Experimenting with visuals, in pursuit of a dreamy, half-asleep, photographic aesthetic. A sort of “fake photorealism”? Blatantly minimalist, unrealistic subjects, presented as if seen through a camera lens… trying to fake depth of field, realtime lighting/shadow, etc as much as possible within the capabilities of Unreal Engine 4 on mid-spec hardware.
Did a little sightseeing earlier with a vaguely daguerreotype-esque tonality:
The game is as experimental and visually eye-catching as I just love (in fact I adore all first-person experimental stuff like this).
And I’ve been following this project since a very first prototype came out.
So good luck with that!..
Wow this is indie game of the year kind of stuff.
Is this still in development? Its really eye catching
Thanks for the kind comments – I’m afraid I’ve been in “deep dev” mode for months now, as I was posting some things that were a little too work-in-progress just for the sake of posting something.
And while I really appreciate the feedback here and elsewhere, and it’s been great to get so much inspiration from others (and to know I’ve inspired a few others as well), I want to both keep what’s to come an unknown quantity, and to resist homogenizing the end result to conform to maximum user-friendliness. I’m all for optimum user experience, but everywhere I look in VR dev I see the same restrictive mechanics designed to not make anyone sick, and to lower the barriers to entry to absolutely nil, which in my mind is the equivalent of designing a scary movie to only be viewed on a tiny screen in a brightly lit room with the volume turned down low. Where’s the fun in that?
I’ll have more to share on singmetosleep, parasomnia, and other projects soon enough… thanks for the interest!
I really hope you haven’t thrown in the towel on this development since I think you have a really strong atmosphere and aesthetic with this prototype. There is so much potential within this kind of abstract language that you could go in almost any direction if you applied actual story driven narrative to the game play. The music is superb and the way it blends into different environment local while highlighting more stark landmarks works wonderfully.
As it currently stands though there isn’t much of any objective in this virtual simulation outside walking and looking at abstract forms which runs dry once you have walked hundreds of virtual yards only to reach a dead end with no conclusion or reward to the time spent other then an up close look at yet another abstract form… This makes walking superfluous other then a means to reach the next viewing opportunity.
There is also a control bug in the most recent build where if you press space bar it locks the camera movement to Z and X with no Y -Y movement…
this simulation could function just the same with no translate movement and a locked camera with full rotational FOV in an animated space full of abstract gimbals and gyroscopes.
I personally would love to see this turned into an actual game along with all this wonderful atmosphere and music. If you decide to do so and want to collaborate feel free hit me up!
Sorry the most recent (now two-year-old) atmosphere prototype is in such a state. It was originally intended to be just a milestone, an early and hastily-released snapshot of a work-in-progress, and just one of a continually improving series to demonstrate the aesthetic and give an impression of some of the otherwise hard-to-verbalize ideas that might be present in the final product. There have been many points since then while working on it that I’ve told myself:
“OK, time to lock down a milestone so people can try it without dealing with some of the bugs in the existing prototype, or try it on newer VR hardware, etc… but just let me polish this and fix that and maybe disable this system and that area that aren’t ready to be shown…”
And before I know it another month has passed and I’m just as far away from a vertical slice that’s ready for consumption.
I’m still working on singmetosleep – albeit not as feverishly as in those first couple months after UE4 was released, and I occasionally get distracted working on side projects – but I still have the same very strong conception as I had in the beginning of what I would like it to be before it’s “done.” I would rather not give away any story details as part of the prototype or in these development notes (such as they are); much more than what has already been revealed will be better shown fully realized in the final product.
I’d like to share more of the technical hurdles I’ve encountered and how I’ve overcome them, various design ideas, what has worked and what has not worked so well, etc. (you know, like a normal game dev log); but I’d prefer opacity over transparency at this point even if it means “going dark” and possibly losing the interest of some players. It’s better (IMHO) than the alternative of dragging the nebulous insides of this thing out into the bright light of day for scrutiny and classification.
Sorry (again) – this isn’t a satisfactory response, is it? But rest assured, the self-replicating machine elves toil tirelessly at their stations, chipping away at this rock to reveal what lies within…
**Bonus post (actual dev log content):
Why all the purple?**
[cut & paste from an informal Q&A a while back]
Q: You have a lot of purple. Does that have a significance in your work?
A: I’ve tried a lot of other shades, and keep coming back to that. I’m going for the starkness of monochrome, but not completely desaturated black/white/gray. This purple/violet shade almost implies a certain darkness, and it doesn’t have as much clearly defined baggage as the common primaries like blue (water, sky, soothing) and red (blood, fire, anger); green and yellow just look like *****; orange… well, orangeish works sometimes:
Q: It depends on the story that you are trying to tell…
A: Purple looks like the future, like a bruise, like a storm cloud, like uncertainty, like strong but complex emotion, like steel. The almost ultraviolet shade I’m going for is suggestive of the very edge of the visible spectrum of light, bringing to my mind at least the idea of the unseen, or at least the presence of more beyond what can be seen. But I temper my expectations knowing that part of the problem with choosing a specific defining tone is the wide color/gamma discrepancy between all the monitors and phones (and head mounted displays) out there: what looks violet on my monitor may look purple on someone else’s, blue on an Oculus Rift, and so on. And importantly, I wanted something distinct that stands out in a sea of thumbnails on the internet… black/purple/white is a color combination I see very little of.
Q: Why not expand your palette?
A: Why not use more complex shapes? I’m rapidly prototyping, nothing is set in stone… Trust me, I’ve tried a lot of color experiments. The color palette I use for this type of architecture and geometry is as flat and bold and minimalist as the shapes themselves. I plan to have some other parts of the game that are less, uh, nightmarey – but this is a fun place to play around in while I experiment with the tools and see what’s possible.
Great to hear you are still working on this project - the last video you posted really captivated me! And I think the purple works well - its a softer take on a monochromatic approach but still visually striking.
I have a question. Other than it simply being an art project, are you planning on adding any specific gameplay or story elements into it? The prototype I played - while pretty - was a bit devoid of actual interactivity it felt like. Seems like the perfect game for puzzles.
Story? Most definitely, but it may need to be teased out a bit…
Gameplay? Certainly no stats, or shooting at enemies, or jump mechanics.
Puzzles? Yes, but perhaps not in the most straightforward sense.
Art project? Most definitely
There’ll be environmental interactivity and problem-solving to a certain degree, though it won’t be a “puzzle game” as much as there will be things to discover as a result of figuring certain things out. Sorry, I know that’s pretty vague But I’m still trying things out to see what can be done (particularly in VR) to tell the story and engage the player, yet avoid a traditional “gamey” feel.
I feel that this medium has so much more potential than the traditional gameplay we’ve seen for the past 20+ years: I hope to someday get to the point that anyone could come to the table with no prior gaming skillset, put on a VR headset, and be blown away by the story and artistry as in many of the best films; but beyond traditional passive media, to allow for the player to put so much more of themselves into the experience, to be in that place, inside that character’s head, to explore at their own pace, to retrace their steps, repeat their mistakes (and perhaps learn from those mistakes and ultimately overcome the obstacles they face); and most importantly to increasingly blur the line, and deepen the parallels, between media and reality, character and player, to discover more about ourselves by learning where we choose to draw that line.
This is precisely the ethos behind my own projects. The term ‘walking simulator’ is often thrown around, sometimes as a criticism for games lacking traditional goal- / interaction-based design, sometimes used by developers to reclaim the term for a ‘thinkpiece’ genre. Whichever way it’s used, it’s a term which does a massive disservice to experiences that singmetosleep appears to be (along with the likes of NaissanceE, KAIRO, etc.).
I’ve just discovered this thread and I love your ethereal / brutalist architecture and choice of palette, plus the audio is superb (currently running in the background as I work on my own project). I’ll follow you on Twitter and keep an eye on this for sure.
Hey, curiosity is eating me. It’s about that really nice looking “vignette blur” post process shader (not sure if it has an official name). It certainly fits “dreamy” environments well, for example I remember that LIMBO had it too. So how did you made it? Any tips?
I know that there’s custom “radial blur” post process material on these forums, and also you can mask gaussian depth of field in the Post Process Volume settings, but it’s not really work that well.
Sorry for the belated reply, I’ve been meaning to describe the blur effect for a long time, and got sidetracked breaking things up, making toggles, etc to help portray the impact of the postprocessing.
I love the Limbo look – no doubt it informed my aesthetic quite a lot though I hadn’t really thought about it at the time. I think you’re only asking about the peripheral blurring, not the peripheral darkening… but I’ll mention both since the synergy of the two effects is important to the surreal look I’m going for. More consciously I’m aiming for something maybe akin to some of the Brothers Quay work, e.g. the video for Can’t Go Wrong Without You: His Name Is Alive • Can't Go Wrong Without You - YouTube. Well, not exactly that, but I can’t think of a better example offhand.
Peripheral darkening (traditional “vignette” effect)
Implemented with the built-in postprocess volume effects. I have a global PPV with the following settings:
Scene Color -> Vignette Intensity: 0.85
Film -> Contrast: 0.75
Scene Color -> Scene Tint: any shade with a Value around 0.5
Importantly, the extreme contrast and dark overall tint significantly increase the impact of the vignette effect:
You can see in the video that there is a major blurring artifact along the right edge of the screen (more noticeable in some scenes than others): a blurred black band pulling inward from the edge. I’ve only noticed this since the changes in 4.9+, and don’t have the technical prowess to track down and resolve the issue (thinking this is more to do with the way my mask is used in the postprocess material than the gaussian blur per se).
Peripheral blurring (dream-like toy lens effect)
I used a PostProcess material to apply a gaussian blur using a texture mask. Thanks to the ingenuity of other devs much better at this stuff than me (thanks @EdWasHere), this was simple enough in older versions of UE4 (but note the bottom of this post for a fix in 4.9+):
For anyone unfamiliar with the technique (or if the above thread doesn’t make it clear), I essentially created a new material using a SceneColor lookup node to combine the rendered output with and without postprocess settings according to a texture mask:
Using a radial gradient texture mask (actual gradient is smoother than this appears) – white passes through postprocess settings (gaussian blur in this case) that this material will be applied to, and black masks the postprocess (unblurred image):
Set the material domain to “Post Process”:
Then create a new unbound PostProcess Volume in your scene, and add the material you just created to the list of Blendables:
Then configure any settings you want to be applied by the mask (in this case a very strong gaussian blur):
Major caveat: As of version 4.9, SceneColor lookups are no longer available in PostProcess materials (you’ll get a material compile error if the material domain is set to “Post Process”). But again thanks to some clever trickery (thanks, @BlueSpud !) we can reenable it by commenting out the sections that throw the error in a single file of the UE4 source code:
Note that the most recent source code version I’ve applied this change to is 4.12, so your mileage may vary if applying this to newer releases.
I read somewhere (sorry can’t find link), the engine devs stated that this attempt to prevent SceneColor lookups for PostProcess materials was implemented for compatibility with some rendering refactoring in 4.9+ specifically for the console rendering path (which benefits console performance). However I have not experienced any issues with PC builds (again only as of 4.12), and performance impact is minimal in my use case.
Really enjoy your videos.
There is a subliminal message apparent when you watch the visuals and hear this excellent music. Thanks.
Would enjoy a 2 or 3 hour long video with this music Is this dark Psymbient, or what do you call it?
Wow, thank you so much for such detailed answer! Honestly, I thought that you have some kind of custom HLSL code inside your post-process material, but it turned out easier, but also much trickier actually. And surprisingly it works with bokeh dof too, but gaussian looks somewhat better in that case. Just too bad that you can’t really use peripheral blur (from now it has an “official name” for me) combined with actual depth of field due to its nature. A big thanks to you anyway, I had strong curiosity for this for a while. It fits with your project really well!
Here’s the whole track (about 15 minutes, be aware it gets loud): http://acatalept.com/music/?noar1h3
What would I call it? “pretty / abrasive music?” Melodic crescendo noise drone? If “dark psybient” is anything like the first few google hits I found, then that’s not it. That track was influenced quite a bit by some atmospheric post-classical drone / noise / etc I’ve been listening to lately, for example:
Siavash Amini: https://room40.bandcamp.com/track/the-wind
Rafael Anton Irisarri: Empire Systems | Rafael Anton Irisarri
Lawrence English: Another Body | Lawrence English (beware, this one is loud right from the start)
or https://lawrenceenglish.bandcamp.com…-s-bitter-grip (I like his genre classification on this page: “post-classical post-minimal post-everything”)
and a number of others.
As for a 2-3 hour video… that’s kind of a primary thrust of this “game”: an extended interactive soundscape exploration. There will be a strong narrative as well (to the degree that I’m able to make it work in the final product)… but you could just as easily ignore that entirely and search for hard to reach / hidden places, discover new sound palettes, and linger or move on as it suits you. Perhaps at some point I’ll put together some longer mixes of the soundtrack elements, or just do an ambient Let’s Play with any voice/effects disabled
No worries. Let me know if you end up implementing this in your work, I’d like to check it out!
The technique is quite easy, aside from the initial learning curve if you haven’t compiled the engine before (it was intimidating my first time). I agree gaussian DOF looks better than bokeh DOF in this specific case, and is much less performance impact than bokeh DOF. I haven’t experimented much with the newer circle DOF, though I’m intrigued by the promise of realistic lens-like DOF behavior (I’m a photography nerd), and I’d like to get some variable focal depth happening at some point to get a more convincing exaggerated shallow DOF look in some cases in addition to the existing look, almost like the “miniature” stop-motion look of Brothers Quay (mentioned above) or the early music videos of Tool, e.g. Sober or Aenema:
Okay, now I don’t know how I supposed not to write completely off topic message how Tool is one of my absolute favourite bands! Especially “Lateralus”, the pure brilliance.
Did some experimenting with realtime raytraced global illumination in UE 4.22 using some test assets from singmetosleep: https://forums.unrealengine.com/deve…on-cheat-sheet