Adaptive loudness control

• US 9,391,575 B1
• Filed: 12/13/2013
• Issued: 07/12/2016
• Est. Priority Date: 12/13/2013
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Yang, Jun
• Primary Examiner(s): Agustin, Peter Vincent
• Application Number:
  US14/106,529
• Time in Patent Office: 942 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: H03G 3/3005 in amplifiers suitable for low-frequencies, e.g. audio amplifiers H03G3/32, H03G3/34 take precedence
• CPC Class Codes: H03G 3/3089 Control of digital or coded signals
• CPC Class Codes: H03G 3/32 the control being dependent upon ambient noise level or sound level
• CPC Class Codes: H03G 9/005 of digital or coded signals
• CPC Class Codes: H03G 9/025 frequency-dependent volume compression or expansion, e.g. multiple-band systems H03G9/10, H03G9/18 take precedence

Abstract

Loudness control is performed by estimating the energy and loudness of an audio signal. Loudness of the audio signal is determined by decomposing the audio signal into multiple frequency bands at different center frequencies. The energy of the audio signal in each frequency band is calculated and converted to a loudness, using a function that models human loudness perceptions of audio energies at different frequencies. The loudnesses are summed to obtain the total loudness of the audio signal. A signal gain is calculated as a function of a loudness setting provided by a human listener, the estimated signal energy, and the total loudness of the audio signal.

Claims

1. A method, comprising:

receiving an input audio signal;

determining a signal energy of the input audio signal;

determining a band energy of the input audio signal for each of multiple frequency bands of the input audio signal;

converting each band energy to a corresponding band loudness to produce multiple band loudnesses;

summing the multiple band loudnesses to produce a signal loudness;

receiving a loudness value specified by a human listener, wherein the loudness value indicates a desired relative audio loudness;

multiplying the loudness value by a ratio of the signal loudness to the signal energy to produce a signal gain;

amplifying the input audio signal by the signal gain to produce an output audio signal; and

providing the output audio signal to a loudspeaker.

2. The method of claim 1, further comprising:

determining whether a zero crossing rate of the audio input signal exceeds a first threshold;

determining whether a dynamic level of the audio input exceeds a second threshold;

determining whether the signal gain exceeds a predefined value;

if the zero crossing rate exceeds the first threshold, the dynamic level does not exceed the second threshold, and the signal gain exceeds the predefined value, reducing the signal gain to the predefined value.

3. The method of claim 1, wherein the converting is performed in accordance with a function that models human loudness perceptions of audio energies at multiple frequencies, wherein the function converts each band energy to a loudness value that represents a perceived loudness of the band energy by a human.

4. The method claim 1, wherein:

the input audio signal comprises a plurality of amplitude values; and

determining the signal energy comprises calculating a square root of a sum of squares of the amplitude values.

5. The method of claim 1, wherein determining the band energy of the input audio signal for each of multiple frequency bands comprises:

filtering the input audio signal with a plurality of band-pass filters to produce a plurality of filtered signals corresponding respectively to the frequency bands, wherein each of the filtered signals comprises a plurality of amplitude values; and

for each filtered signal, calculating a square root of a sum of squares of the amplitude values of the filtered signal.

6. A device, comprising:

one or more processors;

computer-readable media storing instructions that are executable by the one or more processors to perform acts comprising;

receiving an input audio signal;

determining a signal energy of the input audio signal;

analyzing the input audio signal at different frequencies of the input audio signal to determine a signal loudness;

receiving an input signal specifying a loudness value;

determining a signal gain based on the signal loudness and a product of the loudness value and the signal energy; and

applying the signal gain to the input audio signal to produce an output audio signal.

7. The device of claim 6, wherein the signal gain is a ratio of the product over the signal energy.

8. The device of claim 6, wherein analyzing the input audio signal comprises:

filtering the input audio signal with a plurality of band-pass filters to produce a plurality of filtered signals corresponding respectively to different frequency bands;

analyzing each of the filtered signals to determine a corresponding band energy;

converting each of the band energies to a corresponding band loudness; and

summing the band loudnesses.

9. The device of claim 8, wherein the converting is performed in accordance with a function that models human loudness perceptions of audio energies at different frequencies.

10. The device of claim 6, wherein:

the input audio signal comprises a plurality of amplitude values; and

determining the signal energy comprises calculating a square root of a sum of squares of the signal amplitude values.

11. The device of claim 6, wherein analyzing the input audio signal comprises:

filtering the input audio signal with a plurality of band-pass filters to produce a plurality of filtered signals corresponding respectively to different frequency bands, wherein each of the filtered signals comprises a plurality of amplitude values;

for each filtered signal, calculating a square root of a sum of squares of the amplitude values of the filtered signal to produce multiple loudness values; and

summing the loudness values.

12. The device of claim 6, the acts further comprising:

detecting audio silence in the input audio signal;

detecting audio noise in the input audio signal;

limiting the signal gain in response to detecting the audio silence or the audio noise in the input audio signal.

13. A method, comprising:

determining a signal energy of an input audio signal;

analyzing the input audio signal to produce a signal loudness;

determining a preliminary value of a signal gain as a function of the signal energy and the signal loudness, wherein the preliminary value of the signal gain is based at least in part on multiplying a loudness value by a ratio of the signal loudness to the signal energy;

adjusting the signal gain by applying a function to consecutive values associated with the preliminary value; and

applying the signal gain to the input audio signal to produce an output audio signal.

14. The method of claim 13, wherein analyzing the input audio signal comprises:

determining a band energy of the input audio signal for each of a plurality of frequency bands;

converting each of the band energies to a corresponding band loudness; and

summing the band loudnesses.

15. The method of claim 14, wherein the converting is performed in accordance with a function that models human loudness perceptions of audio energies at different frequencies.

16. The method of claim 13, wherein analyzing the input audio signal comprises:

filtering the input audio signal with a plurality of band-pass filters to produce a plurality of filtered signals corresponding respectively to different frequency bands;

analyzing each of the filtered signals to determine a corresponding band energy;

converting each band energy to a corresponding band loudness; and

summing the band loudnesses.

17. The method of claim 13, wherein analyzing the input audio signal comprises:

filtering the input audio signal with a plurality of band-pass filters to produce a plurality of filtered signals corresponding respectively to different frequency bands, wherein each of the filtered signals comprises a plurality of amplitude values; and

for each filtered signal, calculating a square root of a sum of squares of the amplitude values of the filtered signal.

18. The method of claim 13, wherein analyzing the input audio signal comprises:

filtering the input audio signal with a plurality of band-pass filters to produce a plurality of filtered signals corresponding respectively to different frequency bands, wherein each of the filtered signals comprises a plurality of amplitude values;

for each filtered signal, calculating a square root of a sum of squares of the amplitude values of the filtered signal to produce multiple loudness values; and

summing the loudness values.

19. The method of claim 13, further comprising:

detecting audio silence in the input audio signal;

detecting audio noise in the input audio signal;

limiting the signal gain in response to detecting the audio silence or audio noise in the input audio signal.