Method and apparatus for the compression, processing and spectral resolution of electromagnetic and acoustic signals

• US 5,570,305 A
• Filed: 12/22/1993
• Issued: 10/29/1996
• Est. Priority Date: 10/08/1993
• Status: Expired due to Term

• Current Assignee: Quarterhill Inc. (f/k/a Wi-LAN Inc.)
• Original Assignee: Hatim Zaghloul, Michel Fattouche
• Inventors: Fattouche, Michel, Zaghloul, Hatim
• Primary Examiner(s): Envall, Jr., Roy N.
• Assistant Examiner(s): MOISE, EMMANUEL LIONEL
• Application Number:
  US08/173,627
• Time in Patent Office: 1,042 Days
• Field of Search: 364/485, 364/482, 364/484, 364/487, 364/553, 364/572, 364/576, 364/580, 364/726, 364/829, 364/715.02
• US Class Current: 708/203
• CPC Class Codes: H03H 17/02 Frequency selective networks digital computers for complex mathematical operations G06F17/10

Abstract

A signal that may be represented by a band limited periodic function m(t) is transformed into its z-domain zeros using characteristics of the signal in the neighbourhood of fades of the signal, namely a bounds of the fades and the depths of the fades. The z-domain zeros are the zeros of the function in the complex plane, and represent local minimums of the function. The z-domain zeros are used to estimate the discrete spectrum of m(t), and the differential phase of m(t) is derived from the envelope of m(t) using the concept of the z-domain zeros. Also, speech is compressed based on the z-domain zeros of a signal m(t) that represents a speech signal, and knowledge of the location of fades is used in interference reduction at single moving antennas by applying a spectral resolution technique during a fade and applying notch filters to the signal to resolve the signal components.

Claims

1. A method of transforming an input signal m(t) which is describable by a function M(z) having one or more fades corresponding to one or more complex zeroes z_k, for k=1, . . . N where N is the number of fades, the input signal m(t) having values during an interval between 0 and T, into a sequence of data points, the method comprising the steps of:

acquiring the input signal;

low pass equivalent filtering the input signal;

processing the filtered input signal in a signal processor to determine, for each fade, the location of the fade; and

constructing a compressed signal at least partially formed of data points corresponding to the location of each fade.

2. The method of claim 1 in which processing the low pass equivalent of the input signal includes determining, for each fade, the depth of the fade.

3. The method of claim 2 further including storing the compressed signal as depth and location differentials.

4. The method of claim 1 or 2 in which determining the location of each fade includes:

differentiating the logarithm of the envelope of the input signal m(t) for t between 0 and T; and

dividing the interval between 0 and T into subintervals, where each subinterval has a lower bound t_k,1 and an upper bound t_k,2, and each subinterval includes a fade.

5. The method of claim 4 in which:

the lower bound t_k,1 of each subinterval corresponds to the kth local minimum of the derivative of the logarithm of the envelope of the signal m(t);

the upper bound t_k,2 of each subinterval corresponds to the kth local maximum of the derivative of the logarithm of the envelope of the signal m(t); and

the product of (1) the kth local minimum of the derivative of the logarithm of the envelope of the signal m(t) and (2) the kth local maximum of the derivative of the logarithm of the envelope of the signal m(t) is negative.

6. The method of claim 5 in which the location t_k,0 of the kth fade for k=1, . . . , N is estimated as

7. The method of claim 6 in which the depth of the kth fade for k=1, . . . , N is estimated as

8. The method of claim 2 in which the depth and location of each fade are determined by:

taking a Fourier transform of the signal to produce data points corresponding to the Fourier coefficients of the signal;

forming a polynomial using the Fourier coefficients of the signal;

factorizing the polynomial to produce the roots of the polynomial; and

deriving the location and the depth of each fade from the roots of the polynomial.

9. The method of claims 1 or 2 further including, after acquiring the input signal:

forming a derivative signal corresponding to a derivative of the low pass equivalent of the input signal.

10. The method of claim 9 further including:

constructing an expanded polynomial whose roots correspond to the locations of the fades of the input signal; and

estimating the spectrum of the input signal as the coefficients of the expanded polynomial.

11. The method of claim 10 in which the expanded polynomial is constructed in a recursive manner continuously while the input signal is being acquired.

12. The method of claims 1 or 2 in which the input signal is a speech signal and further including:

quantizing the compressed signal and transmitting the quantized compressed signal.

13. The method of claims 1 or 2 in which the input signal is a speech signal having voiced and unvoiced segments and further including:

quantizing segments of the compressed signal according to whether the segment is voiced or unvoiced, and transmitting the quantized compressed signal.

14. The method of claim 1 in which the input signal is formed by at least first and second sinusoids of different frequencies, and further including resolving the first and second sinusoids of different frequencies by applying a spectral estimation technique to the input signal in the neighbourhood of one of the N fades.

15. The method of claim 14 further including:

applying a first notch filter centered on the first frequency to m(t) to produce a signal m(t)₁ ;

applying a second notch filter centered on the second frequency to m(t) to produce a signal m(t)₂ ; and

combining the signals m(t)₁ and m(t)₂.

16. The method of claim 1 further including further processing the compressed signal to calculate the difference in location of successive fades, whereby a phase differential of the input signal may be estimated.

17. The method of claim 1 in which the input signal is a speech signal.

18. The method of claim 1 in which the input signal is data encoded in an electromagnetic signal and further including modulating a carrier signal with the compressed signal for transmission of the carrier signal over a communications channel.

19. The method of claim 1 in which the input signal has an envelope with inphase and quadrature components and acquiring the input signal includes sampling the inphase and quadrature components of the envelope of the input signal.

20. The method of claim 1 in which the input signal is formed by at least first and second replicas of the same signal, and further including resolving the first and second replicas of the same signal by applying a spectral estimation technique to the input signal.

30. The apparatus of claim 18 in which the processor includes means to calculate the difference in locations of successive fades.

21. Apparatus for processing an input signal comprising:

input signal acquisition means having a first output signal;

a low pass equivalent filter connected to receive the output from the input signal acquisition means and having filtered output; and

processing means connected to receive the filtered output for locating the fades of the input signal and constructing a compressed signal comprising the locations of the fades.

22. The apparatus of claim 21 in which the processor includes means to quantize the compressed signal.

23. The apparatus of claim 22 further including a transmitter connected to the processing means.

24. The apparatus of claim 23 further including:

an encoder connected to receive the compressed signal for encoding a carrier signal for transmission.

25. The apparatus of claim 24 in which the input signal is a speech signal and further including a transducer having electrical output corresponding to the speech signal.

26. The apparatus of claim 21 in which the processor includes means to determine the depth of each fade.

27. The apparatus of claims 21 or 26 in which the processor includes means to differentiate the first output signal.

28. The apparatus of claim 27 further including means to estimate the spectrum of the input signal.

29. The apparatus of claim 21 in which the input signal is composed of first and second signals having first and second distinct frequencies and further including:

first and second notch filters having notches at the first and second distinct frequencies and each being connected to receive the compressed signal; and

means to combine outputs from the first and second notch filters.

31. The apparatus of claim 21 in which the input signal is a speech signal and the signal acquisition means includes a transducer having electrical output corresponding to the speech signal.

32. The apparatus of claim 21 in which the input signal has an envelope with inphase and quadrature components and the input signal acquisition means includes a sampler having output corresponding to inphase and quadrature components of the envelope of the input signal.

33. The apparatus of claim 21 in which the input signal acquisition means includes a band pass filter.

34. The apparatus of claim 21 in which the input signal is composed of first and second delayed replicas of a transmitted signal and further including:

a spectral estimator to resolve the delay of each replica; and

means to regenerate the transmitted signal from the delayed replicas.

35. A vocoder for encoding a speech signal for transmission, the vocoder comprising:

a first transducer having electrical output corresponding to the speech signal;

a bandpass filter connected to receive electrical output from the transducer and having filtered output;

processing means connected to receive the filtered output from the bandpass filter and having compressed signal output at least in part corresponding to the locations of the fades of the input signal; and

an encoder for encoding a carrier signal with the compressed signal output.

36. The vocoder of claim 35 in which the processing means includes an analog to digital converter for digitizing the filtered output from the bandpass filter.

37. The vocoder of claim 36 in which the processing means includes a segmentizer connected to receive digitized signals from the analog to digital converter and having output corresponding to segments of digitized speech.

38. The vocoder of claim 35 in which the processing means includes means to quantize the compressed signal output.

39. The vocoder of claim 35 further including:

a decoder for decoding the carrier signal to obtain the compressed signal;

a deprocessor having output corresponding to a reconstructed speech signal reconstructed from the compressed signal; and

a second transducer for producing acoustic signals from the reconstructed speech signal.

40. A system for reducing noise in an input electromagnetic signal having an envelope, the system comprising:

signal acquisition means having electrical signal output corresponding to an estimate of the envelope of the input electromagnetic signal;

a bandpass filter connected to receive the electrical signal output from the signal acquisition means and having filtered output;

a phase differential estimator connected to receive the filtered output from the bandpass filter and having output comprising phase differentials estimated using the locations of the fades; and

an encoder having as input both the phase differentials and the filtered output, and having encoded output for transmission.