Phonexi 3.4

 Phonexi 3.4

Interactive system for identifying the speaker Phonograms Russian speech "Foneksi." Version 3.4 Win XP / 7, 2011. (Certificate of official registration number 2000611270 from 06.12.2000g. ROSPATENT).


Interactive system (software - software) "Foneksi" is intended for forensic expert research phonograms to identify persons on the specifics of their Russian speech.


Can do the following: I / O phonograms speech to / from a PC, a preliminary assessment of the quality of phonograms, the calculation of acoustic and linguistic identification of individualizing characteristics, comparisons of features and a decision on the identification, further training of decision rules.
Allows you to perform forensic examinations of the speaker identification Phonograms speech is not high quality.
Integrated use of acoustic (instrumental) and linguistic (auditory) study phonograms speech make for a complete description of the individual characteristics of the speaker's speech.
Allows the acoustic survey and identification of speakers Phonograms speech as random, and with the same context (phrases and words).
Has the ability of adaptive retraining and customization of decision rules on the features of any spoken language of the audited entity.
For short speech phonograms available means of acoustic research comparable context triads (sounds) with its system of statistical decision threshold.
There is a possibility of adaptive selection of individualizing characteristics and decision rules for different quality compared phonograms speech.
Has in its memory support base sounding linguistic standards, avoiding one-sided approach and subjective errors in linguistic research tracks.
The system has means of investigation and decision thresholds separately for phonograms speech of men and women.
Is provided with context-sensitive file detailed step by step instructions to the user for the identification of studies, guidelines and examples of expert elicitation.
ON "Foneksi" adaptable to various standard I / O unit of speech signals to / from the PC (CSL KayPENTAX, Sound Blaster, etc..), As well as to the various formats of digital files of speech.
Requirements for the quality of phonograms speech used to identify

Duration of phonograms arbitrary speech unknown person - 10 sec; suspect litsa- 100 sec.
The frequency range of speech signals - 300-3400 Hz, and change in the detonation velocity recordings - less than 2%, the ratio of average values ​​of speech signals and noise - not less than 15 dB.
Reliable identification of the speaker

ON "Foneksi" allows identification speaking Phonograms arbitrary speech (with the above quality) for the integrated use of all linguistic groups and acoustic individualizing features with an average error rate of identification at no worse than 0.01.
Requirements for installing the software "Foneksi" to your computer

To install, you must have at least 300 MB of free disk space.
The presence of free disk space for voice files at least 80 GB.
Running the operating system Windows XP / Win7.
Works with the following devices, audio input / output to / from the PC: the standard sound card like Sound Blaster and CSL company KayPENTAX.
When you install on your computer and operating system requires a security key.
Scope of delivery

Laser disk with software "Foneksi."
Documentation in the form of a Help file with the methodology of work and step by step instructions on the CD-ROM.
Key software protection.