-
SPEECH ANALYSIS EXPERT SYSTEM
ACU-EXPERT option OT represents an audio processing shell that is designed to the needs of forensic speaker voice identification and speech analysis.
It is developed hand-by-hand with practicing audio experts in accordance with the state of science and technology and supports all stages of speaker voice ID:- Segmentation and audio text decoding
- Audio suitability assessment
- Channel normalization and noise reduction
- Speaker voices comparison
- Expert report
The following speaker voice ID methods are available:
- Pitch extraction and statistical analysis
- Prosodic patterns comparison
- Formant tracks extraction and analysis
- Phonetic sets comparison
- Matching articulatory events analysis
- Vowel formants statistics collection
- Speaker profiling
The governments of various countries has approved ACUSTEK forensic solutions for government forensic audio laboratories and institutes.
FEATURING
PITCH STATISTICS - voice pitch curve is extracted automatically or manually over cepstrogram. Pitch distributions and their statistical characteristics are compared.
PROSODIC PATTERNS COMPARISON - fast calculated dynamic cepstrograms provide the evaluation of intonation stereotypes degree of similarity.
FORMANT ANALYSIS - search and comparison of formants positions for the matching vowels. Formant track dynamics analysis helps to estimate the similarity of speaking manner.
FORMANT STATISITICS - formant comparison table allows expert to make the conclusion over the basis of acoustic speech production theory.
TEXT SEARCH - the system provides search for matching words (or sllables).
ANALYSIS OF MATCHING ARTICULATORY EVENTS - language independent analysis of matching articulatory events is based on investigation of neurodynamic patterns used for speech production. Matching vocal tract configurations bring the same traces to the spectral layout of the acoustic event.
SPECTROGRAM LAYOUT NORMILIZER - view normalization helps to adjust the image without additional signal processing.