Name: Denoiser Clear Voice, Declicker Clear Voice, Sound Stretcher, VoiceCom, Voice Key, Music Spotting,

Text: Voice of the Future

Speech
Technology
Center

Developers’ Solutions

NOISE CANCELLATION
SPEAKER VERIFICATION
VOICE COMMAND RECOGNITION

Speech Technology Center Limited (STC) is active in the following areas: speech processing and enhancement, voice
recognition, speaker verification, noise filtering, audio signal analysis, logging, monitoring, music and keyword searching in
a sound stream.
STC develops, manufactures and markets software and hardware solutions, standalone devices, DSP boards and software
toolkits.
Founded in 1990, STC has an excellent reputation in the high-tech market, designing algorithms for high performance
speech processing applications. STC provides end-user solutions as well as SDK for PC and DSP platforms.
STC’s scientific team focuses on developing unique know-how in the field of speech/signal technology. All STC products
are based on proprietary patented technologies.
The software libraries and DSP solutions developed by Speech Technology Center are mostly used in such industries as:







Telecommunications
Business software applications
Multimedia software
Home appliances
Mobile accessories
Medical equipment

Noise Cancellation

1

Voice Command Recognition

3

Speaker Verification

4

Music Spotting

5

All the software and hardware modules are offered by STC for
licensing. STC software and hardware modules are easily integrated
into customer’s products. All software development kits include:
32 bit DLL
API
User’s manual
Demo software with the code examples.



All the hardware modules are based either on TI C54xx DSP,
or hyperstone E1-16, 32.
Customers can easily combine STC products for a complete solution.
Customization, engineering and consultancy services are also
available. Technical support during the development stage is provided.

N O I S E C A N C E L L AT I O N
Denoiser Clear Voice SDK (Software Development Kit)
Denoiser Clear Voice SDK is multifunctional, real-time, noise suppression software designed to remove noise and
improve speech clarity.
Voice recognition: Substantial reduction in noise and improvement of voice quality.
Teleconferencing solutions and IP telephony: Enhanced speech quality and superior compression on VoIP.
Sound processing software for Broadcasting and music restoration: Improvement of sound quality in
poor speech and music recordings.
Adaptively removes background noise in real-time.
Adjustable noise filtering parameters.
Overall Signal-Noise-Ratio (SNR) improvement: up to 30dB.
Tonal noise suppression: up to 50dB.
Broadband noise suppression: up to 15dB.
Effective against the following noise sources: AC, office equipment and fans, industrial and vehicle engines,
street traffic, wind, slowly varying music, power supply hum, etc..
Denoiser is also available as an ActiveX control: this provides easier and rapid integration into any application,
includes examples of web application.



Denoiser DSP board
Hardware module with automatic speech enhancement for:
Intercoms, mobile phones, hands-free kits, speakerphones, two-way radios.
Telephony equipment: answering machines.
Sound processing equipment for broadcasting.
Voice recognition devices and dictation equipment.
Sound recorders.
Call-center equipment.
Any communication device operating in noisy environments such as
aircraft cabins, construction sites, factories.








Real-time signal processing, high speed and stability of algorithm
performance.
High quality of processed speech.
Commonly used interfaces supported.
Available for TI TMS C54 and Hyperstone E1-16,32.
Low cost of integration and further production.
Small physical size 50mm x 40mm; customizable upon request.

TI eXpressDSP compliant Denoiser algorithm is available for licensing.

1

N O I S E C A N C E L L AT I O N
Declicker Clear Voice SDK
Automatic real-time removal of pulse-like interference (clicks, pops,
cracks, etc.) for:












Telecommunications: Speech quality improvement for
telephone, satellite and radio channels.
Broadcasting: Quality enhancement for radio transmissions
and audio data.
Music restoration and CD mastering: Audio signal
improvement for vinyl LPs and DAT recordings.
Multimedia software: Enhanced sound quality.
Effective against pulse-like and spike-like distortions of LP
recordings, digital clicks and static discharge cracks of DAT
masters, impulse noise in communication channels.
Adjustable noise filtering parameters.

Denoiser-by-Reference Clear Voice SDK
Significant improvement in music and sound recordings’ quality, using noise pattern
reference for:







Sound processing software for Broadcasting and Music restoration: Accurate
and gentle enhancement of sound quality for poor speech and music recordings,
particularly effective on old recordings.
Multimedia software: Music and sound quality improvement.
Adjustable noise filtering parameters.
Effective against various distortions found on most old recording media.

Sound Stretcher SDK - Variable Speed Speech Playback
Sound Stretcher SDK allows playback of speech in variable fast or slow speed without
distorting voice pitch, so that speech still sounds natural. This gives the listener a chance
to alter the speed and have complete understanding of the content for analysis purposes.
It is an algorithm of real time speech modification for:











2

Multimedia and Broadcasting: More efficient use of advertisement airtime.
Transcription of speech: The possibility to slow down speech rate and still hear
coherent sentences enables the user to perform the task quicker and easier.
Educational software for studying foreign languages: Better understanding of foreign
languages encouraging the learning process.
Up to 3 times faster or slower speech playback speed.
Speed can be changed without voice pitch distortions.
Language-independent. Sound Stretcher can be used for any language regardless of
the user’s accent.
Smooth acceleration/slow-down rate adjustment.

VOICE Voice
COMMAND
RECOGNITION
Command
Recognition
VoiceCom SDK
Voice command control for:















Education: Using voice commands makes the
learning process more enjoyable for both children
and adults.
Notebooks & PDAs: Giving voice commands is
easier than using a cumbersome mouse and a
small keyboard.
Industrial applications: Where a person’s hands
need to be free for other tasks.
Applications for people with physical difficulties.
Language-independent.
Few requirements for microphone quality and
placement location.
Low resource consumption: The technology can
be embedded in any resource-limited device
(mobile equipment, PDAs, etc).
Noise robust: Voice recognition is combined with
Denoiser Clear Voice, powerful noise filtering
technology, in order to get the best recognition
performance in noisy environments.

Voice Navigator: Voice Control of PC Programs
Voice Navigator is a fun application based on VoiceCom software library
that helps you control your computer using the power of your own
voice. Use your PC “hands free” without having to use your keyboard
or mouse for up to 200 commands. Voice Navigator can be used by
people of all ages in a variety of situations.
You can:
Easily open any Windows application, file, and folders purely by
voice commands.
Link quickly by voice to your favorite web pages.
“Press” frequently used hot keys by voice.
Noise Robust: Voice Navigator gives good voice recognition even
in a noisy environment.
Not dependent on any particular language or specific words.

DiVo – Voice command recognition
DSP board
Voice command recognition on a single DSP board to
control by voice applications such as:
Industrial equipment: Voice control raises the
efficiency of machines operation.
Home appliances: Remote control by voice
Car equipment: Drive safely talking to navigation
and other on-board systems.
Electronic toys: kids will enjoy speaking to their
favorite toys.
Car Hands-free kits with such features as:
Fully voice-controlled interface with voice responses.
Voice dialing.
Voice phonebook.
Powerful filtering of car noise.
Far-end echo cancellation, full duplex.
DiVo is available for TI TMS C54xx, Hyperstone E116, 32 platforms.

3

S P E A K E R V E R I F I C AT I O N

Voice Key SDK
Biometric secure speaker verification/authentication
technology. Voice Key verifies a user by a voice
password. Cost effective access restriction for:

















4

Web-based applications: Small size of the engine
allows much more reliable client side verification.
PDAs: Faster and easier to input a voice password
than type it on the small inconvenient keyboard.
On-line information services: User doesn’t need
to remember long alphabetic passwords, using a
natural language pass phrase instead.

Voice Key SDK can be easily combined with other security
technologies, including biometrics.
Only an authorized person says the password in exactly the
right way.
Voice password cannot be stolen or imitated by a possible intruder.
Language-independent.
Safe, reliable and secure voice verification technology with Minimal False Acceptance Rate FAR=0.01 %; Minimal False Rejection Rate FRR= 0.28%; Equal Error Rate EER=1.2%.
Very noise robust. STC speaker verification technology is based
on voice features different from those used by its competitors.

M U S I Voice
C S P Command
OTTING

Recognition

Music Spotting SDK
Real-time identifying of music samples in a
sound stream for:
Advertising Agencies: Checking how many
times and when exactly a promotion jingle
plays back on a number of broadcasting
channels.
Broadcasting & Music Copyright
Organizations: Checking how many times
and when exactly a music piece plays back
on a number of broadcasting channels.









Music Spotting creates music sample templates, puts them into a database and then
searches the inputted sound stream and logs the occurrence of the same music
samples.
Real-time processing.
Low resource consumption. Can be used on common personal computers (PC).
Easy adaptation to any specific application. Depending on the application, speed and
false acceptance/false rejection balance can be adjusted.
High accuracy and recognition stability with different kinds of music and jingles.

5

Speech Technology Center
www.speechpro.com
Headquarters: 4 Krasutskogo, St.Petersburg 196084, Russia
Phone: +7 812 3310665, +7 812 3258848, Fax: +7 812 3279297
e-mail: info@speechpro.com
European office: Speech Tech GmbH
Saargemünder str. 211, Saarbrücken 66119, Germany
Phone: +49 681 9655709, Voicemail/fax: +49 69 25577077
e-mail: info@speech-tek.com
www.speech-tek.com

Document Path: ["1148-speech-technology-center-catalogue.pdf"]

e-Highlighter

Click to send permalink to address bar, or right-click to copy permalink.

Un-highlight all Un-highlight selectionu Highlight selectionh