Some more amazing concept art to reference for a volcanic planet

Mustafar

Mordor

Before I begin painting on the final canvas I need to work out what the final composition of the landscape will be – The generated image is quite interesting and a great base to work with, but I need to interpret a more defined composition from that. I especially want to add more depth to the piece and the computer does not really work in foreground/background, it is mostly just on one flat plane.

• 

• 

• 

Printed the image out and drew over it with a thick pen – trying to figure out what I want to change and how I want to interpret certain things. My main focus is on trying to add some foreground element to the right – I want a closer portion to work with so I can have a bit of closer detail and show off some of my characters/plants or whatever I create.

Next I worked in digital to get a range of values for more depth.

Overlaying the generated image and adding in some colours and photo textures to get a better sense of how the composition will work as a final image.

Keeping it quite rough at this stage – I am using this just as a guide to figure out the overall composition and placements of major elements, perspective etc.

Upon looking at the generated image for Addor my friend noted that it looked quite like the volcanic planet ‘Mustafar’ from star wars. This planet appears to fit quite nicely within the generated description as well.

I also looked into Tatooine – This is a planet in Star Wars which has twin suns, just as Addor does.

• 

• 

I did a bit of research into how twin stars operate – I found there it is not just in the fantasy realm, they actually exist within the great depths of space.

Kepler-47 is a solar system with planets orbiting around two suns – A binary star system. In this system the two suns orbit around their common center.

Research around it seems that typically there is one big sun and one much smaller, less bright one.

Today in class I discussed the current progress of my project – I was feeling a bit unsure about what exactly my final outputs should be as I have a few options to consider. I talked with Tanya about the direction –

• Characters – They don’t feel like they match the descriptive text supplied for the world created.
• Consider how much I go over the original image – Don’t necessarily want to completely cover over all of the AI created image
• Use a digital process to mockup the final physical version? Allow for iterations that can be changed/adapted without destroying the print.
• Frame as I am using AI to enhance my oil painting process to create more interesting compositions and subject etc.
• Just the video to show process, and then just the final artwork as the physical output.
• AI generated artworks that I was exploring don’t really fit within the idea of developing on final image and showing the conceptual process behind that.

Moving forward – While I quite enjoyed my exploration into displaying the AI created art without my collaboration, I don’t think it will have a spot within my final display which should focus on just the final artwork and process behind it.

Having a go at developing some different things from the shapes created around the input of Addor

The generated shape is a bit of a mess as it draws over itself so many times

Here I have tried separating out each instance of drawing that the program does to create the overall shape.

• 

• 

• 

• 

• 

Working from these to create some different creatures and other things, based somewhat off the description assigned for this world

The description mentions chromatic dragons –

Moving forward – These are interesting but I need to push it further to create things which feel like they belong within the same world that has been generated – I need to try and work closer with the descriptive text.

Tanya showed me how I could use illustrator’s colour wheel tool to create some interesting colours – It will take a range of colours and change them based on different colour relationships eg complimentary etc to create a colour scheme.

First I took the image into illustrator and used image trace to convert it into colour shapes with a more limited range – The illustrator adjusted the colours for me.

I played around with different numbers of colours – in the end the best result was reducing the image to 6 colours and recolouring based on that.

Creating the final output image for my large scale painting – going through a similar process to before.

This process aims to give agency to the computer over the process of creation, by allowing equal contribution and decision making.
However, I recognise that collaboration is generally never 50/50.
As this is a speculative exploration, the program that I am working with is still at a fraction of it’s potential.
Therefore, for this collaborative project I have taken on the role of ‘creative director’

This does not mean I make all of the decisions – I have tried to embrace the chance of what the program creates, and wherever possible give it the choice and input over the final output.

Beginning the process

First generating a list of random fantasy names, and picking the first one in the list

1. Addor: A ruined realm of cratered battlefields and oceans of lava, illuminated by two crimson suns. The realm is primarily populated by chromatic dragons. It is also the location of the Fane of Fiery Ruin, a bastion of insane cultists and warlocks.

Taking note of the description generated – This is what the machine has decided for the world, so I will be working with this to create a final image that represents that.

• 

First output from gauGAN

Interpreting and painting over-top – Trying to stay reasonably true to the original image

Using illustrator to adjust colours –

• 

• 

I then used photoshop to apply the colour

And cropped in for the final image

I presented all of my work to a small group with Kerry Ann, and talked about what I aimed to achieve over the break

Reflection – Over the break I set out to create a digital and physical painting in order to decide my final output medium. I was successful in doing this, and decided on physical which is a great step forward. I also managed to create a printed canvas test for class, which was quite engaging to look at.

Feedback –

Kerry Ann had some great points around the creation time of the artworks for myself vs the AI – The AI has the power to create exceptionally fast, it’s sections of the collaborative process are created in less than a minute. My sections take much longer – hours even. During this whole time I am thinking and interpreting while creating, which can be quite engaging in the process.

First quick test at painting overtop of the printed canvas image

The process was quite interesting and very different to any painting I had done before, but lots of fun! It was like going from underpainting straight into detailing – all the values are there already and all I needed to do was to go through and define areas.

I decided to leave it here for now – The test was a great success and I know moving forward that I definitely want to use this technique for the final painting.