Cochlear Implant Simulation and Hearing Loss Simulator

Version 1.05.02 (1.6Mb)

by Qian-Jie Fu, Ph.D.

1. TigerCIS - Tools for Cochlear Implant Simulations and Hearing Loss Simulations

In the latest cochlear implant speech processors, speech signals are bandpassed into several frequency bands, the temporal envelope in each frequency band will be extracted by rectifcation and lowpass filtering. The extracted acoustic amplitude will be converted to electric current by a power function or logrithmic function. The electric current will be delivered to implanted electrode located in the cochlea. Noise vocoder or sinewave vocoder provides important tool to understand the effects of speech processor paramters on speech performance in normal-hearing listeners. The implementation of noise/sinewave vocoder is described as follows. The speech signals are bandpassed into the several frequency channels. The adjustable parameter includes the number of frequency channels. The temporal envelope is extracted in each frequency band by half-rectification and low-pass filtering. The adjustable parameters include the cutoff frequency and slope of lowpass filtering. The extracted envelope is used to modulate a wide-band white noise, which is then filtered with a bandpass filter. The adjustable parameters includes the corner frequencies and the slopes of the bandpass filters.Adjusting these parameter can effectively simulate the spectral shifting by shallow inserton depth and spectral smearing by channel interaction. The modulated noise bands are then summed and the speech level will be adjusted. This hearing loss simulator allows to simulate the hearing loss based on the measured hearing threshold.

2. TigerCIS - Latest Updates:

Add animated GIF in the bottom of the program to link to the website.
Redesign the batch processing so simplify the main window. Also re-organize the button layouts.
Allow the progam to directly download Typical Sounds from web site for demonstration.
Move time stretching function to the general processing. The new version can combine noise, time stretching and noiseband processsing together for more flexible simulations.
Change the icon of the program.
Modify a few buttons in the dialog. Fix one bug in bilateral simulation and add a few bilateral processors in the predefined processors.
Allow noiseband and sinewave vocoders based on FFT approach instead of traditional filterbank methods. This allows very deep and very shallow slopes for filtering. Also it is much faster for very large number of spectral channels.
the program allows real time noise-band or sinewave vocoders.
the program will be updated periodically with new digital signal processing functions.
time stretching and compression is available in this update.
More >>>

3. TigerCIS - How to Use:

Step 1: Lanuch the TigerCIS Program.
Step 2: Load the Audio Sounds.
Step 3: Select the Speech Processing Strategies.
Step 4: Process the Audio Sounds Based on Selected Strategies.
Step 5: Listen to Original and Processing Sounds.
Step 6: Tips and Others.

Step 1: Lanuch the TigerCIS Program. Here is the screen shot for the program.

Step 2: Load the Audio Sounds

Step 3: Select the Strategies

Step 4: Process the Audio Sounds

Step 5: Listen to Processed and Original Sounds

Step 6: Tips and Others

Note: Here is the description for customized corner frequency for the bandpass filters:

File name: example.fil

Format: Each row will have two numbers. Left is the low corner frequency and right is the high corner frequency. Depending on the number of channels, the number of rows should be same as the number of channels. The following shows the customized filters for the 4-channel processor. Analysis and carrier filters have the same format.

200 800

800 1500

1500 3000

3000 6000

Note: Here is the description for customized MAP for speech processors:

channel_number peak_number

analysis_low_cut analysis_high_cut analysis_slope envelope_cut envelope_slope carrier_low_cut carrier_high_cut carrier_slope carrier_type (0/1)

File name: example.map

4 4

100 500 48 200 24 100 500 6 0

500 1000 48 200 24 500 1000 6 0

1000 2000 48 20 24 1000 2000 6 1