How to Use Velocity Modelling for Depth Conversion

Depth conversion and velocity modelling are both necessary to convert the size of objects in 3D space into their physical representation on a 2D screen, and there are many ways to create these effects. This article explores using velocity modelling with depth conversion in order to create an accurate and fast conversion system. We’ll discuss why velocity modelling, as opposed to other methods, works best when it comes to depth conversion, and look at an example of how to implement it with real-time graphics.

What is Velocity Modelling?

Depth conversion is often used by designers and engineers to convert a drawing in one reference frame into a new, different reference frame. One of the easiest ways to do depth conversion is through velocity modelling. For example, imagine you have an object in one position travelling towards another object at 5 km/h. The exact distance it has travelled by some future time is easy: simply multiply it’s current velocity by that time (5 km/h × 1 hour = 5 km). You can use similar techniques with depth as well: find an object travelling towards another object in your scene and add its current depth offset.

The Benefits of Velocity Modelling

Depth conversion—reducing your data from high-depth pixels (such as a 16-bit RGBA file) to low-depth pixels (such as an 8-bit RGBA file)—can reduce image file sizes and offer performance advantages in some circumstances. The benefits of depth conversion can be even greater if you take advantage of velocity modelling techniques. These techniques rely on knowing which areas of an image are more likely to change than others during animation.  Essentially, we’re using information about how these regions move over time to help us compress them more effectively. For example, when compressing interlaced content that contains large amounts of motion blur or temporal antialiasing. So let’s talk about two useful ways you can use velocity models when dealing with depth conversions.

A Simple Guide on How to Achieve Depth Conversion with Velocity Modelling

Here’s a simple guide on how to achieve depth conversion using velocity modelling. In general, an indoor scene is easier to model because there are fewer variables that can affect depth estimation. However, in order to convert an indoor scene from 2D image pixels into 3D real-world coordinates, you need at least a single scan of information about how far each pixel lies from the camera – and at least one such scan is pretty difficult to obtain indoors (when you aren’t walking around with a big sensor). As luck would have it, though, we can get that information by measuring relative velocities between object points in consecutive frames.


The velocity model is an extremely useful tool for transforming between geodetic and topographic coordinates, as well as understanding how these types of systems relate. However, there are many different coordinate reference systems, which makes being able to convert quickly between them an essential skill. If you plan on working with spatial data or are involved in a project that involves multiple coordinate reference systems, it’s important that you learn all you can about using velocity models. If not done properly, your results could be incorrect or worse yet misleading.