Dynamic range control with large look-ahead

Patent metadata

Field	Value
Publication number	US-9608588-B2
Application number	US-201414301269-A
Country	US
Kind code	B2
Filing date	Jun 10, 2014
Priority date	Jan 22, 2014
Publication date	Mar 28, 2017
Grant date	Mar 28, 2017

Abstract

A system and method applying Dynamic Range Control/Compression (DRC) to an audio signal. The dynamic range controller presented here differs from conventional DRC techniques by providing a much larger look-ahead time. In particular, the system and method takes advantage of the look-ahead by analyzing macroscopic loudness changes in the order of seconds as opposed to the microscopic changes most conventional DRCs are designed to control. This approach avoids most of the typical DRC distortions associated with conventional DRC techniques and preserves the micro-dynamics of the audio signal. Gain changes are applied at a rate comparable with manual volume adjustments by mixing and mastering engineers to balance a mix. Ideally, the DRC will approach what a professional sound engineer would do to reduce the dynamic range if there were only a volume control to accomplish the task on the final mix.

First claim

What is claimed is:

1. A method of performing dynamic range control (DRC) with look-ahead on an audio signal, comprising:
generating a loudness estimate for the audio signal;
generating adaptive time constants by (i) pre-smoothing the loudness estimate, and (ii) analyzing the pre-smoothed loudness estimate during a predefined look-ahead time period by estimating amounts of change in the pre-smoothed loudness estimate over time intervals, respectively, in the predefined look-ahead time period, wherein estimating the amounts of change in the pre-smoothed loudness estimate comprises generating one or more decay slopes for the pre-smoothed loudness estimate and generating one or more rise slopes for the pre-smoothed loudness estimate;
smoothing the loudness estimate using the adaptive time constants to produce a shaped loudness estimate;
generating DRC gain values to apply to a delayed version of the audio signal based on the shaped loudness estimate; and
applying the DRC gain values to the delayed audio signal.

2. The method of claim 1 , wherein a decay slope l D (n) for a time index n in the pre-smoothed loudness estimate is equal to w D l D (n−1) when l D (n−1) is greater than the pre-smoothed loudness estimate at the time index n.

3. The method of claim 1 , wherein a decay slope l D (n) for a time index n in the pre-smoothed loudness estimate is equal to the pre-smoothed loudness estimate at time index n when l D (n−1) is less than or equal to the pre-smoothed loudness estimate at the time index n.

4. The method of claim 1 , wherein a rise slope l R (n) for a time index n in the pre-smoothed loudness estimate is equal to w R l R (n−1) when l R (n−1) is less than the pre-smoothed loudness estimate at the time index n.

5. The method of claim 1 , wherein a rise slope l R (n) for a time index n in the pre-smoothed loudness estimate is equal to the pre-smoothed loudness estimate at time index n when l R (n−1) is greater than or equal to the pre-smoothed loudness estimate at the time index n.

6. The method of claim 1 , further comprising:
moving the one or more decay slopes and the one or more rise slopes into look-ahead buffers for the predefined look-ahead time period.

7. The method of claim 6 , further comprising:
determining a loudness change for a decreasing loudness time period in the audio signal based on slow and fast decay slopes; and
determining a loudness change for an increasing loudness time period in the audio signal based on slow and fast rise slopes.

8. The method of claim 7 , wherein the adaptive time constants are generated based on the determined loudness changes such that a first adaptive time constant is selected for a first loudness change and a second adaptive time constant is selected for a second loudness change, wherein the first adaptive time constant is shorter than the second adaptive time constant when the first loudness change is greater than the second loudness change.

9. The method of claim 8 , wherein the adaptive time constants are stored in a time constants buffer and the oldest adaptive time constants in the time constants buffer are used to smooth the loudness estimate.

10. The method of claim 7 , wherein determining the loudness change for the decreasing loudness time period comprises a measurement that begins at a decay crossover point between the slow and fast decay slopes wherein the decay crossover point is the last point in time where the slow and fast decay slopes have equal values, and
wherein determining the loudness change for the increasing loudness time period comprises a measurement that begins at a rise crossover point between the slow and fast rise slopes wherein the rise crossover point is the last point in time where the slow and fast rise slopes have equal values.

11. The method of claim 1 , further comprising:
processing the shaped loudness estimate through a first characteristic function to generate initial gain values; and
processing the shaped loudness estimate through a second characteristic function to generate gain reduction values for the rise and decay slopes, wherein the DRC gain values are generated based on the initial gain values and the gain reduction values.

12. The method of claim 1 , wherein the predefined look-ahead time period is between five seconds and ten seconds.

13. The method of claim 1 , further comprising:
resetting the one or more decay slopes and the one or more rise slopes to the pre-smoothed loudness estimate in response to a reduction in an associated adaptive time constant.

14. An article of manufacture for performing dynamic range control (DRC) with a look-ahead on an audio signal, comprising:
a non-transitory machine-readable storage medium that stores instructions which, when executed by a processor in a computer,
generate a loudness estimate for the audio signal;
generate adaptive time constants by (i) pre-smoothing the loudness estimate and (ii) analyzing the pre-smoothed loudness estimate during a predefined look-ahead time period by estimating amounts of change in the pre-smoothed loudness estimate over time intervals, respectively, in the predefined look-ahead time period wherein estimating the amounts of change in the pre-smoothed loudness estimate comprises generating one or more decay slopes for the pre-smoothed loudness estimate, and generating one or more rise slopes for the pre-smoothed loudness estimate;
smooth the loudness estimate using the adaptive time constants, to produce a shaped loudness estimate;
generate DRC gain values by using a characteristic unit to map the shaped loudness estimate, as its input, to the DRC gain values as its output; and
apply the DRC gain values to the delayed audio signal.

15. A system for performing dynamic range control (DRC) with a look-ahead on an audio signal, comprising:
a device, including a hardware processor, the device being configured to:
generate a loudness estimate for the audio signal;
generate adaptive time constants by (i) pre-smoothing the loudness estimate, and (ii) analyzing the pre-smoothed loudness estimate during a predefined look-ahead time period by estimating amounts of change in the pre-smoothed loudness estimate over time intervals, respectively, in the predefined look-ahead time period, wherein estimating the amounts of change in the pre-smoothed loudness estimate comprises generating one or more decay slopes for the pre-smoothed loudness estimate and generating one or more rise slopes for the pre-smoothed loudness estimate;
smooth the loudness estimate using the adaptive time constants to produce a shaped loudness estimate;
generate DRC gain values to apply to a delayed version of the audio signal based on the shaped loudness estimate; and
apply the DRC gain values to the delayed audio signal.

16. The system of claim 15 wherein the device is configured to move the one or more decay slopes and the one or more rise slopes into look-ahead buffers for the predefined look-ahead time period.

17. The system of claim 16 wherein the device is configured to compute a loudness change for a decreasing loudness time period in the audio signal, based on slow and fast decay slopes, and compute a loudness change for an increasing loudness time period in the audio signal, based on slow and fast rise slopes.

18. The system of claim 17 wherein the device is configured to generate the adaptive time constants based on the computed loudness changes wherein a first adaptive time constant is selected for a first loudness change, and a second adaptive time constant is selected for a second lou

Assignees

• Apple Inc

Inventors

• Baumgarte Frank M

Classifications

• H03G9/025 Primary

  frequency-dependent volume compression or expansion, e.g. multiple-band systems (H03G9/10, H03G9/18 take precedence) · CPC title