I created the Tokyo Penthouse images a couple of years ago, purely as a portfolio piece blending photography and CGI together. I was looking back at them and thought it would be good share these images of the various stages of creating the image - the raw photograph from the studio, the quick render with a 3D model person as a placeholder to help set up the shot, and then the final composition. When you're after lifestyle images of a property that hasn't been built yet, this is the creme de la creme.
Over the last few months, I have been designing a house for a private client in Joshua Tree, California. As the design developed I thought that it was something that would catch a little bit of attention in the press if I showed it in its best possible light (I never imagined it would go viral). I was already using the computer to make simple 3D maquettes and with the design locked off, it was fairly straightforward to build a detailed model of the house on the computer. The context was the real challenge. The client had provided me with a topographical site plan as a PDF, which I had been sketching over. He had also provided me with photographs of the site, but the photographs weren't quite sufficient for me to use to make some nice images of the scheme - I wanted to be able to surround the building with context, in 360 degrees, so that we could get the shadows and reflections across the building. To solve the context problem I devised this rather satisfying workflow which I thought warranted sharing.
First of all I opened the pdf in Adobe Illustrator and exported it as a DWG. I then opened the drawing in Autocad and scaled it to the right size, cleaned it up, and joined fragmented contours into continuous polylines (I later realised that this could have been done automatically in 3DS Max with SiNi plugins). This drawing was then imported into 3DS Max and each contour was raised to its correct height (this was tedious and I'm unaware of any shortcuts).
With the contours set out in 3D space I then used the Terrain Compound Object to create a surface from the contours. This generates a surface that is detailed but an ugly mesh to work with - it's all triangulated.
At this point, I identified some areas around the building that where cliffs and rocks, rather than sand surfaces. I duplicated the terrain surface, drew a 2D polyline around the area that I wanted to be more detailed and cropped the landscape surface down.
Next, I centred the surface's pivot point to the centre of the surface and then added a point helper to the scene and aligned it to the pivot point. This is to help align the detailed surface in the right place, later in the process.
Now I did a quick UVW Unwrap on the surface.
I moved the surface to 0,0,0 in model space and exported it to Mudbox. (I found that if I didn't move it to 0,0,0 first I had difficulty viewing the mesh in Mudbox.) With the UVW mapped, nicely regular mesh in Mudbox I could work detail into it easily, before taking it back into 3DS Max and using the helper to locate it in the right place.
The final step was just to adjust the original, main landscape surface so that the detailed surfaces sat over it well. I made a bunch of rocks in Mudbox and then used Forest Pack Pro to scatter these and vegetation over the landscape.
SiNi's plugin was really the kingpin in this process. If you do visualisation work and aren't familiar with SiNi I recommend that you check them out.
I've just been doing some concept design work in 3DS Max. With the design finished I needed to extract a 2D drawing that I can clean up in Autocad and then pass on to the set builders. It's a little bit of workflow that is always helpful to do with architectural projects, say when you're working on a competition, but can be a bit of a pain.
Well, I'm quite pleased with this workflow in the end. I viewed my 3D model in elevation or plan and drew splines over it using the 2.5 snap. This is pretty quick to do but can leave you with lines floating all over the place in 3D space which are then a pain to work with in Autocad. Scribe by SiNi Software has a flatten spline button, and with a click of a button your splines are all flattened to the C-Plane. Ta da!
This is a little tutorial for SiClone, a neat 3DS Max plugin that I’ve been using recently by SiNi Software - https://www.sinisoftware.com/
If you have any questions or queries please ask away in the comments section below and I'll try my best to help, and if you enjoy the tutorial please sign up to our newsletter so we can keep in touch - http://eepurl.com/cf3NM5
On Wednesday evening I'm presenting my work and, in particular the making of Space Baby, at the 3DS Max London Users Group. I'll be showing some of my workflow and how I created Jack's space suit with Marvellous Designer.
For more information go to - https://www.meetup.com/3DSLondon/
Come along and say hello over a pint.
If you've not been before you should, there's always interesting stuff being shown (this week being a perfect example) and beer - a winning combination!
This visualisation is of a proposed 8 storey office building, with retail and coffee shop at ground floor. You can compare the final image with the computer model below to see what we did in 3D. I prefer to create as much of the image as possible in the 3D model and keep photoshop work to a minimum. Partly that's just how I like to work, but it has the giant benefit that it increases our ability to adopt client changes as late into the process as possible, and minimise the impact of that one final tweak to the design.
I'm working on a project at the minute that needs a good Spitfire model and so having spent some time getting it looking all nice I did this for a bit of fun.
Here's a little test render from the next series of images that I'm working on for Getty Images. They're proving rather fun!
The Spitfire model was scavenged from somewhere on the internet, but was a little temperamental and prone to crashing Max. So having tweaked and polished it I'll add a link to my model as soon as I have a chance.
People often ask me how does CGI work, or how do I create my images and I've always felt like I wasn't giving a particularly good answer so I've made this short video as an introduction for the uninitiated.
And signup for our newsletter to receive a 3D model of the cup - whitakerstudio.co.uk/how-does-cgi-work-newsletter-signup
A free piece of 3D software for you to play with is Sketchup - sketchup.com/
So, last week I finally finished my series of images Space Baby, and they are now selling through Getty Images. I've been chipping away at them between commissions and they've been great fun to create, so I thought you might enjoy a little break down of what went into making them.
Each shot was sketched out before hand. Sometimes this was a quick pencil sketch and sometimes a very dodgy bit of photoshop, as below. As these were never intended for public consumption I didn't worry about spending too much time on them, they were really just for blocking out the principals of the shot - the camera angle, lighting etc.
We did all of the photography work at the brilliant Cloud & Horse studios in north London, and with a rough idea of the shot that we were after we could set up the studio to give us the right lighting conditions. However, unlike when I was shooting Jennifer for the Tokyo penthouse we couldn't be that precise - a toddler just doesn't stand still that long. We had to be far more fluid and adaptable.
Back at the computer I was able to make my selects from the shoot and begin the hard work.
The first step was to place the photograph of Jack behind the model in 3DS Max. I could then move the virtual camera into the right position so that model and photo would align. With the camera setup and locked in position I could then work on the 3D model, adding detail where it would be seen without wasting time on stuff that would be out of shot.
With all the shot modelled up I then went back to Photoshop to mask out the background in the photograph of Jack. In the instance of this photo I had to use 2 or 3 photographs and composite them together to remove any trace of gravity and me holding Jack.
There was then a little back and forth to double check the camera angle and lighting in the model before rendering out the final image of the spaceship. This was then taken into Photoshop where it could be combined with the photograph of Jack.
I render out my computer images as 32-bit images in multiple passes so that I can adjust each element (reflections, refractions, lighting, etc.) in Photoshop. This is particularly important on images like this to bed Jack into the image and adjust the little details that make him look like he's actually there rather than just pasted on top.
Every detail was carefully embellished to give the series a comfortable real world feel. For instance the magnetic letters were modelled from scratch and then textured with a Vray SSS2 material. This material allowed me to give the plastic that slightly translucent feel that you get on some children's toys, which then glows a little when it's back lit.
Jack's spacesuit brought some unique challenges, particularly for someone who specialises in architectural imagery (which is normally hard and rectilinear). In the end I took a model of a small boy and tweaked it to give him the proportions of a toddler. I then rigged the model with a skeleton in 3DS Max and exported that to Marvellous Designer. In Marvellous Designer I could then use the toddler model like a mannequin and slowly craft the spacesuit over him. With the fabric parts of the spacesuit finished in Marvellous Designer I then animated the mannequin back in 3DS Max to move from standing to the pose for each shot. I then took that animated mannequin back to Marvellous Designer and animated it while it wore the spacesuit. The spacesuit was then finally taken back to 3DS Max for texturing, to add the hard elements, and then finally adding it to the rest of the scene.
If you have any questions feel free to ask them below and I'll do my best to answer them for you.
I made this model just before Christmas for an image I was working on and then revisited it last week to get it ready for selling online. While getting it ready for sale I made this little film of it which I find oddly enjoyable and hypnotic if watched on repeat. So here it is for your enjoyment too!
You can purchase the model on turbosquid.
Here's one of the props that I've created for a fun set of images that I've been working on this week. The actual images will follow soon...
I have had a bit of an obsession with Thomas More's Utopia for years now. The text is wonderful, and parts of it have a real poignancy in the modern world. More wrote the book back in 1516 and, for anyone who hasn't read it, it largely comprises of a conversation between More and a traveller, Raphael Hythloday, who has just returned from the island of Utopia.
Almost every made up name in the book is a pun or reference, so Hythloday is a Greek compound meaning expert in nonsense and Utopia is derived from the Greek prefix ou-, meaning not, and topos, meaning place. No-place or nowhere. Utopia doesn't exist. And so computer generated imagery feels like the perfect vehicle to explore Utopia, and 2016 felt like an apt time to start this series as a means of reflecting on the political atmosphere in the UK and Europe.
The image above is the first from the series and is based on the portion of text copied below:
...entry into the bay, occasioned by rocks on the one hand and shallows on the other, is very dangerous. In the middle of it there is one single rock which appears above water, and may, therefore, easily be avoided; and on the top of it there is a tower, in which a garrison is kept; the other rocks lie under water, and are very dangerous. The channel is known only to the natives; so that if any stranger should enter into the bay without one of their pilots he would run great danger of shipwreck.
It is my intention to finally exhibit the series, although it will take a little while to reach that point.
A recently completed image for A-Zero Architects of a house that they're currently working on on the edge of the river Thames. (Car choice was critical for the French client.)
Following on from the image that I shared back in July here is another image from the series.
Since experimenting with 360 videos last week, Facebook have now introduced 360 photographs and it turns out it is incredibly easy to produce and upload cg images, although it does require you to go through a couple of steps. So here is a little tutorial for anyone needing to create 360 content.
We use 3ds Max and V-Ray here at WS so the first part of this tutorial will describe the settings specific to our pipeline, but can be translated to other renderers. The second part will look at what you need to do to prepare your rendered image for upload to Facebook, regardless of software used.
Stage 1 - Render Settings
Set up your camera as you normally would.
Render settings are largely the same as your normal preferences, however under V-Ray > Camera select Spherical Panorama for type and check Override FOV, entering 360.0 for the horizontal override and 180.0 for the vertical override.
Finally, Facebook needs your image to be in a 2:1 ratio with the maximum recommended dimensions being 6000 wide by 3000 high.
Now you can hit render! We normally save out as 32-bit EXR with render passes and you can still do this, editing your image in Photoshop as you see fit before saving out as a jpg.
Stage 2 - EXIF editing
With your image rendered you now just need to add some additional information into the EXIF data so that Facebook interprets it as a 360 image rather than a normal flat 2D image.
For this you need to visit theexifer.net. Upload your image and then click eXif.me. Here we need to enter Ricoh for Camera Make and Ricoh Theta S for Camera Make. This will fool Facebook into believing that the image was taken with a recognised 360 camera.
Now you can download your photo with its corrected EXIF data and upload it to Facebook.
I've just been experimenting with producing content for Facebook 360. This first test is rough but quite good fun I think. Check it out - https://www.facebook.com/WhitakerStudio/videos/1198693146830632/
I'm delighted to present a short film that I've been working on recently - Hechingen Studio
Hope you enjoy it!
Someone once said to me that if you are going to make a jeans company concentrate on making a really good pair of jeans before you start selling t-shirts. I thought it was pretty good advice so I've been concentrating on stills up to now, making sure that they are as seductive and polished as can be. However when I was an undergraduate at uni I was a bit of a geek and used to teach animation on the post-graduate course. A couple of years ago I made a short film for fun with my brother and a bunch of friends and it went on to win a film festival in Canada. So we know a bit about making nice films. We approached this film in the same fashion as we would approach a live footage piece, working up a storyboard and then animatic, before editing and fine-tuning the shots, then we worked up the animation to what you see above.
For the geeky amongst you all animation and modelling was done in 3DS Max, rendering with VRay and post-production in Adobe. Rebus Farm was used to outsource some of the number crunching and the music was found on Musicbed. If you'd like to know any more feel free to ask in the comments section below.
You can read more about the building in this article on Dezeen.
Last year I modelled a lot of beds for one project, and as with most things I definitely improved with practice. I ended up getting the most satisfying results modelling the sheet in Marvellous Designer, then the duvet and pillows back in 3DS Max using cloth simulation, before returning back to Marvellous Designer to model the throws. In this instance I then used VRay fur on one of the throws. The great thing about modelling the throws in Marvellous Designer is that they come into Max already UV mapped making it really easy to change the texture of them.
If it would be helpful let me know and I can put together a tutorial covering the process.
For anyone without the time to start from scratch you can get this bed from Turbosquid.