Apparatus and method for extracting a test signal section from an audio signal

• US 7,680,056 B2
• Filed: 11/23/2005
• Issued: 03/16/2010
• Est. Priority Date: 06/17/2003
• Status: Active Grant

• Current Assignee: Opticom Dipl-Ing
• Original Assignee: Opticom Dipl-Ing
• Inventors: Bitto, Roland, Keyhl, Michael, Schmidmer, Christian
• Primary Examiner(s): Pham; Brenda
• Application Number:
  US11/286,311
• Publication Number:
  US 20060177003A1
• Time in Patent Office: 1,574 Days
• Field of Search: 704/233, 704/275, 704/200, 704/201, 370/241
• US Class Current: 370/241
• CPC Class Codes: G10L 25/69 for evaluating synthetic or decoded voice signals
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10
• CPC Class Codes: H04M 2203/056 non-invasive testing, i.e. in operation testing without service interruption
• CPC Class Codes: H04M 3/2236 Quality of speech transmission monitoring

Abstract

For extracting signal test sections conforming to an auditory test from an audio signal, in particular for non-intrusive tests for a quality assessment of a transmission system, first of all a temporal structure of the audio signal is parsed in order to differentiate an information-carrying section of the audio signal from a preceding non-information-carrying section of the audio signal or a subsequent non-information-carrying section of the audio signal. Hereupon, a test signal section is generated based on the information-carrying section of the audio signal. Thus, test signal sections conforming to an auditory test are obtained, on the one hand determining the comparability with other auditory tests and, on the other hand, reducing a too inaccurate assessment of the transmission system based on system-immanent artefacts.

Claims

1. A device implemented on a computer which is programmed to extract a test signal section from an audio signal, the device comprising, implemented on the computer:

a parser for parsing a temporal or spectral structure of the audio signal in order to differentiate information-carrying sections of the audio signal from a non-information-carrying sections of the audio signal, the information-carrying sections being separated from each other by the non-information-carrying sections of the audio signal, wherein the information-carrying sections and the non-information-carrying sections alternate within the audio signal;

a generator for generating a sequence of test signal sections based on the information-carrying sections of the audio signal; and

a transmitter configured to transmit the sequence of test signal sections to a measurement device for a non-intrusive quality assessment resulting in a quality measurement value per test signal section of the sequence of test signal sections.

2. The device according to claim 1, wherein the audio signal comprises voiced portions and pause portions between the voiced portions, and

wherein the parser is implemented to perform a voice activity detection in order to detect voiced portions as the information-carrying sections arranged in temporal respect between respective two pause portions as the non-information-carrying sections.

3. The device according to claim 1, wherein the audio signal comprises signal portions having a predetermined minimum power and pause portions having a power less than the predetermined signal power, and

wherein the parser is implemented to perform a pause detection to detect the pause portions consecutive in time between each pair of which a signal portion is arranged as one of the information-carrying sections.

4. The device according to claim 1,

wherein the audio signal comprises noise portions and intermediate signal portions, and

wherein the parser is implemented to perform a noise detection in order to detect the noise portions consecutive in time in order to determine the signal portions as the information-carrying section, each of which lies between a respective pair of the noise portions.

5. The device according to claim 1,

wherein the generator is implemented to generate the sequence of test signal sections based on the information-carrying sections of the audio signal by, for each information-carrying section, determining as to whether the temporal length of the respective information-carrying section is smaller than a predetermined maximum length, and, depending on the determination, taking over, in complete, the respective information-carrying section into a respective one of the test signal sections in the case the temporal length of the respective information-carrying section is smaller than a predetermined maximum length.

6. The device according to claim 1,

wherein the generator is implemented in order to generate the sequence of test signal sections based on the information-carrying sections of the audio signal by, for a predetermined test signal section, taking over a continuous part of the audio signal containing at least one of the information-carrying sections of the audio signal in complete and one or several of the non-information-carrying sections, so that the predetermined test signal section comprises a predetermined ratio of temporal length of the at least one information-carrying section and temporal length of the one or the several non-information-carrying sections.

7. The device according to claim 1,

wherein the generator for generating is implemented to generate the sequence of test signal sections by, for a first and a second predetermined test signal section of the sequence of test signal sections, taking over a respective continuous part of the audio signal containing at least a respective one of the information-carrying sections of the audio signal in complete and one or several of the non-information-carrying sections, so that, an endpoint of the first test signal section with regard to the audio signal is in time after a starting point of the second test signal section, so that both the first test signal section and the second test signal section have at least a part of a non-information-carrying section of the audio signal in common.

8. The device according to claim 1,

wherein the parser for parsing and the generator for generating are implemented to transform the audio signal into the sequence of test signal sections such that the test signal sections are of different length, and such that each test signal section is longer than or equal to a predetermined minimum length and shorter than or equal to a predetermined maximum length.

9. The device according to claim 1,

wherein the parser is implemented to output a notification when no information-carrying section is detectable in the audio signal.

10. The device according to claim 1,

wherein the generator is implemented to generate a sequence of test signal sections, and further comprises a loudness measurer for a loudness measurement, in order to obtain a loudness value for each test signal section, wherein the loudness values of the test signal sections may be used by a loudness feedback in order to control a loudness course of the audio signal with regard to feedback.

12. The device according to claim 1,

wherein the generator is implemented to generate the test signal sections such that each of the test signal sections has a respective temporal length which is greater than or equal to a predetermined minimum length and smaller than or equal to a predetermined maximum length.

13. The device according to claim 12, wherein the predetermined minimum length lies between 2 and 12 seconds and wherein the predetermined maximum length lies between 12 and 25 seconds.

14. The device according to claim 1,

wherein the generator is implemented in order to take over at least a part of the information-carrying sections of the audio signal unchanged into the test signal sections.

11. A device implemented on a computer which is programmed to extract a test signal section from an audio signal, the device comprising, implemented on the computer:

a parser for parsing a temporal or spectral structure of the audio signal in order to differentiate an information-carrying section of the audio signal from a preceding non-information-carrying section of the audio signal or a subsequent non-information-carrying section of the audio signal; and

a generator for generating the test signal section based on the information-carrying section of the audio signal,

wherein the audio signal comprises voiced portions and pause portions between the voiced portions, and the parser is implemented to perform a voice activity detection in order to detect a voiced portion as an information-carrying section arranged in temporal respect between two pause portions as non-information-carrying sections, and

wherein the parser is further implemented to subject the voiced portion to a speech detection and to parse the information-carrying section with regard to a speech and/or a sentence context, and

wherein the generator is implemented in order to consider the speech and/or the sentence context when generating the test signal section based on the information-carrying section of the audio signal such that the test signal section has a temporal length which is greater than or equal to a predetermined minimum length and smaller than or equal to a predetermined maximum length.

15. A device implemented on a computer which is programmed to extract a test signal section from an audio signal, the device comprising, implemented on the computer:

a parser for parsing a temporal or spectral structure of the audio signal in order to differentiate an information-carrying section of the audio signal from a preceding non-information-carrying section of the audio signal or a subsequent non-information-carrying section of the audio signal; and

a generator for generating the test signal section based on the information-carrying section of the audio signal,

wherein the generator is implemented

to detect a temporal length of the information-carrying section,

to determine whether the temporal length exceeds a predetermined maximum value, and

in the case of exceeding the predetermined maximum value, to fade out a part at the beginning or at the end of the information-carrying section to obtain a modified information-carrying section whose temporal length is smaller than the predetermined maximum value, and

in a transition range between a faded-out portion and a non-modified portion, to obtain a continuous level change of the modified information-carrying section, and

to generate the test signal section so that the test signal section includes the transition range and the non-modified section.

16. The device according to claim 15, further comprising:

a transmitter for transmitting the test signal section to a measurement device for a quality assessment of a transmission system from which the audio signal may be obtained.

17. A device implemented on a computer which is programmed to extract a test signal section from an audio signal, the device comprising, implemented on the computer:

a parser for parsing a temporal or spectral structure of the audio signal in order to differentiate an information-carrying section of the audio signal from a preceding non-information-carrying section of the audio signal or a subsequent non-information-carrying section of the audio signal; and

a generator for generating the test signal section based on the information-carrying section of the audio signal,

wherein the audio signal comprises a music portion, and

wherein the parser for parsing is implemented to perform a level detection, a loudness detection or a modulation detection in order to differentiate a non-information-carrying section from an information-carrying section.

18. A method for extracting a test signal section from an audio signal, comprising:

a computer parsing a temporal or spectral structure of the audio signal in order to differentiate information-carrying sections of the audio signal from non-information-carrying sections of the audio signal, the information-carrying sections of the audio signal and the non-information-carrying sections alternate within the audio signal; and

the computer generating a sequence of test signal sections based on the information-carrying sections of the audio signal; and

the computer transmitting the sequence of test signal sections to a measurement device for a non-intrusive quality assessment resulting in a quality measurement value per test signal section of the sequence of test signal sections.

19. A digital storage medium having stored thereon a computer program having a program code for performing, when the computer program runs on a computer, a method for extracting a test signal section from an audio signal, comprising:

parsing a temporal or spectral structure of the audio signal in order to differentiate information-carrying sections of the audio signal from non-information-carrying sections of the audio signal, the information-carrying sections of the audio signal being separated from each other by the non-information-carrying sections of the audio signal, wherein the information-carrying sections and the non-information-carrying sections alternate within audio signal;

generating a sequence of test signal sections based on the information-carrying sections of the audio signal; and

transmitting the sequence of test signal sections to a measurement device for a non-intrusive quality assessment resulting in a quality measurement value per test signal section of the sequence of test signal sections.