Chop up long sound files

Often linguists have long recordings and want to analyze subsections of this. If the subsections are indicated in textgrid files, there is a function in Fast Track that can help you get the files ready for use in Fast Track by cutting them with the appropriate padding.

 

If you select a sound and its corresponding textgrid, you will see a new option: "Fast Track > Chop up sound file". If you select this you will see a form appear. Indicate:

• Folder: Cut sounds will be placed in this folder.

• Segment tier: Which tier has segmental information

• Word tier: Which tier has word information.

• Marker: Only segments with this mark at the end of them are extracted. By default this is "_x".

 

The function goes through the textgrid and extract all sounds with "_x" at the end of the segment label. The reason for this is that often we are interested in only a small subset of sounds and analyzing all vowels is time consuming, and then leads a researcher to have to pick through the analyzed data to find what they are interested in. Of course, this function can be modified in any way to suit individual needs, and will likely be refined in the future. For example, right now it only works if you have a word tier, although this is not strictly necessary for the primary purpose of the function.

 

Output

---

The function outputs the cut sound files, one file for each indicated segmental interval. The naming convention for the files is 'filename_XXXX' where filename corresponds to the original large-sound filename, and XXXX corresponds to the number of the file (1st = 0001, 2nd = 0002, etc.)

the function also generates a CSV file containing the following information across columns, with one row for each file:

• file - the file number.

• filename - the filename of the output file.

• vowel - the vowel label as indicated in the vowel tier (the indicator is automatically deleted).

• interval - the interval number for the vowel.

• start - the start of the vowel segment, in seconds.

• end - the end of the vowel segment, in seconds.

• word - the word the segment is in, according to the label in the word tier.

• word_interval - the interval number for the word.

• word_start - the start of the word, in seconds.

• word_end - the end of the word, in seconds.

• previous_sound - the immediately preceding sound ("-" if nothing).

• next_sound - the immediately following sound ("-" if nothing).

• previous_word - the immediately preceding word ("-" if nothing).

• next_word - the immediately following word ("-" if nothing).