Telephonic communications – Echo cancellation or suppression
Reexamination Certificate
1999-01-15
2002-11-19
Isen, Forester W. (Department: 2644)
Telephonic communications
Echo cancellation or suppression
C379S406020, C379S406040, C379S406050, C379S406060, C379S406080, C379S406090
Reexamination Certificate
active
06483915
ABSTRACT:
This application is based on and claims priority from German Patent Application No. 198 01 390.6 filed Jan. 16, 1998, which is incorporated herein by reference.
BACKGROUND OF THE INVENTION
The invention involves equipment and a procedure for the suppression of whole echoes comprised of several partial echoes in telecommunications (TC) equipment, such as end devices, transmission systems or switching equipment, with the help of adaptive FIR (=Finite Impulse Response) filters, which reproduce the whole echo and extract it from the useful signal impeded by the echo, which has been transmitted to the respective TC equipment.
Such a procedure and equipment for adaptive echo compensation is known from the German patent DE 44 30 189 A1.
With the transfer of speech signals via telecommunication lines, acoustic echoes can originate either on the “near end” from the user sending out the signal due to direct sound transfer from the loudspeaker to the microphone of the end device. The “near echo” problem is considerably intensified when several end devices are set up closely together, for example in an office or a conference room with multiple telephone connections since there is a coupling from each loudspeaker signal to each microphone. The multi-channel echo suppression procedure and the associated electronic circuit proposed in the EP 0 627 825 A2 should help in this case. In addition to the near acoustic echo, it is possible that a composite electric line echo is also created from a great variety of partial echoes due to the reflections of the speech signals sent out to positions which are varying distances away in the transmission channel, for example in the case of 2-4 wire hybrids (hybrid circuits) in the end device or in the exchanges and at network interconnections. In this case it is usually so-called “short distance echoes” which are reflected back to the speaker in a time period up to 128 ms. These short distant echoes are differentiated from long distance echoes, which come back to the speaker in a timeframe of up to 640 ms and are especially disturbing due to the large time delay since in the meantime the speaker is already considerably advanced in his speech by the time his own reflected speech signal reaches him again. Such delayed long distance echoes occur for example during intercontinental calls, or during long distance telephone calls transmitted via deep-sea cables or satellites. The network operators which maintain such telephone networks are therefore especially interested in suppressing long distance echoes with large time delays, but also short distance echoes reflected back with higher intensities as effectively as possible.
To this end, for example in EP 0 792 029 A2, echo suppression equipment with an adaptive filter is proposed, including a coarse and a fine detector for near end speech signals which are set up on both sides of the echo suppression equipment and are consequently able to monitor the incoming signal before and after the echo suppression.
An adjustment to the reflected echo value should be carried out in speech pauses on the near end.
DE 44 30 189 A1 quoted at the beginning proposes a cost-efficient procedure which can be used under various acoustic conditions, and which uses a FIR filter whose filter coefficients are determined using the NLMS algorithms. According to this method, echo suppression in a TC network with multiple parallel channels utilizes an adaptive FIR filter for each channel, which is realized in software on a digital signal processor. The adaptive FIR filter is supposed to reproduce the expected whole echo, which may be composed of several individual partial echoes and extract it from the useful signal which has been transmitted to the TC end device and is impeded by the echo.
SUMMARY OF THE INVENTION
In contrast, the purpose of the present invention is to develop the procedure and equipment of the type indicated at the beginning to the point of enabling the effective suppression of echoes, without a large amount of storage space and computational effort, even for a large number of m parallel TC channels which are independent of each other, for example as they are in exchanges.
In accordance with the present invention, this task is solved in that for m parallel, independent TC channels to be served, n homogeneous FIR filters or partial filters, each of which can reproduce a whole or partial echo, are implemented on a single ASIC (=Application Specific Integrated Circuit), whereby m, n &egr; N and especially n≧2; that the whole or partial echoes to be reproduced can be calculated in real time on the ASIC; and that there are provisions for a digital signal processor (=DSP) which controls the ASIC and can carry out the filter settings in the ASIC, especially after the coefficients and time-delays required to reproduce the echo have been calculated in the DSP.
Through the high integration density of an ASIC, the equipment in accordance with the present invention can be kept especially compact. Since the digital signal processor is quite considerably relieved through the integration of the FIR filter on the ASIC, many more parallel TC channels can be served with constant processor effort with effective echo suppression. This allows for the processing of larger time delays in the echoes (for example up to 640 ms), multiple individual echoes for each whole echo and many parallel channels (>2000) with minimal effort.
More advantageously, a number of FIR partial filters are cascaded in order to reproduce a whole echo. At present the demands of TC network operators call for three to five partial echoes to be suppressed per channel. However, this could still increase in the future.
It is especially preferred that the equipment in accordance with this invention be designed in such a way that a variable subset of FIR partial filters can be assigned to a TC channel to be served, as required, depending on the number of partial echoes as well as the size of the delay times. Through this flexible assignment of FIR filters per processed channel according to the number of individual echoes and according to the size of the delay times between the signal and echo, the ASIC can be of an extremely compact design despite the very high number of channels.
An advantageous advanced set-up distinguishes itself in that there are provisions for a macrocell for each FIR partial filter in the ASIC. This macrocell includes a shift register of suitable length for the temporary storage of the sampled speech signals, demultiplexer equipment for setting the effective shift register length, as well as adding equipment for the summation of the sample value weighted with the appropriate coefficients. These macrocells can be precisely the same construction, which simplifies the manufacture of an ASIC with many homogeneous FIR partial filters and therefore makes it less expensive.
It is especially easy to have a cascading of several macrocells that enables any choice of flexible interconnections, for example for reproducing a whole echo, and therefore an especially high capacity utilization ratio of the macrocells.
A preferred advanced set-up provides for a macrocell including one of the coefficient memories approximated by the digital signal processor, which can be directly read by a modulo N counter as well as a digital multiplier unit with an operating frequency of f
A
=N·f
S
, to which the appropriate coefficients from the coefficient memory and the appropriate sample values from the shift register are fed for multiplication, in which case the sample values with a sampling frequency of f
S
have been stored in the shift register. There are provisions for an adder which adds the product values output by the digital multiplier unit to each of the sums of the preceding product values stored in an initial temporary memory via a feedback loop and which can finally be extracted from the useful signal impeded by an echo, which has been stored in a second temporary memory as the reproduced partial echo signal and transmitted to the TC e
Mann Armin
Matt Hans Jürgen
Walker Michael
Alcatel
Isen Forester W.
Singh Ramnandan
Sughrue & Mion, PLLC
LandOfFree
Equipment and processes for echo suppression with adaptive... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Equipment and processes for echo suppression with adaptive..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Equipment and processes for echo suppression with adaptive... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2961158