Categorizer of content in digital signals

• US 7,188,065 B2
• Filed: 11/04/2004
• Issued: 03/06/2007
• Est. Priority Date: 04/24/2001
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Mihcak, M. Kivanc, Venkatesan, Ramarathnam
• Primary Examiner(s): Azad, Abul K.
• Application Number:
  US10/981,165
• Publication Number:
  US 20050066176A1
• Time in Patent Office: 852 Days
• Field of Search: None
• US Class Current: 704/270
• CPC Class Codes: G06F 16/433 using audio data
• CPC Class Codes: G06F 16/634 Query by example, e.g. query by humming
• CPC Class Codes: G06F 16/683 using metadata automatically derived from the content
• CPC Class Codes: G10H 1/0058 Transmission between separate instruments or between individual components of a musical system G10H1/0083 takes precedence

Abstract

Described herein is a technology for facilitating the recognition and categorization of the content of digital signals. This abstract itself is not intended to limit the scope of this patent. The scope of the present invention is pointed out in the appending claims.

Claims

1. A computer-implemented method for recognizing a digital audio signal, the method comprising:

obtaining a digital audio signal;

deriving a categorization value representative of the digital audio signal so that perceptually similar digital audio signals have proximally similar categorization values, wherein the deriving of the categorization value comprises;

transforming the digital audio signal into a digital-audio-signal transform;

randomly dividing the digital-audio-signal transform into multiple chunks, each chunk containing signal data, wherein the dividing is carried out recursively to form hierarchical levels of overlapping chunks;

averaging, for each of the chunks, the signal data to produce corresponding chunk averages;

generating, based in part on the chunk averages, an exponential distribution having multiple distinct quantization levels;

randomly rounding each of the chunk averages to one of the quantization levels to produce rounded values;

hashing a composite of the rounded values;

storing the categorization value in relationship with the digital audio signal.

2. A computer-implemented method as recited in claim 1 further comprising comparing categorization value of the digital audio signal to determine if such value is proximally near categorization values of a group of digital audio signals having proximally clustered categorization values.

3. A computer-implemented method as recited in claim 1 further comprising:

comparing categorization value of the digital audio signal to determine if such value is proximally near categorization values of a group of digital audio signals having proximally clustered categorization values;

grouping the digital audio signal with the group of digital audio signals if the categorization value of the digital audio signal is proximally near the categorization values of the group.

4. A system for recognizing a digital audio signal, the system comprising means for performing functions comprising:

obtaining a digital audio signal;

deriving a categorization value representative of the digital audio signal so that perceptually similar digital audio signals have proximally similar categorization values, wherein the deriving of the categorization value comprises;

transforming the digital audio signal into a digital-audio-signal transform;

randomly dividing the digital-audio-signal transform into multiple chunks, each chunk containing signal data, wherein the dividing is carried out recursively to form hierarchical levels of overlapping chunks;

averaging, for each of the chunks, the signal data to produce corresponding chunk averages;

generating, based in part on the chunk averages, an exponential distribution having multiple distinct quantization levels;

randomly rounding each of the chunk averages to one of the quantization levels to produce rounded values;

hashing a composite of the rounded values;

storing the categorization value in relationship with the digital audio signal.

5. A system as recited in claim 4, the functions further comprising comparing categorization value of a digital audio signal to determine if such value is proximally near categorization values of a group of digital audio signals having proximally clustered categorization values.

6. A system as recited in claim 4, the functions further comprising:

comparing categorization value of the digital audio signal to determine if such value is proximally near categorization values of a group of digital audio signals having proximally clustered categorization values;

grouping the digital audio signal with the group of digital audio signals if the categorization value of the digital audio signal is proximally near the categorization values of the group.

7. One or more computer-readable media having computer-executable instructions thereon that, when executed by a computer, perform acts for recognizing a digital audio signal, the acts comprising:

obtaining a digital audio signal;

deriving a categorization value representative of the digital audio signal so that perceptually similar digital audio signals have proximally similar categorization values, wherein the deriving of the categorization value comprises;

transforming the digital audio signal into a digital-audio-signal transform;

randomly dividing the digital-audio-signal transform into multiple chunks, each chunk containing signal data, wherein the dividing is carried out recursively to form hierarchical levels of overlapping chunks;

averaging, for each of the chunks, the signal data to produce corresponding chunk averages;

generating, based in part on the chunk averages, an exponential distribution having multiple distinct quantization levels;

randomly rounding each of the chunk averages to one of the quantization levels to produce rounded values;

hashing a composite of the rounded values;

storing the categorization value in relationship with the digital audio signal.

8. One or more media as recited in claim 7 further comprising comparing categorization value of a digital audio signal to determine if such value is proximally near categorization values of a group of digital audio signals having proximally clustered categorization values.

9. One or more media as recited in claim 7 further comprising:

comparing categorization value of the digital audio signal to determine if such value is proximally near categorization values of a group of digital audio signals having proximally clustered categorization values;

grouping the digital audio signal with the group of digital audio signals if the categorization value of the digital audio signal is proximally near the categorization values of the group.

10. One or more computer-readable media having computer-executable instructions thereon that, when executed by a computer, perform acts for recognizing a digital audio signal, the acts comprising:

deriving a categorization value representative of a digital audio signal based upon intrinsic characteristics of the digital audio signal, so that digital audio signals that are perceptually similar have categorization values that are proximally similar, wherein the deriving of the categorization value comprises;

transforming the digital audio signal into a digital-audio-signal transform;

randomly dividing the digital-audio-signal transform into multiple chunks, each chunk containing signal data, wherein the dividing is carried out recursively to form hierarchical levels of overlapping chunks;

averaging, for each of the chunks, the signal data to produce corresponding chunk averages;

generating, based in part on the chunk averages, an exponential distribution having multiple distinct quantization levels;

randomly rounding each of the chunk averages to one of the quantization levels to produce rounded values;

hashing a composite of the rounded values;

storing the categorization value in relationship with the digital audio signal;

comparing categorization value of the digital audio signal to determine if such value is proximally similar to categorization values of a group of digital audio signals having proximally clustered categorization values;

grouping the digital audio signal with a group of digital audio signals if the categorization value of the digital audio signal is proximally near the categorization values of the group.

11. A computer-readable medium having stored thereon a data structure, comprising:

a first data field containing a digital audio signal;

a second data field derived from the first field by deriving a categorization value representative of the digital audio signal so that perceptually similar digital audio signals have proximally similar categorization values, wherein the derivation of the categorization value comprises;

transforming the digital audio signal into a digital-audio-signal transform;

randomly dividing the digital-audio-signal transform into multiple chunks, each chunk containing signal data, wherein the dividing is carried out recursively to form hierarchical levels of overlapping chunks;

averaging, for each of the chunks, the signal data to produce corresponding chunk averages;

generating, based in part on the chunk averages, an exponential distribution having multiple distinct quantization levels;

randomly rounding each of the chunk averages to one of the quantization levels to produce rounded values;

hashing a composite of the rounded values;

a third data field functioning to delimit the end of the data structure.