This site is best viewed in portrait mode.
Please rotate your device.
In 2010 we approached the highly respected stills photographer Mark Rogers, whom we had worked with on several film campaigns previously, to shoot one or two images to represent Carnival Studio. Mark introduced us to production designer Ingrid Weir, and we began to formulate a vision for this world and the talented but fringe characters that would inhabit it.
What resulted was a two location shoot (Centennial Park and Weber’s Circus) featuring several circus performers and experienced actors including Maeve Dermody and Panda Likoudis. The series of seven images told a loose story of the performers and their relationships, away from the performance and when they are less guarded. The photos were released on a ViewMaster reel to our clients and elicited an incredible reception. Titled ‘Carnival del los Sueños’ which translates as ‘The Carnival of Dreams’.
In 2015, we decided to embark upon another photoshoot and set about planning within the core creative team of Mark Rogers, Ingrid Weir and Demi Hopkins. While the real circus performers did not return, we had recruited prominent Melbourne actor Damian Walshe-Howling as a sideshow spiv, alongside Panda and Maeve again. The concept with this shoot was to capture single-character portraits of each carnie, eyes to camera, with the carnival troupe life carrying on behind. The intention was to appear as if it were reportage photography where they had been captured in the field by a visiting journalist. We also sought a group portrait which would be in front of a giant illuminated CARNIVAL sign (to be created in 3D modelling)
The key difference with this shoot was instead of shooting on location; we would shoot within a studio environment. Extensive scouting had looked at shooting in Hill End, and various outdoor locations but due to logistics and limited budget we decided it would be easier to shoot in studio and address backgrounds separately.
While on the day, Mark captured some stunning photographs of the carnies, the lack of external clean plate photography to composite our characters into proved to be a significant hurdle. The strength of the first series hinged a lot on the real world we had created – set somewhere in the early 20th century with lush, overgrown grass and a sense of nostalgia and warmth. Without shooting clean plates in multiple locations, the digital post-production stagnated until 2020.
Carnival Studio has a core team of print and digital designers who are supported by in-house retoucher and matte painter, Mike James. His experience heralds from over a decade in commercial retouching at Cream (now Cream Electric Art) before shifting focus to 3D modelling and matte painting for TVCs and feature films at Delux, Method Studios and The Gingerbread Man before settling at Carnival Studio.
Taking advantage of the reduction in work during the Covid shutdown, Mike concentrated on building photo-real environments almost entirely within 3D for our Carnival portraits. The results have seen us create a world not possible for logistical and budget reasons – with the modelling, texturing and rendering of physical structures including various tents, string lights, sideshow banner walls and caravans – as well as the natural environments including grass, trees and terrain.
The 3D environments were composited within Photoshop with talent portraits with global grading and effects. All renders are created for print resolution output.
Carnival Studio is equipped to combine 3D modelling within our key art development, allowing us to enhance or create visuals not possible in the past. An excellent example of this is our recent work on Ellie & Abbie (& Ellie’s Dead Aunt) which can be viewed here.
If you are interested in reading more on the technical breakdown of our Carnival portrait shoot, please read on.
We’d love to chat with you if you’re interested in knowing more, or how these methods can be applied to other productions.
During the post-production process after portrait photography was shot, we set out to address a number of technical considerations and challenges including:
3D ASSET CREATION
Using the original Carnival composites as reference an initial library of 3D assets was developed. These assets included various grasses, trees, sticks, and various other organic debris. Multiple techniques were researched and developed during this stage including procedural modelling and sculpting trees and grasses utilising Maya’s Paintfx and ZBrush’s fibre system as well as experimentation and development of 3D photogrammetry. (3D models extracted from many photographs of an object at various angles).
A lot of the textures used came from texture libraries, particularly for the organic objects, but many were also created – some procedurally using software such as the Substance Suite of tools. Many were also hand-painted, particularly for the Wagon, using Zbrush and Substance Painter. The texture techniques used were all based on the current PBR (Physically Based Rendering) Standards, which is a texture workflow that aims to simulate how light reacts with a model to attempt to simulate real-life materials.
VIRTUAL CAMERA MATCHING
Whilst the original raw files from the shoot contained the Lens and Camera Sensor size information, creating a 3D camera for still images can prove tricky. When footage is moving, a camera tracker in software such as Nuke or After Effects analyses the distances between objects and takes into account lens information and lens distortion and can produce an accurate real-world scaled scene and a virtual camera that matches the height, position, and rotation of the camera at any given frame. This gives a VFX artist the ability to composite 3d objects and 2d images that match the real camera’s perspective.
In the case of using still images, the images were undistorted using the lens description data and horizon lines were drawn as a guide. A virtual camera was positioned and rotated by “eyeballing” the horizon and using simple 3D shapes such as a cube to determine an approximate world position of the virtual camera. After initial concept renders were produced using the eyeballing technique, we decided it was necessary to try to more accurately reproduce the camera’s position, so proxy 3D humans were posed at scale to match the photographed talent in each scene. This method proved to be very useful as the subtle perspective lines and indicators of the 3D humans guided the positioning and slight rotations of each scene camera more accurately.
Whilst going through many iterations, the 3D Lighting of each image was reasonably straightforward as the real characters had pre-determined lighting from the photoshoot. The challenge was to put the characters into outdoor environments that had lighting around them that would feel consistent with the lighting on their bodies. Early on in the process, it was decided that dusk to night-time lighting would suit best as we would be able to use multiple light sources such as string lights as well as the sun or moon to represent the fill and backlighting that each character was shot within the studio. The Rigged 3D humans also proved useful during this process as a guide to position lights as you have an accurate representation of how backlighting affects the characters’ cheekbones for example.