Full duplex speaker-phone providing increased loop stability

Details Full duplex speaker-phone providing increased loop stability Full duplex speaker-phone providing increased loop stability

2000-08-02

2004-09-21

Kuntz, Curtis (Department: 2643)

Telephonic communications

Substation or terminal circuitry

For loudspeaking terminal

C379S387020, C379S388050, C379S390010, C379S390020, C379S420020

Reexamination Certificate

active

06795547

ABSTRACT:

BACKGROUND
1. Technical Field
The present invention relates generally to speaker-phones; and, more particularly, it relates to full duplex speaker-phone technology targeted to reduce loop instability.
2. Related Art
Conventional speaker-phone technology suffers tremendously from instability in the loop formed between the two ends of the speaker-phone. For example, in a fully duplex speaker-phone, whether analog or digital, the loop stability is not guaranteed even though the loop estimation is typically well below the predetermined limit for the loop. One reason that generates this deleterious effect of loop instability within conventional speaker-phones is the low path gain estimations and high gain estimations of any echo cancellers employed within the speaker-phone loop. For example, because the energy of speech is typically concentrated within the relatively low frequency range (i.e., below 1 kHz), the gain estimations are commonly only valid within that spectral range where the energy of the speech is in fact concentrated. In the specific case where the energy spectral density of the speech provided to the speaker-phone is in fact concentrated below the frequency range of approximately 1 kHz, then the gain estimations are valid and operable. However, when the characteristics of the speech signal are such that the energy spectral density is significantly contained above this conventional cutoff of approximately 1 kHz, then the gain estimations are typically invalid leading to undesirable reduced quality in operation of the speaker-phone.
In addition, the mathematical methods employed in the echo cancellers converge much quicker in the lower frequency ranges than in the higher frequency ranges. This is largely because there is very little excitation within the higher frequency ranges of the speech signal. Other portions of conventional speaker-phones suffer from the lack of excitation within the higher frequency ranges of the speech signal. For example, an acoustic echo canceller will typically be unable to track the changes in the acoustic environment at the ends of the speaker-phone that are often unavoidable and inevitable. This inability is largely traceable to the lack of excitation within the higher frequency portion of the speech signal. Perceptually, the effects of these combined deficiencies within conventional speaker-phones will result in instability in the loop of the speaker-phone, and an undesirable effect is fact that there is typically an increased audibility of echoes, specifically at these higher frequencies.
Many conventional methods try to employ a certain degree of attenuation of the various paths of the speaker-phone during the different modes of operation. For example, one conventional method increased the attenuation in the non-active path during single talk by a predetermined amount, such as 20 dB. In this mode, the speaker-phone is practically running at half-duplex. During a double talk mode of operation, the attenuation of the speaker-phone is divided in a certain way between the two paths of the speaker-phone. Typically, the attenuation is only around 6 dB in the receive path and the remainder of the attenuation is applied to the transmit path. However, these conventional methods suffer greatly in overall perceptual quality, in that, double talk detectors are typically highly unreliable and the resultant speech is often choppy or not even heard at all. For the two-band solutions that are used, an echo canceller is typically run in the lower band, and the upper band is run at half-duplex. In addition, for the multi-band (or polyphase) solutions that are used, a large lag is introduced into the system thereby compromising the perceptual quality of the speech signals within the speaker-phone.
Further limitations and disadvantages of conventional and traditional systems will become apparent to one of skill in the art through comparison of such systems with the present invention as set forth in the remainder of the present application with reference to the drawings.
SUMMARY OF THE INVENTION
Various aspects of the present invention can be found in a speaker-phone that performs adaptive filtering on a speech signal. The speech signal is partitioned into a first speech signal corresponding to a first user and a second speech signal corresponding to a second user. The speaker-phone itself includes a main control circuitry, a transmit circuitry, a receive circuitry, and a mode detection circuitry. The transmit circuitry includes a first adaptive frequency dependent attenuation circuitry to performs adaptive filtering on the first speech signal using a first attenuation parameter. The receive circuitry includes a second adaptive frequency dependent attenuation circuitry to perform adaptive filtering on the second speech using a second attenuation parameter. The mode detection circuitry detects an operation mode of the speaker-phone from among a number of operation modes. The number of operation modes includes a receive mode, a double talk mode, a transmit mode, and a silence mode. The main control circuitry operates cooperatively with the mode detection circuitry to adjust the first attenuation parameter and the second attenuation parameter based on the operation mode of the speaker-phone.
In certain embodiments of the invention, the speaker-phone also includes a real time modification circuitry that operates cooperatively with the main control circuitry and the mode detection circuitry to adjust the first attenuation parameter and the second attenuation parameter in real time. The real time modification circuitry includes a number of sliding coefficient sets. The first attenuation parameter and the second attenuation parameter are selected from the number of sliding coefficient sets. The selections of the first attenuation parameter and the second attenuation parameter are based, at least in part, on a characteristic of the first speech signal and a characteristic of the second speech signal. The first adaptive frequency dependent attenuation circuitry adjusts the first attenuation parameter to a minimum predetermined value and the second adaptive frequency dependent attenuation circuitry adjusts the second attenuation parameter to a maximum predetermined value.
A sum of the first attenuation parameter and the second attenuation parameter exceeds a predetermined threshold. The speaker-phone also includes a real time modification circuitry that is communicatively coupled to the first adaptive frequency dependent attenuation circuitry and the second adaptive frequency dependent attenuation circuitry. The real time modification circuitry includes a plurality of sliding coefficient sets. The speaker-phone also includes a programmable sliding low pass filter that is communicatively coupled to the first adaptive frequency dependent attenuation circuitry and the second adaptive frequency dependent attenuation circuitry such that the first attenuation parameter and the second attenuation parameter are selected from the number of sliding coefficient sets. The real time modification circuitry exchanges a third attenuation parameter for the first attenuation parameter and exchanges a fourth attenuation parameter for the second attenuation parameter in real time. The speaker-phone also includes a double talk detection circuitry that is operable to detect the double talk mode.
Other aspects of the present invention can be found in a speaker-phone that performs adaptive filtering on a speech signal. The speaker-phone includes a main control circuitry, and an adaptive frequency dependent attenuation circuitry communicatively coupled to the main control circuitry that operatives cooperatively with the main control circuitry to perform adaptive filtering on the speech signal using an attenuation parameter.
In certain embodiments of the invention, the speaker-phone includes a real time modification circuitry that contains a number of sliding filter coefficient sets, and the main control circuitry is operable to select at least one of the number of sliding filter coefficie

Affiliated with

Also associated with

Rating

Full duplex speaker-phone providing increased loop stability does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Full duplex speaker-phone providing increased loop stability, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Full duplex speaker-phone providing increased loop stability will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3244366