Working with Audacity

Recording files must be in MS WAV format with a specific sample rate – 16 kHz, 16 bit, mono for desktop application.
It’s critical that the audio files have a specific format.
Use the freac software for quick conversion of multiple audio recordings to the required format.

In Audacity find Edit -> Preferences -> Quality :
Set the Dither field to Triangle ( or Shaped).
This is essential , as digital silence is not recognised as “silence”
by the Sphinxtrain trainer. All “silences” must be filled with a dither.

Create the following macro using Tools -> Macros -> New
1. Normalise : Check Remove DC offset and Normalise peak amplitude to -1.0 Db.
2. Truncate silence : Threshold -48Db , Duration 0.01 seconds, Truncate detected silence to 0.01 seconds, Check “Truncate tracks independantly”.
3. Trim Extend : Trim extend Start and End by 1 second.
4. Truncate silence again : Threshold -35 Db, Duration 0.5 seconds, Truncate detected silence to 0(zero) seconds, Check “Truncate tracks independantly”.
5. Trim Extend again : Trim extend Start and End by 1 second.
6. Truncate silence again : Threshold -48 Db, Duration 0.5 seconds, Truncate detected silence to 0.15 seconds, Check “Truncate tracks independantly”.
7. Move focus to next and select.
8. – End –

You will need to download and install the Nyquist plug-ins for Truncate Silence and Trim Extend.
Download and move the file to your Plug-in folder under Audacity .
Activate the plug-in by clicking on : Generate – > Add/Remove Plug-ins
https://wiki.audacityteam.org/wiki/Nyquist_Effect_Plug-ins

This macro removes silences from the beginning and end of the Ayah and any pauses during recitation, leaving 0.15 seconds of silence at the ends.
You may experiment with these settings to find a better result.

Select and drag files into Audacity .
Then run the macro using : Tools -> Apply macro -> Your macro name.
Audacity tolerates around 500 Mb or 1000 files per run (Therefore around 6 -7 runs are required to complete a full Qur’aan).
Beyond that, it may crash.

Save your converted files with File -> Export -> Export multiple.

Another useful tool is Found at Generate -> Silence
Click and drag to select parts of the recording that you wish to remove , then click Generate -> Silence. Then run your macro.
This should produce a properly trimmed .wav file.

Sources

The best audio preparation freeware is available at Audacity

Other handy tools include:

https://www.freac.org/ – for bulk audio conversion to the required audio format for CMU Sphinx

https://www.bulkrenameutility.co.uk/ – for quick and easy renaming of multiple files