Step 2 - Record your Speech with Audacity

Installing Audacity 

If you are using Linux (we will be using the Fedora in this example) use the following command to download and install Audacity (as superuser):

$su
Password:
#yum install audacity  

Starting Audacity

Quiet Environment

Before you begin, you need to make sure that the room you are recording in is as quiet as possible.  In addition, make sure you turn off you speakers while recording - to avoid acoustic feedback in your audio files.

See this FAQ entry for more information. 

Adjust your Microphone 

Next you need to adjust your microphone for optimal pickup of your voice. 

If you have a headset microphone, this should be easy to do.  Your microphone should be a bit to the side and below your mouth (so the microphone won't pick-up your breathing), and no more than a half inch (1-2 cm) away.

See this FAQ entry for more information.

Recording Levels

Now you need to test your recording levels. 

Start Audacity. 

Make sure your microphone volume in Audacity is set to 1.0. 

Then click Record (i.e. the red circle button) and begin speaking in your normal voice for a few seconds, and then click Stop (i.e. the yellow square button). 

Look at the Waveform Display for the audio track you just created.  The Vertical Ruler to the left of the Waveform Display provides you with a guide to your audio levels.  Try to keep your recording levels between 0.5 and -0.5, averaging around 0.3 to -0.3.  It is OK to have a few spikes go outside the 0.5 to -0.5 range, but avoid having any go beyond the 1.0 to -1.0 range, as this will generate distortion.  If necessary, adjust Audacity's microphone volume to keep your audio within the proper ranges.

For more information see this FAQ entry.

Configuring Audacity Preferences

VoxForge collects speech audio at the highest Sample Rate that your Sound Card can support (up to a Sampling Rate of 48kHz, at 16 Bits Per Sample).  You'll need to look at your Sound Card's manual to determine the maximum it supports.  If you don't have your manual, this FAQ entry might help you get the required information (another FAQ entry provides some caveats with respect to your sound card and recording rates).

In Audacity, you set the Project Sampling Rate and Sample Format in your Preferences (under the 'File' menu selection).  Click  the 'Quality' tab:

• set your 'Default Sample Rate Format' by clicking the up/down arrows to change it to 48000Hz;
• set your 'Default Sample Format' to 16-bit.

Next click the 'Audio I/O' tab, and then:

• set your 'Channels' to 1 (Mono).

Then click the 'File Formats' tab, and then

• set your 'Uncompressed Export Format' to WAV (Microsoft 16 bit PCM) or export your audio using FLAC format.

Note: Please only submit audio files in an uncompressed format such as WAV or AIFF or lossless compressed format such as FLAC.

 

Now you need to exit and re-start Audacity to make these Project Setting changes active.  Look at Project rate selector on the bottom left hand corner of the Audacity window, make sure it says 48000. 

See this FAQ entry for details.

Recording your first Audio Track

Prompts file

Each line of the prompts file (which you got in Step 1) corresponds to the transcribed contents of one audio file.  The first column contains the name of the audio file.  The following words on the same line contain the text you will record in the audio file (see example below):

Audacity

So cheer up, and give us your paw

Leave a one second pause before and after you speak (these pauses will help in determining noise levels in your recordings).  Speak normally, not too fast or too slow, and clearly.  Speak as you would if you were reading the text aloud to someone else, with appropriate pauses corresponding to the punctuation in the text. If you are unsure how to pronounce a word, this FAQ entry might help. 

Click the 'Stop' icon when you are completed. 

Don't exhale until after you have clicked stop (most people have a tendency to breath out after speaking a sentence) otherwise your microphones will pick up your breath noises.

Review your waveform to ensure that your recording levels are between 0.5 and -0.5, averaging around 0.3 to -0.3.  It is OK to have a few spikes outside the 0.5 to -0.5 range, but don't have any go beyond the 1.0 to -1.0 range. 

If your waveform looks good, then listen to the track (press 'Play' in Audacity) to make sure your pronunciation is clear and that you do not hear any non-speech noises (i.e. breathing noises, lip smacking, or background noises, ...).

Click here for Tips for Recording VoxForge Prompts with Audacity.

Recording your Remaining Audio Tracks

Record your second audio track by clicking Record again (in Audacity) and say the words in the second line of your prompts file.  Audacity will start a second track leaving your first track untouched.

Review your waveform and listen to each track before proceeding to the next one.  You can select a particular track to listen to by clicking 'solo' on the track menu of that particular track, and then clicking play in Audacity.  Remember to click 'solo' again when you are done, to unselect the track.

Repeat the same process for each line in your prompts file, adding a new track for each line. 

Exporting your Audacity Audio Tracks to Wav files 

When you have finished recording all the prompts in the prompt file, then click 'File' on the Audacity menu, and click 'Export Multiple...'.  The change the following settings on your 'Export Multiple' window:

• Set your 'Export format' to 'WAV';
• Enter your 'Export location' (we will assume it is a directory called 'train');
• In the 'Name files' section, select 'Numbering consecutively';
• Set the 'file name prefix' to the VoxForge file prefix used in your prompts file (in this case it is 'vf5').

Then click 'Export'.

Comments