Methods and apparatus for harmonic source enhancement

1. A system comprising:

an audio analyzer to:

determine a magnitude spectrogram of audio corresponding to a media signal;

generate a time-frequency mask based on the magnitude spectrogram; and

apply the time-frequency mask to the magnitude spectrogram to enhance a harmonic source of the media signal; and

an audio determiner to:

compare the media signal having the enhanced harmonic source with reference media signals to determine an instrument corresponding to the enhanced harmonic source; and

identify a genre corresponding to the media signal based on the identified instrument.

2. The system of claim 1 , wherein the audio analyzer determines the magnitude spectrogram by transforming the audio into a time-frequency representation using at least one of a Fourier transform, or a Constant Q transform.

3. The system of claim 2 , wherein the time-frequency representation includes respective columns corresponding to a spectrum at a particular instant.

4. The system of claim 3 , wherein the audio analyzer determines a Fourier transform at each of the respective columns.

5. The system of claim 4 , wherein the audio analyzer generates an inverse Fourier transform for each of the respective columns.

6. The system of claim 5 , wherein the audio analyzer divides the Fourier transform by an absolute value of the Fourier transform.

7. The system of claim 1 , wherein the system further comprises an audio analyzer to:

transmit the enhanced harmonic source to the audio determiner; and

obtain at least one of a classification of the audio or an identifier corresponding to the media signal from the audio determiner.

8. The system of claim 1 , wherein the audio determiner determines audio equalizer settings for the media based on classification, further including an audio equalizer to adjust equalizer settings of a media output device that output the media based on the determined audio equalizer settings.

9. The system of claim 1 , wherein the harmonic source is a first harmonic source, the audio determiner identifies the genre of the media based on comparison with a second harmonic source of the reference media signals.

10. A non-transitory computer readable medium comprising instructions that, when executed, cause one or more processors to at least:

determine a magnitude spectrogram of audio corresponding to a media signal;

generate a time-frequency mask based on the magnitude spectrogram;

apply the time-frequency mask to the magnitude spectrogram to enhance a harmonic source of the media signal;

compare the media signal having the enhanced harmonic source with reference media signals to determine an instrument corresponding to the enhanced harmonic source; and

identify a genre corresponding to the media signal based on the identified instrument.

11. The non-transitory computer readable medium of claim 10 , wherein the instructions further cause the one or more processors to transform the audio into a time-frequency representation using at least one of a Fourier transform, or a Constant Q transform.

12. The non-transitory computer readable medium of claim 11 , wherein the time-frequency representation includes respective columns corresponding to a spectrum at a particular instant.

13. The non-transitory computer readable medium of claim 12 , wherein the instructions further cause the one or more processors to determine a Fourier transform at each of the respective columns.

14. The non-transitory computer readable medium of claim 13 , wherein the instructions further cause the one or more processors to generate an inverse Fourier transform for each of the respective columns.

15. The non-transitory computer readable medium of claim 14 , wherein the instructions further cause the one or more processors to divide the Fourier transform by an absolute value of the Fourier transform.

16. A method comprising:

determining, by executing an instruction with one or more processors, a magnitude spectrogram of audio corresponding to a media signal;

generating, by executing an instruction with the one or more processors, a time-frequency mask based on the magnitude spectrogram;

applying, by executing an instruction with the one or more processors, the time-frequency mask to the magnitude spectrogram to enhance a harmonic source of the media signal;

comparing, by executing an instruction with the one or more processors, the media signal having the enhanced harmonic source with reference media signals to determine an instrument corresponding to the enhanced harmonic source; and

classifying, by executing an instruction with the one or more processors, a genre corresponding to the media signal based on the identified instrument.

17. The method of claim 16 , further including transforming the audio into a time-frequency representation using at least one of a Fourier transform, or a Constant Q transform, the time-frequency representation includes respective columns corresponding to a spectrum at a particular instant.

18. The method of claim 17 , further including determining a Fourier transform at each of the respective columns.

19. The method of claim 18 , further including generating an inverse Fourier transform for each of the respective columns.

20. The method of claim 19 , further including dividing the Fourier transform by an absolute value of the Fourier transform.

21. A system comprising:

at least one memory;

instructions; and

one or more processors to execute the instructions to:

determine a magnitude spectrogram of audio corresponding to a media signal;

generate a time-frequency mask based on the magnitude spectrogram;

apply the time-frequency mask to the magnitude spectrogram to enhance a harmonic source of the media signal;

compare the media signal having the enhanced harmonic source with reference media signals to determine an instrument corresponding to the enhanced harmonic source; and

identify at least one of a song, an advertisement, a show, or a video game based on the identified instrument.