The other day when I posted these 3D images by Lasse Rode
, I got some questions about how it is done, and as I was wondering about this too I decided to ask Bertrand Benoit
, a professional CG artist. He took some time out of his busy day to explain the technique to me, and also sent me some pictures from his work process that I'd like to share that with you, in case you are as curious as I was.
Me: – Hi Bertrand, nice to hear from you again! (I posted his work before
, so we have been in touch earlier.)
Bertrand: – Hi Emma! I was very happy to see Lasse's images on the blog, not just because they are gorgeous, but also because I had been wondering if you were aware of how big an influence you are on digital artists, especially those like me who specialise in architecture and interiors. Strangely enough, stylists and CG artists don't tend to mingle and although CG artists are acutely aware of the stylists' work, the opposite is not true.
Me: – Thanks, I'm so happy to hear that you find inspiration in the pictures I post! Some of my readers were wondering about the creational process of this kind of pictures, how does it work? Do you start by using the original photo and sketch your picture on top of it?
Bertrand: – Since it is 3D, it is easier to start from plans or blueprints than from photos. This is because the dimensions should be as accurate as possible and these are sometimes difficult to infer from a photo. One can start from a photo, though, as Lasse has done (it is called photo-matching and often used when a CG building has to be integrated into a real photo). Working from photos is also important in terms of getting the materials, the lighting and other object properties right. Most people in this line of business aim for photorealism in their work, so this type of exercise - to work from a photo - is really useful and certainly the best way to improve. The software is a bit like a camera - it can do everything but what counts in the end is the eye and the composition, which is why the best CG artists in the architectural and interior areas (people like Peter Guthrie for instance) are those with a photography background.
The pictures above give us some idea of how the process looks.
Me: – I think I get the general idea, but could you please take us through your process step by step?
Bertrand: – There are a few steps involved in photo matching. The first would be to try and find out (or to guess from experience) what camera settings were used when taking the photo, particularly the focal length, and to ask yourself simple questions about the setup (where would the camera be positioned in the room; how big is this door frame or this sofa supposed to be; what should be the distance between them...). Then you would create a CG camera in an empty scene that matches these settings as closely as possible. The next step would be to load the photo in the viewport of your 3D software and to start sketching a simple 3D scene with placeholder geometry where the different walls and objects in the photograph would be. Once you are done it is just a matter of tweaking your camera to match the image as closely as possible and proceed to build your full scene. It is not an exact science. Rather, a lot of eyeballing is needed. A good instinct for space and dimensions. Some people are just better at it than others. I hope that clarifies it a bit.
Me: – Yes, the process is totally clear to me now! Thank you so much for taking the time to explain it all to me and my readers, I really appreciate it.
Below are some great examples from Bertrand's portfolio
, all modeled after pictures originally created by some great photographers and stylists. The last pictures were inspired by this blog post
, and I actually think the 3D images are better than the originals in this case. I hope you enjoyed the interview!