Master's Project: Procedurally Generated 2D Map

Unturned
Level 0 - Null
Posts: 2
Joined: 04 Jun 2016, 20:17

Master's Project: Procedurally Generated 2D Map

Unread post by Unturned » 04 Jun 2016, 20:20

Hello,

I am in the process of writing my Master’s thesis on Procedural map generation, with a focus on 2D environments. As part of this, I have developed a ‘game’ prototype which takes input from the ‘player’ and creates a map on those specifications, using a combination of the different methods I have researched.
It would help me out if someone with a bit of experience in the industry could help me out and take a look at it and answer some quick questions I have prepared (or in whatever other way you prefer). The response decides which direction I take the final part of the project in, and what conclusions I make.

You can post your answers here, or send them to me on mia@sogaard.us

By the way, large maps take a while to create.

Also, your computer might automatically set the quality to the lowest, so you if you want it higher, you just choose a higher quality in the drop-down menu when you open the build :)

https://drive.google.com/open?id=0B6Noy ... kJlZ0paSUE

Thank you in advance,
Mia Søgaard


Questionaire:

Before playing the prototype:
• Do you have any previous experience creating games?
• When designing games, what are your most common obstacles?
• Do you use any tools when designing levels? Ones that make the process easier?
• Do you have a need for a tool that can help you generate the levels/maps automatically? If yes, what kind of tool would you want?
• What does Procedural Generation mean to you, in terms of creating game maps?

After playing the prototype:
• What is your impression after playing the prototype?
• Could you use any aspects or features of this prototype as a tool in your own work?
• Which features/aspects are missing?
• Which features/aspects are unnecessary?
• Which feature would you be most likely to use?
• Could this product be of in your creative process?
• If this was a tool, would you use the whole product, or only select features? This could be any part, from the overall map shape, to the automatic road generation.
• Would this product be of most use for big or small games? (such as full hour-long games versus quick minute-long games)

User avatar
Henrik LSP
Level 10 - Dark arts student
Posts: 105
Joined: 19 Apr 2011, 20:53

Re: Master's Project: Procedurally Generated 2D Map

Unread post by Henrik LSP » 09 Jun 2016, 10:17

This looks like an interesting project. Are you using Perlin Noise?

Before playing the prototype:
• Do you have any previous experience creating games?
Yes. I have worked in the game industry for years.
• When designing games, what are your most common obstacles?
Designs always turn out to be incomplete. It is almost impossible to figure out before development what will work and how long things will take to develop.
• Do you use any tools when designing levels? Ones that make the process easier?
I am not a level designer, but having worked with level designers, I have found they often need tools for visualizing behind-the-scenes data while the game is running and tools for quickly navigating through the level (aka. cheat mode).
• Do you have a need for a tool that can help you generate the levels/maps automatically? If yes, what kind of tool would you want?
Not currently. If I did, it would probably be an internally developed tool custom built for the specific game. Although source code for such algorithms would certainly be useful.
• What does Procedural Generation mean to you, in terms of creating game maps?
Can be useful for creating variation and replayability.

After playing the prototype:
• What is your impression after playing the prototype?
It creates interesting and varied patterns. Different patterns for different types of terrain, which is good. The water seems to always be created in swamp-like puddle patterns. Including different types of water might be useful (lakes, coastline, river).
• Could you use any aspects or features of this prototype as a tool in your own work?
Not currently.
• Which features/aspects are missing?
Documentation. What are the effects of the sliders?
• Which features/aspects are unnecessary?
The main character, character controls, and walking-around part. I just go straight to free camera and zoom out.
• Which feature would you be most likely to use?
Source code for the algorithm used.
• Could this product be of in your creative process?
Maybe in the future.
• If this was a tool, would you use the whole product, or only select features? This could be any part, from the overall map shape, to the automatic road generation.
I would probably want to take specific features and restructure them for whatever game they need to work in.
I think the prototype is more useful for study and experimentation than for a general-purpose tool.
• Would this product be of most use for big or small games? (such as full hour-long games versus quick minute-long games)
It would be most useful for a game in which the player passes very quickly through the maps, which would mean that a lot of maps would need to be designed. Procedural generation would help.

Unturned
Level 0 - Null
Posts: 2
Joined: 04 Jun 2016, 20:17

Re: Master's Project: Procedurally Generated 2D Map

Unread post by Unturned » 10 Jun 2016, 20:55

Henrik LSP wrote:This looks like an interesting project. Are you using Perlin Noise?

Before playing the prototype:
• Do you have any previous experience creating games?
Yes. I have worked in the game industry for years.
• When designing games, what are your most common obstacles?
Designs always turn out to be incomplete. It is almost impossible to figure out before development what will work and how long things will take to develop.
• Do you use any tools when designing levels? Ones that make the process easier?
I am not a level designer, but having worked with level designers, I have found they often need tools for visualizing behind-the-scenes data while the game is running and tools for quickly navigating through the level (aka. cheat mode).
• Do you have a need for a tool that can help you generate the levels/maps automatically? If yes, what kind of tool would you want?
Not currently. If I did, it would probably be an internally developed tool custom built for the specific game. Although source code for such algorithms would certainly be useful.
• What does Procedural Generation mean to you, in terms of creating game maps?
Can be useful for creating variation and replayability.

After playing the prototype:
• What is your impression after playing the prototype?
It creates interesting and varied patterns. Different patterns for different types of terrain, which is good. The water seems to always be created in swamp-like puddle patterns. Including different types of water might be useful (lakes, coastline, river).
• Could you use any aspects or features of this prototype as a tool in your own work?
Not currently.
• Which features/aspects are missing?
Documentation. What are the effects of the sliders?
• Which features/aspects are unnecessary?
The main character, character controls, and walking-around part. I just go straight to free camera and zoom out.
• Which feature would you be most likely to use?
Source code for the algorithm used.
• Could this product be of in your creative process?
Maybe in the future.
• If this was a tool, would you use the whole product, or only select features? This could be any part, from the overall map shape, to the automatic road generation.
I would probably want to take specific features and restructure them for whatever game they need to work in.
I think the prototype is more useful for study and experimentation than for a general-purpose tool.
• Would this product be of most use for big or small games? (such as full hour-long games versus quick minute-long games)
It would be most useful for a game in which the player passes very quickly through the maps, which would mean that a lot of maps would need to be designed. Procedural generation would help.
Thank you for taking the time to check out my prototype! Your fedback was very useful, and I agree with your observations.
I have to include a player character (as a way to prove that the map is functional), but I think I will have the default settings set to free camera.

As for your questions; yes, I am using Perlin noise! It can be a bit unpredictable when I mix it with several different layers of heightmap-noise, moisture-noise, temperature-noise, etc., but it generates some very interesting maps with a good variation, and an ability to reproduce the same map again, which is extremely useful. It's only real downside is that the 'zones' aren't necessarily of a useful size (I can end up with very small snow-covered areas, for example), which makes dividing the map into levels a little tricky; but for a completely open world, it's great.

- Mia

Post Reply