System and method for providing voice equalization
View Patent ↗Systems and methods for providing voice equalization are provided. In exemplary embodiments, acoustic signals are received from both a near-end and a far-end environment. A power spectrum estimate for the far-end signal and a noise estimate based on the near-end signal are determined. A voice equalization mask based on the power spectrum estimate of the far-end signal and the noise estimate based on the near-end signal is generated and applied to the far-end signal to obtain a modified signal. The modified signal may then be dynamically processed to obtain an enhanced audio signal. The enhanced audio signal is then output.
1. A system for providing voice equalization, comprising:
an energy module configured to determine a power spectrum estimate of a far-end signal, the far-end signal and a near-end signal each representing at least one respective captured sound received by at least one acoustic sensor;
a noise suppression engine configured to determine a noise estimate based on the near-end signal; and
a voice equalization engine configured to generate and apply a voice equalization mask based on the power spectrum estimate of the far-end signal and the noise estimate based on the near-end signal to obtain a modified signal as well as dynamically process the modified signal to obtain an enhanced audio signal.
2. The system of claim 1 wherein the noise suppression engine is further configured to dynamically process a modified far-end signal to generate the voice equalization audio signal, the system further comprising the at least one acoustic sensor.
3. The system of claim 1 wherein the voice equalization engine comprises a stationary noise suppression (SNS) module configured to generate a modification gain (G S ).
4. The system of claim 1 wherein the voice equalization engine comprises a volume control and unmasking (VCU) module configured to generate a global gain (G O ) and a plurality of local gains (G N ).
5. The system of claim 1 wherein the voice equalization engine comprises a mask generator configured to combine a modification gain (G S ), a global gain (G O ), and local gains (G N ) to generate a final modification mask.
6. The system of claim 1 wherein the voice equalization engine comprises a limiter configured to generate a limit gain (G L ) that prevents signal peaks from exceeding a pre-defined value.
7. The system of claim 1 wherein the voice equalization engine comprises an onset smoother configured to generate a smoother gain (G P ) that smoothes a waveform to prevent quick energy changes.
8. The system of claim 1 wherein the voice equalization engine comprises a linear interpolator configured to combine a limit gain (G L ) and a smoother gain (G P ) and interpolating a result to a predetermined sampling rate.
9. The system of claim 1 wherein the noise suppression engine comprises a level difference module configured to determine a level difference between power estimates of a primary acoustic signal and a secondary acoustic signal from a near-end environment.
10. The system of claim 9 wherein the noise suppression engine comprises a noise estimate module configured to determine the noise estimate of the near-end signal based on the level difference.
11. The system of claim 1 wherein the voice equalization engine comprises a mask generator configured to combine a modification gain (G S ), a global gain (G O ), and local gains (G N ) to maintain a constant signal-to-noise (SNR) target.
12. A non-transitory computer-readable storage medium having embodied thereon a program, the program providing instructions to a processor when executed would perform a method for providing voice equalization, the method comprising:
determining a power spectrum estimate of a far-end signal, the far-end signal and a near-end signal each representing at least one respective captured sound received by at least one acoustic sensor;
determining a noise estimate based on the near-end signal;
generating a voice equalization mask based on the power spectrum estimate of the far-end signal and the noise estimate based on the near-end signal;
applying the voice equalization mask to the far-end signal to obtain a modified signal; and
dynamically processing the modified signal to obtain an enhanced audio signal.
13. The non-transitory computer-readable storage medium of claim 12 wherein generating the voice equalization mask comprises generating a modification gain (G S ).
14. The non-transitory computer-readable storage medium of claim 12 wherein generating the voice equalization mask comprises generating a global gain (G O ).
15. The non-transitory computer-readable storage medium of claim 12 wherein generating the voice equalization mask comprises generating local gains (G N ).
16. The non-transitory computer-readable storage medium of claim 12 wherein generating the voice equalization mask comprises combining a modification gain (G S ), a global gain (G O ), and local gains (G N ) to generate a final modification mask.
17. The non-transitory computer-readable storage medium of claim 12 wherein dynamically processing the modified signal comprises generating a limit gain (G L ) that prevents signal peaks from exceeding a pre-defined value.
18. The non-transitory computer-readable storage medium of claim 12 wherein dynamically processing the modified signal comprises generating a smoother gain (G P ) that smoothes a waveform to prevent quick energy changes.
19. The non-transitory computer-readable storage medium of claim 12 wherein dynamically processing the modified signal comprises combining a limit gain (G L ) and a smoother gain (G P ) and interpolating a result to a predetermined sampling rate.
20. The non-transitory computer-readable storage medium of claim 12 wherein determining the noise estimate comprises receiving a primary acoustic signal via a primary microphone and receiving a secondary acoustic signal via a secondary microphone and determining an inter-microphone level difference between the primary and the secondary acoustic signals.