reported Spotting dot bugs in VR

[draconus]

19 minutes ago, SturmVogel said:

when comes in close range they just dissapear

What aircraft at what range?

9 hours ago, PawlaczGMD said:

I see nothing like a spotting dot, they are nonexistent at this setting at all distances as far as I can tell

The dot is a dot - a pixel - that's all. It's black and then it fades.

9 hours ago, PawlaczGMD said:

Frankly, I don't understand why the spotting system always put larger dots at lower resolutions.

It doesn't. Only if you use upscaling from much lower resolution rendering will try to recreate the low resolution image in your native resolution which can make the big low res dot render now as a square of 4 pixels (not tested).

Check the excel tool I made. You can test other resolutions, fov and aircraft wingspan. Note: wings are very thin in this aspect they quickly become invisible over 1nmi. Here's example of F-16 at 1920 res and 90 degrees horizontal fov.

On my current VR setup I cannot see the F-16 cold aspect over 15nm so cannot comment on 30nm visibility.

target size.xlsx

Edited August 12, 2024 by draconus

[peachmonkey]

25 minutes ago, draconus said:

What aircraft at what range?

The dot is a dot - a pixel - that's all. It's black and then it fades.

It doesn't. Only if you use upscaling from much lower resolution rendering will try to recreate the low resolution image in your native resolution which can make the big low res dot render now as a square of 4 pixels (not tested).

Check the excel tool I made. You can test other resolutions, fov and aircraft wingspan. Note: wings are very thin in this aspect they quickly become invisible over 1nmi. Here's example of F-16 at 1920 res and 90 degrees horizontal fov.

On my current VR setup I cannot see the F-16 cold aspect over 15nm so cannot comment on 30nm visibility.

all of these IPD/FoV calculations for spotting is a fruitless idea. There's a 100% certainty you end up with a list that can in no way be unified across all different graphics configs and hardware. You will end up introducing a trade off into each use case that will diminish the end result by half or more, means the end result will end up being "back to square 0".  I guarantee you this.  If it were possible it would've been achieved a looong time ago. DLAA/DLSS/FSR makes this problem even worse.

The way it was solved in all other sims was via a contrast bubble/outline/whatever you want to call it, essentially performing a forced de-antialiasing of the model and purposfully adding more contrast in cases of model/background color parity.

ED does not want to do it "because reasons", as in always looking for the most complicated solution to a simple problem. Such as trying to replicate a "realistic/retina like" model and a rendered object model lighting using the game engine that can't even deliver it on macro scale, and that's also not counting the lack of actual hardware that can display it (monitors/hmd's), or the GPUs to render a 16K image.

But people on this forum will continue chasing the impossible and playing straight into ED's hand of the everlasting "we are still evaluating it" nonsense.

 

[SturmVogel]

54 минуты назад, draconus сказал:

What aircraft at what range?

 

target size.xlsx 13.33 \u043a\u0411 · 0 загрузок

 

This was at ECW Server and I cant measure it, at my personal feelings was about 3-4 nm for Su-25.

 

Tested visibility in offline for MiG-21. Contact shows from 24 km and starts increase angular size untill merge. This is good, better then June DCS versions, when contact has step in angular size increasing.

[draconus]

18 minutes ago, peachmonkey said:

ED does not want to do it "because reasons"

Because that's stupid idea to purposely increase target contrast or make the models size bigger - it's unrealistic. Current solution is almost perfect (too long visibility ranges possible depending on display and zoom, no light glints) but then there come players who now can't see sh... and will still demand easy spotting.

[SharpeXB]

28 minutes ago, peachmonkey said:

The way it was solved in all other sims

All other sims?  

[peachmonkey]

44 minutes ago, draconus said:

Because that's stupid idea to purposely increase target contrast or make the models size bigger - it's unrealistic. Current solution is almost perfect (too long visibility ranges possible depending on display and zoom, no light glints) but then there come players who now can't see sh... and will still demand easy spotting.

Unrealistic?  You're counting the darn pixels on your screen and speak of realism, c'mon, now.

do you speak for the WW2 pilots as well? All WW2 engagements are within 1-2km, and in those I can't see <profanity>. I don't give a hoot about pixels from 10-20km away, there are no computers/rockets in WW2 in case you forgot.

[draconus]

1 minute ago, peachmonkey said:

All WW2 engagements are within 1-2km, and in those I can't see <profanity>.

What aircraft? You know we can test it?

[peachmonkey]

Just now, draconus said:

What aircraft? You know we can test it?

all of them.

and for testing, please try the following configuration (to make sure you're testing what the customers are using):
- Varjo Aero at 39ppd
- DCS with DLAA/quality/sharpness=52

[draconus]

4 hours ago, peachmonkey said:

and for testing, please try the following configuration (to make sure you're testing what the customers are using)

Right away. I can only test what I have. PD1.3, DLSS:quality, sharpness=0.7.

P-51 rear aspect turned black blob from 1 to over 2nmi then more like 1 pixel up to 5 miles, then more faint and gone over 8nmi.

I tried some warplane "dogfight" in TF-51 vs FW-190 for 10 minutes and it was fine too.

Use the mission and test for yourself.

Cannot reproduce any "they just disappear" effect.

[SharpeXB]

25 minutes ago, draconus said:

Cannot reproduce any "they just disappear" effect.

I think it’s only reasonable for claims of “disappearing” aircraft to be documented by screenshots with a label floating over empty space. 

[TAIPAN_]

8 hours ago, draconus said:

Cannot reproduce any "they just disappear" effect.

If your signature is still correct, it's because of the headset difference. 

Rift S being low resolution is one of the headsets known for making visibility easier. 

Varjo Aero being high res seems to get the dot scaled much smaller if it's anything like the Pimax Crystal. That's compounded even worse if they are using eye tracking / quad views. 

A simple ability for people to reduce the size for only themselves would solve the issue, that won't negatively affect other users. Yet I see people wanting to go back to the "1080p to be competitive" nonsense ie wanting to enforce everyone goes smaller dot on the server side settings. 

[YoYo]

29 minutes ago, TAIPAN_ said:

If your signature is still correct, it's because of the headset difference. 

Rift S being low resolution is one of the headsets known for making visibility easier. 

Varjo Aero being high res seems to get the dot scaled much smaller if it's anything like the Pimax Crystal. That's compounded even worse if they are using eye tracking / quad views. 

A simple ability for people to reduce the size for only themselves would solve the issue, that won't negatively affect other users. Yet I see people wanting to go back to the "1080p to be competitive" nonsense ie wanting to enforce everyone goes smaller dot on the server side settings. 

You didnt know his resolution. His default resolution could be higher or lower than default (upscaling or downscaling) so its your imagination only about default settings (of course lower resolution improves visibility, but the same applies to 2D users as 2K and 4K monitor). Im using about 6,3K and also havent a problem to find a target. Practice makes perfect, also in reality, the target can usually be seen at a distance of 3-5 miles usually (in DCS the target is visible for me from 14 nm). I recommend that you read the document that I have attached as well.

Edited August 13, 2024 by YoYo

[draconus]

3 hours ago, TAIPAN_ said:

Rift S being low resolution is one of the headsets known for making visibility easier. 

Let's see Varjo Aero (35 PPD) case for F-16, rear aspect:

distance [nmi]	angular size [degrees]	size [pixels]
1	0,308	10,8
2	0,154	5,4
3	0,103	3,6
4	0,077	2,7
5	0,062	2,2
6	0,051	1,8

and compare to Rift S, same target:

distance [nmi]	angular size [degrees]	size [pixels]
1	0,308	4,4
2	0,154	2,2
3	0,103	1,5
4	0,077	1,1
5	0,062	0,9
6	0,051	0,7

Obviously lower resolution means bigger pixel (lower PPD) but until it becomes 1 pixel higher resolutions respresent the target by more pixels so it's the same angular size^* - exactly the same visibility but better chances to distinguish the type and aspect.

I get that single smaller pixel is harder to notice at further distances but then we're talking about high ranges like 6-8 miles for a small fighter.

^*If it doesn't work like that it should be reported. I'm not here to put anyone down. I want realistic visibility for all VR headsets. I will change my headset too. I'm not keeping it for MP competitivness, just for performance on weaker GPU.

@YoYo You keep repeating the 6,3K for your setup but it's only upscaling res. You get good image quality but that's it - you still have to deal with native res, which is 3600x1920 for Quest Pro. There's nothing to assume on my Rift S settings as I specifically provided info as PD1.3/DLSS:Q in my last test.

Edited August 13, 2024 by draconus

[peachmonkey]

19 hours ago, draconus said:

Cannot reproduce any "they just disappear" effect.

you can check out this post on Reddit along with the included youtube link:

 

https://old.reddit.com/r/hoggit/comments/1eqrdhq/for_vr_users_dcs_is_either_a_game_of_spot_the/

 

 

11 hours ago, TAIPAN_ said:

If your signature is still correct, it's because of the headset difference. 

Rift S being low resolution is one of the headsets known for making visibility easier. 

Varjo Aero being high res seems to get the dot scaled much smaller if it's anything like the Pimax Crystal. That's compounded even worse if they are using eye tracking / quad views. 

A simple ability for people to reduce the size for only themselves would solve the issue, that won't negatively affect other users. Yet I see people wanting to go back to the "1080p to be competitive" nonsense ie wanting to enforce everyone goes smaller dot on the server side settings. 

@TAIPAN_ I sincerely recommend you not to engage with this troll.

Edited August 13, 2024 by peachmonkey

[TAIPAN_]

22 hours ago, draconus said:

 

^*If it doesn't work like that it should be reported. I'm not here to put anyone down. I want realistic visibility for all VR headsets. I will change my headset too. I'm not keeping it for MP competitivness, just for performance on weaker GPU.

@YoYo You keep repeating the 6,3K for your setup but it's only upscaling res. You get good image quality but that's it - you still have to deal with native res, which is 3600x1920 for Quest Pro. There's nothing to assume on my Rift S settings as I specifically provided info as PD1.3/DLSS:Q in my last test.

 

Nice post and data.

I took peachmonkey's report at face value, we can't know everybody's hardware setup and there are other factors. It's moulded by my own experience with the Pimax Crystal, while it has similar lenses to the Varjo Aero the resolution used is a lot higher which is then modified by using QuadViews. I'm no expert but I've been told the supersampling at 200% along with any form of DCS Anti-aliasing tends to anti-alias away some or all of the dot.

When switching to VR and hearing all the talk about "it's so much easier to spot" gave me a WTF moment when I went to the crystal and it actually became harder. I say this as someone who has no issue spotting in another WW1/WW2 sim where I often spot, bounce and kill others before they see me (we are not allowed to mention competitors names, I'm just highlighting it's not a personal/skill issue)

I even got prescription inserts, which did make things look a bit nicer but my vision had nothing to do with the spotting.

I'd be curious to know if @peachmonkey is using any kind of eye tracking or high supersampling.

At the end of the day - the solution I mentioned is still valid for all the posts behind me - that is give the people a way to reduce the size of their own dot without affecting other peoples dot size. The #1 complaint I see is that someone says their own dot is too big and they want to reduce it to how it was before (for them).

Edited August 14, 2024 by TAIPAN_

[YoYo]

23 hours ago, draconus said:

 You get good image quality but that's it - you still have to deal with native res, which is 3600x1920 for Quest Pro. There's nothing to assume on my Rift S settings as I specifically provided info as PD1.3/DLSS:Q in my last test.

The optical effect, for the user, whether you do increased PD direct in DCS or outside upscaling is the same (optical effect, what the eye sees). Not long ago I had in my home a friend who has Varjo Aero, he was very surprised by the picture quality in QP and stated that the image was at least the same as his (his native resolution is higher than mine for sure, 5760x2720).

Edited August 14, 2024 by YoYo

[draconus]

3 hours ago, TAIPAN_ said:

give the people a way to reduce the size of their own dot without affecting other peoples dot size

There are already options for improved dots: on/off or labels: off/dots/full.

[zerO_crash]

On 8/11/2024 at 9:14 PM, draconus said:

@zerO_crashThere are some helpers available for the VR too, like necksafer which also allows you to check 6 with minimal head turn, so it's not like VR players are all clean and realistic clients either.

The only cases I've seen someone want yo use it, are people with some form of neck damage, reduced physical ability or simply age-related. In fact, there is something called "feedback", which they then degrade vs. someone having a 1:1 movement ratio. Still, I'm sure that this can be solved by applying a more restrictive registry check.

 

In general, DCS is a simulator, not Quake. The perception of "competitiveness" is completely dumbfound. IRL, there occur individual restrictions on airframes (lift, speed, temperature, etc...) due to imprecisions in manufacturing (more often damage, but not severe enough to require the change of an expensive component). As such, ED ought to start educating its userbase, such that more realistic implementations can be made down the line, rather than mainting the average sandbox mentality that promotes fooling around.

 

Then there is the concept of accepting flying a more inferior aircraft, and rather using one's head to actively make the best out of it. The good old fighting spirit seems to be completely gone. It's very clearly not a mentality promoted in the west these days. (eye of the beholder).

Edited August 14, 2024 by zerO_crash

[Dallatorre]

Hello, I want to provide my feedback on the current status of spotting in VR. I am the creator of "that" Reddit post, after getting my clearance to post here I wanted to post some unbiased findings and opinions:
The dots are generally fine, I feel they give me a good clue about the enemy positioning when flying a jet with no datalink/hud.
BUT there is a big problem when the jets are closer than the dots:

1. The single-pixel rendering of a plane is problematic for VR headsets, most headsets out there require display compression and image warping for proper projection, resulting in single-pixels that get often culled from the image the eye can see.
2. There should be a distance band where the dots disappear to render the pixel-model where both coexist, to make that transition less jarring, maybe this will mitigate the issue with the zoom making planes disappear/appear.
3. I have seen multiple times planes under 3nm from me disappear, I have it on video, I think this is the biggest issue and people complain "I don't have the dot", it's not for far targets. I think this effect is noticeable only when the engine has to render more than one plane and ground units, in a dynamic environment.

I did some reasoning on the matter, but everything is based on my experience in DCS online which consist of 60-90s scenarios. I envision DCS to not be frustrating in the wrong ways, we are not real pilots, we are not physically fit to fly these jets, some of us have eyesight issues, we should be focusing on air maneuvers and systems, not dig trough settings to get a competitive advantage. I think the Spotting dots are really close to be a true playing field leveler in DCS, while not being "OP". This is what I would like to see tested for feasibility:

• Tiny (current) Spotting dots near always present even at close range where the enemy model would be rendered at less and equal to 4 pixels. That would probably level the playing field between high-cost and low-cost setups.
  • Yes spotting will be generally easier on pre-cold war but at that era it's much more important how you fly an aircraft to score a victory. 
  • Spotting will not be easier in the modern age, as I just need to select a contact in the HSD to have him outlined in my helmet anyway.
  • This will make gameplay on mixed 80s scenario more enjoyable to everyone, where the F-16 still has it's wallhack datalink, but at least now the masochist F-4 can see the enemy diving on him.
  • Consideration: rotorcrafts might be at a serious disadvantage with this, might consider having different dot parameters for those like it's been implemented for ground units.
• Test implementation of a variable-size dot client-size, with a slider. The tradeoff being not able to IFF a target because I made the dots large for my advantage. This would save a lot of computation to scale the dots properly for all the resolutions. Maybe servers could have a setting for "max dot size"?

What brought me here: I see the graphics engine struggle to display the not-so distant planes and being a masochist myself I cry a little when the target I'm tiring to pincer disappears from my displays or I don't see the MIG-21 I am 2-circling.
I have on video a scenario where I placed several low-flying jets, and recorded from the desktop the effect of the single-pixel planes being culled from the rendering engine, while the spotting dots work totally fine at distance.

CPU 13700K
GPU RTX 4090 24GB
RAM 128GB
Headset: Oculus Quest 3 (VDXR)
Settings worth noting: 
	Graphics mostly high
	Draw Distance: Ultra
	Improved Spotting Dots: On
	LOD Slider: 2.0
	Sharpening: no
	Anti Aliasing: no
	DLSS: no
	Foveated Rendering: no
	GODLIKE resolution settings should be something like 3120x3120 at 72hz

Video posted: Recorded from the desktop raw output, not the headset.

1:33 4K Video on YouTube showing the rendered planes disappear.
 

I want to preemptively answer some criticism on my points:
"I have no issues seeing my enemies that's why i dominate on online games, if you give this advantage to everyone I wouldn't be this good anymore, it's unfair" This is more akin to Counter Strike than DCS... you can still dominate if you focus on the actual BFM rather than scouting for the noobs.
"This will make the game too accessible" isn't that good? somebody on Discord was really mad about making the game more accessible.

Edited August 14, 2024 by Dallatorre

[zerO_crash]

On 8/12/2024 at 4:36 AM, PawlaczGMD said:

I set up a test mission with KC135(large) and F18(medium) as spotting targets against the sky at day. Cold aspect, co-altitude, I'm trail behind them and they are at different ranges (5,6,8,10,15,20,25,30 nm etc).

My results at 100% Pimax Crystal resolution, MSAAx2:

improved spotting dots on: large target 8nm, medium target 6nm (I see nothing like a spotting dot, they are nonexistent at this setting at all distances as far as I can tell)

improved spotting dots off: large target 10nm, medium target 8nm (I think I see a slight improvement, and spotting dot is actually rendered)

And you say you can see stuff at around 30 nm? This is a huge difference. My eyesight is good btw.

Frankly, I don't understand why the spotting system always put larger dots at lower resolutions. Wouldn't it be fair to keep the solid angle subtended by the dot constant, regardless of resolution?

For an arbitrary example say that at 1080p you see a 4-pixel(2x2) dot at 20 nm. If Pimax res is 2880x2880 per eye, this is roughly 2.5 times larger than 1080p. Therefore, I should see a 4-pixel dot at 50 nm, and a 25-pixel(5x5) dot at 20 nm, and then a 100-pixel (10x10) dot at 10nm, etc. But this is not the case as dots don't exist for me as far as I can tell. And I remember using Quest 2 and Pico 4, where the dots were huge, as their resolution was much lower. So this is not just my subjective impression, as when I had a lower-res headset, I could see the dots just fine and I thought they were big.

In before spotting an f18 at 6nm is realistic - that's not the point. The point is that it is very significantly easier to spot at lower resolutions. So if it realistic at my settings, it would make it unrealistic at lower resolution. Can we pick a distance that will be the same for everyone? What concerns me here is the imbalance and incentivizing playing the game/sim looking worse, when it can look so good.

If this guy has visibility at 20-30 nm and I at 6nm, this is a huge advantage, as he can set up a merge as he wishes, and I'm racing to even spot him before the merge. Very critical in cold war and earlier scenarios.

 

I said I can potentially spot at 30nm+, but it is not something that is consistent. Typically, the target has to fly high, albeit not too high (in the right part of atmosphere (background) relative to my position). I am not certain this is related to redolution scaling, as HP G2 is 2160x2160, which isn't a big difference vs. your Pimax. This becomes even more true, when you consider Pimax having yhat resolution on bigges screens due to the increased FOV. This taken into account, the PPI would be more similar, if not almost the same. 
 

I have an unusuably good sight, so I'm not sure if all G2-users can see as well. Still, I'd wait to see with the next updates if it improves for you. 6nm sounds incredibly poor to me. I presume that you have confirmed that your Pimax is setup properly?

 

For reference, if you've every looked up at the sky and seen a passanger plane fly, being recall the size of the aircraft and its altitude (skyradar), you can basically make some extrapolations. What is the smallest part of the aircraft that you can reliably discern, and how big is that part of the plane. You'll quickly come up with the fact (assuming you have 20/20 or better eye sight) that 6nm in DCS for a fighter jet-sized target, is actually pretty poor.

[zerO_crash]

@Dallatorre Good video overall. You should however test spotting against sky, not ground. Spotting targets below the horizon (particularly against ground), will always yield worse results than spotting against a blue sky. Afterall, we are talking about maximums here, and in best conditions possible conditions.

 

There does seem, however, to be a problem with the rendering making "jumps" between close/far rendering of an object, depending on how close or far it is away from you. It doesn't seem to be a smooth transition. There is definitely room for improvement with regards to this.

[draconus]

1 hour ago, Dallatorre said:

...when the jets are closer than the dots:

I really don't understand this one. The dots are the jets when far enough to be a pixel. There's no two separate objects.

1 hour ago, Dallatorre said:

• The single-pixel rendering of a plane is problematic for VR headsets, most headsets out there require display compression and image warping for proper projection, resulting in single-pixels that get often culled from the image the eye can see.
• There should be a distance band where the dots disappear to render the pixel-model where both coexist, to make that transition less jarring, maybe this will mitigate the issue with the zoom making planes disappear/appear.
• I have seen multiple times planes under 3nm from me disappear, I have it on video, I think this is the biggest issue and people complain "I don't have the dot", it's not for far targets. I think this effect is noticeable only when the engine has to render more than one plane and ground units, in a dynamic environment.

Can you reproduce it with the available tools and save a track? If there's something wrong devs need to know what aircraft, at what range, on what resolution disappears, but you can't expect devs to fix your problem if your HMD gets blurry because of low link bitrate.

I don't experience "jumps" between dot (or disappear) and rendered aircraft but I understand it may be only visible on high resolution headsets but maybe it's only some aircraft LOD issue.

1 hour ago, Dallatorre said:

Yes spotting will be generally easier...

This is what ED tries to avoid in the simulation part. Accessibility options were always available.

[peachmonkey]

9 hours ago, TAIPAN_ said:

I'd be curious to know if @peachmonkey is using any kind of eye tracking or high supersampling.

At the end of the day - the solution I mentioned is still valid for all the posts behind me - that is give the people a way to reduce the size of their own dot without affecting other peoples dot size. The #1 complaint I see is that someone says their own dot is too big and they want to reduce it to how it was before (for them).

 

I use Varjo Aero with 39ppd configuration, and DCS with DLAA + Quality upscaling (.50 sharpness). I realize my use case is one of the extremes (and requires a serious machine with 7950x3d + 4090), however this config allows me to achieve the best looking visuals in DCS, i.e. very high res picture with smooth presentation. Running without upscaling (i.e. MSAA x2 or x4) also works fine, but the amount of shimmer and sparkling ruins the experience, and running DCS without MSAA anti-aliasing at all makes the game look like a Vegas sparkle fest.

If ED can also focus their spotting efforts on configurations with heavy-antialising then perhaps both VR and monitor camps can reconcile their differences.

Or maybe I just want too much, and instead should scale down my expectations.

[Why485]

What @Dallatorre is talking about, and the "transition area" that a lot of people complain about with the dots is how at close range dots will inelegantly be removed, leaving an aircraft too small to easily be seen underneath it. It looks strange, and has the ironic effect of planes further away being easier to see than near planes. This has been an issue ever since the dots were introduced, and even after they were updated, this hasn't changed because the changes don't affect the minimum range at which dots appear.

A simple solution, and I think this is what Dallatore is suggesting, is to either make the dots disappear closer in, or not at all. Let the aircraft underneath the dot organically appear out of the dot rather than the dot just disappearing. This would completely eliminate that rough transition distance aircraft are too close to get a dot, but too far for them to be rendered in a way that makes them visible in DCS. The tradeoff here is that the distances at which you could identify a target would be reduced, since the dot is visible closer up, but by how much depends on many factors, and zooming in would clear up the image because that would make the aircraft larger on screen than the fixed pixel size dot.

This is a very crude drawing/graph, but it's the easiest way I can visualize this without booting up DCS and taking screenshots. The precise distance this happens will depend on your field of view and the physical size of the object itself. This is not a problem unique to VR. At a mild zoom in 2D, I'd say this happens at roughly 1 to 3 miles, but again this distance will vary dramatically depending on FOV (VR uses far wider FOV than 2D) and the the model in question.

It can happen that various things coincide (resolution, FOV, model, etc.) where the dot disappears at just the right time that the transition isn't noticeable, but I would say that for the majority of players this doesn't happen, and it shouldn't be on the players to figure out some perfect combination of hardware, graphics settings, third party VR rendering tools, etc. to get this right.

Edited August 14, 2024 by Why485

[Dallatorre]

6 hours ago, draconus said:

I really don't understand this one. The dots are the jets when far enough to be a pixel. There's no two separate objects.

but you can't expect devs to fix your problem if your HMD gets blurry because of low link bitrate.

I'm pretty sure that the graphic engine internally plasters the 'DOTS' over the models when they're too small to be a single pixel, and from what i;ve seen in game that is the case, they're 2 separate "entities" or they follow different code flow paths.

And i did remove the HMD from the loop just because i might get this comment. The recording was done PC-side.