An evolutionary step in imaging

Dimension is a world leading volumetric video and 3D capture studio, and first Microsoft Mixed Reality Capture studio partner.

Join our Partner Programme

Video just got real

Dimension captures life in wonderful volumetric detail, providing a step change in the realism that can be achieved for the creation of virtual humans and environments. Inject new life into stories, games and experiences for both immersive and 2D media.

Showreel Studio tour

We capture humans and other living creatures in photo-realistic quality.

Our state-of-the-art volumetric video stage is used for performance capture, and we have a separate scanning stage for 3D character and head capture. Whether creating volumetric video or 3D images, we record shape and detail from every angle, for realistic content that can be used across multiple platforms.

Watch Monk demo

We create environments for ever richer virtual worlds.

Our environment creation services include photogrammetry and set capture, while our artists are experts at creating virtual worlds within Unreal and Unity. We also offer prop scanning and modelling to bring any environment to life with the right style and vibe.

Learn more

Performance capture

New immersive platforms demand a re-think in how to capture human performances so they are believable, emotional, and exciting. Volumetric video enables you to record actors, performers, and animals to create virtual and mixed reality experiences that feel more real, or produce multi-dimensional magic for 2D and touchscreen devices.

Showreel
Watch Ballet demo
Immersive experiences

Immersive experiences

Bring powerful and enchanting performances to life to create stories, games and original experiences for virtual reality headsets. No “uncanny valley”

Mixed Reality Experiences

Mixed Reality experiences

Blur the line between the physical and virtual as characters step through the frame for incredible Augmented and Mixed reality experiences. Ready for mobile and Hololens devices.

Interactive 2D Experiences

Interactive 2D experiences

Volumetric video opens exciting and dynamic new creative possibilities for 2D content. Create amazing interactive 2D touchscreen experiences and impossible camera moves for film and VFX.

Get in touch to find out more about the incredible creative possibilities

3D character capture

Our 3D scanning stage uses 190 cameras to shoot volumetric images of bodies, faces and heads. Shot at maximum resolution, capture photo-realistic 3D stills or hyperreal 3D animation-ready characters for games and interactive.

Wimbledon Crowd

Bullet-time capture

Our team have over 25 years experience of multicam and array based shoots, and pioneered the technique behind the epic Matrix bullet-time shots. Our 90 GoPro camera array is adaptable for both studio and on-location filming. See examples below.

Sky

Why choose Us?
Watch Sky

CAT Footwear

Why choose Us?
Watch CAT

Nike ACG

Why choose Us?
Watch Nike

Environment capture

We are masters at creating virtual environments, using both photogrammetry and modelling. Whether you need to capture the set of a period drama or ancient ruins, build a spaceship in Unity, or a rain forest in Unreal, we have an experienced art and development team to help you create something extra special. See examples below.

Dimension Partner Programme

Create amazing apps and experiences featuring volumetric video.

Watch AR demo Download VR demo

Join our Partner Programme for early access to the Windows Desktop, VR, iOS and Android SDKs. Apply today and be the first to create next generation AR, VR, 3D and 4D experiences featuring volumetric video using Microsoft Mixed Reality Capture.

As part of the programme you will receive:

AR and VR Player Apps

The applications to preview sample captures along with your own volumetric captures for Windows Desktop, Oculus Rift, HTC Vive, iOS and Android.

Windows Desktop, VR, iOS and Android SDK

The tools, plugins, documentation and sample projects to help you build your own applications with Unity and Unreal Engine 4.

Volumetric Capture Samples

A selection of volumetric captures for you to test and have fun with.

FAQs

What is volumetric video capture?

It’s material that looks like video but exists volumetrically in 3D, allowing the user to view performance from any angle, any time. Dimensions volumetric video capture stage uses Microsoft’s Mixed Reality Capture technology.

What are the applications?

Once rendered, we’re able to place the filmed subject into any AR, MR, VR, 360, or 2D video application.

What does volumetric capture offer that green screen doesn’t?

Green screen captures images of people for inclusion in 2D Media like film or print. Volumetric capture however allows inclusion into media requiring 3D assets such as video games, VFX, AR and VR

Does this mean we can re-frame once shot?

Absolutely. As soon as your rendered subject has been imported into real-time 3D software, you can position the model at any angle, anywhere in the scene.

How many cameras are in the rig?

The volumetric video capture stage has 106 synchronized cameras.

Camera specs

4MP global shutter cameras.

Why these cameras?

We have 53 RGB cameras which read and record the colour required for the .png texture maps, and 53 IR cameras that record depth and position in space for creating the mesh.

Lens specs.

Kowa 16mm.

Can I shoot high speed?

We can shoot up to 60fps, although please note that quotes assume shooting at 30fps unless stated otherwise.

What are the limitations of this technology?

Our software relies on light and colour to join up the images captured by each camera. Different materials absorb infrared at different levels and we rely on this information to create a detailed mesh, so we carry out a test day on all wardrobe to ensure maximum capture information and provide guidance for final shoot.

What can clients expect from the studio personnel during pre-production?

Studio personnel are available for pre-tests, pre-production calls and planning, and are available to advise on the best way to set up for the shoot. We always advise a test shoot is done using the talent in the exact wardrobe and make up as per the shoot day. This is to block out the action and help us identify any potential movement or wardrobe issues.

What crew are provided with the studio hire

Technical Director, Studio Manager, Studio Coordinator.

Do you still light your subjects as you would in traditional film making?

Not as dramatically (contrasty) due to a) the computer vision algorithms used when unifying the footage from each camera and b) artistic lighting on the stage being baked into the texture map so there’s no going back. It's better to globally light a subject (flat light) then make the necessary modifications in either the real-time engine, 3D software, or in processing.

How many people can we film on the stage at any given time?

It depends on the action they’re performing. If one subject is occluding the camera’s view of another person, the information won’t exist for the software to build from. We’ve successfully shot 3 people at once before. Once rendered, all subjects in the capture area will live together as one mesh. We can however, bring multiple individual captures into the same scene.

How does Audio capture work?

There are 8 channels of shotgun microphones that are equidistant around the performer in a circle and input, via a mixer, into a pro-tools session live on set. All audio is then trimmed and synchronised with the take. Where sound really matters, an audio engineer can be hired for the shoot who can utilise other options such as lapel mics, boom mics, etc.

How do we review footage?

Whilst filming, a constant 2D live feed can be taken from one of the cameras within the rig. After a take, we’re able to render out a single frame of footage to view in 3D, although rendering time for this will take around 20 minutes.

How much can we shoot in one day?

We can shoot over an hour’s worth of content each day. The limitations come from how much you can afford to render. Within the working day you need to factor in time to decide exactly which takes get sent to the rendering farm. Selects need to be made on the shoot day so that data can be processed overnight.

How do we decide what gets rendered?

Rendering is expensive so be selective. If the shot doesn’t work, bin the take and go again.

Are traditional crew such as a DP, Gaffer, and sound recordist still relevant?

If you decide artistic lighting on stage is going to provide you with better results than global lighting conditions, then yes, a DP can be used to recreate the lighting conditions you want to mimic. However, the lights that are pre-rigged with the stage remain your only source of light. That said, we can change the intensity and colour.

What’s the volume of data created per 1 minute of footage?

10GB of raw footage per second, 10GB per 30s of OBJ, 10MB per second of geometry based MP4 format.

What’s the render time?

Around 24 hours per 6 minute of RAW footage. However, all files go through the render process several times before being presentable to client, so it’s reasonable to expect a turn-around time of around 2.5 minutes of footage a day.

What’s the workflow after the shoot?

The length of the production cycle is dependent on the complexity of a project (i.e. length and type of capture; number of subjects being captured; number of different takes). For the most part though, the cycle is similar to that of a standard video production, which includes pre-production, hair, make-up, wardrobe, etc., followed by a period of time for processing data.

At the end of each shoot day, we work to identify priority takes which are sent off to be processed into a form that can be consumed in your playback engine. As a guideline it often takes a couple of passes to calibrate processing settings for your particular content, after which it’s reasonable to expect a turn-around of between 2.5 minutes of footage a day. Every project though is slightly different, which impacts turnaround time. We can give more accurate estimates based on specific proposals.

What file type do you export once rendered?

In it’s rawest form, our files are OBJ/png. To reduce the size of the files sufficiently enough to work in real-time engines, we create geometry based mp4 files using proprietary software. The complexity of the data stored within the mp4’s is dependant on the intended platform. For VR, we recommend a mesh poly count of 20-40k and a 2k texture map. For mobile we recommend a 10k poly count mesh and a 1k texture map.

What software can read this file type?

3D mp4s need to be read by our specialised plugin which can live in games engines, 3D software such as Maya, desktop, tablet, mobile, VR, MR, and AR headsets. OBJ/PNGs can be read by many digital content creation apps.

How many mp4 files can be read in Unity at any one time?

This depends on the density of the mesh but as a guide, we recommend no more than 6.

Can the rendered files be sent to any post house to work on?

Yes they can providing they have the plug in which we’ll supply to anyone who asks.

Is the output possible to be played directly by mainstream video player or only compatible with Unity

The mp4 files can be viewed via a plug-in for Windows 10 devices through Unity. There are plans to write a plug-in written for Unreal engine, with timing TBD. The OBJ/PNG files can be viewed on a variety of platforms with some development on the clients part.

What is the bandwidth required for the video (within ten MB) and Is it possible to stream with today’s average bandwidth for consumer?

We can compress within ranges typical for HD video. We use h.264 for the mp4 files and can compress to a client’s specifications. Higher compression results in some image degradation. Files shipped on Hololens were compressed at different rates, between 8 Mbits on the low side and 12 Mbits on the high side.

Is it possible to shoot outdoors:

Our volumetric video stage is designed for internal performance capture and is not mobile at this time. We do capture environments outdoors with our photogrammetry services, and Hammerhead offer outdoors multi-cam and array based shoots outdoors. We can also create environments using real-time engines.

What level of accuracy of the model can be achieved?

The apparent "accuracy" of the model is dependent on several factors: the number of polygons used to represent the model, the resolution of texture used and what the camera configuration is able to see. Everything must be balanced for playback. For example, we decimate our mesh from thousands of thousands of triangles down to the order of 20-40k triangles for real-time playback.

What are Microsoft’s thoughts in terms of volumetric video’s strategic value and besides Hololens, how significant is volumetric video to other types of consumer devices?

Realistic human representation is critical in the creation of more personal experiences for Immersive experiences across VR, MR and VR, and Microsoft’s investment in volumetric video capture technology over the last seven years is proof of that. Volumetric video capture carries us past the so called “uncanny valley” seen and experienced with artificial content. For experiences where an authentic human performance is important, volumetric video allows viewers to engage the content in natural and comfortable ways. Before the advent of HoloLens and the proliferation of VR devices, Microsoft spent time working on touch (and mouse based) applications using this content type. Our experience with conventional UI and with a variety of AR and VR driven experiences gives us great confidence that volumetric captured and rendered humans add value to a wide variety of scenarios.

Get in touch

Dimension
Unit 35
Wimbledon Business Centre
Riverside Road
London
SW17 0BA

hello@dimensionstudio.co
+44 (0)203 8890000
@hello_dimension

Dimension is a proud partnership between
Hammerhead, Digital Catapult and Microsoft.

Hammerhead
Digital Catapult
Microsoft