IFC formant tracker

From Phonlab
Revision as of 11:25, 22 July 2015 by John.stockdale (talk | contribs) (→‎Background noise)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

The IFC formant tracker, also known as ifcformant, is an acoustic analysis tool that can be used to help speed up the sometimes tedious process of measuring formant values in an audio file. ifcformant is capable of reading .wav files containing speech and returning the rms, fundamental frequency, and values for the first four formants at 10 ms intervals. ifcformant can also be used in conjunction with Praat TextGrids to greatly accelerate the process of formant analysis. For a description of how to do this, see Meas formants

How to use

Using ifcformant is straightforward. The command has two required arguments: the speaker, and the audio file. The speaker must be stated to be either male, female, or a child. The basic syntax for running ifcformant is as follows:

> ifcformant --speaker=male|female|child [--output=outfile] [--fields=fieldlist] [--fpfmt=floating_point_format] 
[--timefmt=time_format] [--sep=separator] [--sox=soxcmd] [-effects=soxeffects] [--print-header] [0--no-midpt] audiofile

For a detailed description of the optional arguments, use the command

> ifcformant --help 

Troubleshooting

Background noise

ifcformant is often used to analyze audio from non-laboratory sources, such as movies, TV shows, etc. An unfortunate byproduct of these types of sources is the existence of background noise, often music. ifcformant may fail to properly track the formant values in audio files containing large amounts of background noise, so it is preferable to find a source containing as small amount of audio besides the speech as possible. If this is not possible, one workaround is to try to see if the source is stereo, and isolate the channel containing the speech.