This paper presents Part 2 of our study on personalized timbre optimization for stereophonic sound reproduction via earphones, following our previous work [1]. Part 1 established an auditory-model-based framework, hereafter referred to as PTTC-1 (Personalized Timbre Target response Curve), for reproducing the listener`s individual reference for natural timbre. However, its practical application was constrained by the need for high-resolution 3D anatomical scanning, individualized HRTF computation by the boundary element method, and controlled listening tests conducted by trained audio engineers. The present paper introduces PTTC-2, a simplified and fully automated implementation for consumer True Wireless Stereo (TWS) earphones, in which a single smartphone image captures the listener`s head and outer ear, the directionally averaged HRTFs are estimated by machine learning, and a brief in-app listening test determines a listener-specific weighting coefficient .
Two empirical studies are reported. First, the analysis of values selected by 2,850 users, together with a follow-up assessment of 29 listeners, indicates that is a listener-specific parameter that is not strongly biased by program material. Second, a controlled subjective experiment with 24 listeners compared PTTC-1, PTTC-2, and two non-individualized target response curves used in consumer products. PTTC-1 and PTTC-2 did not differ significantly in holistic timbre attributes, and both were preferred over non-individualized target response curves. Principal component analysis further revealed that listener preference is governed by an overall naturalness axis rather than by a high-frequency emphasis axis. These results demonstrate that auditory-model-based timbre personalization can be effectively translated into a practical, consumer-ready technology that preserves the perceptual benefits of the full workflow.
(See document for exact affiliation information.)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
