WebAR – is it production-ready? Wins and pitfalls of AR development

Design | Development - 9th February 2020
By Pixelfield team

We have recently started working with WebAR technology. It is super fun both developers and final users. The augmented reality development is definitely going to be something to think of when you will plan your future product and campaigns. What can you expect? How production-ready is the technology and what are the pitfalls? 

The web AR app performance

One of the main things you definitely need to things about is app performance. We tested on different devices, more powerful Androids are working fine (and we had no issues except for tracking, see below), iPhones (especially not the current lineup) and slower Androids show some stuttering after a while. There’s not much to do about tracking, as that’s handled by A-frame (or another framework) and their algorithms.

Think about the detail of your model

Next one is the model itself and it’s quality. The 3D model can be huge and very detailed, you need to create a simplified version more suitable for the use case, that will still look good. You can get some performance improvements by degrading the quality of the model even more. This is something you need to think about before you start both modelling and development

Camera speed varies based on the tracking type

Camera speed and occasional freezing is a performance issue. Most of it is due to the nature of the experience (marker-based is heavier on hardware), theoretically, this can be improved by having a simpler (and worse looking) model.

Marker-based experience is generally performing worse on iPhones (especially not the latest ones), as they are using the same internals (CPU / graphics chip) to find and process the marker and display the model+lighting+animation. Quick Youtube search will show that most examples of marker-based webAR is laggy, as opposed to markerless surface-based experiences, which uses iPhone’s integrated AR chip to do the AR computing (finding surface etc) and CPU / graphics chip has then more power for the rest.

The technology and object format

You can choose almost any software to create your 3D models and animations. Be careful about which one you are going to choose – it needs to work with the tech you will be used for creating the experience. Not every software supports all the formats. It is almost impossible to fix this in the later stages as the conversions are not reliable.

Test your exports early.

Be careful when you create the skeleton for your objects. Export the animation and see how it works in the augmented reality. Create the final animations, objects and textures only when you are certain that the export is going to be without issues.

Now it’s up to you. What is missing in this article? Let us know in the comments!

Written by
Pixelfield team
Related posts
Protection od product in phase of Proof of Concept
How to protect your idea in the Proof of Concept and MVP phase?
Business | Design | Development - 13th December 2021
By Pixelfield team
MVP and PoC of digital product
MVP and Proof of Concept of digital product
Business | Development | Marketing - 13th December 2021
By Pixelfield team