Differential channel parameters

Here you can download the code for the estimation of differential channel parameters (Binaural Cues) for both the source and the corresponding reverberant signal. Only ILD and IC cues are considered in this simulation.
 
This code was implemented by Thomas Zarouchas and is related to his work:

[T. Zarouchas, J. Mourjopoulos,  "Modelling perceptual effects of reverberation on stereophonic sound reproduction in rooms", J Acoust Am., July 2009, 126 (1).]

Tracking reflections on Room Impulse Responses

Here there are some functions related to the reflections on room responses.

Here you can download a function that calculates the reflection density, based on the formula proposed in [H. Kuttruff, Room Acoustics, Spon Press, London, 4th edition, 2000]
 
Here you can download a function that determines the onset time (in samples) of reflections in an acoustic impulse response. The method is based on the modified kurtosis approach proposed in [John Usher, "An improved method to determine the onset timings of reflections in an acoustic impulse response" JASA Express Letter, 3/2010]

Spectral statistics calculation

Here you can download the code in order to compute several statistical quantities in the frequency domain for signals and room responses:

1. skewness (skew_oct.m)
2. mean (mean_oct.m)
3. standard deviation (std_oct.m)
4. kurtosis (kurt_oct.m)

The above statistical quantities may be measured in octave bands or full band.

Classifiers

In these files you can find matlab contain that contain several implementations of classifiers.

Linear regression:            LinearRegression.m 
                                         LinearRegressionAki.m

Neural Networks  :           neuralNetworks.m

Support Vector Machine: SVM/supportVectorMachine.m


The code was developed with Matlab 7.6.0.324 (R2008a)

Auditory Modelling Toolbox (external link)

 Models included in Auditory Modelling Toolbox

The goal of AMT is to provide a complete set of models that deal with the auditory hearing system. Ranging from the outer ear up to the cortex.

The Auditory Modeling Toolbox (AMT) is a Matlab toolbox intended to serve as a common ground for auditory modeling in Matlab/Octave. 

The AMT is developed by several auditory research groups.
The AMT is Free software, released under the GNU General Public License (GPLv3).

Visit the homepage.

Audio signals and response databases (external links)

Recordings and RIRs

This is a complete dataset of measurements and recordings to facilitate the evaluation of speech and audio enhancement algorithms in realistic environments. The measurements were carried out in an actual meeting room at the conference center of the University of Patras. A dummy head fitted with a mouth simulator was used to simulate a moving speaker.

Go to link

 

Open Air Library

An expanding library of downloadable acoustic impulse responses (IRs) from interesting buildings, spaces and other sources. The library contains also sounds recorded in an anechoic environment

Go to link

Speech datasets

A list of speech datasets with detailed attributes and links to software baselines and evaluation results. Each dataset may be used for one or more applications: automatic speech recognition, speaker identification and verification, source localization, speech enhancement and separation.

Go to link

Aachen Room Impulse Responses

The Aachen Impulse Response (AIR) database is a set of impulse responses measured in a wide variety of rooms. It contains binaural room impulse responses (BRIR) measured with a dummy head, but also with omnidirectional microphones in different locations with different acoustical properties, such as reverberation time and room volume. The distance and the azimuth angle between head and desired source is varied.

Go to link

HRTF and BRIRs database

An eight-channel database of head-related impulse responses (HRIR) and binaural room impulse responses (BRIR) is introduced. The impulse responses (IR) were measured with three-channel behind-the-ear (BTE) hearing aids and an in-ear microphone at both ears of a human head and torso simulator. In addition to the HRIRs derived from measurements in an anechoic chamber, sets of BRIRs for multiple, realistic head and sound-source positions in four natural environments are provided.

Go to link

Binaural distance dependent feature (BSMD STD)

The following function reproduces figures that can be found in Georganti et al. (2013) [1].

 




You can download it from here.

The figures present the computed distance-dependent feature BSMD STD (Binaural Spectral Magnitude Difference Standard Deviation) for audio signals measured at different/receiver distances.

The BSMD STD may be derived from any dual-channel signal  (binaural/stereo recordings). The BSMD STD feature is related to the standard deviation of the magnitude spectrum of Room Transfer Function, which is known to depend on the source/receiver distance. See [2] for more information.

REFERENCES:

[1] E. Georganti, T. May, S. van de Par, and J. Mourjopoulos. "Extracting sound-source-distance information from binaural signals." In J. Blauert, editor, The technology of binaural listening, chapter 7. Springer, Berlin, Heidelberg, New York NY, 2013.
http://link.springer.com/chapter/10.1007%2F978-3-642-37762-4_7


[2] E. Georganti, T. May, S. van de Par, and J. Mourjopoulos, "Sound Source Distance Estimation in Rooms based on Statistical Properties of Binaural Signals", IEEE Transactions on Audio, Speech and Language Processing, Vol. 21 (8), Aug. 2013.

url: http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6508870

REQUIREMENTS:

1) Download audio files from
http://sourceforge.net/projects/audiofordistanceestimation/files/wavFilesGeorganti.zip/download

(copy/paste the above link at your browser)

2) Unzip folder (wavFilesGeorganti.zip)

3) Move folder in the working directory of matlab

 

Single-channel distance estimator

In this folder you can find files related to the following publications:

[1] E. Georganti, T. May, S. van de Par, A. Harma and J. Mourjopoulos, "Single channel sound
source distance estimation based on statistical and source specific features",
126th AES convention, Munich, 7-11 May 2009

[2] E. Georganti, T. May, S. van de Par, A. Harma and J. Mourjopoulos, "Speaker Distance Detection using a Single Microphone", IEEE Transactions on Audio, Speech, and Language Processing, Vol. 19, Issue 7, September 2011.

The code is divided in three main parts:
1. GenerateTrainingData.m
2. GenerateClassifier.m
3. Evaluation.m

Check each seperate m-file for more details about the implementation of the method.
The code was developed with Matlab 7.6.0.324 (R2008a) by:

    Authors  : [1] Eleftheria Gerganti (Audiological Engineer, Phonak, Switzerland)

                     [2] Tobias May (Postdoctoral researcher, DTU, Denmark)

                  
     
 Enjoy it!

Common Evaluation Schemes for audiological algorithms

This section contains the outcomes while working as a postdoctoral researcher within the framework of the ICANHEAR project working on ”Common Evaluation Schemes (CES) for audiological DSP algorithms” at the group of Professor Norbert Dillier (University Hospital Zurich/University of Zurich)

 

The last years, there has been a significant effort in the scientific community for the deeper understanding of the auditory mechanisms and the improvement of the algorithms and the acoustic signal coding schemes, employed in hearing devices. These on-going research efforts lead to the development of novel methods for the enhancement of the acoustical signal; however their efficiency, performance and potential advantage compared to existing solutions should be examined.

This has recently increased the need of a universal, standardized, common methodological framework for the assessment of the various introduced Digital Signal Processing (DSP) algorithms. In the field of hearing aid–related technologies no standard procedures are being followed up to now and the various private/public research insitutes working in this field usually follow their own methodological approaches. However, for many reasons, it would be beneficial to establish some Common Evaluation Schemes (CES) by providing guidelines and series of tools to the scientific community that could assist in this direction.

 

Please check the following website for more information (email me for password) :

 

http://elegeor.wix.com/icanhearces