Development and preliminary evaluation of clear tone: A smartphone-based personalized hearing and communication assistance application

Aim: To develop and preliminarily evaluate ‘ClearTone’, an Android-based hearing-support prototype that combines app-based hearing-threshold estimation, personalized six-band amplification, environment-aware mode switching, and speech-assistance functions.
Materials and methods: A prospective comparative prototype evaluation was conducted using hospital-based clinical audiogram reports and ClearTone threshold estimates. Seventeen (17) patients were tested during three evaluation days. Ten anonymized cases were used for detailed descriptive comparison, while the full paired dataset contained 132 app-clinical observations before outlier filtering. The app tested each ear at 250, 500, 1000, 2000, 4000, and 8000 Hz. Additional engineering evaluation included environment-classifier testing, latency measurement, scripted speech-to-text/text-to-speech testing, and digital signal-processing verification. In this study, pure tones at 250, 500, 1000, 2000, 4000, and 8000 Hz were used as controlled test signals to evaluate the gain applied by the ClearTone DSP pipeline at each corresponding frequency region.
Results: After outlier filtering, 122 paired values were retained. Linear regression produced the preliminary mapping Estimated clinical-like threshold = 13.84 + 0.971 x app threshold, with R2 = 0.932 and residual standard error of 6.14 dB. The binary environment classifier achieved 84% segment-level accuracy on a random test split of 1,509 clips, with an area under the receiver operating characteristic curve of 0.9892. Pixel 7 application-stream latency was approximately 10 ms, while Bluetooth capture and output produced substantially larger delays. Of 30 scripted speech-assistance tests, 24 passed. DSP tests verified frequency-selective gain, mode-dependent compression, summation, and limiting.
Conclusion: ‘ClearTone’ demonstrates the engineering feasibility of integrating personalised hearing support and communication assistance on a smartphone. The findings do not establish diagnostic equivalence, clinical effectiveness, or universal calibration. Larger controlled studies, device-specific electroacoustic calibration, independent validation, lower-latency audio routes, and binaural processing are required before clinical deployment.