Click Here to view all the amazing entries to Rookie Awards 2024
Sara Stanoeska | Digital Matte Painter
Share  

Sara Stanoeska | Digital Matte Painter

Sara Stanoeska
by stanoeska on 26 May 2024 for Rookie Awards 2024

I'm excited to share my entry for The Rookie Awards 2024, featuring a selection of the DMP projects I've worked on within Nuke and Photoshop.

24 2236 3
Round of applause for our sponsors

"Omen" - Stormy Sea DMP

This one is my most recent personal piece, trying to capture a feeling of premonition.

Thematically closely tied to my second project you will see listed here, this one is also based on the concept of rough seas, rocky shores, and stormy skies.

Below you can see the reference board I had assembled while crafting the idea of where I wanted to go with this, which also served as the basis for collecting my library of images I would use for this project.

Of course, I had to begin with choosing a piece of footage as my base plate, and all the reference images found later on had to be respectful of its perspective, lighting, and the composition it would allow for.

I opted for this video I found on Pexels.com, since I loved the powerful dynamism of it, and its potential to emphasize a focal point following the camera movement.

Tracking it in Nuke proved to be a little challenging, since there was no camera or lens information listed on the website, but after trying my hand several times with the 3D Camera Tracker, and getting the lowest possible solve error I could, I also imported 6 user tracks from a 2D Tracker to improve it further.
The remaining visible deviations from the track I did my best to fix by animating several Transfrom nodes at the end of my projection node tree.

As showcased in my Nuke script, I keyed out the sky of the original footage, and replaced it using a skydome projection.
Everything was colour graded to fit the envisioned mood and atmosphere.

The DMP elements painted in Photoshop, and exported as EXRs would later on be projected onto cards and added into the script.

An aspect I found most challenging, and interesting at the same time, was adding the effects of lightning and fog, which I learned how to do while working on this project.

The lightning was a keyed out selection of the highlights in the clouds, plugged into an animated Grade node, and separated by zones using Roto nodes and a Shuffle node to help animate it alternately.

For the fog, I used Noise nodes projected onto overlaid cards, added over the precomp of everything done previously.

This project was definitely a fun learning experience, and the opportunity for me to implement all of my DMP skills into one piece individually directed from start to finish.

Hebridean Landscape

(Final Project for the MattePaint Course)

This piece was the final project done for the 12-Week MattePaint course I attended.

We had complete freedom over the creative choices for this assignment.
In terms of technical execution, I decided to go with a still image in order to be respectful of the short time frame, while being able to showcase my Photoshop skills optimally.

The theme and setting I chose were inspired by my (then) recent travels to Scotland, and all the Game of Thrones filming locations I was lucky enough to be able to visit there.

As a huge fan of the series since a young age, I was immersed into that experience, and decided to buy a book (showcased below) that elaborated on the geographical and historical inspiration behind Westeros and its politics.
There, I found the following quote about the Hebrides, which served as a poetic basis for the concept of this piece to develop:

I assembled this reference board at first to communicate my vision to my mentors, and then, with their help, decided on which of the concepts to go with.

Collecting the images I would use to compose - at first a collaged sketch - and then the fully rendered piece was made easier thanks to having a precise location in mind, which already dictated the geology, climate, types of vegetation, etc.
The majority were from a free reference pack, and some of the images were sent to me by friends living in the UK.
Thankfully, all maintained a consistency in time of day, type of lighting, and similar perspective angles, that made the matte painting possible.

While doing the rough collage concept, I sketched out some explorations of composition, to bring the layout and the focal point in place in order to guide the eye, and serve the story.

The final rendition came in two version:
- One colour graded, and featuring story elements (the birds, the beaten paths, and the group of people walking on the far cliff);
- One without, staying true to the definition of a matte painting.

Below them, you can see the image I used as the base plate.

And finally, you can see the entire process showcased in the breakdown video.

Day to Night Relighting

A day to night relighting of a wide angle shot done in Photoshop, featuring:

- plate cleanup
- sky replacement
- relighting and colour grading
- adding artificial light sources

Plate photography used: mattepaint.com
Location: Banff (Alberta, Canada)

After doing the plate cleanup - removing all the unnecessary details, artificial, and moving objects from the scene, I spent time figuring out how atmospheric perspective works differently at night, and finding ways to implement that into this matte painting.

While darkening the scene (using Levels adjustment layers for the values), I also made sure to be mindful of the different ways different colours appear in moonlight, tweaking each of the colour channels accordingly (using Curves, Selective Colour and Hue/Saturation adjustment layers).

What proved to make me unable to suspend my disbelief of this being a night shot even after that, was how sharp the shadows from the direct sunlight were, compared to references of nighttime photography I was looking at.
In order to address this, I did a frequency separation of the entire image as a smart object, and then used the low frequency layer to soften the shadows, and darken certain areas, in order to match the intensity of moonlight.

At the end, I played around with adding artificial lights of different sources and colours, to subtly add character to the image.

Below is the final result.

Seaside Castle

2D Tracking, CG Integration/Paintover & Plate Cleanup

I decided that if I am to showcase one student project, it has to be this one.
Done as the second assignment within Nuke during the 12-Week MattePaint Course, this one proved to bring the most challenging obstacles, as well as the biggest learning opportunities.

We were given this piece of footage from Pexels.com as a base plate, onto which we were required to integrate provided CG elements (created by Okelsiy Golovchenko), using 2D Trackers, and paintover of said elements done in Photoshop.

I wanted to go a step further, and do a plate cleanup as well. However, the lack of camera and lens information made it difficult to calculate the distortion of the plate footage, and the cleanup patches seemed to be sliding from the tracks at first.

In my attempts to solve this, following my mentors' advice, I tried to first stabilise the footage, so the distortion and warping at the edges of the lens would be more easily noticeable, and then, to manually apply that same distortion onto the cleanup patches using CornerPin nodes.

The results weren't completely satisfactory, and the thing that proved to work best for this in the end, turned out to be separating the patches of the cleanup into smaller parts, and using SmartVectors to match them to the distortion of the plate.

For the water, I had to use RotoPaint to preserve its movement.

As you can see below, this project also required some precomping, since the script was getting a bit heavy, especially thanks to the SmartVector and the RotoPaint nodes.

The actualization of these projects wouldn't have been possible without the support and guidance of my mentors Conrad Allan, Ian Vicknair, Nikola Angelkoski, and instructors Angeliki Gkaleridi and Samuel Cleall.

You can reach me

- Via email: [email protected]

- On LinkedIn: linkedin.com/in/sara-stanoeska-2b6772179/

- Through my ArtStation profile: artstation.com/s_stanoeska

Thank you so much for your consideration!


Comments (3)