Case study · Ericsson Research
Testing spatial audio in VR
A VR system with real-time 3D spatial audio and an Android companion app. I built it, set it up across devices, and designed the user tests that showed how well it worked.
- Role
- R&D Intern, Audio team
- When
- Jun – Dec 2025
- Methods
- User testing, metrics design
- Outcome
- 2 granted patents
Project 1 · Granted patent
HRTF personalization with real-time ITD adjustment
Spatial audio only feels real if it fits your own ears. This project lets people tune 3D sound to themselves inside the headset, just by listening and pointing.
The problem
In VR, sound is placed in 3D space using filters called HRTFs. They model how sound reaches our two ears. But every head and ear is different, so a standard filter can make sounds seem to come from the wrong direction. The usual ways to personalise it are slow or hard: measuring with microphones in the ears for hours, scanning the ears, or taking body measurements. None of these work when someone just puts on a headset.
The idea
Turn calibration into a short listening task inside VR. The user hears a sound and points to where it seems to come from. If they point too far or not far enough, the system adjusts the timing difference between the two ears (the ITD) and plays the sound again. It repeats until the sound lands where it should. The user never takes off the headset or measures anything.
Why it matters for users
- No equipment and no measuring. Everything happens with the headset and headphones people already wear.
- Fits into the experience. It can run as a short setup step before a game or a meeting, or any time sound feels off.
- Simple to do. The only task is "point where you hear it", which anyone can understand.
What I did
- Built the entire 3D side of the project on my own: a VR framework where users interact with objects and hear real-time 3D sound from them.
- Created the interactive scenes in Unity and Godot that the calibration runs in.
- Set up and tested the system across headsets and audio hardware.
- Designed and ran user sessions, measuring how accurately people located sounds and how natural the audio felt.
Outcome. The method was granted as an Ericsson patent, with me as a co-inventor.
Project 2 · Granted patent · Shown at MWC 2026
Smooth transition for immersive audio
Immersive phone calls let you hear where each voice comes from. I built an Android app that makes that spatial sound visible and interactive, and it was used in live demos at Mobile World Congress 2026.
The context
A new 3GPP standard for mobile calls, the IVAS codec (Immersive Voice and Audio Services), brings spatial audio to everyday calls. Instead of a flat voice, you can hear people and sounds around you in 3D. The challenge is that sound is invisible. When you show the technology to people, especially in a busy, noisy setting, it is hard for them to understand what is happening and why it matters.
What I built
- A new Android app, built from scratch in Kotlin with Android Studio, to visualise immersive audio during a call.
- An interactive avatar that represents the audio in space. Users can interact with it and immediately hear how the sound changes, so the 3D effect becomes something you can see and play with.
- Integrated the app with the spatial audio and tested it together with the headsets and audio hardware.
Outcome. The work was used in live immersive calls between Barcelona and Stockholm over Ericsson's network at MWC 2026, in a focus session with other 3GPP partners. It also led to a granted patent, with me as a co-inventor.
What I took away
A good visual turns an abstract technology into an experience people understand in seconds. Designing for a live demo also taught me to keep interactions simple and robust, because there is no second chance in front of an audience.