SVG to PNG via Command-Line (Inkscape)

SVGs are awesome, but not everything supports them. Unity especially. For SVGs stored on your Linux server, this is a quick Command-Line solution.

____________________________________________________________

Get To Your Files

Whether you're running this one file at a time or in batches with a script, you'll want to start in the folder your svgs are in.
If you don't want to...
I think you can indicate the path within filename, but I'm not sure.

____________________________________________________________

Install Inkscape

Inkscape has to be installed on your server for this to work. This should be totally easy, but if you might want to check the docs.
Essentially, you enter this as root.

apt-get update

apt-get install inkscape

____________________________________________________________

Run the Command

This will call inkscape to read the svg file and export it as a png. If you want to branch out, here's a good reference
Run this code on all the svgs This will make a png copy in the same directory. :D

inkscape filename.svg --export-png=filename.png

If you want to define what the 'empty' background color is, try it this way

inkscape filename.svg  --export-background=ffffffff --export-png=filename.png

____________________________________________________________

Enjoy your PNGs :D

____________________________________________________________

All blog posts by UnityGirl are inspired by my work with Brunelleschi: Age of Architects on the Aesop Games team. Download the current Unity client On IndieDB

Unity Tips: Combined Mesh into Highlight (part 2)

This is Part 2 of making a single-mesh highlight for a several-mesh object. In the Previous Post we discussed how to combine the meshes into a single mesh. Now all we have to do is use that new mesh as a highlight.

____________________________________________________________

____________________________________________________________

Shaders
Because Shaders are connected to Materials that effect 1 mesh at a time, we had to combine all those meshes into one outline-able mesh. Now we just need the outline.

Though shaders are a fairly deep hole to jump into, the Unity Community is there for us with a few commonly desired shaders. I found several useful resources Here. Select Silhouette-Outlined Diffuse and copy the variant that's right for you. There's Standard, Bumped, and Outline Only. I like Outline Only but you may have other needs.

____________________________________________________________

New Shader
Copy the Shader script. In Unity, make yourself a new Shader, open, and paste shader script.

____________________________________________________________

Now The Outline
This part's super simple

• Make yourself a new Material.
• Assign the new Outline Shader
• Adjust color to preference. I like bright greens and blues.

____________________________________________________________

The Outline-Object
Finally, we combine these two techniques into a complex outline.

• Take your original multi-mesh object, with it's "empty" parent-object
• From there make an empty game-object child with scaled about 1 to 1.
• Assign the Mesh you made earlier
• Assign your new Outline Material
• Adjust to preference

____________________________________________________________

Tadaa!

With these steps we've completed the Multi-Mesh Object Outline path. I hope you enjoy your outlines derived from this method. 

____________________________________________________________

All blog posts by UnityGirl are inspired by my work with Brunelleschi: Age of Architects on the Aesop Games team. Download the current Unity client On IndieDB

Unity Script: Oculus Selection with Coherent UI Web Views (3 of 3)

The last cool thing in this Oculus Pointer Trilogy post is how to work with Coherent UI Web Views using the Oculus Pointer we made in post 1

To see how to select 3d GameObjects (like buttons) check out post 2

____________________________________________________________

Raycasts

If you read Post 2, you'll remember I mentioned that Raycasts are kind of backward from the way you normally think of input going (player-out). Fortunately Coherent UI is pretty much self-contained so we can put the Raycast in the UI script.

Basically: we're going have Coherent UI respond to the pointer, rather than having the pointer trigger Coherent UI.

Why?
A combination of reasons coming together to a single solution

• Unity mouse function actually IS a back-end Raycast
• Coherent UI calculates what you clicked on based on the view texture's x/y coordinates
• RaycastHit.textureCoord is the only way I know to find and send the texture x/y

____________________________________________________________

Coherent UI System

This is the script that comes with your Coherent UI system that we'll be adding to in order to do this little trick. I won't be showing the full altered script, but I'll give you the placements for our entered code snippets.

____________________________________________________________

Public Variables
The first thing to note when you're looking at the Coherent UI System script in the editor is that almost all of it's values are private. We can't see them, therefore can't check mid-UI navigation.
It's also worth noting all those specialized CoUI variables don't share well between functions.

You're going to want to add a few public variables for you to work with inside and outside the script. I've made comment lines on either side to make it obvious where I've changed the script.

____________________________________________________________

Finding Your Assets
We're going to do this in two ways. OnLevelWasLoaded(), which we'll add, and through Start(), which we'll add to. This will tell the script whether or not to look for Oculus interaction

Here I have it looking for any of the Oculus prefabs I use that could interact with Coherent UI. If any are present, expect to interact with an Oculus pointer

• Make sure to use your own asset names

____________________________________________________________

void TrackInputFocus ()
This is a void() that runs every Update cycle. Find it, we're going to put our Raycast here.

The script:

Note

• See that hanging "} else {" at the bottom? That closes "if (oculus) {"
• don't forget to close if after the end of the next (original) raycast 

The Logic:
What we're doing here is

1. Defining a ray from the right eye out through the pointer. This defeats the distortion caused by the Oculus program.
2. Making a Raycast with that ray
3. If the raycast hits a CoUI view, it records the x/y coordinate hit on the view texture and responds accordingly
4. If the mouse is clicked while the raycast is pointed, it will click that point on the view.

____________________________________________________________

Mouse Position

There's one last thing we need to do at the end of the script to update the mouse position in Oculus mode.

____________________________________________________________

Final Step
There's a special way to set up your Oculus pointer-using scenes.

1. Remove the OVR Manager script from your Oculus player
2. Make a GameObject tagged MainCamera that does not have a camera component on it
3. Put the OVR Manager script on that

This arrangement was found by trial and error. I know only that it has something to do with Coherent UI needing a Main Camera and getting bugs doing it other ways.

____________________________________________________________

This is actually remarkably simple and mostly copied from other bits of the same function. The really tricky part was identifying the texture coordinate technique. The rest is just variable management. :D

Tadaa!

____________________________________________________________

All blog posts by UnityGirl are inspired by my work with Brunelleschi: Age of Architects on the Aesop Games team. Download the current Unity client On IndieDB

Unity Tips: Lighting vs Ambient Light

Simple solution to a basic processing issue. Especially if you don't have Unity Pro and / or haven't mastered baked lighting, lights slow your game down.

_______________________________________________________________

Lighting Your Scene

Point 1
Every individual light rendered on every texture has to be processed.

Point 2
Beginning scenes are hella dark. You HAVE to use lighting to be able to see your assets.

Point 3
You don't actually have to use lights to fix this problem.

• Hit Edit
• Pan down to and select Render Settings 
• Render as in how things show up
• Click the dark gray box labeled Ambient Light. 
• I usually go with a very light gray, but pale yellow is good for simulated sunlight
• blue is good for night or dream scenes
• green to enhance a forest glade
• Best Part: pick a light color and you won't need lights to see!

_______________________________________________________________

Also in Render Settings

You might want to look around further in this menu area for other useful ways to tweak your scene.

• Skybox - a special texture set that creates the illusion of 3d sky
• many skybox sets are available in the assets store
• you can also follow a guide on how to make your own skyboxes
• Fog - a player view shortening and somewhat magical effect, especially set to odd colors
• obscures overhead map views, but great for immersion
• Halo and Flare - camera effects I haven't actually played with yet.

_______________________________________________________________

Secret in Quality Settings:

Go to Edit -> Project Settings -> Quality and you'll find an interesting setting.
"Pixel Light Count" Turning this up increases the number of pixel lights that render completely, but has performance repercussions.
I turn this up for high-quality urban promo shots and down for everything else.

____________________________________________________________

All blog posts by UnityGirl are inspired by my work with Brunelleschi: Age of Architects on the Aesop Games team. Download the current Unity client On IndieDB

____________________________________________________________

Wondering where Part 3 of the Oculus Pointer Trilogy is? I'm not done with it yet. :D

Unity Script: Oculus Selection Logic (2 of 3)

Oculus selection is something I've seen a lot of examples of in demo games, but almost no documentation on as to how to manage your selection. Here's some basic Unity logic I've learned while working on Brune's Oculus integration.
See how we made the pointer in post 1

____________________________________________________________

Goal

Selection: plain and simple. Choosing between game objects, which way to look, and what to interact with. Mainly this post is about how to use a pointer moved by the mouse (last post) that sends clicks where you want them to go.

____________________________________________________________

Obstacles:

Note the curve, and how each eye sees further to one side

Camera Distortion: It's important to keep in mind that the Oculus uses 2 in game cameras distorted together to simulate your eye-based vision inside the Rift. This also means that it's easy to aim at something with "one eye" and not really hit it within the distortion.







Mouse Function: I just haven't managed to get the windows mouse-pointer function in Unity to work inside the Rift. That said, OnMouseDown does work if referenced indirectly such as

public bool triggered;

void OnMouseDown () {
       Application.LoadLevel("MyScene");
}

void Update () {
       if  (triggered) {
          OnMouseDown();
       }
}

I use this to have Oculus and regular mouse function in the same script with the same click-response code.

____________________________________________________________

How to Trigger:

Building text and highlight appear when building is triggered

1. Make a cube child object of the pointer named "PointCollider" (or whatever)
2. give it a cube collider set to "Is Trigger" and a rigid body set to "Is Kinematic"
3. transform it such that it makes a long thin line through pointer object on out into the world as far as you need. (make sure it goes far enough backward to tough the player)
4. Put a collider and this snip of code on your click-responsive assets:

void OnTriggerEnter(Collider col) {
// this part keeps your script from responding to all collider interactions, only your pointer
       if (col.transform.name == "PointCollider") {
             triggered = true;
      }
}

void OnTriggerStay(Collider col) {

       if (col.transform.name == "PointCollider") {
             triggered = true;
      }
}

void OnTriggerExit(Collider col) {
       if (col.transform.name == "PointCollider") {
             triggered = false;
      }
}

____________________________________________________________

Example In Use

I've done it slightly different in this JS example, using if (Input.GetMouseButtonDown(0)) because OnMouseDown is linked to a GUI confirmation irrelevant to the Oc player.

____________________________________________________________

Raycasts

If you ask around about what to use for generic direction-based selection, much of the Unity community will suggest Raycasting. This is an invisible line you send out from a specified origin along a specified angle. If it hits something, it records what it hit.

Why I don't use them if I can help it:

a) conditionals for raycast hit can only be calculated in the script that sends the raycast. This means that

• button-like scripts are useless 
• trigger area scripts are useless
• you are mostly reliant on tag-management to sort your responses
• effecting other objects is a logistical pain

b) selecting what the raycast hits is a pain in the butt.

c) programming for Raycasts involves programming with catcher-scripts or backward-pointers that I find not-so-fun.

Why Raycasts are useful anyway and you'll see them later on:

They are the only way (that I know of) to select a point on a TEXTURE and not just a gameObject. Therefore we will be using a Raycast to handle the Web Vewier later.

____________________________________________________________

Wanna see how your Oculus pointer can select specific points on a Coherent UI web view? Tune in to final post of this trilogy next time! :D

____________________________________________________________

All blog posts by UnityGirl are inspired by my work with Brunelleschi: Age of Architects on the Aesop Games team. Download the current Unity client On IndieDB

Unity Script: Oculus 3d Pointer (1 of 3)

One of the biggest obstacles to developing for the Oculus is simulating mouse. This post will take us through the first of three basic steps to get Oculus mouse integration working. We'll create a 3d mouse-object that follows the mouse-movement. Next time we'll add click-integration using triggers in post 2 and we'll finish with click-integration for Coherent UI web views.

____________________________________________________________

Myriad Limitations:

There are a couple of very good reasons why we can't just use the mouse the normal Unity way.

1. Oculus won't display the windows mouse-cursor
2. You can't hijack/alter the windows mouse-cursor at-all easily
3. The windows mouse-cursor aims wrong through the Oculus distortion -even if you can see it-
4. Oculus ALSO doesn't display Unity GUI, which only responds to windows-mouse anyway. (suggest 3d menu objects as well)

____________________________________________________________

The Solution:

A Unity-Object 3D Pointer

That's right. We're gonna take a Unity asset and turn it into our pointer with a few simple scripting tricks. We'll start with your player-camera and your pointer. I use a red sphere-object as default but you can use whatever works for you.

• Create your object
• Parent object under Right Eye Anchor
•  this means it'll stick to the front of the head when you move
• Transform thusly

• It should look something like this with the pointer selected

• Finally, add this script that will link the pointer-movement to your physical mouse movement. 
• This responds like MouseLook and doesn't need the windows-mouse

____________________________________________________________

The Script

For Copying

using UnityEngine;

using System.Collections;

public class Pointer_Mock : MonoBehaviour {

public GameObject targetCam;

// Use this for initialization

void Start () {

targetCam = transform.parent.gameObject;

}

void Update() {

transform.LookAt (targetCam.transform);

float localX = transform.localPosition.x;

float localY = transform.localPosition.y;

float h = (localX += (Input.GetAxis("Mouse X") * 0.05f));

float v = (localY += (Input.GetAxis("Mouse Y") * 0.05f));

transform.localPosition = new Vector3 (h, v, transform.localPosition.z);

}

}

____________________________________________________________

All blog posts by UnityGirl are inspired by my work with Brunelleschi: Age of Architects on the Aesop Games team. Download the current Unity client On IndieDB

Unity Tutorial: Easy Mesh Combining For Highlights

Guess what devs? After months of head scratching I've put together a quick and dirty way to both
a) Combine meshes of a Unity-build Prefab into a single mesh and
b) Outline the damn thing OnMouseOver like many games in the past have done.

Best part? I did it out of found script pieces and didn't have to ask anyone directly! :D

____________________________________________________________

The Beginner's Problem:

Whether you're in this post for Performance or Graphics, you've probably determined that you've got way too many meshes per-object. Why do we have these problems? Lack of expertise being shared around effectively.

One thing I've noticed is this Mesh issue is something that mostly afflicts beginner to intermediate devs. If you're really hardcore you've been making your models in Blender or Maya from the start and know a bunch of Mesh / Shader / Camera based tricks to optimize both appearance and performance.

When I've seen the beginner Devs ask things like "how do I highlight the whole prefab like in diablo?" the most common answer is "Why don't you just make your whole model in Blender/Maya and import it?"
Sorry guys, but this is NOT REALLY HELPFUL for those of us who have most of our game models worked out in Unity basic shapes and myriad imported meshes.

____________________________________________________________

Mesh Combining

Your Prefab:

• The Parent Object (one everything else is inside) MUST BE EMPTY 
• No mesh
• scripts OK
• Take out all mesh-ed objects you don't want included in the single mesh
• ex: things you want to move separately from the mesh
• Place at (0,0,0) (you can write a compensation so this isn't necessary, but I didn't bother)

The Theory:

What we're doing here follows some basic steps

1. Make an array of all meshes inside your Prefab
2. Make a new Mesh that will be placed on the Parent Object
3. Combine array of meshes into single mesh on Parent Object
4. Save mesh you've created to a folder (this part took extra research)

The Steps:

How to Implement

1. Make sure Line 36 directs to a real folder you have
1. Mine is directed to the folder Experimental under Assets
2. Place Prefab at 0,0,0
3. Place script on empty Parent Object
1. this will add Mesh Filter and Mesh Renderer assets
4. Press Play
5. Pres Play again to stop. The Mesh should be saved in the designated folder
6. At this point you can do whatever you want with the mesh (or make it again by hitting Play again) In a following post I'll be using this Mesh to create a building outline.

____________________________________________________________

The Script

For Copying

using UnityEngine;
using System.Collections;
using UnityEditor;

[RequireComponent(typeof(MeshFilter))]
[RequireComponent(typeof(MeshRenderer))]

public class CombineMeshes : MonoBehaviour {

GameObject highlight;

void Start() {
MeshFilter[] meshFilters = GetComponentsInChildren<MeshFilter>();
CombineInstance[] combine = new CombineInstance[meshFilters.Length];
int i = 0;
while (i < meshFilters.Length) {
Debug.Log (meshFilters[i].gameObject.transform.name);
combine[i].mesh = meshFilters[i].sharedMesh;
combine[i].transform = meshFilters[i].transform.localToWorldMatrix;
//meshFilters[i].gameObject.active = false;
i++;
}



transform.GetComponent<MeshFilter>().mesh = new Mesh();
transform.GetComponent<MeshFilter>().mesh.CombineMeshes(combine);
transform.gameObject.active = true;

saveMesh ();
}

void saveMesh() {
Debug.Log ("Saving Mesh?");
Mesh m1 = transform.GetComponent<MeshFilter>().mesh;
AssetDatabase.CreateAsset(m1, "Assets/Play Assets/Experimental/" + transform.name + ".asset"); // saves to "assets/"
AssetDatabase.SaveAssets();
}
}

____________________________________________________________

Why Not Blender / Maya?

Why not just re-make every asset of your game carefully in the detailed mesh programs? You're more than welcome to but here's my assessment

Blender

• Blender is a huge free complex program that really is great. It's just so complex and great that I don't have time to learn it all.
• To be short: the learning curve is too steep. I'd need more time than I have to make assets I'm not happy with. 
• And to be honest, what little I have learned to make in Blender looks sloppy

On the other hand, I highly suggest Blendswap.com where real Blender artists occasionally deign to share their works with us script plebs. If you have the time and love it, please learn Blender and share with the rest of us. :D

Maya

• In shortest terms, it costs money. If you've got it, I've heard it's great but still time consuming.

____________________________________________________________

The Journey

I started this journey in an attempt to take building prefabs (made mostly of textured Unity cubes) and highlight/outlining the entire building on mouse-over for selection and info purposes.

Simple right? Anyone who's done this research knows I'm kidding. You can try to do this a couple ways.

Most Common Suggestion: Outline Shaders

Unlikely Suggestion: Post Processing Effects

Even Less Likely: Eldritch Rituals with Camera Values

My list is also in order of complexity. From my research I figured there are two ways to take clusters of meshed objects and highlight only the outside

1. combine the meshes into one and use an outline shader
2. somehow identify the cluster of objects with camera effects and have the camera try to place an outline on the right lines.

Somehow the first seemed simpler so I looked into combining meshes.

THIS was suggested to be done in a couple of ways

1. Export the whole thing with a bought-in asset thingy to a new file type
2. Use a bought-in ($55) mesh combining package
3. Use mesh.CombineMeshes,

Not buying something combined with THIS excellent resource

http://wiki.unity3d.com/index.php/Shaders

Swayed me toward the solution we chose today.

______________________________________________________________

If you haven't already figured out how I turn this mesh into an outline, stick around for the future post where I explain it. :D

______________________________________________________________

All blog posts by UnityGirl are inspired by my work with Brunelleschi: Age of Architects on the Aesop Games team. Check out our Crowd Funding page(s)!