Making stimuli


Recording sounds

Insert instructions for scheduling in the P-lab / using that equipment

Editing sounds

We use Audacity to edit sound files for my experiments. There are four ways we typically edit my sound files. Select the best sample of a given stimulus, Remove background noise, Add or remove silence if necessary, and Export as a .wav file.

Select the best sample

First, select the file that you want to work on.

Right click to open the sound file you want to work on with Audacity.

With the sound file open, listen to each of the recorded examples of the sound and select the one that sounds best to you.

To select in Audacity you just click and drag around the sound that you want. In this example, I've selected the first one. Make sure that you get at least a little bit of the silence around the sound. You will need it to remove background noise later.

With the best sound selected, hit cmd + c to copy it to the clipboard. Open a new audacity file by hitting cmd + n. Then paste the selected sound file into the new audacity file with cmd + v.

Remove background noise

To remove background noise, we first need to zoom in to our sound stimulus. Use cmd + 1 to zoom in. (Note that cmd +3 zooms out and cmd + 2 returns the sound stimulus to its original size.)

Next we need to select a part of the silence to give Audacity a "noise profile". Do this by clicking and dragging around the silence that you want to select.

Click Effect > Noise Reduction to open the "noise reduction" dialog box.

With your silence already selected and the "noise reduction" dialog box open, click the Get Noise Profile button.

Next, go back to the sound stimulus and select the entire audio sample.

Select Effect > Noise Reduction again to return to the "noise reduction" dialog box. This time select OK.

You can adjust the dB of noise reduction, the sensitivity, and the smoothing. For my experiments, the default selections are appropriate. For reference, those default selections are:

  • Noise reduction (dB) : 12

  • Sensitivity: 6.00

  • Frequency smoothing(bands): 0

Add or remove silence

To remove silence (or any section of the stimulus), select the portion of want to remove and click delete.

To add silence, place the cursor where you want to add the silence. Here I've chosen the end of the stimulus.

Select Generate > Silence... from the Audacity menu.

In the "silence" dialog box, enter the amount of silence you want to generate. Here I have entered 30 seconds of silence.

Click OK to finish generating silence.

Export final stimulus

To export the final stimulus, select File > Export Audio... or File > Export Selected Audio... if you have selected a smaller portion of a file to export.

Enter a file name and select WAV format.

Select Save. You can enter meta-data if you wish to. I often leave this part blank.

Click OK to finish exporting your stimulus.

Synthesize sounds

Synthesize a short list of sounds

To synthesize sounds, you can use the say command in the terminal. For example, to synthesize a single sound, you would type

say -v Alex -r 250 "say this please" -o ~/Desktop/say-this.wav --data-format=LEF32@32000

Each parameter in this command are:

  • say is the command for Text-to-speech

  • -v Alex use voice (-v) Alex

  • -r 250 use a rate of 250 ms

  • "say this please" is the text you want to synthesize

  • -o ~/Desktop/say-this.wav is the path and name of the file you want to save

  • --data-format=LEF32@32000 is the format of the data (required for WAV files)

When I want so synthesize a short list of words or sentences with exactly the same parameters, I use my python script.

The parameters in the script are:

  • path is the name of a folder I want to create to save my sound files in

  • voice is the voice I want to use

  • rate is the speech rate

  • wordList is a list of words or sentences that I want to synthesize

To download the script, navigate to the directory you want to save the script in. For example, the Desktop.

cd ~/Desktop

then use curl to download the script.

curl -o

To run the script, first edit with the parameters you want to use. Then navigate to the script's directory. For example, if you saved on your Desktop you would navigate to the Desktop

cd ~/Desktop

And then run the script using


Synthesize a long list of sounds

When I need to synthesize a long list of sounds, or a list of sounds with different parameters, I use my script. This script reads parameters from a .csv file for each individual sound file.

Synthesize with natural reader

Instructions coming soon...


Finding images

When you are looking for images to use as stimuli, generally any open-source image search tool works well. You can see some sources we recommend on the Favorite tools page.

Editing images

There are two ways in which I typically edit image stimuli: removing background and re-sizing. I like to edit images with Inkscape, an open-source vector graphics editor. To show how to remove background and re-size images in Inkscape, we will edit this picture of doughnuts.

Removing background from images

To remove background images, I like to outline the part of the image that I want and use the clip tool. I created a short (silent) YouTube tutorial to show you how to do this. Click the image below to watch it.

Re-sizing images

Using Inkscape

To re-size images with Inkscape, you can create a background of your desired size and export it with your image. I created a (silent) YouTube tutorial for how to do this. It is available by clicking the image below.

Using Preview

It is also reasonably simple to edit images using Preview on a Mac. To begin, open the file you want to edit with Preview. Click and drag around the image to select the area you want.

You can further adjust by clicking and dragging the blue circles. Once you have the selection square the size you want, drag the square to adjust the position. Then you can hit cmd + k to crop the image. Once the image is cropped, hit Tools > Adjust size....

In the Adjust size dialog box, close the lock icon by clicking on it. This locks the aspect ratio to make sure the image does not get distorted. With the lock closed, adjust the size of the image. When the lock is closed, you only need to adjust one dimension (height or width). The other dimension will automatically adjust to preserve the aspect ratio.


Instructions coming soon...

Last updated