Dynamic sweet spot calibration

Patent metadata

Field	Value
Publication number	US-2020008002-A1
Application number	US-201816025986-A
Country	US
Kind code	A1
Filing date	Jul 2, 2018
Priority date	Jul 2, 2018
Publication date	Jan 2, 2020
Grant date	—

Abstract

A technique for dynamic sweet spot calibration. The technique includes receiving an image of a listening environment, which may have been captured under poor lighting conditions, and generating a crowd-density map based on the image. The technique further includes setting at least one audio parameter associated with an audio system based on the crowd-density map. At least one audio output signal may be generated based on the at least one audio parameter.

First claim

What is claimed is:

1 . A method comprising:
receiving an image of a listening environment; generating a crowd-density map based on the image; and setting at least one audio parameter associated with an audio system based on the crowd-density map, wherein at least one audio output signal is generated based on the at least one audio parameter.

2 . The method of claim 1 , wherein setting the at least one audio parameter comprises:
determining a target location in the listening environment based on the crowd-density map; determining values for the at least one audio parameter to configure the audio system to produce a sweet spot at the target location; and setting the at least one audio parameter based, at least in part, on the values.

3 . The method of claim 2 , wherein the location for the sweet spot corresponds to a centroid of the crowd-density map.

4 . The method of claim 1 , further comprising enhancing the image of the listening environment via a convolutional neural network to generate an enhanced image, wherein the convolutional neural network is trained with (i) a first training image of the listening environment illuminated by a first level of light and (ii) a second training image of the listening environment illuminated by a second level of light greater than the first level of light.

5 . The method of claim 4 , wherein generating the crowd-density map further comprises:
detecting, via at least one machine learning algorithm, at least one physical feature of individual persons included in the image of the listening environment; determining a crowd density based, at least in part, on the physical features of the individual persons; and generating the crowd-density map using the crowd density.

6 . The method of claim 1 , wherein the at least one audio parameter comprises at least one of phase and power of a signal of the audio output.

7 . The method of claim 1 , further comprising dynamically modifying the at least one audio parameter in response to a change in the crowd-density map.

8 . The method of claim 1 , wherein determining the location for the sweet spot in the listening environment further comprises converting pixel locations in the crowd-density map to a real-world coordinate system based, at least in part, on physical dimensions of the listening environment.

9 . A system comprising:
a memory storing an application; and a processor that is coupled to the memory and, when executing the application, is configured to:
receive an image of a listening environment;
generate a crowd-density map based, at least in part, on the image; and
set at least one audio parameter associated with the system based, at least in part, on the crowd-density map.

10 . The system of claim 9 , wherein the processor is further configured to:
determine a sweet spot location in the listening environment based, at least in part, on the crowd-density map; determine at least one value for the at least one audio parameter to configure the audio system to produce a sweet spot at the sweet spot location; and set the at least one audio parameter based, at least in part, on the at least one value.

11 . The system of claim 9 , wherein the processor is further configured to enhance the image of the listening environment via a convolutional neural network to generate an enhanced image, wherein the convolutional neural network is trained with (i) a first training image of the listening environment illuminated by a first level of light and (ii) a second training image of the listening environment illuminated by a second level of light greater than the first level of light.

12 . The system of claim 9 , wherein the processor is further configured to dynamically modify in real-time the at least one audio parameter in response to a change in the crowd-density map.

13 . The system of claim 9 , wherein the crowd-density map comprises a heat map.

14 . A non-transitory computer-readable storage medium including instructions that, when executed by a processor, causes the processor to perform the steps of:
receiving an image of a listening environment; generating a crowd-density map based, at least in part, on the image; determining a location of a centroid or other substantially central distribution of the crowd-density map, wherein the location is with respect to the listening environment; and determining at least one audio parameter based, at least in part, on the location.

15 . The non-transitory computer-readable storage medium of claim 14 , wherein the instructions, when executed by the processor, further cause the processor to perform the step of transmitting the at least one audio parameter to an audio system.

16 . The non-transitory computer-readable storage medium of claim 14 , wherein the instructions, when executed by the processor, further cause the processor to perform the step of determining at least one value for the at least one audio parameter to configure the audio system to produce a sweet spot at the location.

17 . The non-transitory computer-readable storage medium of claim 14 , wherein the instructions, when executed by the processor, further cause the processor to perform the step of enhancing the image of the listening environment to generate an enhanced image via a convolutional neural network, wherein the convolutional neural network is trained with (i) a first training image of the listening environment illuminated by a first level of light and (ii) a second training image of the listening environment illuminated by a second level of light greater than the first level of light.

18 . The non-transitory computer-readable storage medium of claim 14 , wherein the processor is further configured to dynamically modify in real-time the at least one audio parameter in response to a change in the crowd-density map.

19 . The non-transitory computer-readable storage medium of claim 14 , wherein the crowd-density map is generated via a neural network.

20 . The non-transitory computer-readable storage medium of claim 19 , wherein the neural network is trained based on a plurality of training images that include to a plurality of different listening environments, a plurality of different crowd densities, and a plurality of different lighting conditions.

Assignees

• Harman Int Ind

Inventors

• Sahay Pratyush
• Sai Srinivas Kruthiventi Subrahmanyeswara
• Dasgupta Arindam
• Chakraborty Pranjal
• Majumder Debojyoti

Classifications

• G06N3/08

  Learning methods · CPC title

• H04S7/302 Primary

  Electronic adaptation of stereophonic sound system to listener position or orientation (H04S7/301 takes precedence) · CPC title

• G06V20/53

  Recognition of crowd images, e.g. recognition of crowd congestion · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title