Google Daydream

Google aims for ultra high resolution, virtual reality in focus

From screens with ever-thinner resolution to more natural images, this is probably the only solution to offer virtual realities total realism. Google and LG, a specialist in OLED technology known for its contrasts and the colors it offers, announce the study of the screen with the highest resolution in the world. With a diagonal of 109 mm, its definition should be 5500 x 3000 pixels or, once this value is reduced to its dimensions, a resolution reaching 1443 pixels per inch (DPI) or 56 points per millimeter. Nearly five times better than what a photo-quality paper print offers.

Why such a surge of definition ?

For several years now, the screens that equip our smartphones offer a definition that is constantly growing. Almost all are now “full HD” compatible and therefore have a definition of 1920 x 1080 pixels, or even more on some models. Brought back to the dimensions of the screen, about fifteen centimeters diagonal at best, one can question the usefulness of such a race to definition. Indeed, in normal use, it becomes completely impossible to distinguish pixels with the naked eye. But it would be to forget a little quickly the new uses for smartphones and, in particular, their use at the heart of applications of virtual realities.

In this case, very many applications directly use a smartphone by dressing it with a headset, inexpensive, containing for the most part a set of lenses. Their vocation is to allow visual accommodation at the very small distance that separates the screen from the eyes, a bit like what happens at the viewfinder of a camera. The Google Daydream accessory is a perfect illustration of such devices. But then there is indeed a problem of resolution. Already, each eye has only half the screen to render the image. Let’s not forget that for a raised vision it is necessary to display two images side by side on the screen. Each image thus has only a resolution of 960 x 1080 pixels.

On the other hand, the magnifying effect created by the lenses of the helmet significantly increase the apparent surface of the image. Our eyes, instead of perceiving a screen a few centimeters diagonal, have the feeling of observing an image of several meters. It is obvious that its resolution is in free fall and the pixels become particularly visible. By moving to a definition of 5500 x 3300 pixels, each eye would perceive an image of 2250 x 3300 pixels, the equivalent of the image offered by digital cinemas. The pixelation of the image fades to give way to a feeling of naturalness.

Handling such a definition is not within the scope of any processor

Still, increasing the definition of the screen is not limited to a display problem. Because the image represents more than 18 million pixels to manage. However, for a good fluidity, the tendency is to frame the movement at 120 frames per second. A point that is equivalent to saying that under these conditions the image processor of the smartphone must be able to handle 120 times 18 million pixels, that is, more than two billion operations every second. Computing power that is not within the reach of any image processor.

However Google here has an original approach to relieve the image processor. It could limit this need for computing power by a “foveal” approach to the problem. Behind this denomination lies the fact that the periphery of our field of vision is not very demanding in terms of definition, color and movement. In other words, his analysis is quite succinct. Only its central part, that corresponding to the fovea-hence the name of this approach -, finely analyzes the perceived image. By detecting the gaze of the wearer of the virtual reality headset, the foveal processing would then concentrate the computing power of the image processor in the center of its field of view, treating in a “coarser” way the peripheral areas of the image.

Let us point out that while this approach is interesting, it is not the first to exploit the ” weaknesses of our senses “. For many years, MPEG for video, or MP3 for music have used the same kind of tricks to practice little perceptible compressions, making it possible to drastically reduce the bitrate requirements to transmit a video or audio stream.

Go to our cases Get a free quote