In the example, breakAfterEvery is 0, because there are only 12 trials. To allow her some breathing time, you can insert a break after every so many trials. Since the carrier phrase is concatenated with the stimulus before it is played, it should have the same sample rate as the stimulus files.Ī new stimulus will arrive as soon as the listener makes her choice. If you want to have a short silence before every stimulus (and before the carrier phrase), you supply a non-zero initialSilenceDuration, as in the example. If you want only an introductory phrase before the stimulus, and nothing after the stimulus, you do something likeĪnd of course you supply the file listenTo.wav in the directory Sounds. If you do not want a carrier phrase, you do: In the example, we see that the stimulus will be inserted between the sounds in the files weSayTheWord.wav and again.wav, both of which are expected to be in the directory Sounds. The stimuli can be embedded in a carrier phrase. Since Praat supports the forward slash "/" as a directory separator on all computers, you can run the exact same experiment on Macintosh, Windows and Unix computers, independent of the type of computer where you have created your experiment. Or you can put the whole experiment on a CD and run the experiment directly from the CD. The advantage of using relative file paths is that you can move your whole experiment (experiment file plus sounds) from one computer to another without changing the experiment file, as long as you put the experiment file in the same directory as where you put the directory Sounds. If you have an older Macintosh and the stimuli are in My disk:My experiment:Stimuli, you writeįileNameHead = "My disk:My experiment:Stimuli:"īut relative file paths will usually be preferred: they are more portable. If you have a Macintosh (OS X) or Unix computer and the stimuli are in /Users/mietta/Sounds/Dutch, you writeįileNameHead = "/Users/mietta/Sounds/Dutch/" For instance, if you have a Windows computer and the stimuli are in the directory D:\Corpus\Autumn\Speaker23, you can writeįileNameHead = "D:\Corpus\Autumn\Speaker23\" Such a path depends on your computer and on your operating system. Instead of a relative path, you can also supply a full file path. In other words, "Sounds/heed.wav" is a relative file path. In this example, the experiment will look for the sound files in the directory Sounds, which has to be in the same directory as your experiment file. They all have to have the same sample rate. You can also use AIFF files, in which case fileNameTail would probably be ".aiff", or any other type of sound file that Praat supports. Hence, the 4 stimuli are expected in the following 4 files: You can see that the stimulus names are heed, hid, hood, and hud. The names of the sound files containing the stimuli must be identical to the names of the stimuli, bracketed with fileNamehead and fileNameTail. Thus, the listener is confronted with 12 trials. ![]() This experiment will play 4 different stimuli to the listener, each 3 times. PauseText = "You can have a short break if you like. RunText = "Choose the vowel that you heard." The order of the elements in this file cannot be changed. The first two lines have to be typed exactly as in this example, the rest depends on your stimuli, on your response categories, and on the way the experiment is to be presented to the listener. ![]() The following is an example of such an experiment file. My estimation is that that is fine for 90 percent of the perception experiments that phoneticians perform.Īn experiment is defined in a simple text file, which we call an experiment file. If you do not need those extra capabilities, you can use the simpler method of Praat. With these programs, you can measure reaction times as well, or you can let the stimulus depend on the subject's previous responses. If you require more from your experiment design, you use a dedicated program like Presentation or E-prime instead of Praat. Because of the limited possibilities, it is also quite easy to set up the experiment. The advantage of using Praat for this is that it is free, it works on Windows, Unix, and Macintosh, and the whole experiment (experiment file plus sound files) is portable across computers (you can run it from a CD, for instance). ![]() For discrimination, you can have simple same-different choices, or more intricate things like AXB, 4I-oddity, and so on. `Simple' means that for identification, the subject hears a sound and has to click on one of a set of labelled rectangles (optionally, you can have the subject give a goodness-of-fit judgment). With Praat, you can do simple experiments on identification and discrimination. ![]() One of the types of objects in P RAAT, for running a Multiple Forced Choice listening experiment.
0 Comments
Leave a Reply. |