Open Speech Platform User Guide OSP Manual

User Manual:

Open the PDF directly: View PDF .
Page Count: 34

PROJECT SPONSOR NATION INSTITUTE OF HEALTH, NIDCD.

THIS WORK IS SUPPORTED BY THE NATIONAL INSTITUTE ON DEAFNESS

AND OTHER COMMUNICATION DISORDERS, OF THE NATIONAL INSTITUTE OF

HEALTH (NIDCD/NIH). R01DC015436: "A REAL-TIME, OPEN, PORTABLE, EX-

TENSIBLE SPEECH LAB" TO THE UNIVERSITY OF CALIFORNIA, SAN DIEGO.

HARINATH GARUDADRI

OPEN SPEECH PLATFORM

USER GUIDE

OSP TEAM

published by osp team

openspeechplatform.ucsd.edu

Redistribution and use in source and binary forms, with or without modiﬁcation, are permitted provided

that the following conditions are met:

1. Redistributions of source code must retain the above copyright notice, this list of conditions and the

following disclaimer.

2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the

following disclaimer in the documentation and/or other materials provided with the distribution.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND

ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WAR-

RANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.

IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,

INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,

BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA,

OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LI-

ABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR

OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF

THE POSSIBILITY OF SUCH DAMAGE.

First printing, June 2019

Contents

Abstract 8

Setting up the RT-MHA 9

2.1Installing the Software 9

Release 2019a Test and Validation 11

2.2Connecting Audio I/O Devices 11

2.2.1On OS X machines 11

2.2.2On Linux machines 12

2.3Setting up Audio I/O 13

2.3.1Internal Audio Codecs 13

2.3.2External USB Audio Codecs 14

Verifying the RT-MHA is Working Properly 16

3.1Internal Audio Codecs 16

3.1.1Testing with audio ﬁles 16

3.1.2Live Testing with headphones 17

3.1.3Live Testing with Andrea SB-405 Super beam headphones 17

RT-MHA Command line arguments 18

4.1OSP commands 18

4.1.1-h for Help 18

4.1.2-F<option> for choosing Sampling frequency 18

4.1.3-r to enable/ disable Rear mics 19

4.1.4-a for enabling/disabling Adaptive Feedback cancellation (AFC) 19

4.1.5-d<attenuation in dB> to set the attenuation factor 19

4.1.6-T<option> for selecting AFC algorithm 19

4.1.7-N<option> for choosing noise estimation type 20

4.1.8-S<option> for enabling/disabling spectral subtraction 20

4.2OSP modes of control 20

4.2.1TCP daemon interface 20

4.2.2Playback mode interface 21

4.2.3Playback mode with TCP daemon interface 21

4.2.4Command Line interface 21

4.3Recommended Settings for ANSI 3.22 Testing 22

4.4Recommended Settings for Testing with Webserver 22

Setting up the Embedded Web Server (EWS) 23

5.1Running RT-MHA on Local Computer 24

5.2Connecting to OSP Remotely 24

Verifying the EWS is Working 31

6.1Researcher Page 31

6.24AFC 32

Bibliography 33

List of Figures

2.1

Terminal output after the

./install

command. You will need to en-

ter your password at various times. snip represents a large portion

of the terminal output deleted in this ﬁgure. This process can take

30 −90

minutes, depending on your computer and network speed.

The script ends with instructions on invoking osp and ews in the ter-

minal. 10

2.2USB Audio Input Conﬁguration. 12

2.3USB Audio Input Conﬁguration. 12

2.4

List of devices displayed by running the

pa_devs

command 13

2.5

This is what your ﬁnder window will look like when you’re looking

up Audio MIDI Setup. 13

2.6

This is what the Audio MIDI Setup panel should look like when you

select this device for sound output. 14

2.7

This is what the System Preferences>Sound should look like when

you select this device for sound output. 14

2.8

This is what the System Preferences>Sound should look like when

you select this device for sound input. 15

5.9

"This is what your Finder window should look like when you have

selected your Hard Drive 25

5.10

Here are all the ﬁles you have made visible with this command, with

the etc folder highlighted 25

5.11

Here are all the ﬁles you have made visible with this command, with

the etc folder highlighted 26

5.12 This is how you Open the httpd.conf ﬁle using Atom 26

5.13

This is what it should look like after you delete the # on the appro-

priate line 27

5.14

This is what it should look like after you delete the # on the appro-

priate line 27

5.15

This is what your Terminal window will look like when you use the

whoami command. The output is highlighted. Your output will be

your username, not tzubatiy. 28

5.16

This is what your Atom window should look like when have added

your name to the user line, in this case this is my name. 28

5.17

This is what your Atom window should look like when have added

these lines of code in the right place. 28

Abstract

This version corresponds to Release 2019a. We present all of the

necessary tools that need to be installed to make the Open Speech

Platform (OSP) run smoothly, and how to make sure they are work-

ing properly. The OSP is comprised of a real time master hearing

aid (RT-MHA) portion and an Embedded Web Server (EWS) por-

tion. This guide is organized based on these two chapters, so ﬁrst we

present the tools to get the RT-MHA all set up, and then we present

the setup process for the EWS.

https://github.com/nihospr01/OpenSpeechPlatform-UCSD.

Setting up the RT-MHA

2.1Installing the Software

1.Download 2018c

. Download the latest release from

https://

github.com/nihospr01/OpenSpeechPlatform-UCSD

. You can either

“clone the software on your computer” or “download ZIP ﬁle to

your computer.”

(a)

If you wish to clone, type

in to your terminal, to make sure

you are in the local directory. Then press clone in github, and

copy the line of code generated into your terminal. Once it has

cloned, type

cd nameofdirectory

(for example cd OpenSpeechPlatform-

UCSD). Then, type

git pull

to get the latest version. This is the

preferred approach.

(b)

When you download a .zip ﬁle from github, you will have to

manually download newer versions with bug ﬁxes.

2.Install 2018c

Open a terminal.

cd ./Software/

and

./install

This script does several things.

(a)

Identify the operating system (OS) on your computer – cur-

rently OS X and Linux.

(b)

Install software packages including portaudio (for realtime

audio input/output); MySQL (a relational database), PHP (a

server side scripting language) and other packages for LAMP

software stack (LAMP stands for Linux, Apache, MySQL and

PHP).

(d)

Finally, it installs osp in

/usr/local/bin/osp

and a script to

invoke ews in /usr/local/bin/ews.

If everything went well, your screen will look similar to Figure 2.1.

install.png

Figure 2.1: Terminal output after the

./install

command. You will need to

enter your password at various times.

snip represents a large portion of the

terminal output deleted in this ﬁgure.

This process can take

30 −90

minutes,

depending on your computer and

network speed. The script ends with

instructions on invoking osp and ews in

the terminal.

Release 2019a Test and Validation

This chapter describes sanity tests to validate your versions of

osp

and ews.

2.2Connecting Audio I/O Devices

There are many audio input/output options for OSX and Linux

computers. One inexpensive option is the Andrea Communications

3D Surround Sound Recording CANS. They are supra-aural headsets,

with left and right mics.

You can also use a high end audio device such as Zoom TAC-8or

Focusrite Scarlett 2i2.

2.2.1On OS X machines

1. Plug the headsets in to an available USB port.

On Mac computers,

Open Audio MIDI Setup

: This can be found

Finder | Applications | Utilities

, as shown in Figures 2.2

and 2.3.

Select 48,000 Hz option on both screens and levels to 1..0as

shown.

If you use a high end audio interface box (such as Zoom TAC-8)

and if you have 98,000 Hz and 24 bit option, you should choose

this.

Figure 2.2: USB Audio Input Conﬁgu-

ration.

Figure 2.3: USB Audio Input Conﬁgu-

ration.

2.2.2On Linux machines

On Linux machines the default audio is not set. Run the command

pa_devs

to give the list of the available devices and the device

numbers they correspond to

Run the command

osp -input_device x -output_device y

where xand yare values attained from the previous command.

With reference to Figure 2.4, the command issued is

osp -input_device

6 -output_device 6

Can we use white background in the terminal shots for the

document? Black is good for code development, but I am not sure for

documents

Figure 2.4: List of devices displayed by

running the pa_devs command

2.3Setting up Audio I/O

2.3.1Internal Audio Codecs

It is possible to test proper functioning of RT-MHA using the internal

audio codecs built into a MacBook.

1.Open Audio MIDI Setup

: This can be found in Finder: Applica-

tions > Utilities, as shown in Figure 2.5.

Figure 2.5: This is what your ﬁnder

window will look like when you’re

looking up Audio MIDI Setup.

2.Select primary input

: In Audio MIDI Setup, right-click (CTRL+click)

Built-in Microphone on the left hand side. Select "Use This Device

For Sound Input. Change the format to 96,000 Hz."

2.3.2External USB Audio Codecs

1.Select primary output

: In Audio MIDI Setup, click on the Zoom

Tac-8, click on Input, select the format 96,000 Hz. Then click

Output, and also select 96,000 Hz."

Figure 2.6: This is what the Audio MIDI

Setup panel should look like when you

select this device for sound output.

2.Edit Sound Preferences

: Open System Preferences>Sound and

select the Output tab. Click on Zoom Tac 8. Select the Input tab

and also click on Zoom Tac 8.

Figure 2.7: This is what the System

Preferences>Sound should look like

when you select this device for sound

output.

Figure 2.8: This is what the System

Preferences>Sound should look like

when you select this device for sound

input.

Verifying the RT-MHA is Work-

ing Properly

3.1Internal Audio Codecs

3.1.1Testing with audio ﬁles

In this section, we will use audio samples provided with the OSP

software to test the Real time Master Hearing Aid.

In OSP we have provided 96kHz sample speech ﬁles for testing

the software. All of this is found in the following directory: OSP >

Matlab > InputFiles-96kHz.

On the terminal run the following command to make OSP process

audio ﬁles and play them through the speaker. Here we use the -p

argument to play sound through the software.

osp -F1 -M1 -a -r -d-40 -p ~/Downloads/OSP/matlab/Input\ Files-32k/TIMIT_female_sa2.wav

This would play the audio ﬁle "TIMIT_female_sa2.wav" through

the entire OSP process, and you would hear the sound through the

speaker/headphones.

We have also provided the option of playing the sounds with NH,

N2, N4and S2settings. These can be executed using the following

command:

osp -F1 -M1 -a -r -d-40 -p ~/Downloads/OSP/matlab/Input\ Files-32k/TIMIT_female_sa2.wav -s1

Here the argument -s<setting_number> gives control to the user

to choose the gain settings for OSP and play the audio with those

settings. Setting_number are set the following way: 1= N2,2= N4,

3= S2and 0= NH.

3.1.2Live Testing with headphones

1.For live testing

you need a headset with microphone. We found

headphones commonly used for making calls to be adequate. Plug

this into your computer.

2.Make sure you have set the volume on your computer to the

maximum allowed.

Enter the following command on the terminal to start osp in

command line interface mode.

osp -F1 -M1 -a -r

Here choose F1if your system supports 96kHz or else choose F0

for 48kHz.

Next you can increase the gain to 10dB on all bands by typing in

"g10". This way your own voice would be heard louder through

the headphones.

3.1.3Live Testing with Andrea SB-405 Super beam head-

phones

1.For live testing

you need a headset with microphone. We found

headphones commonly used for making calls to be adequate. Plug

this into your mac at the USB port.

2. Make sure the mute switch is in the "off" position.

3.Make sure you have set the volume on your computer to the

maximum allowed.

Enter the following command on the terminal to start osp in

command line interface mode.

osp -F1 -M1 -a -r

Here choose F1if your system supports 96kHz or else choose F0

for 48kHz.

Next you can increase the gain to 10dB on all bands by typing in

"g10". This way your own voice would be heard louder through

the headphones.

RT-MHA Command line arguments

In this section we explore the various functions of the RT-MHA on

the command line interface. There are multiple starting conﬁgura-

tions and these conﬁgurations are speciﬁed using arguments.

4.1OSP commands

4.1.1-h for Help

This would show all the available options/conﬁgurations the RT-

MHA can be executed in, as shown Figure??.

osp -h

4.1.2-F<option> for choosing Sampling frequency

The ﬁrst thing you should do is choose your sampling frequency

using the extension -F

This should correspond with the sampling

frequency you picked in step 1.11 and 1.12. The Sampling rates

supported are 0,48kHz and 1,96kHz.

To set the sampling rate to 48kHz, type the command below into

your terminal and press Enter:

osp -F0

Notice how in the terminal window it now says:"Choosing 48kHz

sampling frequency"

To stop running OSP, type on the keyboard "control"+"c" do

this now, and we will continue verifying that everything is work-

ing properly. You will need to do this step after every step in this

section

To set the sampling rate to 96kHz, type the command below into

your terminal and press Enter:

osp -F1

Notice how in the terminal window it now says:"Choosing 96kHz

sampling frequency"

Type on the keyboard "control"+"c" to stop running OSP before

going on to the next step.

4.1.3-r to enable/ disable Rear mics

Open a Terminal window, type in the line below and press Enter

osp -r

This is the recommended default setting

as we do not have rear

mic support yet. Notice how in the terminal window, it says "Dis-

abling rear mics".

Type on the keyboard "control"+"c" to stop running OSP before

going on to the next step.

4.1.4-a for enabling/disabling Adaptive Feedback cancella-

tion (AFC)

In a Terminal window, type in the line below and press Enter:

osp -r -a

You are now running OSP from the command line with the rear mics

disabled and AFC off. On the ﬁrst line in the terminal window it now

says: "4AFC has been disabled"

Type on the keyboard "control"+"c"

to stop running OSP before going on to the next step.

4.1.5-d<attenuation in dB> to set the attenuation factor

To run OSP with an attenuation factor (let’s say -40dB attenuation),

type the line below into a terminal and press Enter:

osp -r -a -d -40

you can also not add a space between the -d and the -40 as such:

osp -r -a -d-40

Notice how in the terminal window it now says: " Attenuation fac-

tor is -40.000000 dB"

Type on the keyboard "control"+"c" to stop

running OSP before going on to the next step.

4.1.6-T<option> for selecting AFC algorithm

To start OSP with a predeﬁned AFC type use the extension -T

The types of AFC supported are ’0’ = FXLMS, ’1’ = PNLMS and ’2’

= SLMS. For example, say you want to set the setting to SLMS. Then

you would type the line below into your terminal and press Enter:

osp -T2 -r

Notice how in the terminal window it now says: Setting adaptation

type to SLMS

Type on the keyboard "control"+"c" to stop running

OSP before going on to the next step.

4.1.7-N<option> for choosing noise estimation type

To run OSP with a predeﬁned Noise Estimation algorithm, use the

extension -N-noiseExtension

The types of Noise estimation algorithms RT-MHA supports are

’1’ = Arslan Power Averaging Procedure, ’2’ = Hirsch and Ehrlicher

weight averaging procedure and ’3’ = Cohen and Berdugo MCRA

Procedure.

For example, say you wanted to set the Noise Estimation algo-

rithm to the Cohen and Berugo. To do this, type the line below into

your temrminal and press Enter:

osp -N3 -a -r

Notice how in the terminal window it now says: Cohen and Berdugo

MCRA Procedure Noise estimation method chosen

Type on the

keyboard "control"+"c" to stop running OSP before going on to the

next step.

4.1.8-S<option> for enabling/disabling spectral subtraction

To run OSP with spectral subtraction, use the extension -S1or -

By setting spectral subtraction to ’-S0’, spectral subtraction is

disabled; by setting to ’-S1’, spectral subtraction is enabled.

osp -S1 -a -r

Notice how in the terminal window it now says: Spectral subtraction

enabled.

Type on the keyboard "control"+"c" to stop running OSP

before going on to the next step.

4.2OSP modes of control

4.2.1TCP daemon interface

-t starts the application in TCP daemon mode. In this mode, the

application is waiting for a connection on port 8001 from an OSP

client. Once a client is connected, all parameters (such as -g50 and

-g80 gains, attack/release time, knee points, etc) come from the client.

The command line version of the application cannot be changed

manually. To stop the application once it’s started in TCP mode, the

user must hit CTRL+C.

osp -t

Notice how in the terminal window it now says:"Using TCP daemon

mode" and "waiting for a connection from client..."

4.2.2Playback mode interface

-p<playback_ﬁle> starts the application in playback-only mode. The

audio is not enabled, and instead, audio data is read from <play-

back_ﬁle> into the application where the RT-MHA algorithm is

applied to the audio data, and this is played out in the hearing aid/

speakers/ headphones. This mode can be used for verifying the

functionality of the software.

osp -p ~/Downloads/OSP/matlab/Input\ Files-32k/TIMIT_female_sa2.wav

Notice how in the terminal window it now says:"Playback mode spec-

iﬁed, setting up ﬁle /Users/kc/Downloads/OSP/matlab/InputFiles-

32k/TIMIT_female_sa2.wav"

4.2.3Playback mode with TCP daemon interface

-l starts the application in TCP daemon and

4.2.4Command Line interface

When neither of -t, -l or -p<wav ﬁle location> have not been speciﬁed.

OSP can be interacted with on the command line. The following

commands are can be used to interact with OSP on the command line

(terminal).

’g <gain_value>’: If the operator chooses to adjust the gain to

10dB, he/she can simply enter ’g 10’ followed by Enter on the

command line.

’s’ If the operator wants to turn on/off AFC, he/she can simply

enter ’s’ followed by enter to toggle AFC mode.

’p <spectral_subtraction_parameter>’ If the operator chooses to

adjust the spectral subtraction to 0.5, he/she can simply enter ’p

0.5’ followed by Enter on the command line. The range for spectral

subtraction parameter is from 0to 1.

4. ’c’ to save the converged afc ﬁlter coefﬁcients.

5. ’f’ to toggle feedback.

6. ’q’ to quit the application.

4.3Recommended Settings for ANSI 3.22

Testing

For ANSI 3.22 testType the following line of code into your Terminal

and press Enter:

osp -a -r -M0 -d-40

Here the attenuation factor is -40. Please use -40dB when using with

the breakout board (bob) and jellybeans.

4.4Recommended Settings for Testing with

Webserver

textFor testing with webserver, Type the following line of code into

your Terminal and press Enter

osp -t -a -r -M1 -d-40

Setting up the Embedded Web Server

(EWS)

1.Enter the following line of code into your Terminal

to make sure

you are located in your home directory:

2.Type the following line of code into your Terminal and press

Enter to navigate to the folder that the EWS is locate in:

cd downloads/osp/Embedded\ Webserver

3.Type the following line of code into your Terminal and press

Enter to get Composer

brew install composer

4.Type the following line of code into your Terminal and press

Enter to install Composer

composer install

5.Type the following line of code into your Terminal and press

Enter to get NPM

brew install npm

6.Type the following line of code into your Terminal and press

Enter to install PHP

brew install php

7.Type the following line of code into your Terminal and press

Enter

cp .env.example .env

8.Type the following line of code into your Terminal and press

Enter

php artisan key:generate

9.Type the following line of code into your Terminal and press

Enter to install React

npm install react

10.Type the following line of code into your Terminal and press

Enter to install React-dom

npm install react-dom

11.Type the following line of code into your Terminal and press

Enter to install NPM

npm install

12.Type the following line of code into your Terminal and press

Enter to run NPM

npm run dev

5.1Running RT-MHA on Local Computer

1.Enter the following line of code

In the same Terminal Window

in which you ended the previous step (npm run dev), type in the

following line of code and press Enter

php artisan serve

2.Press the ’Run’ Triangle

in the Upper left hand corner of your

XCode window to run the OSP code

3.Open a new Terminal Window

Type in the following line of code

and press enter:

php artisan socket:start

5.2Connecting to OSP Remotely

The following section is speciﬁcally for running OSP on a Mac and

connecting to it remotely.

2.The ﬁrst thing that you should do is download a text editing

software such as Atom or Sublime

To download, go to this link

and press download: https://atom.io/

3.When it ﬁnishes downloading double click on the downloaded

item

, follow the instructions to move it into your Applications

folder, and open a new window in Atom when it is ﬁnally loaded.

4.Open a new Finder Window and double click on your Hard

Drive to see the contents stored there.

Figure 5.9: "This is what your Finder

window should look like when you

have selected your Hard Drive

5.Press following three together: "command"+"shift"+">"

Once you

do this, you should see some grayed out folders appear in the

Finder window, that were not there before

Figure 5.10: Here are all the ﬁles you

have made visible with this command,

with the etc folder highlighted

6.Double click the grayed out folder "etc", then go into the folder

titled "Apache", and ﬁnd the ﬁle titled "httpd.conf"

Figure 5.11: Here are all the ﬁles you

have made visible with this command,

with the etc folder highlighted

7.Right click on this httpd.conf ﬁle and select Open With> Atom

To open the ﬁle in your text editor. When you do this, it will open

the contents of this ﬁle for you to edit using Atom.

Figure 5.12: This is how you Open the

httpd.conf ﬁle using Atom

8.Once you’re looking at the ﬁle in Atom, and can see all of the

text, press the following buttons together "Command"+"f" and

then search for the text in the line below:

#LoadModule php7_mod ule l i b e x e c / a p ac he 2/ l i b p h p 7. s

When you get to this line,

delete the hashtag (#) at the beginning

of the line to complete this step

Figure 5.13: This is what it should

look like after you delete the # on the

appropriate line

9.Press "Command"+"f" and then search for the text in the line

below:

Once you ﬁnd the line,

delete the hashtag (#) at the begin-

ning of the line to complete this step

#LoadModule rewrite_module libexec/apache2/mod_rewrite.so

Figure 5.14: This is what it should

look like after you delete the # on the

appropriate line

10.Open a new terminal window and type in the command below

Copy or remember whatever the output of this command was

for the next step. For your reference in the images, my mine is

tzubatiy.

whoami

Figure 5.15: This is what your Terminal

window will look like when you use

the whoami command. The output is

highlighted. Your output will be your

username, not tzubatiy.

11.Go back into your Atom window, press "command"+"f" and then

search and then search for the text in the line below:

when you

ﬁnd it, replace the _www with the ouptput of step 9

User _www

Figure 5.16: This is what your Atom

window should look like when have

added your name to the user line, in

this case this is my name.

12.On the line after "Group _www", copy in the following code:

ServerName localhost

DocumentRoot "home/downloads/OSP/embedded\ Webserver/public"

AllowOverride All

Options Indexes MultiViews FollowSymLinks

Require all granted

</Directory>

</VirtualHost>

Figure 5.17: This is what your Atom

window should look like when have

added these lines of code in the right

place.

13.In your Atom window press the buttons "Command"+"s"

to save

the changes you have made, and close the ﬁle.

14.Type the following line of code into a new Terminal window

and press Enter:

sudo apachectl restart

15.Type the following line of code into a new Terminal window

and press Enter:

sudo chgrp -R www-data __PATH__/embeddedwebserver/

16.Type the following line of code into a new Terminal window

and press Enter:

sudo chmod -R 775 __PATH__/embeddedwebserver/storage/

17.Type the following line of code into a new Terminal window

and press Enter:

sudo chmod -R 775 __PATH__/embeddedwebserver/bootstrap/

18.Type the following line of code into a new Terminal window

and press Enter:

sudo apachectl restart

19.To connect to the OSP and control the RT-MHA REMOTELY, use

the extension -t

This starts the application in TCP daemon mode. In this mode, the

application is waiting for a connection on port 8001 from an OSP

client. (See section 3.2)

Once a client connects, RT-MHA can be

controlled remotely, such as the EWS, controlled using the web

apps described in section 4.

To launch OSP in daemon mode, type the line below into your

terminal and press Enter:

osp -t -T2 -d-40 -a -r

Notice how in the terminal window it now says:"Using TCP in

daemon mode" and "waiting for connection from client..."

20.Open a new terminal window, copy in the following line of code

and press Enter:

21.Type the following line of code into the Terminal window from

step 20 and press Enter:

cd Downloads/OSP/Embedded\ Webserver

22.Type the following line of code into this same Terminal window

and press Enter:

php artisan socket:start

23.Conﬁrm that you are connected to the same wireless network on

this Mac, as on the device you are trying to connect with

24.Open a browser window on the device you are trying to connect

with, and enter your IP address as the URL to start

Verifying the EWS is Working

For verifying that the system is working properly, we recommend

using Apple earbuds. Put them into your computer before adjusting

volume in step 2below.

6.1Researcher Page

1.If you still have OSP running, press "command"+"c" Otherwise,

proceed to the next step

2.Open a new terminal window, copy in the following line of code

and press Enter

osp -t -T2 -d-40 -a -r

3.Copy the following line of code in the same Terminal Window

in which you ended the previous step, and press Enter

php artisan serve

4.Open a new Terminal Window, copy in the line below and press

Enter:

cd Downloads/OSP/Embedded\ Webserver

5.Type in the following line of code into the Terminal window

from the previous step (4.1.4) and press Enter:

php artisan socket:start

6.Open your browser, and go the following URL

http://localhost:8000/researcherpage

7.Increase the volume on your computer to the maximum

8.Add at least 5db to each cell in the the G65 row

9.Press transmit and listen to the sound that plays

10.Make another change, let’s say removing those 5db and listen

again

11.As soon as you have heard a difference between two parameter

states, you can be sure that the software is working as intended.

6.24AFC

1.If you still have OSP running, press "command"+"c" Otherwise,

proceed to the next step

2.Open a new terminal window, copy in the following line of code

and press Enter

osp -t -T2 -d-40 -a -r

3.Copy the following line of code in the same Terminal Window

in which you ended the previous step, and press Enter

php artisan serve

4.Open a new Terminal Window, copy in the line below and press

Enter:

cd Downloads/OSP/Embedded\ Webserver

5.Type in the following line of code into the Terminal window

from the previous step (4.2.4) and press Enter:

php artisan socket:start

6.Open your browser, and go the following URL

http://localhost:8000/4afc

7.Click on the four choices and the play symbol

and make sure

you are hearing the appropriate word

8.Once you hear the words playing on click

, you know that the

4AFC app is working properly

Bibliography

Harinath Garudadri, Arthur Boothroyd, Ching-Hua Lee, Swaroop

Gadiyaram, Justyn Bell, Dhiman Sengupta, Sean Hamilton, Kr-

ishna Chaithanya Vastare, Rajesh Gupta, and Bhaskar D. Rao, "A

Realtime, Open-Source Speech-Processing Platform for Research in

Hearing Loss Compensation", In Signals, Systems and Computers,

2017 51st Asilomar Conference, IEEE, 2017.

James M. Kates, "Principles of digital dynamic-range compres-

sion," Trends in Ampliﬁcation, vol. 9, no.2, pp. 45-76,2005.

Shilpi Banerjee, The compression handbook: An overview of the

characteristics and applications of compression ampliﬁcation, 4th

Edition, Starkey Laboratories, 2017.

4. James M. Kates, Digital hearing aids, Plural publishing, 2008.

Toon van Waterschoot and Marc Moonen, "Fifty years of acoustic

feedback control: State of the art and future challenges," Proceed-

ings of IEEE, vol. 99, no. 2, pp. 288-327,2011.

Johan Hellgren, "Analysis of feedback cancellation in hearing

aids with Filtered-x LMS and the direct method of closed loop

identiﬁcation," IEEE Transactions on Speech and Audio Processing,

vol. 10, no. 2, pp. 119-131,2002.

Hsiang-Feng Chi, Shawn X. Gao, Sigfrid D. Soli, and Abeer Alwan,

"Band-limited feedback cancellation with a modiﬁed ﬁltered-X

LMS algorithm for hearing aids," Speech Communication, vol. 39,

no. 1-2, pp. 147-161,2003.

Donald L. Duttweiler, "Proportionate normalized least-mean-

squares adaptation in echo cancelers," IEEE Transactions on

Speech and Audio Processing, vol. 8, no. 5, pp. 508-518,2000.

Jacob Benesty and Steven L. Gay, â ˘

AIJAn improved PNLMS al-

gorithm,â ˘

A˙

I in Proceedings of IEEE International Conference on

Acoustic, Speech, and Signal Processing (ICASSP), pp. 1881-1884,

2002.

10.

Constantin Paleologu, Jacob Benesty, and Silviu CiochinÄ ˇ

C, "An

improved proportionate NLMS algorithm based on the l0norm,"

in Proceedings of IEEE International Conference on Acoustic,

Speech, and Signal Processing (ICASSP), pp. 309-312,2010.

11.

Arslan, Levent, Alan McCree, and Vishu Viswanathan. "New

methods for adaptive noise suppression." Acoustics, Speech, and

Signal Processing, 1995. ICASSP-95., 1995 International Conference

on. Vol. 1. IEEE, 1995.

12.

Hirsch, Hans-GÃijnter, and Christoph Ehrlicher. "Noise estimation

techniques for robust speech recognition." Acoustics, Speech, and

Signal Processing, 1995. ICASSP-95., 1995 International Conference

on. Vol. 1. IEEE, 1995.

13.

Cohen, Israel, and Baruch Berdugo. "Noise estimation by minima

controlled recursive averaging for robust speech enhancement."

IEEE signal processing letters 9.1(2002): 12-15.

14.

Lee, Ching-Hua, Bhaskar D. Rao, and Harinath Garudadri. "Spar-

sity promoting LMS for adaptive feedback cancellation". In Signal

Processing Conference (EUSIPCO), 2017 25th European, pp. 226-

230. IEEE, 2017.

Open Speech Platform User Guide OSP Manual

Navigation menu

Versions of this User Manual:

Views

Navigation