A real time compensation system for spatial audio: Development of a functional prototype
Translated title
A real time compensation system for spatial audio
Author
Lerke, Nikolaj Villefrance
Term
4. Term
Education
Publication year
2019
Submitted on
2019-05-28
Pages
73
Abstract
Denne afhandling præsenterer et realtidssystem til hovedtelefoner, der forbedrer rumlig lyd (spatial audio) ved at kompensere for lytterens personlige øre-respons – altså hvordan ens ører former den lyd, der når trommehinden. Normalt kræver måling af øre-respons komplekse laboratorieopsætninger, som ikke er praktiske for almindelige brugere. Målet var derfor at udvikle et enkelt kompensationssystem baseret på overkommelige komponenter. Med udgangspunkt i en mindre undersøgelse af øre-respons blev der udviklet en prototype. Hardwaren byggede på en modificeret AIAIAI TMA-2 hovedtelefon, og softwaren bestod af tre delsystemer implementeret i MATLAB Simulink. Test af prototypen viste lovende resultater, men der er plads til forbedringer: De adaptive delsystemer bør tilpasse sig hurtigere (bedre konvergens), og nøjagtigheden af niveauestimeringen i peak‑EQ-delsystemet (en equalizer, der justerer bestemte frekvenser) kan forbedres.
This thesis presents a real-time headphone system that improves spatial audio by compensating for each listener’s personal ear response—the way an individual’s ears shape incoming sound. Measuring ear response has typically required complex lab setups that are impractical for everyday consumers. The goal was therefore to create a simple compensation system using affordable components. Based on a small study of ear responses, a prototype was built. The hardware was based on a modified AIAIAI TMA-2 headphone, and the software consisted of three subsystems implemented in MATLAB Simulink. Prototype testing showed promising results, but there is room for improvement: the adaptive subsystems should adjust more quickly (faster convergence), and the accuracy of level estimation in the peak‑EQ subsystem (an equalizer that targets specific frequencies) can be improved.
[This abstract was generated with the help of AI]
Documents