This page presents a more detailed overview of the contents of the two possible output file formats of FAVE-extract: Plotnik files and tab-delimited text files.

Plotnik files

Plotnik files are plain text files designed to work with the Plotnik vowel analysis program. Most (but not all) of the data in a Plotnik file is delimited by commas. Plotnik files consist of three main parts:

  1. a header
  2. data for individual vowel measurements
  3. mean data for vowel classes
screen shot of the beginning of a plt file

The above screen shot shows the beginning of an example Plotnik file. The first two lines of the file are header lines containing background information about the speaker and the number of tokens included in the file, respectively. These two header lines are then followed by a large body of individual vowel tokens, with one vowel measurement per line.


The first line of the header contains the background information entered by the user in the "Speaker Information" fields of the FAVE-extract interface, separated by commas: speaker name, age, sex, ethnicity, years of schooling, location, year of recording.

The number in the second line of the header specifies the number of tokens (individual vowel measurements) contained in the file. In the example screen shot above, the file contains 34 individual vowel measurements.

Individual vowel measurements

The two header lines of a Plotnik file are followed by a number of lines containing measurement data for individual vowel tokens, with one vowel measurement per line. The following information is included for each vowel token:

F1, F2, F3, Plotnik vowel code, stress and duration, WORD -style code- /nFormants/ time stamp <formant tracks>

F1, F2, F3
First, second and third formant (if measurable) [in Hertz]
Plotnik vowel code: x.xxxxx
Plotnik vowel codes have the format x.xxxxx. The digit(s) before the period indicate the vowel class of the vowel in question, while the five digits following the period encode information about the manner, place of articulation and voicing of the following segment; preceding segment; and following sequence (complex coda, following syllables), respectively. (See the Plotnik documentation or this Plotnik Cheat Sheet for more details.)
Stress and duration: s.dur
Stress and duration are listed in the format s.dur. The digit before the period indicates the stress level of the vowel (1 for primary, 2 for secondary, 0 for unstressed), while the numbers after the period list the vowel's duration [in milliseconds].
Word containing the vowel in question
-style code-
If a style tier is present in the analyzed TextGrid, and the entries on the style tier match the codes for stylistic levels given in section 3.4 of the PNC transcription guidelines, FAVE-extract automatically converts these entries to the corresponding Plotnik style codes. Plotnik style codes consist of single digits (1 to 7) enclosed by a hyphen on either side. (See the Plotnik documentation or this Plotnik Cheat Sheet for more details.)
If the Mahalanobis formantPredictionMethod was chosen, this number indicates the value of nFormants that was chosen for the final measurement. Otherwise, this number corresponds to whichever value was set by the user for nFormants in the FAVE-extract interface.
time stamp
The time stamp gives the point of measurement [in seconds], i.e. where in the vowel the formant measurements were taken.
<formant tracks>
The formant "tracks", enclosed in angled brackets at the end of each data line, contain formant measurements for both F1 and F2 at 20%, 35%, 50%, 65% and 80% of the vowel's duration, respectively: <F1@20%, F2@20%, F1@35%, F2@35%, F1@50%, F2@50%, F1@65%, F2@65%, F1@80%, F2@80%>
screen shot of a data line

For example, the first line in the screen shot above lists measurement data for the vowel in the word "BUT". The first, second and third formants are measured as 460.4 Hz, 1296.5 Hz, and 2388.5 Hz, respectively. The Plotnik code for this vowel class is "6", and the vowel in question is followed by a stop ("1") which is apical ("4") and voiceless ("1"). The vowel is preceded by a oral labial ("1"), and followed by neither a complex coda nor additional syllables ("0"). At the word level, this vowel has primary stress ("1"), and the duration of the vowel is 50 milliseconds. The style code tells us that this token comes from a reading passage ("-5-"). The point of measurement is 0.853 seconds into the recording.

Detailed information about the format of Plotnik data lines, as well as possible values for vowel and style codes etc. can be found in this Plotnik Cheat Sheet.

Vowel means

screen shot of the beginning of the vowel means data in a plt file

At the end of the file, a blank line separates mean data for all vowel classes from the preceding individual token measurements. The format for the vowel class means is as follows:

Plotnik vowel code, number of tokens for that vowel class, mean F1, mean F2, mean F3, F1 standard deviation, F2 standard deviation, F3 standard deviation <mean formant tracks>

where <mean formant tracks> contains the mean F1 and F2 values at five selected points over the vowel's duration: <F1@20%, F2@20%, F1@35%, F2@35%, F1@50%, F2@50%, F1@65%, F2@65%, F1@80%, F2@80%>.

For example, in the screen shot above, the first line after the blank contains the mean formant data for Plotnik vowel class "1" (short-i). There are 166 short-i tokens in the file, with a mean F1 of 439 Hz, a mean F2 of 1650 Hz, a mean F3 of 2424 Hz, and standard deviations of 38 Hz, 157 Hz and 177 Hz for F1, F2 and F3, respectively. The mean formant values for F1 and F2, measured at 20% of the vowel's overall duration, are 438 Hz and 1647 Hz; at 35% of the vowel duration, 442 Hz and 1656 Hz, and so on.

(For more detailed information on the format of Plotnik files, please see the documentation that comes with the Plotnik program.)

Tab-delimited .txt files

Tab-delimited text files contain the same information as the Plotnik files, with individual items listed in tab-separated columns.

screen shot of options file

The following information is listed in the columns of the tab-delimited text file:

ARPABET vowel code (two letters, ASCII)
Stress level (1 for primary, 2 for secondary, 0 for unstressed)
Word containing the vowel in question
F1, F2, F3
First, second and third formant (if measurable) [in Hertz]
B1, B2, B3
Bandwidths of the first to third formants (if measurable) [in Hertz]
Time stamp (point of measurement) [in seconds]
Beginning of the vowel [in seconds]
End of the vowel [in seconds]
Vowel duration [in seconds]
Plotnik code: vowel class
Plotnik code: manner of articulation of the following segment
Plotnik code: place of articulation of the following segment
Plotnik code: voicing of the following segment
Plotnik code: preceding segment
Plotnik code: following segments (complex coda/following syllables)
Entry on the style tier (if present)
Automatic detection of shortened or monophthongized glides (s for shortened glides, m for monophthongs). Currently implemented for AY only.
Formant settings: number of formants to search for. (This is mainly interesting if you are using the Mahalanobis formantPredictionMethod.)
F1@20%, F2@20%, F1@35%, F2@35%, etc.
Formant "tracks": formant measurements at 20%, 35%, 50%, 65% and 80% of the vowel's duration

(For more information on the Plotnik codes cd, fm, fp, fv, ps and fs, please see page 3 of the Plotnik Cheat Sheet.)