neidu3

joined 7 months ago
MODERATOR OF
[–] neidu3 3 points 1 hour ago* (last edited 1 hour ago)

That makes me think of how it would possibly limit the spectrum of colors we could see.

Say, if we orbited a red giant instead; like you said, red would be the center color in the spectrum that constitutes our white. Our perception of color would revolve around red as a baseline, and the visible spectrum would go a bit up and a bit down from there.

Since the width of the spectrum can only be guessed, let's presume that it would be as wide as our vision here on planet earth.

Therefore, space Newton would use a prism and discover these colors:

  • infra-infra-infra red.
  • infra-infra-red
  • infra-red
  • red
  • orange
  • yellow
  • green

So our TV remotes and motion sensors would emit ultra-green light. Or as some crazy Sci fi writers would call it: blue.

[–] neidu3 14 points 5 hours ago

Not really. Someone will disagree with me eventually, but that doesn't mean I'm interested in debating or earning their approval somehow.

[–] neidu3 5 points 6 hours ago

I used to work with a brit who sometimes said "having a bubble" for "having a laugh".

[–] neidu3 2 points 6 hours ago

It's not littering if it's lithobraking.

[–] neidu3 4 points 9 hours ago

Dr Semmelweis

[–] neidu3 2 points 10 hours ago* (last edited 8 hours ago)

Oh she's there, it's just that she's inside the space capsule.

 

From Dean Hall on Discord:

You can see the SDF font gauge is working great and there is almost no aliasing whether the gauge is large or small. Might be hard to wrap your head around, but in drawing these gauges no mesh or texture* is involved at all. (technically, a texture is used for the SDF font but not in a traditional texture sense). This has a lot of benefits I covered earlier. You can see animated mechanical rollers for representing numbers in this demonstration as well. These gauges are generated entirely from runtime "mod data" meaning the way we make them is the same way modders do. The shader itself is able to access a global shader binding that is very efficient when it comes to transfering data. Individual shaders/gauges don't need special bindings for data, they have access to all the data. Allowing gauges of incredible complexity, for only the per-pixel cost of the shader itself.

 

From Dean Hall on Discord:

This shows the early progress on multiwindow rendering, showing me "undock" the settings menu seamlessly. When the settings window is dragged outside of the main viewport, a new application window gets created. This has tremendous use for many situations, especially so when rendering different cameras such as when doing interplanetary transfers. Hopefully this video gives a good indication to people the power of this functionality.

 
 

From Dean Hall on discord:

A lot of progress has been made improving a lot of our rendering process for many things, and terrain is one of them. This video shows off progress that has been made, with a great demostration of the seamlessness we now have with spherical billboarding. It is important to remember this video is taken live, realtime, from within the simulation at realistic scales (although the height of lunar features is exagerated for testing).

This video on discord is at 30FPS and and HD only, for a 60FPS and full HD download at: https://drive.google.com/file/d/1bb6F98k1nTESsR0I3YJJDuRwZWC-q9E8/view?usp=drive_link

(content creators welcome to use as always. attribution is not required)

HD version mirrored here

[–] neidu3 4 points 13 hours ago* (last edited 13 hours ago) (7 children)

Took me a while to realize that Caliber is roughly inch/100. Once I did I no longer needed to memorize them.

[–] neidu3 7 points 15 hours ago* (last edited 10 hours ago)

I've been to MUC a lot. It's big, but fairly easy to navigate. You'll be fine. I don't remember the terminal names, but you'll probably be arriving and departing at different terminals.

Also, looking up airport maps in advance is always helpful: https://www.munich-airport.com/airport-map-261352

Also, since you're flying business class, you'll have lounge access included in your ticket, so once you've found your departure gate I suggest heading to the nearest Lufthansa or Star Alliance lounge for free drinks, snacks, and generally better waiting locations than out in the terminal.

[–] neidu3 16 points 15 hours ago* (last edited 13 hours ago) (2 children)

Not really some can be pretty big, but it's usually pretty clearly marked, and and there's lots of signs and staff around that can help you if needed. Don't hesitate to contact staff if you're unsure - they're there for you. Hell, you can even ask them if you're sure for that extra confirmation.

Tip: Some things to be aware of: in large international airports some times you need to move between terminals. And this CAN make it seem like you are leaving the airport. It might even involve small trains. Trust the signage.

When I land at a layover airport, my usual strategy is to find the gate of my next departure so that I know that I'm through all security checkpoints and transfers. Then I backtrack a little until I find a good place to wait nearby. Usually a bar, eatery, or a lounge. That way I know how much time I need when the flight is about to board.

Source: Over hundred international flights over the past couple of years.

Which airports are you flying via?

[–] neidu3 11 points 1 day ago

Brown noise or bluegass.

[–] neidu3 3 points 1 day ago* (last edited 1 day ago)

If I was there in NY I'd probably be #13

 

Requirements:

  • Domain registration (duh)
  • Privacy.
  • DNS included.
  • Preferably based in the EU, but it's not a hard requirement
 
 

And are there any you would support if you had just that little bit of extra disposable income?

 

From Dean Hall, playing around with navball config:

This new system allows you to write fragment shaders to "draw" widgets on the screen. All you need to do is write it in XML:

        <Gauge>
            <Component Id="Navball">
                <AnchorUv X="0" Y="1" />
                <PivotUv X="0.5" Y="0.5" />
                <SizePixels X="256" Y="256" />
                <OffsetPixels X="240" Y="-200" />
                <Rotation Degrees="0" />
            </Component>
            <Component Id="TBar">
                <AnchorUv X="0" Y="1" />
                <PivotUv X="0.5" Y="0.5" />
                <SizePixels X="48" Y="48" />
                <OffsetPixels X="240" Y="-200" />
                <!--<Rotation Degrees="45" />--> <!-- would need to use GaugeRotVert to use this -->
            </Component>
        </Gauge>

and these reference defined "GaugeComponents" can get reused:

<Assets>
    <GaugeComponent Id="Navball">
        <Vertex Id="GaugeVert"/>
        <Fragment Path="Shaders/Navball.frag"/>
        <Gauge Id="Fragment"/>
    </GaugeComponent>
    <GaugeComponent Id="TBar">
        <Vertex Id="GaugeVert"/>
        <Fragment Path="Shaders/GaugeTBar.frag"/>
    </GaugeComponent>
</Assets>

which this means is you can write fragment shaders (as an option) to make UI elements. THis is good, but also a little dangerous. If they are small - they are really chheap. but the bigger they get, the more expensive they are
In thiis screenshot i've "unlocked" them so you can see the borders

with this unlocked, I can customize it live, which helps me choose the values to put in the XML

you can resize manually, but in that slider is useful for more precise changes around the pixel etc...

Adjustments seem really fast and simple, both sensible tweaks, and not so sensible ones:

 

Keep in mind that this is an extremely early state of the game as it's pre-alpha. It's more akin to a tech demo at this stage.

13
Procedural Navball Demo (files.catbox.moe)
submitted 1 month ago by neidu3 to c/kittenspaceagency
 

From Dean Hall:

So I've been working on moving away from always using meshes and textures for things like navballs, and doing it procedurally - in this case in a fragment shader alone. Why do this? Well, that's actually a good question. Not one I have a great answer to. Primarily, it started (like much in this project) because I personally don't like the way games look today. Everything is all washed out the heck, its all like someone has spread petroleum jelly over the screen (shakes fist at Tim Sweeney angrily). Bizarely, my time making a lot of complex avionics for my Stormworks vehicles also factored in this.

"What if", I reasoned, "we could make it completely in the fragment"? Not only will it scale really well to any resolution - the nav ball will never have texture "pinching" at the poles, and modders could very easily replace any widget whenver they want with extreme ease. An added bonus you can see in this video, I can move the widget around very easily. To be fair, this would be true (to an extent) for a mesh + texture combo - but you can see it starts to look really great even when I scale it up. This has actually been a good example for me of the paradigm shift we face using BRUTAL. A lot of things get approached from first principles, which means you start asking all kinds of really weird questions and you end up asking what should i do?.

So whats the catch? Well, another really good question. I'm not a graphics programmer, but I'm tempted to say that this procedural shader willl always be less performant than doing it with a mesh + texture combo. However, whether you will notice the performance hit remains to be seen. This is because the entirety of drawing the procedural navball is drawn in the "fragment" shader, with essentially nothing at all from the CPU. No texture, no mesh (well, I do tell the GPU there is a quad - but it doesn't exist, it is just used to make sure the fragment gets called). This has all been quite a ramble.

If I do end up continuing down this path successfully, I can make a lot of widgets like this, which can then be used for IVA gauges and the like, and then rendered on a separate framerate.

view more: next ›