Posted on April 03, 2012
For my qualifying research paper, I automatically measured a great deal of data using a Praat script. Having made mine by putting together pieces of other scripts from the web, I thought I'd post mine so others can use it for their own work. It extracts the duration, pitch max and pitch minimum for segments marked up with a textgrid. It also changes the pitch settings to be more conservative and exclude creaky voice measurements.
Automatically Concatenating Sound Files:
Using Terminal, the command line program on Macintosh computers, to combine a series of sound files into one concatenated sound file (huge thanks to Kevin McGowan for telling me this was possible and writing the code in the create.sh file).
For example: (01aC.wav + SIL920.wav + 01bC.wav + SIL400.wav + 01cC.wav) --> 01C.wav
One of the most useful processes that I was able to take advantage of in the preparation and running of my dissertation experiments is the automatic concatenation of sound files. My experimental stimuli involved combined three sentences, each with its own sound file, with intervening pauses of different durations. Subsequent experiments involved different combinations of sentences and pauses, meaning I had a lot of concatenating to do. Fortunately, Kevin McGowan told me of a way to do so automatically. All I needed with a bit of code (which he provided), Mac's command line program Terminal, the SoX application Sound eXchange (free to download), and a text file that describes which files should be combined with which other files. I'll describe how it works, with the examples below. First, create a folder that has the two files create.sh and list.txt, along with two folders (one where the sound files to be concatenated are housed, and the other for where the resulting files will be put). I'm calling the source folder with the files to be concatenated DoingConcatenating and the output folder Concatenated. Then, open Terminal (presumably a similar program exists on a PC). Then, make your present working directory the folder where create.sh and list.txt are located. You can see what the present working directory is by typing pwd and hitting return. You can change the directory by typing cd followed by the path to the folder of your choice. You can find this path by copying the folder you want to make the present working directly into the command line after "cd" and hitting return. Now if you type "pwd" your present working directory should be the right one. Now you will want to get your list.txt file ready, which involves creating a line with each set of files you want to concatenate in a line delimited by tabs (see the example list.txt). The names of the sound files should be the same as the names of the sound files in the source folder. When your list.txt file has all the right file names, tab-delimited, you can go back to Terminal and type create.sh and hit return. If it worked, you should see a bunch of commands running, and then your concatenated sound files should be in your output folder. Please let me know if you take advantage of this!
Getting Things Done, by David Allen. This is a book about improving one's productivity.
750words.com. This is a website for freewriting.
Jason Kahn's website describing statistical modeling (http://jmkahn.web.unc.edu/modeling).
Materials on d' (d prime), part of signal detection theory; thanks to Diogo Almeida for introducing me to these: (http://wise.cgu.edu/sdtmod/index.asp),