"Corona Renderer will be the best archviz renderer on the market"

Ondra Karlik worked in architectural visualization field for years, until he decided in 2009 to make Corona - new rendering engine. After few years the single-man project expanded, today Corona is developed by team located in two offices. Corona Renderer is getting more and more attention from arch-viz community and it's just the beginning of a way - developers promise more features and more supported software. They want to "make the best arch-viz renderer on the market". Will they succeed? We speak with Ondra Karlik and Atam Hotovy from Corona Renderer team.

Top image: "Bart Square" by The Boundary in association with Iain Banks.

Evermotion: When the idea to make a new renderer come to your mind?

Ondra: I was doing 3D graphics as an artist for few years, but then I started college and switched to programming. When I had to pick up a topic for my bachelor thesis in 2009, a friend suggested doing a renderer. It was perfect for me, as I could utilize my old hobby. Back then it was of course never meant to become a commercial production renderer, that came later.

Weren't you afraid to enter (quite crowdy) renderer's market? Was it smooth or harsh start?

Ondra: Everything happened very naturally. In the beginning I was developing the renderer only for fun, research, and school purposes. Then I started posting pictures to some forums, which soon caught attention of other users. Eventually they asked me to send them a build so they could try it themselves. More and more users joined, and I started releasing the software publicly on our website (as opposed to sending it to friends over ICQ and email). At some point, people started using it also for commercial jobs instead of just goofing around. That is when I realized the project has a commercial potential, which would allow me to
continue the development even after school.

What was the most difficult for you when you started your company?

Adam: For me personally the hardest part was actually the gradual termination of my artist activities (I used to work as archviz artist for almost 10 years) and juggling responsibilities between my archviz studio and newly established company. Everything else went more or less according to plan. We did not need any external funding as we have used all our savings to kickstart the company. Nobody was pressuring us for example to release unfinished product.

Ondra: For me it was dealing with a ton of new responsibilities - setting up the online shop, making contracts with render farms, programming licensing system, setting up offices, etc. I had to stop development of new renderer features for about 4 months just to deal with this. Fortunately this is now over and I can again focus on development of Corona.

How difficult is to sell this kind of software?

Adam: So far it went rather good. We did not need any kind of traditional marketing to sell the product yet. We have decided to continue with the “word of mouth” strategy which Ondrej successfully started. If you create a good product in first place, people will share the information. And you have a free marketing right there. But we do plan to gently push our marketing strategy further and test the waters in upcoming months.

Ondra: most problems we encountered were purely technical ones - online store solutions not flexible enough, rejected payments from customers outside of US/EU, complex tax laws, etc.

How does your FairSaaS model work for you?

Adam: In the beginning we were not sure how our model would be accepted. People from CG industry had experience with subscription pricing from companies like Autodesk and Adobe and they seemed to not like it. We wanted to offer our product with fair price and conditions (both for us and our users).

Ondra: We were surprised how much better FairSaaS licenses sell compared to the traditional perpetual “Box” license we also offer. Even when purchasing Box licenses, customers still opt to get a subscription for software updates. To illustrate, less than 1% of all sales are “Box without subscription”. So few people go for this option that we are actually considering removing it from our store.

Are you satisfied with current sales?

Adam: I can answer this question with yes and no at the same time. It depends on how you look at it and what your goals are. We are thankful that we were able to clear our debts and stabilise the company financially. We are also happy that we can invest the rest of money into the development of Corona Renderer - by hiring experienced developers which can be costly.
With more money we could certainly progress faster, developing more plugins and side projects.
Ondra: Adam said it well. I am satisfied with the sales, but I want them to be bigger at the same time, so we can start some fun side projects ;)

How does piracy affect your business? And how do you fight it?

Adam: Thanks to the fact that most of our licenses are monthly sold subscriptions, we can measure impact of every event or decision almost in real time. Since the release of a crack, our revenue dropped by almost 25% as many people immediately cancelled their subscription. It also means that we have lost some money intended for possible new developers which delays our progress. So short answer is yes, it does affect us, as we are still quite a small company and we can feel it.

As for “fighting” it, I personally try to reach the warez community to better understand it and start a discussion and possibly to find a good solution which will fit everyone.

Ondra: From the technical standpoint, I had to make cracking at least a bit harder than flipping single byte in the file. I had no experience with it, and our main goal was that it must not make the product worse for paying customers, and it must not slow down the development. Considering these constraints, I am satisfied with the result. I was quite surprised how long it took until the first pretty limited attempt appeared on some Chinese website.

How many people are involved in Corona development? Is it a costly business?

Ondra: Yes. Developing real-production software is a slow, painful, risky, and expensive process, especially when working with obscure APIs such as 3ds Max. We have 3 core programmers and several more working on plugins or specialized tasks such as scripts.

Adam: Main expenses are fees related to salaries of the developers/support staff, everything else (offices, computers, coffee, …) is quite cheap. We do not need fancy offices or “startup” perks. We want to give our employees decent pay and an interesting project to work on.

You have a nice and live community on your site. How do you use the community feedback during development?

Ondra: It is essential for us. Corona was literally built on user feedback. But lately we have started to feel the need to filter feature requests more. It seems that after all really essential features got implemented, people started requesting more and more exotic ones. While we want to help those people achieve their goals when rendering, we want to keep Corona fast, lean and clean.

Why artists choose your product?

Adam: Our goal is to create a simple to use, yet powerful render engine. Our intentions are then reflected in our software itself - Corona Renderer does have a simple UI which is free from any unnecessary buttons, knobs and arbitrary constants - we do not believe that the artist should spend time trying to understand the mathematics behind rendering algorithms and trying to find the “correct” values which will work in their scenes. Corona Renderer should work almost with every scene on default settings. This way the artists can focus on what is important - their art. I do believe that the simplicity of our product, combined with speed and visual quality is what artists find appealing.

What features are special and why?

Adam: One of our best features apart from simplicity is the interactive mode. It does support all the function of the final frame rendering as it uses 99.9% same code. So what you see in interactive mode is also what will be rendered in final frame, without any guesswork. It is a really powerful tool in Corona’s arsenal. Download the demo and see for yourself. :)

Who is the target customer for Corona?

Adam: Corona Renderer is widely used in arch-viz at the moment and it is also where we aim with the current version. We do have many users also from other fields, but at this stage we will keep focusing on archviz and product design. One of our long term goals is also VFX, but I have to admit that it will require some time.

What are current hardware (and software) limitations that are the most cumbersome in development of fast and high quality industry grade renderer?

Ondra: The CPU performance when running Corona is limited by the randomness and unpredictability of the algorithms used (such as ray tracing/path tracing). From the point of view of the computer, rendering is a chaotic process. The CPU never knows what will happen next, so it cannot prepare correctly. One pixel difference can mean missing one object and hitting another one kilometer away, with entirely different shaders. GPUs have even harder time with this, as they are specifically built for very predictable workload (which is why their performance under optimal workload is so high). The biggest software problem for us is having to use various badly documented and potentially unstable third party APIs. Building a rock-solid renderer on top of a partially undocumented and sometimes unstable API is an extra challenge.

What is the future of Corona Renderer?

Ondra: It seems as we add more and more features, user focus shifts from feature requests more to stability and speed. So that is the immediate plan - fix all outstanding bugs, and try to make it even faster.

Adam: Short term plan is to make it the best archviz renderer on the market. Later on we would like to get to VFX and possibly we would like to start on other closely related projects. We would like to experiment with real time, cloud rendering and possibly a 3D app.

More supported software? Maybe Oculus support?

Ondra: We want to finish the Maya and Cinema4D plugins as soon as possible. We are also looking at other applications such as Sketchup, Modo, and Lightwave. Virtual reality gets requested by our users lately, so we might try that as well.

Are there any new functionalities that you can talk about?

Ondra: While the next version (1.3) was supposed to be centered on bugfixes, we could not resist adding a few new features. So expect an animated Corona Proxy, Corona Layered Material, object-space reflection anisotropy, vector displacement, and more.
We also have some long-term plans, but we will keep them private for few more months.

Do you plan to expand your product line or will you focus entirely on Corona development in the next months / years?

Adam: We have recently published our Corona Renderer for 3ds Max roadmap, so users can see what lies ahead in the nearest future. We are not putting everything there, as we would like to have some surprises as well. We will certainly start working on small side projects this year which will be closely related to Corona Renderer. For example our own material repository and possibly a model bank. From licensing and pricing point of view we would like to introduce “Personal Learning Edition” version and/or one cheaper license option.

Do you plan to expand your crew or make more offices?

Adam: We have two offices at the moment. Big developer office in Prague and a smaller office in Prostejov where all the administration, accounting and billing happens. We do not plan offices in more cities, but I have a feeling that we might have to move to a bigger space later this year as the capacity in the Prague office is starting to be limited.

Ondra: We are always looking for talented programmers, as they are in short supply. So if you are reading this, know C++, and want to live and work in Prague, do drop us a line ;).

Do you plan some cooperations with other CG companies?

Ondra: We are already cooperating with many, such as render farms, studios, Intel’s Embree team, AMD FirePro team, and with other software developers in order to make Corona compatible with their plugins. We will certainly continue with this in the future. And if any other opportunity arises to cooperate on an interesting project, we will of course be interested.

Thank you for an interview!