Method and Apparatus for Detecting Audio Signals

• US 20110194702A1
• Filed: 04/25/2011
• Published: 08/11/2011
• Est. Priority Date: 10/15/2009
• Status: Active Grant

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Inventors: Wang, Zhe
• Granted Patent: US 8,050,415 B2
• Time in Patent Office: Days
• US Class Current: 381/17
• CPC Class Codes: G10H 2210/046 for differentiation between music and non-music signals, based on the identification of musical parameters, e.g. based on tempo detection
• CPC Class Codes: G10H 2250/235 Fourier transform; Discrete Fourier Transform [DFT]; Fast Fourier Transform [FFT]
• CPC Class Codes: G10H 2250/571 Waveform compression, adapted for music synthesisers, sound banks or wavetables audio compression G10L19/00
• CPC Class Codes: G10L 25/81 for discriminating voice from music

Abstract

A method and an apparatus for detecting audio signals are disclosed. The input audio signal is detected to determine whether it is a background frame. The detected background signal is further detected according to a music characterization value and a decision rule. Therefore, background music can be detected, and the classifying performance of the voice/music classifier is improved.

Claims

1. A method for detecting audio signals, the method comprising:

dividing an input audio signal into multiple audio signal frames;

detecting each of the audio signal frames to determine whether the audio signal frame is a background signal frame;

adding a step value to a background frame counter when a background signal frame is detected;

obtaining a music characterization value of the background signal frame, and adding the music characterization value to an accumulated background music characterization value; and

comparing the accumulated background music characterization value with a threshold when the background frame counter reaches a preset number, and determining that the input audio signal is background music if the accumulated background music characterization value fulfills a threshold decision rule.

2. The method according to claim 1, wherein obtaining the music characterization value of the background signal frame comprises:

obtaining a spectrum of the background signal frame;

obtaining positions and energy values of local peak points in at least a part of the spectrum;

calculating a normalized peak-valley distance corresponding to every local peak point according to the position and energy value to obtain multiple normalized peak-valley distance values; and

obtaining the music characterization value according to the multiple normalized peak-valley distance values.

3. The method according to claim 2, wherein calculating the normalized peak-valley distance of each of the local peak points comprises:

for each of the local peak points, obtaining a minimum value among four frequencies adjacent to the left side of the local peak point and a minimum value among four frequencies adjacent to the right side of the local peak point;

calculating a difference between the local peak point and the minimum value among the four frequencies adjacent to the left side, and a difference between the local peak point and the minimum value among the four frequencies adjacent to the right side; and

dividing a sum of the two differences by an average energy value of the spectrum or an average energy value of the part of the spectrum to generate the normalized peak-valley distance.

4. The method according to claim 2, wherein calculating the normalized peak-valley distance of each of the local peak points comprises:

for each of the local peak points, calculating a distance between the local peak point and at least one frequency to the left side of the local peak point, and calculating a distance between the local peak point and at least one frequency to the right side of the local peak point; and

dividing a sum of the two differences by an average energy value of the spectrum or the part of the spectrum to generate the normalized peak-valley distance.

5. The method according to claim 2, wherein obtaining the music characterization value according to the multiple normalized peak-valley distance values comprises:

selecting a maximum value of the normalized peak-valley distance values as the music characterization value;

or

adding up at least two maximum values of the normalized peak-valley distance values to obtain the music characterization value.

6. The method according to claim 2, wherein the threshold decision rule comprises a rule wherein the accumulated background music characterization value is greater than the threshold.

7. The method according to claim 1, wherein obtaining the music characterization value of the background signal frame comprises:

according to a spectrum of the background signal frame, obtaining a first position of a frequency whose peak-valley distance is greatest among all local peak values on the spectrum;

according to a spectrum of a frame before the background signal frame, obtaining a second position of the frequency whose peak-valley distance is the greatest among all local peak values on the spectrum of the frame before the background signal frame; and

calculating a difference between the first position and the second position to obtain the music characterization value.

8. The method according to claim 7, wherein the threshold decision rule comprises a rule wherein the accumulated background music characterization value is less than the threshold.

9. The method according to claim 1, wherein:

the threshold is adjusted according to a protection frame value, such that if the protection frame value is greater than 0, a first threshold is applied, and if the protection frame value is not greater than 0, a second threshold is applied.

10. The method according to claim 1, wherein after determining that the input audio signal is background music, the method further comprises:

identifying a preset number of audio frames after a current audio frame as the background music.

11. The method according to claim 10, further comprising:

decreasing a preset protection frame value by 1 when the background signal frame is detected; and

applying a first threshold if the protection frame value is greater than 0, and applying a second threshold if the protection frame value is not greater than 0, wherein the first threshold is less than the second threshold if the threshold decision rule indicates that the accumulated background music characterization value is greater than the threshold, and wherein the first threshold is greater than the second threshold if the threshold decision rule indicates that the accumulated background music characterization value is less than the threshold.

12. A coder, comprising:

a background frame recognizer configured to detect each input audio signal frame of a plurality of input audio signal frames and to output a first detection result indicating whether the audio signal frame is a background signal frame; and

a background music recognizer configured to detect the background signal frame according to a music characterization value of the background signal frame once the background signal frame is detected and to output a second detection result indicating that background music is detected, wherein the background music recognizer comprises;

a background frame counter configured to add a step value to the counter once the background signal frame is detected;

a music characterization value obtaining unit configured to obtain the music characterization value of the background signal frame;

a music characterization value accumulator configured to accumulate the music characterization value of the background signal frame; and

a decider configured to determine that the accumulated music characterization value fulfills a threshold decision rule when the background frame counter reaches a preset number and to output the second detection result indicating that the background music is detected.

13. The coder according to claim 12, wherein the music characterization value obtaining unit comprises:

a spectrum obtaining unit configured to obtain a spectrum of the background signal frame;

a peak point obtaining unit configured to obtain local peak points in at least a part of the spectrum; and

a calculating unit configured to calculate a normalized peak-valley distance corresponding to each obtained local peak point to obtain multiple normalized peak-valley distance values and to obtain the music characterization value according to the multiple normalized peak-valley distance values.

14. The coder according to claim 13, wherein the normalized peak-valley distance of each obtained local peak point is calculated as follows:

for each obtained local peak point, obtaining a minimum value among four frequencies adjacent to the left side of the local peak point and a minimum value among four frequencies adjacent to the right side of the local peak point; and

calculating a difference between the obtained local peak value and the minimum value among the four frequencies adjacent to the left side, and a difference between the local peak value and the minimum value among the four frequencies adjacent to the right side, and dividing a sum of the two differences by an average energy value of the spectrum or an average energy value of the part of the spectrum to generate the normalized peak-valley distance.

15. The coder according to claim 13, wherein the normalized peak-valley distance of each obtained local peak point is calculated as follows:

for each obtained local peak point, calculating a distance between the obtained local peak point and at least one frequency to the left side of the obtained local peak point, and calculating a distance between the obtained local peak point and at least one frequency to the right side of the obtained local peak point; and

dividing a sum of the two differences by an average energy value of the spectrum or the part of the spectrum to generate the normalized peak-valley distance.

16. The coder according to claim 12, wherein the music characterization value obtaining unit comprises:

a first position obtaining unit configured to obtain a spectrum of the background signal frame and to obtain a first position of a frequency whose peak-valley distance is greatest among all local peak values on the spectrum;

a second position obtaining unit configured to obtain a spectrum of a frame before the background signal frame and to obtain a second position of the frequency whose peak-valley distance is the greatest among all local peak values on the spectrum of the frame before the background signal frame; and

a calculating unit configured to calculate a difference between the first position and the second position to obtain the music characterization value.

17. The coder according to claim 12, further comprising:

an identifying unit configured to identify a preset number of audio frames after a current audio frame as the background music.

18. The coder according to claim 17, further comprising:

a threshold adjusting unit configured to;

decrease a preset protection frame value by 1 when the background signal frame is detected; and

apply a first threshold if the protection frame value is greater than 0, and apply a second threshold if the protection frame value is not greater than 0, wherein the first threshold is less than the second threshold if the threshold decision rule indicates that the accumulated music characterization value is greater than the threshold, and wherein the first threshold is greater than the second threshold if the threshold decision rule indicates that the accumulated music characterization value is less than the threshold.

19. The coder according to claim 12, wherein:

the decider is further configured to determine that the accumulated music characterization value does not fulfill the threshold decision rule when the background frame counter reaches the preset number and to output a third detection result indicating that non-background music is detected.