Thursday, February 13, 2014
Powerful speech analysis tool
Praat is a piece of software that helps phoneticians analyze, synthesize, and manipulate speech.
Praat enables you to create a source signal from scratch or from an existing speech signal, but also to create a filter or to extract it from an existing source. You can manipulate (change, adapt) both the source and the filter before doing the actual synthesis, which combines the two.
Praat enables you to create a source signal from scratch or from an existing speech signal, but also to create a filter or to extract it from an existing source. You can manipulate (change, adapt) both the source and the filter before doing the actual synthesis, which combines the two.
Here are some key features of "Praat":
Speech analysis:
· spectral analysis (spectrograms)
· pitch analysis
· formant analysis
· intensity analysis
· jitter, shimmer, voice breaks
· cochleagram
· excitation pattern
Speech synthesis:
· from pitch, formant, and intensity
· articulatory synthesis
Listening experiments:
· identification and discrimination tests
Labelling and segmentation:
· label intervals and time points on multiple tiers
· use phonetic alphabet
· use sound files up to 2 gigabytes (3 hours)
Speech manipulation:
· change pitch and duration contours
· filtering
Learning algorithms:
· feedforward neural networks
· discrete and stochastic Optimality Theory
Statistics:
· multidimensional scaling
· principal component analysis
· discriminant analysis
Graphics:
· high quality for your articles and thesis
· produce Encapsulated PostScript files
· integrated mathematical and phonetic symbols
Programmability:
· easy programmable scripting language
· communicate with other programs
· (the sendpraat source code)
· create hypertext manuals with sound I/O
Speech analysis:
· spectral analysis (spectrograms)
· pitch analysis
· formant analysis
· intensity analysis
· jitter, shimmer, voice breaks
· cochleagram
· excitation pattern
Speech synthesis:
· from pitch, formant, and intensity
· articulatory synthesis
Listening experiments:
· identification and discrimination tests
Labelling and segmentation:
· label intervals and time points on multiple tiers
· use phonetic alphabet
· use sound files up to 2 gigabytes (3 hours)
Speech manipulation:
· change pitch and duration contours
· filtering
Learning algorithms:
· feedforward neural networks
· discrete and stochastic Optimality Theory
Statistics:
· multidimensional scaling
· principal component analysis
· discriminant analysis
Graphics:
· high quality for your articles and thesis
· produce Encapsulated PostScript files
· integrated mathematical and phonetic symbols
Programmability:
· easy programmable scripting language
· communicate with other programs
· (the sendpraat source code)
· create hypertext manuals with sound I/O
What's New in This Release:
· Create Sound as pure tone
· First steps towards variable-substitution-free scripting: the do, do$, writeInfo and appendInfo functions
· Create Sound as pure tone
· First steps towards variable-substitution-free scripting: the do, do$, writeInfo and appendInfo functions
Program Information
Developer: Paul Boersma/David Weenink
Price and Added Date
GPL / $0.13-02-2014
Download Links
.
