Down-sampling and converting sounds to use them in Asterisk

Applies to VoipNow 3.5!

Starting with VoipNow 3.5, it is possible to use sounds at higher quality. Before this latest version, all the sounds from Asterisk were encoded at 8KHz. From now on, all sounds files are updated to higher quality 16 KHz. Such sounds are already available for download here.

Step-by-step guide

Getting the sounds

Let's say we want to install a 16KHz French sounds pack recorded by June Wallack, on VoipNow 3.5. For that we need to download the proper archive.

mkdir /root/french

cd french

wget http://downloads.asterisk.org/pub/telephony/sounds/asterisk-core-sounds-fr-sln16-current.tar.gz

wget http://downloads.asterisk.org/pub/telephony/sounds/asterisk-extra-sounds-fr-sln16-current.tar.gz

Extract sounds from both archives and arrange those sounds in the correct format for a sound pack. Basically, all the sounds that are not part of a folder must go to a fr folder that you need to create.

mkdir fr

mv *.sln16 fr

In all the folders, except the newly created fr, there must be a fr sub-folder where all the sounds must go. For instance, all the sounds from the dictate folder must go in the fr sub-folder that must be created under dictate. At the end, the directory structure must have the following structure:

├── dictate
│   └── fr
├── digits
│   └── fr
├── followme
│   └── fr
├── fr
├── ha
│   └── fr
├── letters
│   └── fr
├── phonetic
│   └── fr
├── silence
│   └── fr
└── wx
    └── fr

Converting to the required format

In order to convert from the slin16 format to wave, we can use sox. The following command converts a sound to the proper format used by Asterisk:

sox -t raw -r 16k -e signed-integer -b 16 -c 1 <input>.sln16 output.wav

There are a lot of sounds in those folders. Using the following script, all sounds will be converted to the desired format:

#!/bin/bash
for i in `find /root/french -name '*sln16'`
    do
        final=${i:0:${#i} -6}.wav
        echo sox -t raw -r 16k -e signed-integer -b 16 -c 1 $i $final
        sox -t raw -r 16k -e signed-integer -b 16 -c 1 $i $final
        rm -rf $i
done

Once the sounds are converted, you need to move them to the folder used by Asterisk to store sounds and set the correct owner for those files:

cp -Rap * /var/lib/asterisk/sounds/

chown -R asterisk: /var/lib/asterisk/sounds/

At this point, you need to select French as Phone language in the Edit User section.

Down-sampling files the required format

If sounds are recorded in a professional way or with a professional recorder that is not able to record uncompressed wave at 16KHz, you may use the following command to down-sample those files:

sox sound.wav -e signed-integer -c 1 -b 16 -r 16k sound-down.wav

For an automated process, the following script will down-sample all the wave files from the /root/sounds folder (assuming they are stored there) to the required format used by Asterisk:

#!/bin/bash
for i in `find /root/sounds -name '*wav'`
    do
      final=${i:0:${#i} -4}-16k.wav
      echo sox $i -e signed-integer -c 1 -r 16k -b 16 $final
      sox $i -e signed-integer -c 1 -r 16k -b 16 $final
	  rm -rf $i
      mv $final $i
done

Show all results

Space shortcuts

Page tree

Step-by-step guide

Getting the sounds

Converting to the required format

Down-sampling files the required format

Related articles