Based on the design of digital hearing aids TMS320VC5416DSP (1)

0 introduction

With the development of society and people's growing concern for patients with hearing impairment, hearing aids development gradually attention.

But since the hearing disorders causes vary, their hearing loss situation exists quite a big difference, making every patient the compensation for hearing aids have different requirements. At present, the modern hearing aid technology into the era of digital hearing aids. At the same time, various effective improve the efficiency of digital hearing aids and digital signal processing algorithms also get more attention. In this proposed based on the design of digital hearing aids TMS320VC5416, meet deaf patients on listening needs.

System composition and working principle

1.1 system composition

Technical requirements based on hearing aids, use TI company TMS320C5416 C54X series products (hereinafter referred to as C5416) and digital encoder TLV320AIC23 (hereinafter AIC23).

Digital encoder AIC23 is a TI company launched a high-performance stereo audio Codec chip, A/d and D/A conversion parts integrated on the chip, using advanced σ-△ oversampling technology, built-in headphone output amplifier.

AIC23DSP Codec working voltage and C5416 core and i/o voltage is compatible, you can achieve with C54x serial port for seamless connection, power consumption is very low, making the AIC23 is a very ideal audio analog devices that can be applied to the design of digital hearing aids.

System structure as shown in Figure 1, includes DSP module, audio processing modules, JTAG interface, enclosure and power supply modules, etc.

Analog voice signals through the MIC or IANE IN input AIC-23, after die/number converted by MCBSP serial input C5416, after the actual algorithm for treatment and compensation be deaf patients need a voice signal, and then by number/die cast AIC23, through speakers or headphones output sound signal.

1.2 C5416 and AIC23 interface design

Figure 2 is an interface with AIC23 C5416 schematic.

Because the sample output AIC23 is serial data, hence the need to coordinate and matching of DSP serial Transfer Protocol, is the most suitable for MCBSP speech signal transmission. Be the first 22 feet AIC23 MODE then terminated, receive from the DSP SPI serial data formats. Digital control interface (SCLK, SDIN, CS) and MCBSPl connection, control word total of 16-bit, start by high traffic. The digital audio port LRCOUT, LRCIN, DOUT, DIN, connected with MCBSP0 BCLK. In the works, DSP-based mode, the mode AIC23, i.e. BCLK clock signal generated by the DSP.

Serial clock by BCLKX0, BCLKR0 parallel to AIC23 of BCLK clock so that you can send and receive data to generate serial clock signal.

Input/output sync signal LRCIN and LRCOUT, used to start the serial data transfer and receive DSP frame synchronization signal.

BFSX0 and BFSR0, BDR0 and BDX0 respectively and AIC23 DIN and implement DSP DOUT connections and digital communication between AIC23.

2 system implementation

The basic characteristics of 2.1 voice

The sound is a wave that could be the ear to hear the sound vibration frequency of 20 Hz ~ 20 kHz.

Voice is the voice of a person, he is to issue pronunciation organ, with some syntax and meaning of sound. Voice of the vibration frequency of up to 15 kHz.

Voice as it is divided into different incentive forms: unvoiced sound, voiced sound, the sound of blasting.

While the basic characteristics of the sound is determined by genes cycle and formants and other factors. When fat voiced sound, the air flow through the vocal fold vibration of glottis, quasi periodic excitation pulse train. This pulse train cycle is called the "gene cycle" the countdown to the "gene frequency".

The human channel and nasal road can be seen as a non-uniform interface channel, Channel Canal resonance frequency is called formants.

Change channel shape produces a different sound. Formant used sequentially increased multiple frequency indicated if F1, F2, F3, and so on, called first formant, the second formant, etc. In order to improve the quality of voice reception, must be in as many formants. Actually, the first three formant is the most important, specific conditions for people.

2.2 speech enhancement

In a real application environment, the voice will be varying degrees of affected by environmental noise interference.

Speech enhancement is for noisy speech processing, noise reduction, improvement of the auditory environment.

The actual voice experience interference may include the following categories: