Personal communication device as a hearing aid with real-time interactive user interface

A digital hearing aid system, comprising:

• a headset with a microphone acquiring an analog input speech signal and an earphone for outputting a processed analog speech output signal;
• a personal communication device in communication with the headset, wherein the personal communication device comprises:
• an input amplifier amplifying the analog input speech signal acquired by the microphone of the headset;
• an analog-to-digital converter to convert an amplified analog input speech signal from the input amplifier to a digitized input speech signal;
• a digital-to-analog converter to convert a processed digital speech output signal to an analog output signal;
• an output amplifier amplifying the analog output signal and outputting the processed analog speech output signal to the earphone of the headset;
• a digital processor interfaced to the analog-to-digital converter and the digital-to-analog converter;
• and
• a graphical user interface in communication with the digital processor;
• wherein the digital processor of the personal communication device is configured to process the digitized input speech signal using Fast Fourier Transform (FFT-based) analysis-synthesis for a quantile-based noise suppression to reduce background noise and a sliding-band dynamic range compression to compensate for frequency-dependent hearing loss and reduced dynamic range, in order to improve speech quality and intelligibility for hearing-impaired listeners, wherein the processing for the quantile-based noise suppression comprises the steps of:
• (i) performing a dynamic quantile tracking based noise estimation by estimating a histogram for each frequency index of spectrum of the digitized input speech signal, wherein the histogram is estimated by dynamically tracking a plurality of quantile values for frequency indices of the spectrum, obtaining a quantile value corresponding to peak of the histogram by finding a quantile value having a minimum difference between neighboring quantile values, and estimating noise at each frequency index by using peak of the histogram wherein the quantile value is calculated by applying an increment or a decrement on its previous value, with the increment and decrement selected to be a fraction of a dynamically estimated range of frequency index such that the quantile value approaches a sample quantile over a number of successive analysis frames;
• and
• (ii) performing spectral subtraction of the noise at each frequency index from the spectrum to obtain enhanced complex spectrum;
• and the processing for the sliding-band dynamic range compression is carried out by calculating a frequency-dependent gain function in accordance with a level in a band centered at each frequency index and a piecewise linear relation between an input power and an output power on a dB scale for each spectral sample and the gain function is used for modification of the spectrum,
• and wherein the graphical user interface provides a plurality of control settings to facilitate setting and fine tuning of frequency-dependent parameters for the signal processing in an interactive and real-time mode, wherein the graphical user interface provides a plurality of control settings to the user for setting parameters of the dynamic quantile tracking based noise suppression and the sliding-band dynamic range compression process.