Telephonic communications – Supervisory or control line signaling – Signal receiver
Reexamination Certificate
2001-02-02
2004-02-17
Mei, Xu (Department: 2644)
Telephonic communications
Supervisory or control line signaling
Signal receiver
C379S215010
Reexamination Certificate
active
06694011
ABSTRACT:
BACKGROUND OF MY INVENTION
1. Field of the Invention
My invention relates generally to the detection of voiceband signaling tones sent by a network component to a subscriber device or sent by a subscriber device to a network component. More particularly, my invention relates to methods and apparatus for reliably detecting voiceband-signaling tones in the presence of near-end and far-end speech.
2. Description of the Background
The Public Switched Telephone Network (PSTN) has long used combinations of discrete voiceband frequencies to carry signaling tones between an analog subscriber device and a serving switch. A common example is touch-tone-dialing where signaling tones are sent from a subscriber device to a tone detector on the serving switch to initiate a phone conversation. Other services, such as Calling Identity Delivery on Call Waiting (CIDCW), Analog Display Services Interface (ADSI Services), and Calling Number Delivery are initiated by a serving switch or service node sending voiceband signaling tones to a detector on a subscriber device.
Voiceband signaling tones, whether initiated by a serving switch, a service node, or a subscriber device, are best transmitted under controlled conditions where extraneous “noise,” such as near-end and far-end speech, music, etc. (hereinafter collectively referred to as speech) is not present and therefore cannot interfere with the recognition of these tones. For example, a subscriber's handset is muted while entering touch-tone digits, and services such as Calling Number Delivery occur while the subscriber device is in the on-hook state.
However, an increasing number of applications have emerged where voiceband signaling needs to occur under uncontrolled noisy conditions where near-end and far-end speech can mask a signaling tone or imitate a tone. As a result, the services these tones are intended to activate either fail to activate or are falsely activated. Such service reliability problems directly impact customer satisfaction and therefore service marketability.
For example, CIDCW and ADSI services require tone detection at a subscriber device and can occur asynchronously at any time during a call. These services are initiated by a serving switch/service node sending a CPE Altering Signal (CAS signal) to a subscriber device thereby triggering the device to temporarily exit voice operation and enter data mode operation. Because these services can be activated any time during a call, the CAS detector memo within the subscriber device needs to remain active for the duration of the call and must successfully discern a valid CAS signal from any speech present on the interface emanating from the near-end or far-end. Similar issues exist for services such as voice-mail, where tone detection at a network device is required.
The imitating and masking of signaling tones is formally referred to as “Talkoff” and “Talkdown”. Talkoff occurs when near-end or far-end speech erroneously triggers the signaling tone detection system (i.e., the tone detector accepts an imitation signal). Talkdown occurs when a signaling tone is sent but the signaling tone detection system fails to recognize the tone because it is masked by near-end or far-end speech (Note that for the remainder of this discussion I assume that far-end talkdown is not an issue because the device sending the tone typically mutes the far-end prior to transmission.).
Talkoff and talkdown are difficult to mitigate because the two issues simultaneously exist in many environments and they have an inverse relationship. To successfully deal with talkoff, a tone detection system must resist imitation signaling tones produced by speech by rejecting “non-pure” signals. To successfully deal with talkdown, a tone detection system must recognize weak signaling tones in the presence of speech and therefore accept “non-pure” signals as valid tones. As a result, good talkoff performance is usually achieved by sacrificing talkdown performance, and good talkdown performance is usually achieved by sacrificing talkoff performance.
The difficulty in mitigating talkoff and talkdown is illustrated by the concept of “guard-action,” which is the typical means utilized by signaling tone detection systems to deal with talkoff. Under this concept, a tone detector measures a “signal-to-guard” ratio to determine the purity of a signal. Signal-to-guard ratio is the ratio of the power present at a signaling-tone-frequency to the power of speech in one or several designated “guard-bands”. Guard-bands essentially provide a relative measure of the speech present on a line. Talkoff is circumvented by requiring a large signal-to-guard ratio because a large ratio ensures that a detected tone is relatively pure with respect to the guard-band(s) and is not a simple imitation that will trigger the detector. Talkdown, on-the-other-hand, is circumvented by requiring a small signal-to-guard ratio because a small ratio ensures that an actual tone signal is not lost among the interfering speech. As a result, good talkoff performance creates talkdown because a valid signaling tone superimposed on speech may fail to satisfy the talkoff signal-to-guard ratio requirement. Good talkdown performance creates talkoff because an imitation signaling tone created by speech can easily satisfy the lax talkdown signal-to-guard ratio requirement.
Prior-art tone detection systems have been prone to reliability problems because rather than eliminate talkoff and talkdown, they have attempted to simultaneously balance/address talkoff and talkdown. This balancing ultimately leads to tradeoffs and thereby compromised performance of the overall tone detection system and the associated services that it supports.
Tone detection within a subscriber device, such as a phone or caller-id box, will be used as the basis for the remainder of this discussion. Nonetheless, my invention is also applicable to tone detection within a network device. As a base reference,
FIG. 1
depicts a subscriber device
100
common to the prior-art and my invention. Subscriber device
100
comprises: (1) 2-wire access interface
108
, which interconnects subscriber device
100
to serving switch
120
and service node
122
; (2) hybrid
104
, which converts the 2-wire access interface to a 4-wire access interface consisting of send path
101
and receive path
102
; (3) balance network
106
, which reduces the reflection of near-end speech from send path
101
on receive path
102
; and, (4) other device components
110
, which are not related to the operation of my invention, but are common in many communication applications.
FIG. 2
illustrates a first prior-art system where tone detector
201
is placed across 2-wire access interface
108
. In this location, tone detector
201
must simultaneously deal with talkoff and talkdown and therefore balance the two issues, inherently leading to non-ideal detection performance. More important, this balance is extremely difficult and expensive to achieve because no effort has been made to reduce the strength of the near-end voice incident upon the tone detector and the detector must therefore balance two extremes—significant guard-band to resist near-end talkoff and significantly reduced guard-band to prevent near-end talkdown.
FIG. 3
illustrates a second prior-art system where tone detector
301
is placed across receive path
102
of the balance hybrid. Advantageously, this solution uses the transhybrid loss to attenuate the strength of the near-end speech incident upon the tone detector thereby addressing both near-end talkoff and near-end talkdown. However, due to the complexity of speech and the difficulty in achieving a perfect match between the balance network and the impedance presented by the local-loop interface, the hybrid does not completely eliminate the near-end speech. Specifically, because network impedance can vary widely depending on loop composition, loading, switch type, and the presence of any parallel subscriber equipment, hybrid
104
can produce less than 2 dB of transhybrid loss. As a result, a tone
Falk James W.
Farbanish Glen
Giordano Joseph
Harold Jefferey
Mei Xu
LandOfFree
Method and system for reliably detecting voiceband signaling... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for reliably detecting voiceband signaling..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for reliably detecting voiceband signaling... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3304667