Artist Spotlight: David Hirst

With over 20 years of experience behind him, David Hirst is no stranger to the nuances of media production.

His creative flair was honed through a background in illustration, having completed a degree in technical illustration before embarking on a colorful and varied career—first in video games, then in film production, across a number of projects, studios and roles.

The last fifteen years has seen David specialise in lighting and look development for the visual effects industry. In Vancouver, David worked for MPC as their Head of Lighting, recruiting, hiring and training lighting teams on nearly 30 projects. About a year ago, he made the move back to the UK where he resides in Colchester, working as CG Supervisor on a number of varied projects from the London office.

“A couple I’m most proud of include The Life of Pi, and Pokémon Pikachu Detective, which I worked on whilst at MPC Vancouver,” he tells us. “For Life of Pi we needed to create a sequence where the freighter Tsimtsum sank in a tropical storm.This was one of the most technically challenging projects that I’d been involved with as head of the lighting department. It took a lot of artistic and technical ingenuity between all the departments to make the project achievable, believable and also renderable. ”

Moving to a CG Supervisor role on Pokemon: Detective Pikachu presented its own unique challenges, David explains. Adapting such iconic, well-loved characters into a 3D world was no mean feat. 

“It took the talent of a large team to make this possible and I’m extremely proud of the work we all achieved together on this project,” David tells us. “One of my personal favorites achievements was working closely with the asset development teams to get the look of Pikachu that felt realistic and believable, but also be true to the character from the cartoon.”

Below, we catch up with David to discuss how he’s adapted to emerging technologies whilst working on projects such as these, and what he thinks is in store for the future of look development and lighting.

Q: Given your experience and provenance in the VFX industry—specifically look development and lighting—what tech trends have you seen emerge over time, and how have these affected your approach to projects?

A: I guess the biggest change over the last 10 years is the development in physically based rendering techniques. This technology introduced physically based area lights, image based lighting and energy conservation to shading. It also introduced the widespread adoption of path tracing within the industry. 

Before this, lighting and shading had complex pre-pass setups with shadow maps, baked occlusion or colour bleed and other processing steps. Although these worked well, the complexity of setup and inflexibility of these techniques made it difficult for the artist to make changes. At MPC we worked with Pixar in this transition to a fully physically plausible rendering solution. This change removed the old limitations and vastly reduced the complexity of lighting and look development setups. This in turn freed the artist to make changes quickly and have a more interactive experience. The combination of these changes resulted in a jump to photorealism not achieved before.

The other transitions that happened around the same time was the adoption of Katana as the main lighting and lookdev tool. Katana also made some significant changes to our workflow and introduced a number of key technologies, such as full top-down proceduralism in lookdev and lighting. Shaders or lights could be defined at high level within the scene graph and then subsequently edited through the scene hierarchy at sequence or shot level. The ability to handle scene complexity through deferred loading also let us deal with extremely large data sets not possible defore. Katana is also a powerful referencing tool set of live groups, enabling the development of flexible workflows for artists to work, share and collaborate on projects.

Q: What has been your experience navigating these shifting trends whilst leading whole teams? How have the teams you’ve led adapted to these changes?

A: I've been extremely lucky throughout my career to work with so many talented and generous people. So when I had the opportunity to hire teams for projects, I’d look for a balance of people that wanted to work together and share that learning experience.

Artists and Technical directors are always looking for better tools or better ways to get results. So in the earlier days of my career, the artist often had a fairly technical background to work successfully as a lighter. This is reflected in the people I’d hire for the projects like Life of Pi, which had a setup which was extremely challenging for a lighter to work with. As we have transitioned into a more streamlined pipeline, the artist can now concentrate on doing more creative tasks, with less technical setup. I’d have less to worry about them being able to quickly understand technical requirements of the job.

Q: What’s your take on modern lighting and rendering technologies in relation to real-world cinematography principles?

A: Lighting is heavily dependent on the tools available. In the past, lighting would often have to use a fairly large number of lights to represent the natural and subtle interaction of light bouncing on a set. Now, as technology has improved, we can get lighting results much closer and quicker to reality. This in turn has freed up artists to be more creative. So the cinematography principles of on set lighting can be used in a very similar way as  in real life.

Quote from David Hirst

Q: Can you give some insight into your experience working with Katana? How and when did you start learning it? Do you use it in combination with any other software?

A: I first got involved in testing Katana fairly early on when introduced to MPC. The biggest issue for us when we first started using Katana was to try and make it work like our old pipeline. This was a mistake in not really understanding the true potential of what Katana can bring to a studio. Once we started to really understand its potential, it became really exciting to see the way Katana could handle scene complexity. This was great timing for us as scene complexity was exploding at the same time with more and more ambitious projects.

Since Katana's introduction we have used it together with Pixar's Renderman exclusively on all projects. Both the Foundry and Pixar teams were committed to creating the best products possible and it was very easy to work with both of them as we made the transition quite fluid.

Q: What opportunity for Look Development do you find in Katana that you do not find in other tools?

A: Katana’ s hierarchical design and tools set gives it a number of very desirable features for look development. These features include baking lookfiles and the child material workflows. With Katana it's possible to set up scenes with just a few master shaders then create child material that propagate the master material network, but then also have additional edits layered on top to become separate assignable materials. This is a workflow that lets you quickly set up a material library from just a few master materials. Also this material library can then be baked into a lookfile for easy referencing, and the starting point for another lookdev scene.

Q: What are you enjoying most about the version of Katana you’re working in, what are you excited about in the upcoming Katana releases, and what would you like to see beyond these?

A: Katana has always had a high level of flexibility for lighting and look development. With the latest version I see much improvement in the interactive rendering workflow and the tools for better organising material networks. Other areas that could help smaller studios is better farm management setup both local and cloud integrations that can take easily the advantages of incremental style rendering features like checkpointing, time based rendering and render recoverer.

Q: With the unveiling of Katana 4.0, are there any features that you are looking forward to trying out?

A: The Katana 4.0 announcement had a couple of features that really immediately stood out for me. The first is the possibility of support for multiple simultaneous renders. This is a feature that could lead to significant improvement to look development and lighting workflows. For example, an artist in lookdev could quickly render a number of different wedges to test shader values, getting the results back simultaneously to compare results. Or a lighting artist could work on Key and subshots across a whole sequence at the same time, quickly finding the best lighting angles that work across the whole sequence and then if necessary focusing on the sub shots that may need additional tweaks.

In regards to the Artist-focused Lighting Mode, I can see this really helping the artist to stay in the creative mode without having to jump focus from the viewport to the gaffer to make lighting adjustments. Both of these new features highlight the artist focused improvements in this latest version of Katana.

Q: What facets of USD are you currently using as part of your workflow, if any, and what benefits do these bring?

A: Although I’ve not used USD in production. I’ve spent quite a bit of time recently authoring USD files in Houdini for testing in Katana. The workflow potential for this technology is pretty exciting. I could see a smaller studio being able to leverage this in a workable solution that would have needed considerable investment in the past. It still feels like early days for this technology, but like openEXR or Alembic, it's picking up industry support and I see it becoming more and more relevant with further developments.

Also with shared film projects now being quite common within the industry, a USD pipeline has a potential cost saving when negotiating the exchange of assets and scenes between different VFX studios, if both parties are building pipelines around USD methodologies.

Q: Do you have tips and tricks for burgeoning artists looking to invest in Katana for look development and lighting? 

A: Learn the basics in color theory and the lighting composition. Learn how you can use these techniques to enhance the mood or direct the attention of the audience. Also in conjunction with this, always find and use reference as part of your workflow, either as direct reference or for inspiration.

When working on look development, it's very important to really look closely at the subject matter. Do you understand the lighting and surface interaction? What material is the subject composed of? Does it have multi layers of materials like a clear coat, for example, on a car.

How does the surface feel if you could touch it—is it smooth or rough? Can you see the results of how the reference is constructed? Such as subtle seams from casting, stress points, machineering that can create anisotropic properties or welding marks, and so on. 

Then what's the conditions of those materials and how have they aged? Do they have rust—have they accumulated dirt or dust? Is the reference showing signs of wear and tear or interaction like grease from hands that could add authenticity to your look development?

Keep rendering and comparing to your reference. It's subtleties that make a difference between something looking right or wrong. Next, learn how best to replicate those observations with the tools available to you. Keep an eye on energy conservation when building your shaders. With layered shaders it's possible to break the rules of energy conservation without knowing. So make sure that your software can correctly deal with this situation or manage these layers yourself. 

Q: Do you have any wider tips for artists looking to break into the VFX industry, and specifically look development and lighting? What type of opportunities should artists be looking for?

A: I ran the Lighting Academy for MPC in Vancouver for a number of years and I’d say that this was a great first step into the industry. It gave the artist an opportunity to work closely with experienced industry artists and learn on the job. So finding a studio that is offering that kind of opportunity would be a good start.

From a reel perspective I think you should have just a few good pieces. It's not about quantity, but the quality of those pieces of work, as I know how difficult and how little time a film school student has to get a reel together. Also when it comes to the interview, try to let your passion for the subject show through and prepare yourself for the kind of questions you can expect to be asked.  

Q: What do you think is in store for the future of look development and lighting? How does Katana fit into this, in your opinion?

A: Katana has already got a great start with its lookdev tools sets. I think the next area of interest is how to create more agnostic material libraries. Can we create material and bindings that could work across various renderers and software platforms? There are already a number of initiatives in this area. 

I’m also excited by the latest technologies being introduced to improve rendering time, either by more intelligent sampling schemes, AI driven denoising technology or the introduction of hardware-accelerated ray tracing into mainstream graphics cards. These are all technologies that would work closely with the flexibility of Katana in scene graph construction. 

Want to try your own hand at lighting?

Get a 30-day free trial

Need more time to learn? You’ll be able to extend this trial at the end of your 30-day period.