Comparison between VOCALAB and PRAAT

 

Benefits of PRAAT :

• Free
• Well known in the world of sound processing
• Internal powerful programming language
• Calculates spectrogram
• Extraction of fundamental and formant
• All classical operators ( Jitter, Shimmer , HNR , etc. ..)
• Well suited to the discovery of sound/speech characteristics
• Using the programming language, you may pilot statistical analysis rapidly

Disadvantages of PRAAT :

• No real time feature (recorded first , then display and apply computational tools)
• Not always easy to use : several commands to chain in order to obtain a result
• Complex menus, most commands without interest for speech therapists
• Indicators of voice quality ( shimmer , jitter , APQ , PPQ ) poorly adapted to the context of speech theraphy
• Indicators not easy to comparable to MDVP results using similar computational algorithms
• Noise-sensitive algorithms, delicate algorithm setting requiring good knowledge of sound processing
• No pathological voice database

 

Comparison between VOCALAB and PRAAT

PRAAT VOCALAB  
Free Low cost Cost
Web-based support Available; dedicated trainings organized regularly Training for speech therapists
Signal processing-oriented interface Speech-therapy-friendly interface with expert therapist on-line resources User-friendly interface
Step-by-step voice analysis with state-of-the-art signal processing algorithms. Real-time voice analysis using speech-therapy optimized signal processing Quick on-device voice analysis
Numerous options, choices of algorithms and parameters Strict minimum Speech processing parameters
no real-time features. Based on recorded samples. Real-time spectrogram for voice analysis. "before/after" speech representations for objective progress analysis Immediate audio playback with visual feedback
No media library High-quality voice recordings, media-library of 100 sound High-quality voice recordings
none 100 sounds in media library and on-line data base of more than 250 pathological voice samples Data base of voices
May be programmed Handling of ambient noise with advanced filtering algorithms Automatic cancelation of the ambient noise
Many indicators ; user’s defined parameters may be programmed 5 indicators of voice quality optimized for speech therapy Voice Stability Analysis
none Dedicated indicator Vowel Attack evaluation
Vary depending on indicators and operating procedure Fixed to 1.0 Pathology thresholds
No figure for illustrating the vocal range extraction, No phonetogram available Analysis of pitch, range, generation of relevant pictures as well as phonetogram extraction Recording and analysis of text
No direct evaluation Dedicated screen, automatic computing of averages and pathology index Analysis of S/Z, A/Z ratio
Available through text Available through dedicated graphics Analysis of mean pitch and pitch range
Non dedicated interface but feasible Measurement of the longest you can hold a note, including A/Z and S/Z ratio for in-depth pathology analysis. Measurement of the longest you can hold a note - Maximum Phonation Time (MPT)
No dedicated interface, but can be measured Singing pitch range analysis, conversion into note range, illustration of voice mechanism using spectrogram Singing pitch range
May be programmed using internal software resources, but no dedicated interface to speech therapy. History of voice indicators, vocal range, pitch, MPT, before, during and after therapy for immediate visualization of progresses in therapy. History of pitch, jitter, shimmer, pitch range, singing pitch range and MPT
Statistical analysis can be performed with good knowledge of internal language and Capability of contributing to the on-line data base of /a/ and /siren/, and interactive Capability to capitalize analysis results and audio
algorithm control. discussions with speech therapy experts. files
not targeted to speech therapy, no dedicated support in relation with voice therapy Capability of automatic generation of speech-therapy report for the compilation of all speech indicators Speech therapy report

 

Written by E. Sicard, 11.03.2014