Browse Prior Art Database

Tone Detection Process

IP.com Disclosure Number: IPCOM000116064D
Original Publication Date: 1995-Jul-01
Included in the Prior Art Database: 2005-Mar-30
Document File: 6 page(s) / 222K

Publishing Venue

IBM

Related People

Shvadron, U: AUTHOR

Abstract

The detection of a tone in a signal is of importance in many DSP applications which require a computerized control of a telephone line. Signals such as dial tone, busy tone, DTMF tone, etc. can be detected using digital processing approach in order to enable the computerized control. This enables applications such as automatic dialling of computerized modem and fax, and application of remote access to a computer via a touch-tone phone.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 25% of the total text.

Tone Detection Process

      The detection of a tone in a signal is of importance in many
DSP applications which require a computerized control of a telephone
line.  Signals such as dial tone, busy tone, DTMF tone, etc. can be
detected using digital processing approach in order to enable the
computerized control.  This enables applications such as automatic
dialling of computerized modem and fax, and application of remote
access to a computer via a touch-tone phone.

      The tones used in a telephone environment are well defined
through international standards such as the CCITT recommendations,
the ETSI, and others.  Usually each country would define its own
specifications for telephony.  It is of great importance to use a
robust algorithm which will be able to detect any telephone tone in
any country based on the known specifications.

      Current analog hardware techniques usually utilize several
analog filters such as band-elimination filters and notch filters for
each frequency.  This hardware also contains amplifiers, AGC circuits
and decision logic.  This technique has three main disadvantages:
  1.  Each tone must have its own circuitry.
  2.  Each telephone line should be analyzed separately.
  3.  Frequency shifts cannot be measured within the required
accuracy.

      Many digital schemes have been proposed in order to overcome
the disadvantages of analog hardware.  In most cases the method used
for tone existence determination is a notch filtering process.  This
process consumes a lot of CPU power since the filers are sometimes
very narrow in the frequency domain hence very long in the time
domain.  This process is not reliable in distinguishing a tone from
human voice and also it is almost impossible to determine small
frequency shifts in the input tone.

      Another approach uses the well known FFT algorithm to analyze
the frequency domain of the signal.  The drawback is that for high
resolution analysis long windows should be used which again requires
many computations.

The process described below provides the following advantages:
  1.  The tone specifications (frequency range and volume, etc.) are
       the only parameters needed for the detection process.
  2.  The amount of calculations needed is in the order of the window
       length (M) itself which might be very short while maintaining
       infinite resolution of the analysis.
  3.  It is easy to determine very small shifts in the frequency of
the
       analyzed tone.
  4.  It is possible to distinguish between a pure tone and human
       voice.

      A tone is defined by the range of frequencies where it should
be detected and by a range of frequencies in which if a tone exists
it should not be detected as a valid tone.  The process defines a
window which permits accurate tone detection within the valid range.

      A multiplication of a window by a pure tone (sine wave) in...