Determination of a common fundamental frequency of harmonic signals

• US 8,108,164 B2
• Filed: 01/26/2006
• Issued: 01/31/2012
• Est. Priority Date: 01/28/2005
• Status: Expired due to Fees

• Current Assignee: Honda Research Institute Europe GmbH (Honda Motor Company)
• Original Assignee: Honda Research Institute Europe GmbH (Honda Motor Company)
• Inventors: Joublin, Frank, Heckmann, Martin
• Primary Examiner(s): Barbee, Manuel L
• Application Number:
  US11/340,918
• Publication Number:
  US 20060195500A1
• Time in Patent Office: 2,196 Days
• Field of Search: 702/69, 702 75- 76, 702/106, 702189-191, 704205-208
• US Class Current: 702/75
• CPC Class Codes: G10L 25/90 Pitch determination of speech signals

Abstract

Techniques are provided for determining the time course of the fundamental frequency of harmonic signals, wherein the input signal is split into different frequency channels by band pass filters. Distances between crossings of different orders are determined, and a histogram of all these distance values for each instant in time is calculated. The distance values build a peak at the distance corresponding to the fundamental frequency. An example application of this technique is separation of acoustic sound sources in monaural recordings based on their underlying fundamental frequency. Application of these techniques, however, is not limited to the field of acoustics. These techniques can also be applied to other signals such as those originating from pressure sensors.

Claims

1. A non-transitory computer readable medium comprising computer executable code which when executed by a computer performs the steps of:

receiving a harmonic signal representing sound from multiple sound sources;

splitting the harmonic signal representing sound from multiple sound sources into a plurality of frequency channels;

determining, for each frequency channel in the plurality of frequency channels, distance between crossings of different orders including higher order crossings;

entering the distance crossing at all values between a current crossing and a precious crossing;

storing the distance in a three-dimensional representation together with a related time and frequency; and

calculating a histogram of the determined distances from the three-dimensional representation of different channels for every instant in time, indicating how often a certain distance value is identified; and

determining the fundamental frequency by identifying a maximum peak in the histogram and the distance associated with the maximum peak and multiplying the associated distance with a sampling rate.

2. A computer program product embodied on a non-transitory computer readable medium when executed performs the steps:

receiving the harmonic signal representing the sound from multiple sound sources;

splitting the harmonic signal representing sound from multiple sound sources into a plurality of frequency channels;

determining, for each frequency channel in the plurality of frequency channels, distance between crossings of different orders including higher order crossings;

entering the distance between crossing at all values between a current crossing and a previous crossing;

storing the distances in a three-dimensional representation together with a related time and frequency;

calculating a histogram of the determined distances from the three-dimensional representation of different channels for every instant in time, indicating how often a certain distance value is identified; and

determining the fundamental frequency by identifying a maximum peak in the histogram and the distance associated with the maximum peak and multiplying the associated distance with a sampling rate.

3. The computer program product of claim 2, wherein the crossings comprise one of:

a maxima;

a minima; and

a constant.

4. The computer program product of claim 2, wherein a band pass signal where center frequencies of band passes are in a harmonic relation or close to a harmonic relation is used to calculate the histogram.

5. The computer program product of claim 2, wherein an entry of the histogram is weighted with energy of an underlying band pass signal to make a distance of the fundamental frequency more discernable.

6. The computer program product of claim 2, wherein independent weights are used for a plurality of crossings of different orders in calculating the histogram.

7. The computer program product of claim 2, wherein determined distances resulting from unresolved harmonics are integrated in the histogram.

8. The computer program product of claim 2, further comprising evaluating an evidence value for a band pass signal to originate from the fundamental frequency for the instant in time, wherein a minimum distance between a crossing distance corresponding to the fundamental frequency and those corresponding to the band pass signal is used as the evidence value.

9. The computer program product of claim 2, further comprising suppressing peaks at multiples and integer fractions of a distance corresponding to the fundamental frequency, wherein a maximum value corresponding to the fundamental frequency at the instant in time is used to suppress the peaks at the multiples and the integer fractions at the instant in time.

10. The computer program product of claim 2, wherein the method is applied for separation of acoustic sound sources in monaural recordings.

11. The computer program product of claim 2, wherein said higher order crossings include at least one of a second, third, fourth, fifth or sixth order crossings.

12. The computer program product of claim 2, wherein said step of calculating a histogram further comprises the step of using only filter channels having center frequencies that have a substantially harmonic relationship with each other.

13. The computer program product of claim 2, further comprising the step of combining the distances of the plurality of frequency channels by determining the distances between the current zero-crossing and at least two previous zero-crossings.

14. The computer program product of claim 2, wherein said step of calculating a histogram further comprises the step of using only filter channels having center frequencies that have a substantially harmonic relationship with each other.

15. The computer program product of claim 2, further comprising the step of combining the distances of the plurality of frequency channels by determining the distances between the current zero-crossing and at least two previous zero-crossings.

16. The computer program product of claim 2, further comprising the step, prior to the storing step, of calculating a difference between the current crossing and the previous crossing.