Name: Clear Voice SDK

Text: CLEAR VOICE SDK
Set of 3 Software Developer Kits (SDK)
for Automatic Speech Enhancement & Noise Filtration

u
u
U

Denoiser - Automatic speech enhancement and
noise filtration
Denoiser-by-Reference - Automatic noise
filtration of stable noise
Declicker - Automatic pulse like noise filtration

Advantages
Adaptive filters of Clear Voice Package dramatically
improve communication quality in telephone, satellite
and radio channels, and can be used to increase the
accuracy of speech recognition systems.
u
u
u
u
u
u

Continuous and adaptive removal of background
noise from speech
Improves voice quality and intelligibility
Decreases subscribers' listening fatigue
Reduces time of communication due to
eliminating of double-checking
Increases voice compression systems quality
Improves performance of automatic speech
recognition systems

SDK Package
32 bit DLL (dynamic link library) for MS
Windows XP/2000/9x/ME/XP;
User’s manuals;
Sample demo software with sources in
Visual C++.
SDK for other OS and ANSI C source codes
are available by special order. Prices will be
quoted upon request.
Free demo is available at the web site.

Applications
u
u
u
u
u
u
u
u
u
u
u
u

Improvement in communication quality in Telephone,
satellite and radio channels.
Enhancement for Radio transmitting and Audio
Data
Superior compression on VoIP and PC telephony;
Call Centers;
Improving poor quality recordings of
music/speech
Sound networks;
Large scale IC for communication applications;
Voice recognition systems;
Sound recorders;
Dictation devices;
Voice logging systems;
Audio products;

This is a Denoiser sample software screenshot. In
the waveform field (lower left) there is a waveform of
the source signal (lighter), the processed darker one
(lower right). The difference being, is the filtering of
the background noise level, is clearly seen.

Functional Overview
Clear Voice processes sound signals in digital PCM 16
bit format and removes automatically additive noise
from transmitted (live) or pre-recorded signals with
minimal reduction in signal quality.
u
u
u
u
u
u
u
u

Uses only single input, for example, hands-free car
microphone
Automatically separates useful signal (like
speech) from background slow varying noise
Programmable noise reduction parameters
Up to 30 dB signal-to-noise improvement
Single tone reduction > 50 dB
White noise reduction > 15-20 dB
Hiss and rumble additional filters, frequency
response forming filter
The procedure is self-adjusting but the controls of the
Clear Voice give to the advanced user an opportunity
to choose the best degree of noise reduction and to
tune the method for a specific task.

u

Denoiser is designed specifically to cancel slow
varying in amplitude and frequency broadband noise
such as, for example, office noise, street noise, car
noise, hiss and colored white noise in sum with
narrow band tones such as hum in
communication channels, noise of home
and industrial equipment, slow music etc.

u

Declicker removes from the input signal
pulse-like interferences such as pikes,
clicks, cracks, etc. It gives very good
results from the removal of clicks from old
vinyl records and digital clicks or static
discharge cracks from DAT masters. It
can even mask long duration tape dropouts.

u

Denoiser-by-Reference enables the
user to reduce an arbitrary broadband
stable noise in audio signals (tape hiss,
surface noise of vinyl records, concert hall
acoustic noise, etc.). It is developed
mainly for the purpose of audio signal
quality improvement in audio archives
processing and CD mastering.

STC is a leading speech and signal technology
developer. It has an extensive portfolio of proprietary
know-how and a number of ready-made products in
the field of modern speech technology, especially for
speech enhancement and voice extraction. Contact
us and we will try to solve any problem in the speech
or signaling area.
u

u
u

u

u

The algorithms can be incorporated into any PC
16/32 bit Windows 95/98/NT application which
processes or uses voice or any sound signal.
STC can license the right to manufacture, use and
sell products incorporating Clear Voice algorithms.
Clear Voice may be licensed for use in software,
hardware or integrated circuits (ICs) for
communication applications. The form of the
transfer may include software for PC or DSP, DLL,
source code, hardware, devices, etc.
PC Windows version demands 17% on 133MHz
Pentium and 100 KB internal RAM. May be easily
imported to fixed point or floating point DSPs.
Also available
- Noise Cancellation DSP board
- Denoiser TI eXpress DSP compliant algorithm
- Denoiser ActiveX control

Declicker sample software screenshot. Each SDK
has two kinds of settings - universal (degrees of
noise cancellation - hard ....soft) and advanced (fine
tuning of all algorithm parameters).

Document Path: ["1123-speech-technology-center-brochure-clear.pdf"]

e-Highlighter

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

Un-highlight all Un-highlight selectionu Highlight selectionh