I was introduced to Praat last year for another project I was a part of. We were using this software to measure the latencies in between responses for a Verbal Fluency task. Praat is essentially a phonetic software that allows its users to analyze a speech sample. It can do anything from spectral analysis to cochleograms. In this post I will walk you through three features I’ve learned and used for this current project.
- Measuring Pitch
After uploading the sound file on Praat and clicking “view” you’ll see a waveform on the top half of your screen and a spectrogram on the bottom half. Sometimes in you’ll see a series of dots, blue lines and yellow lines. The dots represent formants, the blue lines represent the speakers pitch and the yellow line represents the intensity. If you’re interested in measuring the average pitch of a segment of the recording you can simply highlight it using your cursor. Once you’ve highlighted the desired segment, select the “pitch” tab on the menu bar and select “get pitch”. This will give you an average pitch in Hz of the segment you’ve selected. You can also get the minimum and maximum pitch of the segment.
2. Measuring Intensity
Getting the intensity of a segment is similar to getting the pitch. You select the “intensity” tab and click “get intensity”.
3. Measuring duration
You can see what time the duration of a segment, when segment begins (onset) and when it ends (offset) in two ways. Either record manually the times on the top left and right hand side of the selection written in red. The time on the left hand side it the onset, the time in the middle is the duration time of the selection and the time on the right is the offset time of the stimulus.
You could also select “query” and select “get start of selection” to get the onset and “get end of selection” to get the the offset. I personally prefer doing it this way because copying and pasting the values leaves less room for human error.
This software has become an integral part for the destining stage of the project I’m currently doing with Dr.Shafer because we’re choosing the recordings for the stimulus based on the pitch, intensity and length. This is important because participants are asked to discriminate between the target stimulus from non-target stimulus under several types of condition. One of the conditions involves two female speaker, when recording the stimulus we have to make sure they’re close enough in pitch and intensity but not too close.