System for improving speech intelligibility through high frequency compression

• US 8,219,389 B2
• Filed: 12/23/2011
• Issued: 07/10/2012
• Est. Priority Date: 04/20/2005
• Status: Active Grant

• Current Assignee: Malikie Innovations Limited (Key Patent Innovations Limited)
• Original Assignee: QNX Software Systems Limited (Canada) (Blackberry Limited)
• Inventors: Hetherington, Phillip A., Li, Xueman
• Primary Examiner(s): He, Jialong
• Application Number:
  US13/336,149
• Publication Number:
  US 20120095759A1
• Time in Patent Office: 200 Days
• Field of Search: 704/201, 704/205, 704/226, 381/316, 381/321
• US Class Current: 704/201
• CPC Class Codes: G10L 21/02 Speech enhancement, e.g. noise reduction or echo cancellation reducing echo effects in line transmission systems H04B3/20; echo suppression in hands-free telephones H04M9/08
• CPC Class Codes: G10L 21/0232 Processing in the frequency domain
• CPC Class Codes: G10L 21/0264 characterised by the type of parameter measurement, e.g. correlation techniques, zero crossing techniques or predictive techniques
• CPC Class Codes: G10L 21/0364 for improving intelligibility

Abstract

A speech enhancement system that improves the intelligibility and the perceived quality of processed speech includes a frequency transformer and a spectral compressor. The frequency transformer converts speech signals from the time domain to the frequency domain. The spectral compressor compresses a pre-selected portion of the high frequency band and maps the compressed high frequency band to a lower band limited frequency range.

Claims

1. A system, comprising:

a computer processor;

a frequency transformer configured to convert a speech signal into a spectrum of frequencies; and

a spectral compressor regulated by the computer processor and coupled with the frequency transformer, where the spectral compressor is configured to define a lower cutoff frequency within a frequency passband having a passband upper frequency limit, where the spectral compressor is configured to compress a pre-selected high frequency band of the speech signal between the lower cutoff frequency and a frequency component above the passband upper frequency limit, and where the spectral compressor is configured to map the compressed high frequency band to a lower frequency range below the passband upper frequency limit in response to a determination that a signal-to-noise ratio of the speech signal in the lower frequency range before compression is less than a signal-to-noise ratio of the speech signal in the lower frequency range after compression.

2. The system of claim 1, where the spectral compressor is further configured to output the speech signal without compression of the pre-selected high frequency band in response to a determination that the signal-to-noise ratio of the speech signal in the lower frequency range before compression is higher than the signal-to-noise ratio of the speech signal in the lower frequency range after compression.

3. The system of claim 1, further comprising a gain controller configured to apply a variable gain to the compressed high frequency band based on a background noise level present in the speech signal.

4. The system of claim 3, where the gain controller is configured to select a level for the variable gain based on a slope of a noise floor present in the compressed high frequency band of the speech signal and a slope of a noise floor present in an uncompressed frequency portion of the speech signal.

5. The system of claim 3, where the gain controller is configured to select a level for the variable gain that substantially aligns a slope of a noise floor present in the compressed high frequency band with a slope of a noise floor present in an uncompressed frequency portion of the speech signal.

6. The system of claim 1, where the pre-selected high frequency band comprises a larger range of frequencies than the lower frequency range.

7. The system of claim 1, where the spectral compressor is configured to apply a non-linear compression basis function to the speech signal.

8. The system of claim 1, where the spectral compressor is configured to compress a first portion of the speech signal above the lower cutoff frequency without compression of a second portion of the speech signal below the lower cutoff frequency.

9. The system of claim 1, where the speech signal comprises a highest frequency component that is greater than a passband upper frequency limit, and where the spectral compressor is configured to compress and map at least a portion of the speech signal above the passband upper frequency limit to the lower frequency range below the passband upper frequency limit.

10. The system of claim 1, where the pre-selected high frequency band comprises a portion of the speech signal between about 2,800 Hz and a highest frequency component that is higher than 5,000 Hz, and where the spectral compressor is configured to compress and map the compressed high frequency band to the lower frequency range between about 2,800 Hz and about 3,600 Hz.

11. A method, comprising:

identifying a frequency passband having a passband upper frequency limit;

defining a lower cutoff frequency within the frequency passband;

receiving a speech signal having a frequency spectrum, a highest frequency component of which is greater than the passband upper frequency limit;

calculating a signal-to-noise ratio of the speech signal in a first frequency range between the lower cutoff frequency and the passband upper frequency limit; and

compressing a portion of the speech signal spectrum in a second frequency range between the lower cutoff frequency and the highest frequency component of the speech signal into the first frequency range between the lower cutoff frequency and the passband upper frequency limit in response to a determination that the signal-to-noise ratio of the speech signal in the first frequency range before compression is less than a signal-to-noise ratio of the speech signal in the first frequency range after compression.

12. The method of claim 11, further comprising outputting the speech signal without compression of the second frequency range in response to a determination that the signal-to-noise ratio of the speech signal in the first frequency range before compression is higher than the signal-to-noise ratio of the speech signal in the first frequency range after compression.

13. The method of claim 11, further comprising applying a variable gain to the compressed speech signal spectrum based on a background noise level present in the speech signal.

14. The method of claim 13, further comprising selecting a level for the variable gain based on a slope of a noise floor present in the compressed speech signal spectrum of the speech signal and a slope of a noise floor present in an uncompressed frequency portion of the speech signal.

15. The method of claim 13, further comprising selecting a level for the variable gain that substantially aligns a slope of a noise floor present in the compressed speech signal spectrum with a slope of a noise floor present in an uncompressed frequency portion of the speech signal.

16. The method of claim 11, where the act of compressing comprises regulating a spectral compressor by a computer processor.

17. A non-transitory computer-readable medium with instructions stored thereon, where the instructions are executable by a processor to cause the processor to perform the steps of:

identifying a frequency passband having a passband upper frequency limit;

defining a lower cutoff frequency within the frequency passband;

receiving a speech signal having a frequency spectrum, a highest frequency component of which is greater than the passband upper frequency limit;

calculating a signal-to-noise ratio of the speech signal in a first frequency range between the lower cutoff frequency and the passband upper frequency limit; and

compressing a portion of the speech signal spectrum in a second frequency range between the lower cutoff frequency and the highest frequency component of the speech signal into the first frequency range between the lower cutoff frequency and the passband upper frequency limit in response to a determination that the signal-to-noise ratio of the speech signal in the first frequency range before compression is less than a signal-to-noise ratio of the speech signal in the first frequency range after compression.

18. The non-transitory computer-readable medium of claim 17, further comprising instructions executable by the processor to cause the processor to perform the step of outputting the speech signal without compression of the second frequency range in response to a determination that the signal-to-noise ratio of the speech signal in the first frequency range before compression is higher than the signal-to-noise ratio of the speech signal in the first frequency range after compression.

19. The non-transitory computer-readable medium of claim 17, further comprising instructions executable by the processor to cause the processor to perform the step of applying a variable gain to the compressed speech signal spectrum based on a background noise level present in the speech signal.

20. The non-transitory computer-readable medium of claim 19, further comprising instructions executable by the processor to cause the processor to perform the step of selecting a level for the variable gain based on a slope of a noise floor present in the compressed speech signal spectrum of the speech signal and a slope of a noise floor present in an uncompressed frequency portion of the speech signal.

21. The non-transitory computer-readable medium of claim 19, further comprising instructions executable by the processor to cause the processor to perform the step of selecting a level for the variable gain that substantially aligns a slope of a noise floor present in the compressed speech signal spectrum with a slope of a noise floor present in an uncompressed frequency portion of the speech signal.