature upgrade on the terrain tech/graphics engine instead of the evolution that we keep seeing, so i hope something like that is in the pipeline.

Maybe even partner up with one of the great game engine makers (Unreal, crytek). As the major player in milsim Eaglee Dynamics should be able to open some doors.

Edited May 6 by Wait4It

[Bosun]

On 5/3/2024 at 6:14 AM, GEIST said:

Thanks for your response. Good points made on your end. This one however I have to disagree. Colorspaces are hugely a complex topic and depend on lots of factors to get it right. Assuming everything is RGB and using linear transformations will get you where we are. CMYK was just an example if anything was based off photo reference there is a chance conversions went wrong - e.g. compare DCS to a photo of a magazine image printed some time ago and say the colors look correct. That's where your perception part comes in but that's not grounded in any data - perception of color is relatively uniform (unless a person has some health related issue) - scale on the other hand is very subjective.

I am not trolling. I see we don't see eye to eye on this but that's all good.. different perceptions I suppose.

 

I'm a graphic designer. 

If you're taking a digital photo for reference, it will be in RGB until you print it. Printing an RGB image through a physical printer on paper is where you get distortion, because color printing requires CMYK bands.  You'll convert that file prior to printing, to run CMYK values.  That's the only time you'll convert to CMYK, and it'll be a separate copy of the file.  

As long as everything stays digital - no conversions between CMYK or RGB are needed, as none are ever done. 


As for Color Theory being complex - it can be given the right discussion or context. We had entire courses on Color Theory that lasted full semesters.  You can make it as complex as you want.  How complex it needs to be given any specific context, however, is fairly limited.  

[Bosun]

On the sense of scale being off
This is a measured and known phenomenon with low-detail assets in a digital 3d gamespace.  Our eyes perceive scale by continually resolving details that are smaller the closer we get to something. When details stop resolving at a close distance, your eyes stop being able to judge or properly contextualize the scale.  The extreme example of this would be Jane's F-15 game from the late 80s - with the simple grid-lines for land, and the pyramids that were triangle-line shapes.  In that gamespace, you'd never be able to tell how far away you were...when you were close to them.  Same for the ground in early games like Aces over Europe, or Aces of the Pacific.  The ground wasn't resolved enough, so while it was somewhat ok to determine your altitude up high, because they had small dots along the land that you could scroll past....up close those dots were spaced far apart and telling your altitude was impossible without the instruments.  

Finer, more modern versions of that confusion come in when flying over peaks in the new Kola map, or flying low over towns in Caucuses.  It's much more subtle now, but your sense of scale feels off because your eyes don't have the finer detail being resolved, even when going fast, that would better lend you to discern how fast you're actually travelling.  The result is that you appear to be going slower than you really are, making the world objects seem larger than they actually are.  

The real eye-trick of this is to fly along and feel how big the trees in Caucasus' map seem to be at low altitude, and then to go land beside one and compare its' scale to a real tree.  The same for the buildings.  You'll notice the 'scale' is actually fairly on par - but why do they seem bigger, or that you're moving slower, at altitude?    Because of that aforementioned lack of detail, tricking your brain into thinking something is larger than it really is, because the finer details that would exist and inform your perception that would be scrolling by in between those larger assets, aren't there.   Those larger assets lack references to place them contextually in size. 

I've often thought, flying through Caucuses, that houses, buildings and other things are 'off scale.'  Sometimes, they actually are. Some of the buildings legimately are too big or small.  But a lot of the mis-perception is that lack of smaller detail giving those assets context within the landscape.  

Fixes for that?  Finer resolution on terrain tile mapping, and higher-res ground textures at lower altitudes, and more 'gak' as it were - more small objects that can place and contextualize the scale of your larger assets. 

 

(And that's just the rendering details - there's an entire other layer to this, which is in-game camera aspect and 'lens' size.  For example, Star Citizen's camera is notorious for being too 'fisheye', so that even when zoomed in, it is hard to judge distances and scale properly.  That isn't an asset issue - its a camera issue in how they frame with the lens and the equivalent mm size they're simulating.) 

Hope that all makes sense.

Edited May 6 by Bosun

[Bosun]

Overall though, I agree with the original post:  We need more detailed landscape engines here. 

Games like MSFS have really pushed the bar, and having 2010 map-technology now is showing its age. MSFS has much higher detail and is still performant, so this isn't an impossible goal. 

The real question is whether or not DCS would require an entirely new engine to support such a shift in detail.  The new Kola map is an example where, given OrbX' other projects in MSFS, their quality is quite high and I can only imagine that the trees and buildings we see in their Kola map being as refined as any other on Caucauses is only due to limitations of the engine itself. 

[Qcumber]

Not quite what the OP was suggesting with this post but I think a new Caucasus map could offer an opportunity to expand the area south and east to meet the Syria map and upcoming Iran map. 

[draconus]

15 hours ago, Bosun said:

(And that's just the rendering details - there's an entire other layer to this, which is in-game camera aspect and 'lens' size.

Yes, it all is important, not just model details (they are decent enough for most objects, maybe apart from the mountains, but they were never considered crucial part of the sim), but most important are correct model sizes displayed on a device with correct camera focal length.

To even start comparisons you have to put yourself in the same display environment, meaning usually VR with correct IPD. Then the game has to have correct size models. Now put yourself next to a man, a vehicle, a tree and a building - it's all correct size in DCS. Other test you can do with flight time over known distance with known speed - also correct.

15 hours ago, Bosun said:

Games like MSFS have really pushed the bar, and having 2010 map-technology now is showing its age. MSFS has much higher detail and is still performant, so this isn't an impossible goal.

You can't compare them directly like that. DCS has much more to do with both the terrain (object DM, IR data, seasons, terrain interaction with weapons and units on all surfaces) and other simulated aspects (avionics, sensors, FM, weapons, other air, ground and sea units) than civil flight sims. It's always a compromise between quality, simulation depth and performance. Year after year, as PC performance allows, we get more details, bigger areas, better textures, more in depth simulation. And no, it's not 2010 tech, it's in constant development.

It's easy to put out highly detailed mesh and cover it with high res textures, normal and displacement maps. Now make it a huge area with whole simulation on top and it's unplayable on current top end hardware. You can't expect high quality without impacting performance. Genius new engines, optimisations and smart graphic tricks can only get you so far.

14 hours ago, Qcumber said:

Not quite what the OP was suggesting with this post but I think a new Caucasus map could offer an opportunity to expand the area south and east to meet the Syria map and upcoming Iran map. 

There's no upcoming Iran map. Maybe you meant Iraq? Even then that'd be Turkey/Armenia/Azerbaijan/Iran map - I don't find it very popular. Please take it to https://forum.dcs.world/forum/339-dlc-map-wish-list/

23 hours ago, Wait4It said:

Maybe even partner up with one of the great game engine makers (Unreal, crytek).

ED is not stubborn or stupid. They did try to use other engines but found them lacking to the given tasks and had to go with their own engine. They try to implement different new techs as they come out.

Edited May 7 by draconus