Systems and methods for determining a repeatogram in a music composition using audio features
A dataset representing repeated sounds within a musical composition recorded on an audio track may be constructed. An audio track duration of an audio track may be partitioned into partitions of a partition size. A current partition may be compared to remaining partitions of the audio track. Audio information for the current partition may be correlated to audio information for remaining partitions to determine a correlated partition for the current partition from among the remaining partitions of the track duration. The correlated partition determined may be identified as most likely to represent the same sound as the current partition. This comparison process may be performed iteratively, for individual ones of the remaining partitions. Correlation results of the comparison process may be recorded to represent the partition time period of the correlated partition as a function of partition time period of the current partition.
1. A method for constructing a dataset for representing repeated sounds within an audio, comprising:
obtaining audio information for an audio track recorded over a track duration;
partitioning the track duration into multiple partitions, individual partitions occurring at different time periods along the track duration;
comparing one or more given partitions of the multiple partitions to one or more other partitions of the multiple partitions to determine one or more correlated partitions for the one or more given partitions, individual ones of the correlated partitions being identified as most likely to represent the same sound as individual ones of the given partitions;
recording the one or more correlations between the one or more given partitions and the one or more correlated partitions; and
organizing the recorded correlations to represent the time period of the one or more correlated partitions as a function of the time period of the one or more given partitions.
2. The method of claim 1 , wherein the audio track is generated from a media file, the media file including audio and video information.
3. The method of claim 1 , wherein the audio information including an audio energy representation.
4. The method of claim 3 , wherein the audio energy representation is filtered into a frequency band to produce a frequency energy representation, the frequency energy representation representing individual energy samples associated with sound in the frequency band captured on the audio track.
5. The method of claim 3 , wherein the comparison includes correlating individual energy samples in the frequency band of the one or more given partitions with individual energy samples in the frequency band in the one or more other partitions.
6. The method of claim 3 , wherein the audio energy representations of the individual partitions are transformed into frequency space in which energy is represented as a function of frequency.
7. The method of claim 6 , wherein individual transformed representations of the individual partitions include identifying pitches of harmonic sound and determining magnitudes of harmonic energy at harmonics of the harmonic sound.
8. The method of claim 7 , wherein the comparison includes correlating pitch of the harmonic sound and harmonic energy of the transformed representation of the one or more given partitions with pitch of the harmonic sound and harmonic energy of transformed representations in the one or more other partitions.
9. The method of claim 1 , wherein the one or more other partitions are selected for the comparison based on one or more sizes of the multiple partitions.
10. The method of claim 1 , wherein the one or more other partitions are selected for the comparison based on a stopping criteria.
11. A system for constructing a dataset for representing repeated sounds within an audio, comprising:
one or more physical processors configured to machine readable instructions to:
obtain audio information for an audio track recorded over a track duration;
partition the track duration into multiple partitions, individual partitions occurring at different time periods along the track duration;
compare one or more given partitions of the multiple partitions to one or more other partitions of the multiple partitions to determine one or more correlated partitions for the one or more given partitions, individual ones of the correlated partitions being identified as most likely to represent the same sound as individual ones of the given partitions;
record the one or more correlations between the one or more given partitions and the one or more correlated partitions; and
organize the recorded correlations to represent the time period of the one or more correlated partitions as a function of the time period of the one or more given partitions.
12. The system of claim 11 , wherein the audio track is generated from a media file, the media file including audio and video information.
13. The system of claim 11 , wherein the audio information including an audio energy representation.
14. The system of claim 13 , wherein the audio energy representation is filtered into a frequency band to produce a frequency energy representation, the frequency energy representation representing individual energy samples associated with sound in the frequency band captured on the audio track.
15. The system of claim 13 , wherein the comparison includes correlating individual energy samples in the frequency band of the one or more given partitions with individual energy samples in the frequency band in the one or more other partitions.
16. The system of claim 13 , wherein the audio energy representations of the individual partitions are transformed into frequency space in which energy is represented as a function of frequency.
17. The system of claim 16 , wherein individual transformed representations of the individual partitions include identifying pitches of harmonic sound and determining magnitudes of harmonic energy at harmonics of the harmonic sound.
18. The system of claim 17 , wherein the comparison includes correlating pitch of the harmonic sound and harmonic energy of the transformed representation of the one or more given partitions with pitch of the harmonic sound and harmonic energy of transformed representations in the one or more other partitions.
19. The system of claim 11 , wherein the one or more other partitions are selected for the comparison based on one or more sizes of the multiple partitions.
20. The system of claim 11 , wherein the one or more other partitions are selected for the comparison based on a stopping criteria.