Why485

Like I mentioned before, the values for the imposters changed dramatically through the patches. Here's a changelog of the imposter settings throughout the patches: The values are listed as such: maxsize/minsize/alphaExp Minsize is the smallest the imposter can get on screen. This is the most obvious in function. The higher the value, the bigger stuff will look when far away. This value is measured in pixels. AlphaExp is a transparency. The higher the value, the more transparent, and thus the harder to see the object is. A value of 0 means a black object is almost completely black at a distance. Maxsize is a little more complicated to explain, and less obvious in its effects. It's both the size an object needs to be for it to switch to an imposter, and the size at which the imposter itself is rendered, and then scaled down based on distance. This value is measured in pixels. 1.5.0: Normal 12.0/3.0/0.5 Enlarged 12.0/5.0/0.25 1.5.1: Small: 20.0/6.0/0.0 Medium: 20.0/8.0/0.0 Large: 20.0/10.0/0.0 1.5.2/2.0: Small: 2.0/4.0/0.75 Medium: 4.0/6.0/0.5 Large: 7.0/8.0/0.15 What's important to note about 1.5.2/2.0 is that the maxsize is unusually low in all settings. This is why, despite the minsize being generally larger than 1.5.0's settings, things are much harder to see in 1.5.2. Having such a small maxsize that's also lower than minsize is also the reason why imposters seem so inconsistent now. They seem to change size and shape much more dramatically than they should. This is because with so few pixels to work with, the imposters are basically being drawn as literal lines and dots rather than things that resembles planes or tanks. It's also worth mentioning that the only change to the system itself between 1.5.0 and now (as in, engine/rendering level changes) has been enabling the imposter system for missiles in 1.5.2 and 2.0. Other than that, the system driving this whole thing has gone completely unchanged since its initial implementation, which is unfortunate because tweaking these settings can only get you so far. There are no settings that can fix the inherent drawbacks with a 100% pixel size based system. The simple addition of the imposter being modified based on its distance to the camera (which is info it already has because alphaExp compares it against a hardcoded range value), would be a game changer and solve so many issues.

This is because the whole system is based on pixel sizes. It's a very simplistic system that makes for exactly the kind of unintuitive behavior you're talking about. The system needs to be just a little more complex than it is now. I've made many posts in the past on how to expand the system to make it into something very good. Make the imposters the same apparent size regardless of resolution. You can't even begin to discuss what looks realistic when everybody is seeing different sized imposters. Separate values for ground and air targets. Right now, settings that work well for air targets make ground targets too visible. Imposters should fade with distance in a much smarter way. Right now, the alphaExp setting is more like a minimum alpha that reaches its minimum value around 1km, which is fairly useless. The way I would do this, is have a minimum distance (in meters) at which the imposter is opaque, and then a maximum distance (in meters) at which the imposter then becomes fully transparent. A smarter system would prevent targets at 40km being easier to see than targets at 10km. And then, stuff that's not required, but would be nice: Ideally, I'd prefer if this was a non-optional system that could not be turned on/off. It would just be set to something realistic and be non-editable. If it MUST remain an option, it should be exposed to mission editors. Right now, the mission saves whatever your current setting is and uses that. This is not made apparent anywhere. If this remains an optional setting, you should be able to see what setting a server is running before joining it.

The settings have changed dramatically through the patches. Right now the visibility imposters on all settings except the largest are smaller than they've ever been. On top of that, all servers now (in a very unwieldy way) enforce a model visibility setting, with most of them just defaulting to off. So even if you had yours set to large, you won't see anything anyway. Honestly, I've basically stopped playing DCS at this point. It is so crushing to see one of the longest standing problems in DCS finally be fixed, and then for the fix to effectively be removed. I can only hope that ED will continue to tweak it, maybe even actually fix it, but now that it's been pushed to live I wouldn't count on it.

The model visibility setting does not draw planes/vehicles any further or more often than it already does. If you can see a model's sprite, the engine would have already been drawing it regardless of what model visibility setting you use. I've heard some say that are some issues with either AMD or Crossfire/SLI configurations that are causing performance issues, that may be what you're talking about in terms of FPS drop. There are all kinds of opinions on how well spotting works without any option on, and I fall squarely on the "it's unrealistically difficult" side. The problem is that everybody has a different baseline of both conception (i.e. I think it should be harder) and perception (i.e. I play 4k and I can see just fine). I don't think anybody denies that the current system isn't perfect, but it's still a hell of a lot better than the 1.2 "dark ages." Unfortunately it's basically disabled across all servers at this point for multiple reasons, one of which being the resolution thing which is a completely fixable problem. We can't even begin to discuss what's realistic or not when everybody is seeing something different.

There were no changes to the imposter system in the latest 2.0 patch as far as I can tell. It'll definitely work, but I'm not sure what it'll look like. I suspect the imposters will look blurry in an artificial way because they were "rendered" at 1080p, but blown up to 4k resolutions. The solution in this thread isn't really what ED should go with because of little issues like that, but it's the best I can do without having access to the engine code that translates from the imposters.lua to the shader itself. They should apply this fix on the engine level, not the shader level. It'd be just as trivial to apply in engine code though, as all I'm effectively doing is multiplying a number by the resolution. ED just needs to want to do it. I know how to work around the blurry imposters problem when viewed at >1080p resolutions, but I don't want to edit the imposters.lua because this mod is more of a proof of concept (to show how important this is and what it looks like) than anything else. It is. Whatever setting that the person editing the mission happened to have when they saved it will be applied to everybody on an online server whether they meant to do that or not. There is no way to disable this apart from, theoretically, hand editing the options file in mission file. It's not something I've tested, but it should work based on my understanding of how it's enforced.

I just fixed the resolution dependency problem. Here's the thread for the fix over in the mods forum along with a download. Resolution INDEPENDENT (scales with resolution) Resolution DEPENDENT (uses a fixed amount of pixels) The reason this is such a big deal is that you cannot even begin to decide what setting is realistic for seeing targets at a distance when nobody is even seeing the same thing. With this, the settings in the imposter.lua (which decide how big they are at each setting) are normalized to 1080p, and scale appropriately up or down with changes in resolution. Something else worth noting is that I took those comparison screenshots above with an imposter.lua that is similar to what the initial 1.5.0 values were. This is not what imposters look like in 1.5.2 or 2.0. Right now, the settings ED has in for the system are very strange and you'll see imposters that don't look like planes, but often lines or a series of dots. This is because the maxsize is both very low, and lower than the minsize. I don't like the current model visibility settings because they're prone to flickering and are in general inconsistent in appearance. If you run at high resolutions (>1080p) these problems become more pronounced.

I made the model visibility imposters resolution independent with literally one line of code in the imposter shader. This means that they’re the same size regardless of what resolution you play at. Previously, the lower your resolution, the bigger and easier to see the imposters were. Conversely, the higher your resolution, the smaller and more difficult they were to see. Resolution INDEPENDENT (scales with resolution). Size of the impostors is exaggerated for clarity. Resolution DEPENDENT (stays a fixed size, regardless of resolution) The first gif is with my fix. The second is what happens normally. You can see how dramatic a difference this makes. The reason this is such a big deal is that you cannot even begin to decide what setting is realistic for seeing targets at a distance when nobody is even seeing the same thing. With this, the settings in the imposter.lua (which decide how big they are at each setting) are normalized to 1080p, and scale appropriately up or down with changes in resolution. Attached is a JSGME compatible version of the mod. If you have JSGME, just put the "Resolution Scaling Imposters" in your JGSME mod folder. If you want to do it the old fashioned way, then extract the contents of the "Resolution Scaling Imposters" into your DCS folder. Something else worth mentioning: I took the screenshots above with an imposter.lua that is similar to what the initial 1.5.0 values were. This is not what imposters look like in 1.5.2 or 2.0. Right now, the settings ED has put on the game are very strange and you'll see imposters that don't look like planes, but often lines or a series of dots. This is because the maxsize is both very low, and lower than the minsize. I don't like the current model visibility settings because they're prone to flickering and are in general inconsistent in appearance. FixedImposters.zip

The sprite you see is an approximation of what the plane looks like from that aspect, which means aspect has a very big effect on how visible the sprite is. Sprite sizes are around the size they were originally in 1.5.0, but with weird settings (maxsize > minzsize) that cause the effect of aspect on spotting to be exaggerated because of the effect that maxsize has on what the sprite looks like. It's also why the new imposters seem to flicker and shimmer more than they used to. From another one of my posts, here's the exact settings: Here are the model visibility setting updates. The 1.5.1 values are listed in parenthesis. Small maxsize 2 (20) minsize 4.0 (6.0) alphaExp 0.75 (0) Medium maxsize 4 (20) minsize 6.0 (8.0) alphaExp 0.5 (0) Large maxsize 7 (20) minsize 8.0 (10.0) alphaExp 0.15 (0) For reference, in 1.5.0, here were the settings: Normal maxsize 12 minsize 3.0 alphaExp 0.5 Enlarged maxsize 12 minsize 5.0 alphaExp 0.25 The system is arguably "working as intended" but I don't like the 2.0 settings because of the flickering effect they tend to have. The original 1.5.0 settings, while slightly smaller in general, were much more consistent.

The sprites in 1.5.2/2.0 have peculiar settings that tend to emphasize this effect. It's "normal" in the sense that that's just how they look in 2.0 but I don't think it looks good or realistic.

I don't know about everybody else, but when I say FC3 jet I mean FC3 like the F-15C, Su-25T, and Su-27. A SFM is simply unacceptable on a newly released plane in my opinion, but I would be perfectly fine with a jet on the level of the current FC3 jets.

I recall hearing that ED themselves would never do FC style jets ever again, but no source. Either way, it doesn't surprise me. Personally I have no problem with adding more Flaming Cliffs level jets as long as their flight models aren't SFM. DCS level jets have development cycles on the order of 2-4 years, which severely cuts down on variety in combat and mission design. I would of course prefer a full on DCS jet over a FC jet, but I also have no problem with sacrificing fidelity for variety a la the IL-2 survey sim model. After all, that's what Lock On, the game that got me hooked on this franchise years ago was, and it's not like adding more FC level jets makes the currently flyable and planned DCS level jets go away. I think there's real value in variety, and honestly when it comes to combat workload, it's not all that different flying a FC3 level jet and pressing hotkeys/HOTAS binds versus clicking them in the cockpit. As was said earlier in this thread, I'd rather have a FC3 F-15E than no F-15E at all. Besides, it's possible that with a FC3 jet laying the groundwork, any of them could be later expanded into a full module.

My biggest argument against the MiG-29A as it is right now, is still that it uses an ancient SFM and I don't want somebody's first impression of DCS to be a SFM. If ED were to make the MiG-29A free, it should happen after the eventual flight model revision. Other than that, I'm fully behind this idea as somebody who has been playing DCS since LOMAC. It would also be good for the MiG-29A to get its own series of brief training missions like the Su-25T's. Those are fantastic because they're all short and to the point. You can complete all of them in an hour or two and know all you need to get started.

I've never understood what it was for either and it's just a nuisance. It'd be nice to see it removed.

And? What's your point? I bought LOMAC way back when, then Flaming Cliffs after that, then Flaming Cliffs 3 (rebuying it, basically) and I have absolutely no problem with what you just said. This isn't about me. This isn't about current users at all. It's about getting more players into DCS, and I believe the more people playing and buying DCS, the better. More players to play with/against in MP, and more paying customers to fund the game as a whole.

If they go with the MiG-29A, there is no extra development time that isn't already scheduled or in progress. The MiG-29 is already going to get a new flight model, and updated exterior model. If this were to happen, that's when it should happen, and is not new work.

I disagree. Neither modules are representative of DCS nor do they present the game in a way that I think is terribly convincing to a new player. The TF-51 is a prop plane. DCS was not built for props, and is not a good WWII game. The TF-51 has no weapons either, so even if DCS was a decent WWII game, the player would never be able to do anything interesting with it other than fly around. The only thing it has going for it is that it is fully simulated, but if you want to play DCS for the jets, then what you learn systems-wise in the TF-51 is largely irrelevant and/or uninteresting. Meanwhile, the Su-25T is a FC3 jet, which means it is very simplified from a systems perspective and thus does not represent the detailed jets in the game. It is also a ground attack aircraft, and thus completely misses out on the interesting mechanics and interplay that comes in air to air combat that is featured in FC3, MiG-21, and very soon the Mirage 2000C. If a player is interested in fast jets or air to air combat, there is nothing representative in the game to give them a taste for that or how cool it could be. My ideal free aircraft for DCS is one that doesn't exist. At least not yet. I think some kind of very simple, but multi-role and fully systems modeled plane capable of both rudimentary air to air and air to ground would be best. Something with a radar. Something like the upcoming F-5, actually, even if it doesn't have radar guided missiles. The MiG-29A is honestly not a bad choice. It's fine in air to air with perfectly good missiles (R-27ET/ER and R-73) and a working radar. It can also do rudimentary air to ground work with an assortment of bombs and rockets. It's not specialized for it, nor can it use some of the cooler weapons, but it doesn't need to be if the Su-25T is around. Saying that it'll be "too effective" is silly argument against the idea. The point is to get people to want to try and buy the DCS modules, which I think such a plane very well could. My main point of contention against using the MiG-29A as a free jet is if it uses the current SFM, which should absolutely not be used in something mean to represent DCS World.

It does. Once the imposter reaches its minimum size it won't get any smaller no matter what happens. If you raise your FOV, it stays the same size, but its apparent size grows relative to everything else. It's why when I talk about the sizing of the imposters I always use two terms. There's the actual pixel size (which is constant once it reaches its minimum) and then there's the apparent size, which is how big the imposter looks in the world. There's a difference because your reference (the surrounding area) doesn't necessarily scale in pixels the same as an imposter would with field of view or distance.

FYI, no changes to the imposter stuff this patch.

The new updater uses BitTorrent to download. It needs settings to limit download/upload rate like any other torrent client or else it hogs all your bandwidth in a very inconvenient and inefficient way.

That seems highly unlikely. This is exactly why it should be in the game. Literally every other sim out there has some solution to the visibility problem and until 1.5, DCS was the only one that didn't. It's been a point of contention with the engine for a decade now. Please, stop posting and end the circular and illogical argument that you just won't let go. I'm really tired of you filling up the thread with your biased, short sighted, and second-hand views on a feature that you are not even interested in using. Others were arguing for closing the thread. Maybe that's appropriate if you just won't stop and let people discuss it.

If the values were just slightly different, what you just said wouldn't happen. The people who are claiming the sky is falling because the current settings, which are the opposite of what the old settings were, are very short-sighted and making the assumption that the current values are how it's going to stay. ED is probably messing around to try and find good settings. I guarantee this is going to get reduced, closer to how it originally was, and then made a server setting on top of that.

Wow, well. Man, it's too bad that it's in its final form and never going to change from here on out. What a missed opportunity.

Having spent a lot of time experimenting with different values for the imposter settings, I've already laid down what I consider my feedback on the feature in a previous post, and what I think needs to be done to make it as good as it can be.

This is again getting too complicated. The reason resolution even matters at all is because the setting gives different results depending on your resolution. If the minsize scaled with your resolution, then everybody would get pictures that are close enough that you don't have one resolution with a huge advantage over the other. Before 1.5, it was high resolutions with the big advantage. After 1.5 with model visibility on high, it's now the low resolutions. As I said before, there should be a base agreed upon value, probably at 1080p since it's the most common resolution. For example if they decide a 6.0 minsize is what 1080p should run at, then if you run the game at 1440p your minsize will be at 8.0. If you run 4k, then your proportional minsize would be 12.0. With this, everybody gets roughly the same level of visibility, regardless of resolution. However, I do wonder that if at 4k you might want a slightly lower setting than the proportion would suggest, as that's a very large jump in pixel density over 1080p. Unfortunately I don't have a 4k screen so I can't test it myself. When it comes to server restrictions, then you can leave it as a simple on/off flag like the majority of the difficulty options. Once ED settles on reasonable values for the system, then I believe most servers will force the option on, as not only is it more realistic to have than not, but it also helps to level the playing field between people who play at low resolutions, and those at very high resolutions.

I think that's a little too complex for a gameplay setting. If it were up to me, the only option I would expose to the user would be the percentage size option as that's the most important one. Maybe have separate options for air and ground, but that's also something that could just be a hardcoded and derived value such as always being 50% of the size of air targets. Those are the options that should be available for server enforcing, and available for the player to modify. Having it set as a percentage of your current resolution is absolutely critical, as (and I'm going to keep saying it as long as it's true) it's very bad to have a system like this so sensitive to what resolution your monitor is running. Smaller resolutions need to have smaller values for minsize or else you get the reverse of 1.2 where the lower your resolution, the easier you can see targets. VR is a special case, and it's an option that should only appear if you have a VR device plugged in, and those settings should only be used when VR is in use. Honestly, I'm not too concerned about restricting VR users beyond preventing them from setting completely silly values. The rest of the values should be set by ED.