Canon EOS VR System technology explained

Immersive VR content is easier to shoot, edit and experience than ever before. The innovative dual lenses and VR software at the heart of the Canon EOS VR SYSTEM – alongside the growing availability of untethered VR headsets – are transforming the way in which stereoscopic VR videos and stills are created and consumed.

The roots of Virtual Reality run deep – all the way back to 1838, in fact, when the first stereoscope was conceived, although we had to wait until the mid-1980s for the first VR head-mounted display (HMD) to become available. Since then, advances in imaging technology and computing power have made more realistic VR content accessible.

It was a step-change in imaging technology – the launch of the high-performance Canon RF mount, with its large diameter and short back focus – that made the EOS VR SYSTEM possible. Canon’s complete solution for capturing stereoscopic 180° VR includes a line-up of stereoscopic VR lenses, software and plug-ins that help more people to create high-quality immersive content more often. Virtual reality is an increasing part of many of our lives, be it for training, education or entertainment, and the EOS VR SYSTEM facilitates a more seamless content creation process.

Here we'll run through some VR basics and explain the difference between VR formats, what a dual lens is, and why interpupillary distance is an important consideration when you’re looking to maximise the 3D effect.

Two popular VR formats are 180° and 360°. With 360° VR, the viewer is surrounded by the environment and can look in all directions, including behind them. "Street view" apps normally use this type of presentation. The 360° content can be captured in various ways, depending on the end use. It can be done via a consumer "action"-type camera with two back-to-back lenses and sensors, or a professional setup using larger cameras and rigs. In either case, though, the output has to be stitched together to generate the 360° environment, and this can result in visible stitch lines, parallax problems or other alignment issues. Shooting stereoscopic 360° VR is even more complex, requiring more expensive and bulkier equipment, such as a dedicated multi-lens/sensor all-in-one camera or a multi-camera rig arrangement.

A head-mounted display (HMD) or VR headset will provide a more immersive experience when viewing 360° VR or 180° VR. Giving users the freedom to explore VR content in 360° has limitations, though. It’s harder to direct a viewer to a point of interest, for example, or to keep their attention focused on a story or message. Creating a 3D effect in a 360° VR environment is also more complex and costly as it requires a left and right image for each camera view/position.

For content creators who want to immerse the viewer in an environment but also focus their attention in the right place, such as in a training film, concert or documentary, 180° VR is the preferred format. Unlike in 360° VR, there’s no need to worry about what’s behind the camera with 180° VR, as only the forward-facing (half sphere) part of the environment is visible. It makes the VR shooting workflow much easier, since there’s no need for the camera operator to "hide" somewhere and remotely control the camera. Compared with capturing 360° content, the camera operator can be more mobile when shooting 180° VR, and lighting, microphones and other equipment can be positioned behind the camera.

The Canon EOS VR SYSTEM has been designed to further streamline the capture of 180° VR while ensuring professional image quality. Using a single camera and a single "dual" lens, two images are recorded side-by-side on a single sensor, providing a single file that simplifies the VR conversion and editing process.

Broadly speaking, there are two types of VR content: monoscopic and stereoscopic. Both types of VR can deliver 180° or 360° views. But where monoscopic provides a 2D image, stereoscopic delivers a more realistic and immersive 3D experience.

Monoscopic is the easiest type of VR to capture and display. It is created from a single image which is shot from one direction using a single lens or array of lenses. Monoscopic’s 2D view means that specialist equipment is not required to view the content, and it can be explored on a flat screen, such as a computer monitor, mobile phone or tablet. It’s for this reason that monoscopic VR is widely used in real estate and tourism, as well as "street view" applications.

VR headsets can also be used to fill the viewer’s field of view with monoscopic content, but there will be no 3D effect. The same single image will be displayed to both eyes, so there’s no sense of depth.

Stereoscopic content captured by the RF 5.2mm F2.8L DUAL FISHEYE or RF-S 3.9mm F3.5 STM DUAL FISHEYE lens produces a more realistic and immersive experience. It uses two images of the same scene – one for each eye – which are captured simultaneously but from slightly different angles. These discrete images can be directed to each eye in a VR headset, where the difference in shooting angle creates a parallax and a sense of 3D depth.

Two lenses are required to capture the separate images needed for immersive stereoscopic and 180° VR content. This normally entails a dedicated stereoscopic camera or separate cameras and lenses in specialist rigs, which are often difficult and time-consuming to align. Canon’s stereoscopic and VR lenses streamline this process, with their dual lens design enabling compatible EOS R System mirrorless cameras and Cinema EOS video cameras to record immersive stills and video.

A dual lens makes for a more efficient shooting and post-production workflow. Not only are the left and right images perfectly aligned and captured in sync on the camera sensor, but two electro-magnetic diaphragms (EMDs) – one for each lens – ensure precise, coordinated aperture adjustments for consistent exposure across both images. The result is a single file that doesn’t require the left and right image outputs to be matched in post-production – the colours and brightness level for both images will be the same, helping to cut down the grading required.

The distance between the lenses in a dual or dual fisheye lens plays an important role in determining how far the stereoscopic or VR lens needs to be from the subject to create the best 3D effect.

The distance between the centres of the dual lenses in a VR or stereoscopic lens is known as the interpupillary distance or baseline length. The average interpupillary distance of human eyes is about 63mm, so lenses that have a baseline length of 60mm, such as the Canon RF 5.2mm F2.8L DUAL FISHEYE and RF-S 3.9mm F3.5 STM DUAL FISHEYE, give a natural and realistic 3D effect.

Lenses with a baseline length of 60mm produce the strongest 3D effect at shooting distances of 0.5 metres to 2 metres. Focusing on subjects that are closer or further away will result in a noticeable reduction in the 3D effect.

A larger baseline length will be required to optimise the 3D effect for subjects further away, while a lens with a small baseline length provides a stronger separation of the two images at shooting distances below 0.5m. The RF-S 7.8mm F4 STM DUAL offers a baseline length of just 11.8mm, which maximises the 3D effect at a shooting distance of 0.2m to 0.5m, making it ideal for creating dynamic close-ups.

While the baseline length determines where the depth perception will be the strongest, the angle of view of the lens dictates how much of a scene it’s able to capture horizontally, vertically and diagonally (corner to corner) – although it's usually just one that’s quoted: as these lenses create two circular images, the angle of view measurement is the diameter of the circle, which is the same in every direction.

VR lenses with short focal lengths can capture a wider viewing angle than those with longer focal lengths, and this can create greater immersion when viewed with a VR headset. The Canon RF 5.2mm F2.8L DUAL FISHEYE captures an expansive 190° angle of view, for example, which fills the field of view of a person viewing your content.

The field of view – or the portion of the scene that’s visible to the viewer at any one time – is affected by several factors, including the lens’s angle of view and the slight loss that occurs during the process of converting the VR file. The captured field of view of the RF-S 3.9mm F3.5 STM DUAL FISHEYE, for example, is 144°, with the outputted field of view being 140° after the file is processed in Canon's EOS VR Utility. The viewing angle of a VR headset also has an impact on what can be seen at any one time. Headsets that offer a wide field of view enable the viewer to see more of the scene without having to move their head.

With an outputted field of view of 180°, the Canon RF 5.2mm F2.8L DUAL FISHEYE pulls more into view than the RF-S 3.9mm F3.5 STM DUAL FISHEYE. However, the latter’s slightly shallower angle of view makes it easier to frame a scene without unwanted objects such as tripod legs appearing in shot. The RF-S 7.8mm F4 STM DUAL has a much narrower outputted field of view of 60°, which is equivalent to the central angle of view of human eyes. This makes for a natural, comfortable viewing experience, as well as making it easier to frame shots without distractions, similar to shooting photos and videos with a standard lens.

The more focused field of view outputted by the RF-S 7.8mm F4 STM DUAL creates spatial content rather than frame-filling immersive 180° VR. Both spatial video and immersive video are able to deliver a 3D viewing experience via a VR headset. They are both capable of a convincing sense of depth, but while immersive content fills a viewer’s field of view and places the viewer within the scene, spatial content is displayed in a smaller window in front of the viewer – like watching a 3D TV screen in a virtual environment.