Meet Martin Duerr

Where are you from originally, and where are you located now?

I was born in Munich, Germany and have my office now 40 km west from the Bavarian capital near the lake Ammersee. For my work on Models from Mars I switch between Munich and Boston periodically.

OK, what was your very first job Martin? I mean before 3D.

After studying graphic design in Munich I was hired by Giesecke & Devrient, a printing company, and received a special eduction in banknote design and banknote printing. There I worked on banknote designs for Asian and African countries. This was a funny time and I learned important lessons about imparting high-precision detail and how to conceptualize my work before I really start with a new project. After that I then worked in several different advertising agencies before I got addicted by 3D...

You have been involved with 3D for like 25 years! Tell us, what was the first commercial 3D modeling system you ever used? Was it a CAD product or what?

Oh, I had a lot of those first tools in my hand, Evans and Sutherland stuff for example, but the very first one I did really serious 3D projects with was Swivel 3D made by Paracomp. That ran on a Mac II. Later I worked with Strata but soon graduated to Lightwave 3D on the Amiga platform (version 3.5). Since then, in addition to modo, I also have gained experience with Maya and other popular 3D applications.

In what ways have you personally been involved in 3D?

I’m teaching speed modelling at the BAF (Bavarian Academy for Television) and I have been in the very first jury for the Animago years ago (1997). As mentioned before I worked as a modeler and concept artist for film and TV projects as well on some game titles. I held lectures about 3D scanning and modeling and I acted as a product designer for software tools. And you will be interested to know that I conduct customized training sessions with modo for the car industry and other high-end clients.

How did you come upon modo?

Well, I think I did so in the only logical way (laugh). As an old Lightwave user it was really an inevitable step to jump onto the modo wagon. When you do modeling as a serious job you simply can’t bypass modo. modo fits like a glove to my 3D modeling workflow, whether the project might be small or huge. And I can tailor it to my needs without losing control over everything. A really nice app.

So tell us about Models from Mars and your role there?

Models from Mars (MFM) was originally initiated by Clint Clemens, a famous action and car photographer. We met two years ago and developed the business over the period into what you see now. While Clint is the one dedicated to imaging in MFM process, I’m responsible for the modeling part of our software development. We have specialized in generating lightweight 3D objects that originate from densely sampled point clouds. As you see in the Salve Courtyard example, our proprietary software outputs automatically a very simple low-poly object (.obj format, readable by almost all 3D apps) and a corresponding displacement map pinned to UV coordinates.

In this case, we used a laser scanner with six different scan positions and a proprietary solution for the cloud alignment and the matching of the photos which have been shot with a hi-res DSLR camera.

With that we achieve a sub-pixel accurate matching of geometry and photos. Alternatively we can provide clients with normal maps if the object will be used in a real-time environment. In opposite to other point-cloud tools around which lead to huge polygonized objects, we do what we call an “information split.” Customers get a lightweighted, low-poly object and a displacement map. This enables them to work with point-cloud scans without the burden to edit millions of polygons. And it’s very easy to show those results in modo.

Can this scanning and topology re-creation be done on smaller objects?

Sure. We can go from kilometers to micrometers. For us as Models from Mars it does not depend how big or small an object is, we just process point-clouds whatever 3D scanner generates them. We even can work with sonar data. Take a look at the candy bar example. The originally scanned chocolate bar had about 18 million points, while the end result has 227 polys. The displacement map is a 4K map, the same size as the color map. The scan was done with a white light scanner.

This image of the car with the canyon wall behind it…how was that one?

This was one of the very first images we used to promote Models from Mars. The original scan has about 70 million points. We ended up with a low-poly object of about 3000 polygons. The funny thing with that is, that almost none of the people we sent it get the fact that the rock it is all about. They concentrate on the car and tell us that they have seen such composited car renderings often in the past. When we tell them that the rock is a scan and then we show the naked object, they are completely amazed.

I sent the scene without the car to a friend of mine to get his reaction. He called me and asked me why I would send him a photo of a rock. He had seen such images before, but I wanted to show him something new... A good example that proves the quality of our process.

So, I have to ask why modo?

modo is the most flexible 3D tool around and it’s ultrafast in terms of workflow and interactivity. When we show to clients the results our software is able to output (object and maps) we have to convince them in seconds not in hours. So whether we use displacement or normal maps to show that our objects are usable in high-res media like print or in realtime applications we can do that on the fly in modo. No complicated setup, no hassle in previewing the files. Simply switch between the previz rendering and the OpenGL viewport. Customers get immediately what we want to show them.

What other tools are in your arsenal?

All apps you need to stay within a professional environment and to communicate with the clients seamslessly. Lightwave, Maya and some other 3D apps. As well as a lot of 2D apps for image editing and manipulation.

What are you excited most about in modo 401?

Speaking as a Windows user: 64 bit! Speaking as a modeler and texturer: all those new features about texturing, lighting and modeling. I know it sounds a little bit general, but modo is known for the fact that you wont be distracted by useless features from Luxology. Everything there is well designed and elaborate. So any new feature we get was well worth the wait for 401. Please keep in mind that I was really happy with 302! modo is like a Swiss-Army knife, once you use it, you never want to be without it.

Looking forward, how do you see the business of scanning affecting every day users of 3D software?

It will change the 3D business effectively, for sure, especially with our toolset. Today most 3D scanning is a dead end, especially for the 3D guys out of the media, entertainment and visualization worlds. They would like to use 3D scans but end up frustrated, because they can’t and do not want to edit millions of polygons. This is not only a tedious and time consuming task, sometimes the results are impossible to edit. It’s like having a Ferrari, but no key to drive it. So a lot of people would like to use “smart” 3D scans in terms of usability, interactivity and convertibility. That’s what we specialize in. With our way of an information split, you get a lightweighted object and an appropriate map (displacement and/or normal). And, because we get the detail directly from the point-cloud and not from a polygonized scan we don’t lose detail. So we take the pain out of the way using 3D scans. In modo with these results, you can simply add or remove detail on the scanned scenes, very interactively and without pushing gigabytes of data. Isn’t that how we all want to be creative?

Thank you Martin!