Apparatus and method for audio content analysis, marking and summing

• US 20060133624A1
• Filed: 12/17/2003
• Published: 06/22/2006
• Est. Priority Date: 08/18/2003
• Status: Active Grant

• Current Assignee: Nice Systems Limited (Nice Ltd)
• Original Assignee: Nice Systems Limited (Nice Ltd)
• Inventors: Freedman, Ilan, Waserblat, Moshe, Aharoni, Gili, Bachar, Aviv, Eliam, Barak
• Granted Patent: US 7,546,173 B2
• Time in Patent Office: Days
• US Class Current: 381/119
• CPC Class Codes: H04H 60/04 Studio equipment; Interconnection of studios
• CPC Class Codes: H04S 1/007 in which the audio signals are in digital form data reduction aspects thereof based on psychoacoustics G10L19/02
• CPC Class Codes: H04S 7/00 Indicating arrangements; Control arrangements, e.g. balance control

Abstract

An apparatus and method for the analysis, marking and summing of audio channel content and control data, the apparatus and method generating a summed signal carrying combined audio content, marking and summing data in the summed signal.

Claims

1. An apparatus for the analysis, marking and summing of audio channel content and control data, the apparatus comprising:

an at least one audio channel marking component to extract from an at least one audio channel delivering a signal carrying encoded audio content signal-specific characteristics and channel-specific control information, and to generate from the extracted control information and signal characteristics channel-specific marking data;

an at least one audio summing component to sum the signal delivered via the at least one audio channel into a summed signal, and to generate signal summing control information; and

an at least one marking and summing embedding component to insert the generated marking data and summing data into the summed signal;

thereby, generating a summed signal carrying combined audio content, marking and summing data into the summed signal.

2. The apparatus of claim 1 further comprising:

an at least one embedded marking and summing control data extraction component to extract marking and summing data and spectral feature vectors data from the decompressed signal;

an at least one audio channel recognition component to identify at least one audio channel from the uncompressed signal associated with the extracted marking and summing control data; and

an at least one audio channel separation component to separate the decompressed signal into the constituent channels thereof;

thereby, enabling for the extraction and separation of previously generated summed signal.

3. The apparatus of claim 1 further comprising an at least one spectral features extraction component to analyze the signal delivered by the at least one audio channel and to generate spectral features vector data characterizing the audio content of the signal.

4. The apparatus of claim 1 further comprising a compressing component to process the summed audio signal including the embedded marking and summing information in order to generate a compressed signal.

5. The apparatus of claim 1 further comprising an automatic number identification component to identify the origin of the at least one audio channel delivering the signal carrying encoded audio content.

6. The apparatus of claim 1 further comprising a dual tone multi frequency component to extract traffic control information from the signal delivered by the audio channel.

7. The apparatus of claim 1 further comprising an at least one group of digital signal processing devices to provide for audio content analysis prior to the marking, summing and compressing of the signal, the group of digital signal processing devices comprising any one of the following components:

a talk analysis statistics component to generate talk statistics from the audio content carried by the signal;

an excitement detection component to identify emotional characteristics of the audio content carried by the signal;

an age detection component to identify the age of a speaker associated with a speech segment of the audio content carried by the signal; and

a gender detection component to identify the gender of a speaker associated with a speech segment of the audio content carried by the signal.

8. The apparatus of claim 2 further comprising a decompression component to decompress the summed signal.

9. The apparatus of claim 2 further comprising at least one digital signal processing devices for content analysis, the group of the digital signal processing devices comprising any of the following components:

a transcription component to transform speech elements of the audio content of the signal to text; and

a word spotting component to identify pre-defined words in the speech elements of the audio content.

10. The apparatus of claim 1 further comprising at least one storage unit to store the summed and compressed signal carrying audio content and marking and summing control data.

11. The apparatus of claim 2 further comprising at least one content analysis server to provide for channel-specific content analysis of the signal carrying audio content and an at least one content analysis database to store the results of the content analysis.

12. A method for the analysis marking and summing of audio content, the method comprising:

analyzing an at least one signal carrying audio content and traffic control data delivered via an at least one audio channel to generate channel-specific control data, and signal-specific spectral characteristics;

generating channel-specific marking control data from the channel-specific control data and the signal-specific spectral features vector data;

summing the at least one signal carrying audio content into a summed signal; and

generating summation control data; and

embedding the channel-specific control data, the segment-specific summation data, and the signal-specific spectral features vector data into the summed signal;

thereby, generating a summed signal carrying combined audio content, channel-specific control data, segment-specific summation data, and spectral features vector data into the summed signal.

13. The method of claim 12 further comprising the steps of:

extracting the marking and summing data from the summed signal;

identifying the at least one channel-specific signal within the summed signal; and

separating the at least one channel-specific signal from the summed signal;

thereby providing a channel-specific signal carrying channel-specific audio content for audio content analysis.

14. The method of claim 12 further comprising the step of compressing the summed signal in order to transform the signal to a compressed format signal.

15. The method of claim 12 further comprising the step of decompressing the summed and compressed signal.

16. The method of claim 12 further comprising the step of storing the summed signal carrying audio content and marking and summing control data on a storage device.

17. The method of claim 12 further comprising the step obtaining the summed signal from the storage device in order to perform audio channel separation and channel-specific content analysis;

and storing the results of the content analysis on a storage device to provide for data mining options for additional applications.

18. The method of claim 12 wherein the marking of the at least one audio channel is performed in accordance with the traffic control data carried by the at least one signal.

19. The method of claim 12 wherein the separation of the summed signal is performed in accordance with the traffic control data carried by the at least one signal.

20. The method of claim 12 wherein the marking of the at least one audio channel is accomplished through selectively marking speech segments included in the at least one signal associated with different speakers.

21. The method of claim 12 wherein the separation of the summed signal is accomplished through selectively marking speech segments included in the at least one signal associated with different speakers.

22. The method of claim 12 wherein the embedding of the marking and summing control data in the summed signal is achieved via data hiding.

23. The method of claim 22 wherein data hiding is performed by pulse code modulation robbed-bit method.

24. The method of claim 23 wherein data hiding is performed by code excited linear prediction compression method.