3D Depth Map Based Stereoscopic Shader

BlueSkyKnight, firstly thankyou so much for your efforts and gift of this wonderful addon for reshade. I have only just discovered it after one of the members of our Virtual Pinball group was asking how to add anaglyph 3D to Visual Pinball X and I found your solution.

To give you a bit of background, Visual Pinball X is free software with hundreds if not thousands of community created pinball tables, both recreations of real world table and also unique creations.

A lot of players have created physical virtual pinball cabinets, consisting of usually 2 or more screens. Playfield screen is in vertical orientation to mimic the layout of a real pinball, and a backglass screen in landscape orientation. They also often have a separate DMD screen for scoring.

For the most part, the majority of users have systems/tables that are setup to run with Windows in landscape mode, but the playfields are actually rotated 270 degrees to match the physical orientation of the screen.

I'm attempting to get your 3D shader to run anaglyph 3D but have hit a problem which perhaps you might consider looking into for a future release. With your current software the separation of anaglyph images is set for landscape orientation screens, ie side by side. However this is an issue for pinball players as when they physically rotate their screens to play, it then causes the separation to be visually vertical ie: over/under. Which I'm assuming would cause issues and kill the 3D effect (I haven't got a pair of anaglyph glasses to test yet).

A solution for us is to run windows in portrait mode, and then it seems to work, however this requires each and every table to have it's POV (point of view) file edited to match the windows rotation.

My question therefore is would it be possible for you to implement an option to make the anaglyph separation over/under, rather than side by side? This would mean Visual pinball users could keep their existing setup without having to makes lots of edits to POV files.

Thanks for listening.

Gravy wrote:

My question therefore is would it be possible for you to implement an option to make the anaglyph separation over/under, rather than side by side? This would mean Visual pinball users could keep their existing setup without having to makes lots of edits to POV files.

Thanks for listening.
 

Firt of all also from me a big thank you for this great addon.

I strongly support Gravy's suggestion to add rotation to the addon. Me and most others use Windows NOT in Portrait mode, because it messes up many generic things with visual pinball and related frontends and and and.

I already checked this out and the results are really stunning, but running Windows in portrait mode is a no-go for me, so unfortunately most of us can't use your addon.

Thanks for listening and keep it up!

Hello! I translate through an online translator, so there may be errors in the text.

First of all, of course, thanks for this shader! This is really something incredible!

But I have a question, and I do not know if it is possible to implement my proposal somehow.

I'm using a simple VR Box for a smartphone, I've tried different shader versions and different settings, and I'm able to achieve a good image and depth effect, but close objects in many games are heavily distorted. Moreover, it is certain individual objects, and not everything that is nearby.
That is, the relief of the earth itself, the corners of buildings and so on, even close up, are displayed without distortion, but other objects disappear by half.
This can often be seen by approaching various pillars.

At the same time, the rest of the picture is excellent, without bifurcation and distortion.

This has to be fixed mainly with three sliders: ZPD Boundary, Deep Map Adjustment, and Weapon ZPD.

But unfortunately, these settings affect not only nearby objects, but also the entire scene.
And I always wanted to be able to adjust the settings only for close objects.
The closest setting to this function is: Weapon ZPD, but unfortunately it also affects the overall depth, and brings distant objects closer (reducing the overall depth effect)

Newer versions of the shader have more options for fine tuning, but this problem still bothers me.

Is it even possible to add the adjustment of close objects to the shader at all, without affecting the entire scene?

Thanks!

https://imgbox.com/oq3Pvu2D

Hello,

are there some tricks to get the depth map of some games?
I mean it seems not possible to get the depth map of games like Automobilista 2, Wreckfest, ...

Thanks for your help.
 

OlivierMDVY wrote:

are there some tricks to get the depth map of some games?

This falls under a more general Reshade question. If you're having problems with a particular game, try to find a forum post about it. If there isn't one, go ahead and make a new post about that specific game, probably in one of the troubleshooting sections. Start by disabling Depth3D, and try the built in DisplayDepth.fx. This shows the depth buffer that Reshade sees. If you can't get the depth map by going through the options there, that's where the problem is. Try disabling anti-aliasing within the game to see if that fixes things (usual culprit), or there might be another in-game effect messing with depth buffer access. Worst case, run the game at it's lowest possible settings (most graphical features disabled, no scaling, etc) and see if that gets you access. If it does, enable other features a couple at a time to determine what's causing the problem.

Within SuperDepth, If you can't get the depth buffer with either Depth_Map = 0 (normal) or 1 (reversed), it usually has to do with the Reshade not getting access from the game.

Hi BlueskyKnight,
I installed the latest version of Reshade V5.0.2 on my new Windows 11 setup, still using the old philips WOWvx TV screen. And I find that the Depth3D_WoWvx.fx won't compile. It states a few errors:
(760, 14): error X3013: no matching function overload for "Weapon"...
(905, 5): error X3004: undeclared identifier 'RE'
(1284, 4): error X3004: undeclared identifier 'RE'
(1339, 4): error X3004: undeclared identifier 'RH'

I realize that I might be the only person in the world using this shader, , but it's pretty good with the TV I have, would be great if you can resolve the errors or give me a hint how I could fix it myself. Thanks a lot if you can spare some time!

I will fix this.

Hello, not sure if this is the right place to report things back and mention bugs, but I remember using this shader like a year or more ago with great success setting it to over/under as it seems to give clearer image in halved resolution aspect.

Anyways I tried to do same with Depth3D shader (as vr one lacks over/under option) and while SBS works perfectly over/under seems completely messed up, like it makes no sense the direction shader separates the stereopair, it seems diagonal and too far for the settings I have put. I tried all options that allow to move the images about but considering the separation direction is wrong at the core even somewhat aligning them doesnt give 3d impression.

And Ive looked at each image in the HMD separate eye, each is fine by its own, its that they do not match as stereopair no matter what adjustments Ive tried.
Games Ive tried it with are Exanima and Eldern ring and both work perfectly in SBS mode.

TL/DR: over/under seems to be broken

Hi BlueSkyDefender,

I've just registered here purely to say thank you ever so much for your awesome Depth3D shader!

I when I previously had a 3D TV I used Tridef but always had limited and variable results with it. After recently getting a 3D TV again I turned to Depth3D and almost instantly have achieved 3D in my games better than I though my TV was even capable of! Your depth customisation options are so powerful, and I've tweaked some games to near perfection in next to no time. Depth3D has made me fall back in love with 3D again, so much so that I'm pretty much ditching my PS4 and going back to PC gaming purely on the strength and capability of your work. I'm even inspired to make a YouTube channel to try and keep 3D gaming alive using Depth3D - it's too good to not spread the word!!

I'm very poor due to health and employment circumstances so I'm afraid that I can't make a donation right now, so I hope my words will be enough thanks for now. I'm ever so grateful and you've given me the ability to enjoy my games in a way that I never thought was possible. I'm a game modder myself so I know first hand how much work goes into any gaming project, so you also have my admiration and respect.

Thank you ever so much again,

Chris.

Before that, I had already written here about the problem of near-field objects that moved out and disappeared. With the release of the new version, this problem has been solved, thank you! Either I had the wrong settings, but I don't always have free time to check everything exactly, so I can't say right now.

Now I want to share my observation. When configuring SuperDepth3D_VR in Doom Eternal on different versions of the shader (2.7.1 and 3.2.0), I noticed some feature, which is why the old version seems better to me.
The case turned out to be in "ZPD_Separation". I hope I can explain it clearly.
When adjusting ZPD_Separation in the new version of the shader, the depth is enhanced more evenly, but it seems to me less realistic.

When viewing the depth in black-and-white mode, you can see that in the old version the gradation of the depth of the foreground is stronger, and from an average distance the image is already quite white (which means its remoteness). And the further away, the smaller the difference in depth. And in fact, from an average distance, the image is almost white (distant). Reducing the gradation of the depth of distant objects and its amplification of nearby ones creates a more realistic perception. At the same time, it is comfortable for the eyes (the stereo pair is not too far apart).
And vice versa - a uniform gradation of the depth of all plans, creates a feeling of looking into a toy house, or simply excessive proximity of objects near the character.
In the new shader, it was not possible to adjust such a ratio of near and far as in the old one, and the image still turned out to be less comfortable.

I recorded a small video example. I have corrected SuperDepth3D_VR+.fx and Overwatch.fxh of both versions (2.7.1 and 3.2.0), so that they do not conflict with each other, and so that they can be quickly switched by changing presets. The video shows the contrast difference in the two versions and how the change in the ZPD_Separation setting affects.


In any case, thank you for all the work done!

There is a conflict here. For non VR screens that toyification effect is wanted.

But, for VR it will mess up scale. So maybe I can add a toggle to disable it globally on the VR shader only.

The Next version of the shader will be moving over to a manual method for balancing. Since I was able to profile over 200 games.

Hi BlueSkyDefender,

I'm still absolutely loving your Depth3D utility within Reshade and I'm getting some awesome 3D results. However during my setting up of Depth3D, my experiments have found a bit of a problem that I can't find the settings within Depth3D to fix. From what I gather, Depth3D seems to have the capability to do what I'm trying to do, but I just don't know if my approach is possible in Depth3D's current form.

I don't like asking for you to spend time on my feedback as I'm already very grateful for what you have given us already, but I think you might be rather interested in my findings as I've used Depth3D in a way that's quite different from the default settings, which basically transforms the 3D experience quite significantly. I don't mean any disrespect with my words or video so I sincerely hope you don't take it that way, I'm just providing my feedback as it could change how Depth3D works "as standard" (as in without much or even any setting up).

I thank you ever so much in advance if you are able to have a quick look at my video please, but either way, thank you very much for Depth3D as it currently is - playing games in 3D is awesome and it should stay that way forever! 

So why is it different from Display Depth? Display Depth also has it's own linearization of said buffer.
You can't rely on it to give a "proper" value or making it like "Raw Depth."

The depth buffer in SuperDepth3D is a bit more complicated that that one and peeping it does nothing but confuse people.
I also hide information in it and in it's channels it's also fully Dynamic.

Other information here about SupperDepth3D:
SuperDepth3D Linearization of of the z-buffer has to account for many different changes and objects in your image.

It actively has to scan the screen to check for intrusions and things that will cause issues in the image. This means I have to chang the depth buffer in such meticulous manner.
To do this a standered system was created for many games where stored values are used and set in a quiet way on startup.

ZPD adjust for Zero Parallax Distance/Plane so if you push it too far it lets things intrude the screen boundary. This also effects the depth buffer and make no longer a 0 - 1 range.
The option under it lets it set a condition that allows the shader to push the objects back in at a set rate.

This also means there is information the shader reads in the depth buffer you can't see with your eyes. So do not rely at all on the the debug view at all. What you see is not what you get.

I myself don't use it for creating profiles. I only use it for checking for missing items or to see if it's broken or offset.

The weapon hand options Can affect depth. But was made for near plane and should be adjusted for that. So most of the time ignore them unless it's a FPS.

Auto Depth was made to prevent extreme pop-out when a object fills your entire screen. This lets you focus on things. So it's a protection value.

TLDR:
There are a lot of misconceptions in the video about the shader because of the debug view. We can't read the depth buffer the way the Shaders does and
since the shader reads depth in a Dynamic way. What you see is not what you get.

More information:

So looking at your video. All the depth is pushed back to the background and there is little no no depth in the foreground. This will lead to strain when looking at the distance.
Also incorrect when it comes to a stereo image.

This is the output of BL 3 Side By Side in First person in the new update.

BL3 50 Divergence

BL3 75 Divergence

BL3 100 Divergence

Thank you so much BlueSkyDefender for very kindly taking time out to view my video and provide your perspective on my feedback for SuperDepth3D (SD3D). I'll offer this response to your last 2 posts just as discussion and I'm not complaining or criticising about anything - Depth3D is obviously your project and I don't want to seem like I'm saying that it should be changed or be anything else than it currently is.

Display Depth also has it's own linearization of said buffer. You can't rely on it to give a "proper" value or making it like "Raw Depth."

Understood, but my enquiry of having Depth3D's depth map look alike the "Raw Depth" was based on the fact that as it currently stands Depth3D already does a very good job in being able to almost universally recreate games' "Raw Depth". The consistently high quality of this method you've applied across so many games means that I use only one personalised profile as my standard settings (and only adjust whatever required settings for the SD3D depth map to be detected). Again I know your depth map and the raw depth are two different systems, but you've already achieved such a close emulation of raw depth in Depth3D that maybe with a little adjustment (if possible) then we could quite well have "proper" "raw depth".

SuperDepth3D Linearization of of the z-buffer has to... actively scan the screen to check for intrusions and things that will cause issues in the image.... To do this a standered system was created for many games where stored values are used and set in a quiet way on startup.

I'm unsure what you mean by "a quiet way" - if you mean like a safe or highly compatible way then I understand, but again using the principle that my slightly adjusted settings and approach for raw depth emulation are just as "quiet" and translatable / compatible over so many games, then just in theory my settings / approach are just as harmless, non-problematic, and won't break anything. I'm not directly suggesting that my approach should be used as standard as I fully respect your methods & approaches to 3D and I can see why people like SuperDepth3D as it is, plus it's obviously your program and project, but instead all that I'm saying is that it would just be so awesome to have it possible for SD3D depth map view to be able to effectively emulate raw depth map (as it already nearly does) just for the SD3D users that might want to setup 3D that way in their games. I recently came across some per-game customisations that you made (was it called overwatch or something?) but I couldn't work out if you had configured 3D effect per game or if this was "just" a compatibility measure. As there doesn't seem to be any game specific pre-defined settings as standard with Reshade builds, I am left to presume that it's up to each user to try and find the best settings for themselves.

... there is information the shader reads in the depth buffer you can't see with your eyes. So do not rely at all on the the debug view at all. What you see is not what you get. I myself don't use it for creating profiles. I only use it for checking for missing items or to see if it's broken or offset.

If you're referring to your "Depth Map View" as the "debug view" then yes we are seemingly using that same system for different purposes. While I totally agree that the SD3D depth map (view) does not define the final 3D image in it's entirety and by itself, there is an obvious and clear correlation between what information is in the SD3D depth map and how game world geometry detail is available to be seen, relied upon, and used to calculate the final 3D image. In very basic terms, if the SD3D depth map doesn't contain a great deal of detail / depth / variance, then likely neither will the 3D effect itself. Said with upmost respect, a major principle of my video was to show that the default settings have limited scene depth range compared to my different approach (as demonstrated in my SBS presentations & Photoshop analysis of SD3D depth map). This difference of depth map approaches applied into normal gameplay should've been very apparent in the Borderlands 3 benchmark? Put simply and whether intentional or not, you can tell by just a glance at a SD3D depth map how the basis of the 3D effect will be presented, hence why I emphasise the importance in its role. I again agree that this is not all that is required for 3D, but to me it's still a very important & significant part of how SD3D interprets & uses game geometry for the final 3D effect. My overall point here is that I think you might be underselling the significance of the SD3D depth map and consequently may not be using the depth map / debug view to it's full potential and benefit toward the 3D effect. I've worked in that view a lot as it's the basis of my approach and I've achieved a result that (even currently flawed) still offers significantly (almost double) 3D depth range potential & capability than default.

The weapon hand options Can affect depth. But was made for near plane and should be adjusted for that. So most of the time ignore them unless it's a FPS.

Totally understand what you made and use these options for and there's obviously nothing wrong with that, but I respectfully say that you might be disregarding that in reality, and again seemingly unintentionally, these "weapon hand" options are very powerful and they effectively unlock the ability for SD3D to help to see and calculate significantly more scene depth than "typical" 3D does. This single customisation aspect alone is why I love SuperDepth3D so much and why I rate it way more than any other 3D software. Building on what you have already achieved, surely my suggestion for a slightly more accurate / detailed linearisation of SD3D depth map would only offer even more potential for better 3D in terms of customisation, personalisation, versatility and 3D effect in general? As I said in the video, I totally get how and why you've setup your program as you have as it's an approach along the lines of 3D movie industry standard, but if you analyse 3D films closely the use of 3D varies shot by shot - some are as in SD3D (mainly "depth" with close-range "pop-out") some are just "depth" with no "pop-out", and some have a broader and more equal distribution of both (typically in close-ups). As games are "shot" or presented very differently in that we have vast environmental / world scenery to look at a lot of the time, my approach suggestion has been mentioned purely to compensate, accommodate, and adapt to this environment. 

So looking at your video. All the depth is pushed back to the background and there is little no no depth in the foreground. This will lead to strain when looking at the distance.

While I respect your opinion, I think that we again aren't literally seeing everything in the same way. Although I totally agree with you in that if you compare a very close plane / gun 3D range your approach has loads of depth and mine has none other than being in the extreme ("popout") foreground, I'm unsure if you're recognising or just seeing that my settings redistribute both "depth" and "pop-out" systems and provide more 3D depth range in general. Looking at your Borderlands 3 pictures it seems that your 3D preference is to (primarily) focus on the extreme close plane / gun region and fore / middle-ground, whereas (as my video showed) my preference is a simple 50/50 split over the entire foreground and background, or entire scene in general. It also seems that I'm different to you in that I don't look at 3D as being either "depth" or "pop-out", I simply literally see it all as depth (potential) being just foreground depth and background depth regardless of the actual scene's depth and focal point. Either way, both area systems offer depth so to me, discarding "pop-out" as potential useable scene depth is not only a waste of potential and results, but perhaps is a single-track approach for what in reality can be approached & used in different ways. So although I can logically understand your comment of all the depth being pushed to the background, I simply and literally don't see it that way and I still believe that it can (potentially and subjectively) offer and be so much more as my video demonstrates. As for eye strain, I can't comment for VR, but I've not experienced any of this with my settings using 3D TVs, and in fact I find that most games (both first & third person) require a focal point in / around the vertically middle of the screen - which as I almost always don't need to look at the extreme foreground / gun area is another reason why I push the 3D effect further into the scene, so that I can literally see and appreciate the 3D effect more. Also, as my earlier video shows, there's currently a 20% loss of SD3D depth map information at this crossover point transition (50% jumping straight down to 30%) plus around a further 10% loss for the extreme foreground gradient issue, and if anything I regard and am addressing this matter not only for visual improvement but to improve any eye strain or even extra / unwanted neurological interpretations & adaptions of a currently imperfect presentation (using my approach). 

As for "little" to "no depth" in the foreground - your description isn't anywhere as bad as what I actually see. If you compared the regular 3D view of my Borderlands 3 benchmark (occasionally pausing it and analysing the more linear and progressive nature of my settings) I personally see that the general scene foreground is rather prominent, effective, natural, and has a good level of detail & range of 3D - and considering that this is with a currently "flawed" depth map emulation in the foreground (60% depth map loss / inaccuracies), imagine how much better and effective the 3D effect would be if it wasn't flawed? This matter was the origin and reason of my feedback. As I haven't been able to distribute the ZPD / weapon / foreground depth range equally and consistently, the 3D effect is also unequally inconsistent in it's depth progression, so my video was just a demonstration of potential and not final result hence why you won't be that impressed by it!

The following 2 short videos below show how much 3D / depth I've already been able to achieve with ZPD options covering the foreground - in the first video the 3D effect within the tree village area is all within both the ZPD and foreground convergence systems, and the second video is simply another demonstration of this compared to your latest SD3D's release (in response to your pictures of Borderlands 3 that you kindly provided)...

Also incorrect when it comes to a stereo image.

Er, how and / or why? What is correct and what is incorrect? This I don't understand as it works for me!? If you're referring to the Borderlands benchmarks then I did some clever editing to show 2 side-by-side 3D demonstrations simultaneously (hence 4-way imagery) but everything else is just "normal" Half SBS 3D?!

Two quick last things to mention:

Although my settings and approach won't suit everyone's preferences, and some could argue that I'm spreading the 3D effect too thin and across too wide or far a range (which are valid subjective perspectives) I still think that my way is at the very least a strong alternative and mainstream approach to 3D compared to how its normally implemented and therefore perhaps stands some reasonable merit for consideration (of if my approach could be made possible within SuperDepth3D). Obviously if this isn't possible then it is what it is, but I remain convinced / hopeful that this is possible as we're already so very close to perfection and SD3D already runs a system that seems perfectly capable of my approach, and I would just love to see this happen if at all possible please?

My last point is (and this might be an obvious 3D related fact) that I've simply found that my settings work better when using a bigger display / TV compared to a smaller one. I have a 55" Sony 4K 3D TV which has good 3D, and I also have the same model TV but 65" (which is the TV used to setup SD3D) and this size increase provides a significantly better 3D experience. So if anyone thinks that my settings are somewhat weak or lacking in 3D, then smaller TV sizes might just be a factor there.

I won't persist with my feedback any more unless further discussion is ok, but I'm just clearly very eager to find out how SuperDepth3D progresses as I'm on the verge of committing to a GPU upgrade purely for 3D gaming and I hope I'm making the right choice here. Either way, thank you very much once again BlueSkyDefender for SuperDepth3D itself, and for all your time and help on this forum! 

Here's my .ini code used for the above Borderlands 3 videos, using the very latest build of SuperDepth3D. My approach to 3D might make more sense if games are played using it instead of just watching & analysing videos!

I can't edit my previous post for some reason, so for a quick update / correction:

The code I posted and therefore the videos contain a small but significant error, I used the wrong perspective setting for some reason. Please disregard those 2 videos and tomorrow (Monday) I'll post a video using the correct code, which is...

1. There is no "Raw Depth." Linearization Has to be done for any of this to work.

2. When there is a profile for a game it's more then just basic settings It can have specific fixes for the game. This is what OverWatch is doing.

3. You can't read the information in the depth buffer view or debug view the same way the shader can.

4. Your settings are removing information near the screen and it shows in your videos.

5.You have no depth near and around the camera. This why it's incorrect. You removed any precision in the depth buffer when you use the weapon hand option. You know how I know this? because that's how I programed it.

6. You do see depth in your image. But, not to scale and not how we see depth IRL. 

It's basically flat near the screen then after a few Feet it start to Diverge. I will show it below here so you can see it by overlaying both images so everyone can see it.

Image with no mask showing the location of broken depth.

This is the image with masked out area that has problems with little no to no deapth.

Oh ya compare that to the base profile.

You have depth near the cam and diminishes out to the distance.Depth perception is limited to around 20 - 40 feet or something like that, we are all different so there is not exact distance. But, it's around there. The problem with your profile is that it exaggerate this depth at the the distance causing it have improper output at a distance. Also it remove detail and depth near the cam so your weapon hand is flat.  This also extended to the world near you. This is not an opinion I am just stating a fact. 

Also beyond a certain point the brain uses depth clues.  www.hitl.washington.edu/projects/knowled...A.1.c.DepthCues.html

If you want more information about how we see depth at large distances read this  jov.arvojournals.org/article.aspx?articleid=2191614

But, because of this I prioritize Depth near the cam to create depth based on our natural human visual system.

Two extra things:
If you adjust existing games that have a profile from Overwatch. You risk breaking the fix. Keep that in mind.
If you over adjust the image to break it like in your examples. The Shader will also limit things automatically since it's out of range.

Thank you BlueSkyDefender again for your replies, it truly is very much appreciated.

Meant respectfully and nicely, I did understand you first time that there is no Raw Depth available within SuperDepth3D. I sincerely and absolutely appreciate that you've taken time out to show me where my settings are flawed in the 3D output. I won't mention Raw Depth anymore (after this post)!

You are right in that my current settings aren't setup with ocular three dimensional depth perception primarily in mind, and although I had a basic grasp of your principles I can now see why you've seemingly setup SD3D to work on that basis. That said, I've followed the path that I've taken for valid a reason that after much though I can now explain well (enough). I'll have to mention some stuff that you seem to know anyway, but it will make (some) sense in the end and I'm now convinced I have a good point somewhere...

Ocular depth perception (in a general sense) is obviously a real life triangulation between two eyes and depth / focal point within our limited depth perception range and capabilities. The further away something is the less depth information is present due to the reduction of angle within the focal triangulation point, therefore (in extremely basic terms) the nearer to the eyes an object is the more we perceive that object in 3D, and (comparatively speaking) the further to the eyes an object is the more we perceive that object in 2D.

This three dimensional depth perception going into two dimensional depth perception is clearly your choice of 3D effect and what you call "correct", and I do get it - it's (in part) how our eyes work and it's also no doubt why your program is called SuperDepth3D after all! However... 

I think it's important here to distinguish the difference between three dimensional depth perception and general distance depth perception. Your comment of depth perception being around 20-40 feet (if correct) would only apply to being able to see an object accurately in all three dimensions. In reality though this is not where depth perception stops - we have continued working triangulation further than that relatively near distance, which gives our eyes the ability to perceive discernible distances well beyond that plane, and perhaps even as far as the eye can see. This is where the balance of scale and relativity shifts - at further distances we lose our ability to define an object in three dimensions but we still retain an ability to determine the depth / distance of something compared to and on a bigger scale, such as bigger object in the environment or the environment itself.

For example, if you look at a tree close up you'll obviously see loads of three dimensional detail, but if you look at a tree that's 2 miles away in the distance, even though we can't make out three dimensional details within the trees branches and leaves anymore due to the reduction of angle of triangulation for such a relatively small object, we still have enough distance perception to triangulate roughly how far away the tree is, especially in relation to the scale of the surrounding environment.

So remembering those distinctions between three dimensional depth perception and distance depth perception, you could possibly sum up that SuperDepth3D maybe, solely, and strictly focuses on recreating three dimensional depth perception, but does not recreate any further information beyond that in order for further distance perception to be possible? I think now that this is why I personally find that using the default settings the environment becomes prematurely flat and distance which should be visually discernible has no depth information. So this is why I've attempted to extend SD3D's 3D distance / range depth in my settings by effectively borrowing some 3D effect from the closer three dimensional perspective, albeit executed with a fair bit of imbalance.

I will say that while I see your approach as being an understandably strict / theoretical recreation of three dimensional depth (obviously being a 3D utility), in hindsight I admit that my settings so far are a strict application of purely distance depth and somewhat lacking in perspective depth. I will also say though that although my settings were incorrect for three dimensional perspective depth, distance-wise they were realistic & fully utilised the environment in a 3D capacity so in that specific regard only, those settings were correct and I guess it was why I found those visuals appealing. As a side note - one reason why I had this Raw Depth emulation mindset was thinking that real life effectively had a Raw Depth Map, and if that could be emulated in game then using the literal perspective we had in the game it would present as a natural and convincing parallax effect, but again I was all about distance and not considering perspective enough.

So moving forward (and I do apologise for continuing on) - I'll now experiment with settings to try and find a perspectively realistic foreground to background method (with realistic three dimensional depth perception) but also contain as much distance depth as possible while hopefully not compromising too much on either. Your default settings are a strong basis to work from (or even the most capable settings already) so it (hopefully) won't take long to tweak. It better not take long because I'm actually going insane working all this out.

Lastly, I did read of Overwatch, but I'm admittedly confused about what this is and how it's used. If it's within Reshade then I've not come across it, but I'll look into this more as well as it sounds like it's useful and may even have some answers in there!

Thank you again BlueSkyDefender, you have the patience of a saint for putting up with me 

Quick edit: the first link you shared is broken, but after reading the second article then I can see why including "perspective clues" areas as actual 3D might not necessarily be accurate or realistic, but I personally would like the 3D effect to stretch out a bit further if possible to provide an emulation if not recreation of further depth / distance as I feel it would provide a bit more consistency in range and 3D effect as opposed to stopping or stalling out in the middle-ground area? Thanks again anyway for sharing! 

As always, thanks for this! It just continues to get better over time. I did notice that the latest version combined with Reshade 5.3.0 provides less depth that prior to the update. I can adjust for that of course, but thought I should mention it. I performed an in-place update in Greedfall and had to immediately adjust a couple settings. I think the depth was maybe half what it was before updating.

Is there an way to get Beyond Two Souls to work? I use this with Detroit Become Human and it worked perfectly. No usable depth however, out of Beyond Two Souls with any of the array of options I've so far tried.