Spatial disassembly processor

• US 7,630,500 B1
• Filed: 04/15/1994
• Issued: 12/08/2009
• Est. Priority Date: 04/15/1994
• Status: Active Grant

• Current Assignee: Bose Corporation
• Original Assignee: Bose Corporation
• Inventors: Beckman, Paul E., Arnold, Finn A.
• Primary Examiner(s): Lee; Ping
• Application Number:
  US08/228,125
• Time in Patent Office: 5,716 Days
• Field of Search: 381/1, 381/2, 381/17, 381/18, 381 19- 23, 381/27, 375/122
• US Class Current: 381/18
• CPC Class Codes: H04S 2400/05 Generation or adaptation of centre channel in multi-channel audio systems
• CPC Class Codes: H04S 7/30 Control circuits for electronic adaptation of the sound field

Abstract

A method of disassembling a pair of input signals L(t) and R(t) to form subband representations of N output channel signals o₁(t), o₂(t), . . . , o_N(t), wherein t is time. The method includes the steps of generating a subband representation of the signal L(t) containing a plurality of subband components L_k(t) where k is an integer ranging from 1 to M; generating a subband representation of the signal R(t) containing a plurality of subband components R_k(t); and constructing the subband representation for each of the plurality of output channel signals, each of those subband representations containing a plurality of subband components o_j,k(t), wherein o_j,k(t) represents the k^th subband of the j^th output channel signal and is constructed by combining components of the input signals L(t) and R(t) according to an output construction rule: o_j,k(t)=f(L_k(t),R_k(t)) for k=1, 2, . . . , M and j=1, 2, . . . , N.

Claims

1. A method of processing a pair of input signals L(t) and R(t) representing left and right channels of a stereo audio signal, characterized by a predetermined spectral balance and predetermined spatial balance to form subband signals representative of N output channel signals o₁(t), o₂(t), . . . , o_n(t), wherein N>

2 and t is time, the output channel signals to be reproduced over spatially separated loudspeakers, said method comprising;

generating a first subband signal representation of the signal L(t), said first subband signal representation containing a plurality of first subband frequency sample components L_k(t) where k is an integer ranging from 1 to M;

generating a second subband signal representation of the signal R(t), said second subband signal representation containing a plurality of second subband frequency sample components R_k(t); and

combining said frequency sample components of the input signals L(t) and R(t) according to an output construction rule o_j,k(t)=f(L_k(t),R_k(t)) for k=1, 2, . . . , M and j=1, 2, . . . , N to provide the output subband signal representation for each of said plurality of output channel signals, each of said output subband signal representations containing a plurality of output subband signal components o_j,k(t), wherein o_j,k(t) represents the k^th subband output signal component of the j^th output channel signal,

wherein the output construction rule establishes the following relationship for at least some of the subband signal components L_k(t) and R_k(t) and output subband signal components o_j,k(t)

$<br><br>L_k<br><br>\left(t\right)<br><br>+<br><br>R_k<br><br>\left(t\right)<br><br>=\underset{j=1}{\overset{N}{\sum <br><br>}}<br><br><br><br><br><br>o_{j,k}<br><br>\left(t\right)<br><br>$

and reproducing the N output channel signals with N output speakers while preserving said predetermined spectral balance and said predetermined spatial balance of said input signals.

2. The method of claim 1 further comprising generating time-domain signals representative of the output channel signals, o₁(t), o₂(t), . . . , o_n(t), from their respective output subband signal representations.

3. The method of claim 1 wherein the output construction rule is subband specific, i.e., o_j,k(t)=f_j(L_k(t),R_k(t)) for k=1, 2, . . . , M and j=1, 2, . . . , N.

4. The method of claim 2 further comprising additionally processing one or more of the time-domain signals.

5. The method of claim 4 wherein the step of additionally processing comprises combining the N output channel signals to form two channel signals for playback over two loudspeakers.

6. The method of claim 4 wherein the step of additionally processing comprises combining the N output channel signals to form a single channel signal for playback over a single loudspeaker.

7. The method of claim 3 wherein the construction rule is also output channel-specific, i.e., o_j,k(t)=f_j,k(L_k(t),R_k(t)) for k=1, 2, . . . , M and j=1, 2, . . . , N.

8. The method of claim 1 wherein the output construction rule is further defined such that when the output channel signals o₁(t), o₂(t), . . . , o_n(t) are reproduced over N spatially separated loudspeakers, a perceived loudness of the kth subband signal component of the output channel signals is the same as a perceived loudness of the k^th subband signal representations of the left and right input channel signals L(t) and R(t) respectively when the left and right input channel signals are reproduced over a pair of spatially separated loudspeakers.

9. The method of claim 1 wherein the output construction rule also establishes the following relationship for at least some of the subband signal components L_k(t) and R_k(t) and output subband signal components o_j,k(t):

${<br><br>L_k<br><br>\left(t\right)<br><br>}^2+{<br><br>R_k<br><br>\left(t\right)<br><br>}^2=\underset{j=1}{\overset{N}{\sum <br><br>}}<br><br><br><br>{<br><br>o_{j,k}<br><br>\left(t\right)<br><br>}^2.$

10. The method of claim 1 wherein the output construction rule is further defined such that when the output channel signals o₁(t), o₂(t), . . . , o_n(t) are reproduced over N spatially separated loudspeakers, a perceived location of the kth subband output signal component of the output channel signals is the same as the localized direction of the kth subband signal representation of the left and right input signals L(t) and R(t) respectively when the left and right input signals L(t) and R(t) respectively are reproduced over a pair of spatially separated loudspeakers.

11. The method of claim 1 wherein the pair of input signals L(t) and R(t) are processed in accordance with a short-term Fourier transform to provide said first and second subband signal representations.

12. The method of claim 1 wherein the pair of input signals L(t) and R(t) are processed in accordance with a discrete cosine transform to provide said first and second subband signal representations.

13. The method of claim 1 wherein the pair of input signals L(t) and R(t) are processed in accordance with a Hartley transform to provide said first and second subband signal representations.

14. The method of claim 1 wherein the input signals L(t) and R(t) are processed with an array of bandpass filters to provide said first and second subband signal representations.

15. The method of claim 1 wherein the input signals L(t) and R(t) are processed in accordance with a wavelet decomposition.

16. The method of claim 1 wherein the input signals L(t) and R(t) are processed in accordance with a filterbank decomposition to provide said first and second subband signal representations.

17. The method of claim 1 wherein the step of processing of the L(t) input signal comprises:

sampling the L(t) input signal to provide a sequence of L(t) input signal samples;

grouping the latter samples into overlapping blocks;

applying a window function signal to each of said overlapping blocks to provide a corresponding plurality of windowed blocks; and

processing each windowed block in accordance with a fast Fourier transform to provide the first subband signal representation of the L(t) input signal.

18. The method of claim 17 wherein the blocks overlap by a factor of substantially ½

.

19. The method of claim 17 wherein each block contains about 2048 samples.

20. The method of claim 17 wherein the window function signal is representative of a raised cosine function.

21. The method of claim 17 and further comprising zero padding each block before processing each windowed block in accordance with a fast Fourier transform.

22. The method of claim 17 further comprising processing said subband signals representative of said N output channel signals to provide time-domain representations of the output channel signals, o₁(t), o₂(t), . . . , o_n(t).

23. The method of claim 22 and further comprising processing the first subband signal representation in accordance with an inverse short-term Fourier transform to provide time-domain representations of the output channel signals, o₁(t), o₂(t), . . . , o_n(t).

24. The method of claim 1 wherein the subband-specific construction rule is chosen so that the subband representation of the output signal o(t) is the correlated portion of the input signals L(t) and R(t).

25. The method of claim 1 wherein said construction rule is of the form o_k(t)=α

_kL_k(t)+γ

_kR_k(t) and wherein α

_k and γ

_k are weighting factors, the values of which depend upon k.

26. The method of claim 1 wherein said construction rule is of the form o_k(t)=α

_kL_k(t)+γ

_kR_k(t) and wherein α

_k and γ

_k are weighting factors, the values of which depend upon the values of L_k(t) and R_k(t).

27. The method of claim 1 wherein said construction rule is of the form o_k(t)=α

_kL_k(t)+γ

_k(t) and wherein α

_k=γ

_k.

28. A spatial disassembly system comprising,

first and second input terminals for receiving first and second input signals L(t) and R(t) representing left and right channels of a stereo audio signal, respectively characterized by predetermined spectral balance and predetermined spatial balance,

a spatial disassembly processor having a plurality of N outputs greater than two, constructed and arranged to

disassemble signals on said first and second inputs including subdividing the signals on said first and second inputs into a plurality of M frequency sample subbands L_k(t) and R_k(t) where k is an integer ranging from 1 to M, and

provide a corresponding plurality of output signals o₁(t), o₂(t), . . . , o_n(t), on said plurality of outputs derived from the frequency sample subbands of the disassembled signals according to an output construction rule o_j,k(t)=f(L_k(t),R_k(t)) for k=1, 2, . . . , M and j=1, 2, . . . , N,

each of said output subband signal representations containing a plurality of output subband signal components o_j,k(t), wherein o_j,k(t) represents the k^th subband output signal component of the j^th output channel signal,

wherein the output construction rule establishes the following relationship for at least some of the subband signal components L_k(t) and R_k(t) and output subband signal components o_j,k(t):

$<br><br>L_k<br><br>\left(t\right)<br><br>+<br><br>R_k<br><br>\left(t\right)<br><br>=\underset{j=1}{\overset{N}{\sum <br><br>}}<br><br><br><br><br><br>o_{j,k}<br><br>\left(t\right)<br><br>$

and

a corresponding plurality of electroacoustical transducers coupled to a respective one of said plurality of outputs for creating a sound field representative of the first and second input signals on said first and second input terminals preserving said predetermined spectral balance and said predetermined spatial balance of the first and second input signals.

29. Apparatus in accordance with claim 28 wherein said spatial disassembler includes a frequency domain spatial disassembly processor.

30. Apparatus in accordance with claim 29 wherein said spatial disassembler includes a fast Fourier transform signal processor in a signal path between an input terminal and said frequency domain spatial disassembly processor.

31. Apparatus in accordance with claim 30 and further comprising,

a decomposer coupled to an input terminal for decomposing the input signal on said input terminal into overlapping blocks of sample signals, and

a first window processor in the signal path between said fast Fourier transform processor and said decomposer for processing the overlapping blocks of sampled signals with a window function.

32. Apparatus in accordance with claim 31 and further comprising,

an inverse fast Fourier transform processor in the signal path between said frequency domain spatial disassembly processor and an output.

33. Apparatus m accordance with claim 32 and further comprising,

a second window processor in the path between said inverse fast Fourier transform processor and the latter output for processing the output of the inverse fast Fourier transform processor in accordance with a window function,

a block overlapper in the path between the second window function processor and the latter output for overlapping signals provided by the second window function processor and combining the overlapped blocks to provide an output signal to an associated output terminal.

34. A method of processing a pair of input signals L(t) and R(t) representing left and right channels of a stereo audio signal, characterized by a predetermined spectral balance and predetermined spatial balance to form subband signals representative of N output channel signals o₁(t), o₂(t), . . . , o_n(t), wherein N>

2 and t is time, the output channel signals to be reproduced over spatially separated loudspeakers, said method comprising;

generating a first subband signal representation of the signal L(t), said first subband signal representation containing a plurality of first subband frequency sample components L_k(t) where k is an integer ranging from 1 to M;

generating a second subband signal representation of the signal R(t), said second subband signal representation containing a plurality of second subband frequency sample components R_k(t); and

combining said frequency sample components of the input signals L(t) and R(t) according to an output construction rule o_j,k(t)=f(L_k(t),R_k(t)) for k=1, 2, . . . , M and j=1, 2, . . . , N to provide the output subband signal representation for each of said plurality of output channel signals, each of said output subband signal representations containing a plurality of output subband signal components o_j,k(t), wherein o_j,k(t) represents the k^th subband output signal component of the j^th output channel signal,

wherein the output construction rule establishes the following relationship for at least some of the subband signal components L_k(t) and R_k(t) and output subband signal components o_j,k(t);

${<br><br>L_k<br><br>\left(t\right)<br><br>}^2+{<br><br>R_k<br><br>\left(t\right)<br><br>}^2=\underset{j=1}{\overset{N}{\sum <br><br>}}<br><br><br><br>{<br><br>o_{j,k}<br><br>\left(t\right)<br><br>}^2$

and reproducing the N output channel signals with N output speakers while preserving said predetermined spectral balance and said predetermined spatial balance of said input signals,

wherein the output construction rule is subband specific, i.e., o_j,k(t)=f_j(L_k(t),(R_k(t)) for k=1, 2, . . . , M with at least two of the subbands having different steering algorithms.