Processing resonances in a digital sound signal using sub-band sub-sampling
This disclosure relates to signal processing of music and in particular the removal of unwanted artifacts from a music signal. A method for attenuating resonances in a sound signal which uses sub-sampling of at least one sub-band of the original signal to save computing resources is set out. The method includes at least steps in which an incoming signal block is filtered into at least two different frequency sub-band signals, at least one of the sub-band signals is subsampled, a resonance analysis step is performed for each sub-band signal, a step resulting in a filter response, a sub-band signal from each sub-band is convolved with the corresponding filter response to produce a filtered sub-band signal, for each subsampled sub-band, the filtered sub-band signal is upsampled, and finally the upsampled sub-band signals of subsampled sub-bands and filtered sub-band signals of any non-subsampled sub-bands are combined to form a result signal block.
1 . A method for processing resonances in a digital sound signal, the method comprising the steps of:
receiving a first block of new sound samples of the digital sound signal,
filtering the first block of the digital sound signal into at least two sub-band audio signals representing different frequency bands of the audio content in said first block of the digital sound signal, and
performing at least the following steps S1 to S7 for at least a first sub-band signal of said at least two sub-band audio signals:
S1) sub-sampling the sub-band audio signal,
S2) adding the sub-sampled signal to a first buffer for the corresponding frequency band,
S3) performing a resonance analysis step for the first buffer for the corresponding frequency band, thereby producing a filter response for the corresponding frequency band,
S4) convolving the sub-sampled signal or a first complete block of the first buffer with the filter response for the corresponding frequency band,
S5) adding the convolution result to a second buffer for the corresponding frequency band,
S6) upsampling a first complete block of said second buffer for the corresponding frequency band, and
S7) adding the upsampled result to a result buffer for the corresponding frequency band; and
summing first complete blocks of said result buffers for each frequency band to form an output block of an output digital sound signal.
2 . The method according to claim 1 , wherein at least the following steps are performed for at least a second sub-band signal of said at least two sub-band signals:
adding the sub-band signal to a first buffer for the corresponding frequency band,
performing a resonance analysis step for the first buffer for the corresponding frequency band, thereby producing a filter response for the corresponding frequency band,
convolving the sub-band signal or the first complete block of the first buffer with the filter response for the corresponding frequency band, and
adding the convolution result to a result buffer for the corresponding frequency band.
3 . The method according to claim 1 , wherein the step of performing resonance analysis on a buffer further comprises the steps of:
performing a fast Fourier transform on a plurality of latest blocks of the buffer,
converting the resulting complex response into a real response,
overpass filtering the real response to retain only peaks of said real response,
forming a resonance response based on said filtered real response, and
forming the filter response based on the resonance response.
4 . The method according to claim 3 , further comprising the step of windowing the data in the buffer before the step of performing a fast Fourier transform.
5 . The method according to claim 3 , further comprising the step of clipping the filtered real response after the step of overpass filtering to remove any signal in said filtered real response that is lower than a predetermined threshold.
6 . A sound processing device for processing resonances in a digital sound signal, comprising:
one or more processors; and
one or more computer-readable storage media comprising instructions stored thereon that, when executed by the one or more processors, cause the one or more processors to perform the steps of:
receiving a first block of new sound samples of the digital sound signal,
filtering the first block of the digital sound signal into at least two sub-band audio signals representing different frequency bands of the audio content in said first block of the digital sound signal, and
performing at least the following steps S1 to S7 for at least a first sub-band signal of said at least two sub-band audio signals:
S1) sub-sampling the sub-band signal,
S2) adding the sub-sampled signal to a first buffer for the corresponding frequency band,
S3) performing a resonance analysis step for the first buffer for the corresponding frequency band, thereby producing a filter response for the corresponding frequency band,
S4) convolving the sub-sampled signal or a first complete block of the first buffer with the filter response for the corresponding frequency band,
S5) adding the convolution result to a second buffer for the corresponding frequency band,
S6) upsampling a first complete block of said second buffer for the corresponding frequency band, and
S7) adding the upsampled result to a result buffer for the corresponding frequency band; and
summing first complete blocks of the result buffers for each frequency band to form an output block of an output digital sound signal.
7 . The sound processing device according to claim 6 , wherein the storage media further comprises instructions stored thereon that, when executed by the one or more processors, causes the one or more processors to perform the following steps for at least a second sub-band signal of the at least two sub-band signals:
adding the sub-band signal to a first buffer for the corresponding frequency band,
performing a resonance analysis step for the first buffer for the corresponding frequency band, thereby producing a filter response for the corresponding frequency band,
convolving the sub-band signal or the first complete block of the first buffer with the filter response for the corresponding frequency band, and
adding the convolution result to a result buffer for the corresponding frequency band.
8 . The sound processing device according to claim 6 , wherein storage media further comprises instructions stored thereon that, when executed by the one or more processors, cause the one or more processors to perform in the step of performing resonance analysis on a buffer the steps of:
performing a fast fourier transform on a plurality of latest blocks of said buffer,
converting the resulting complex response into a real response,
overpass filtering the real response to retain only peaks of said real response,
forming a resonance response based on said filtered real response, and
forming the filter response based on the resonance response.
9 . A non-transitory computer-readable storage media comprising instructions stored thereon that, when executed by a processor, cause the processor to perform the steps of:
receiving a first block of new sound samples of an output digital sound signal,
filtering the first block of an output digital sound signal into at least two sub-band audio signals representing different frequency bands of the audio content in said first block of an output digital sound signal, and
performing at least the following steps S1 to S7 for at least a first sub-band audio signal of the at least two sub-band audio signals:
S1) sub-sampling the sub-band audio signal,
S2) adding the sub-sampled signal to a first buffer for the corresponding frequency band,
S3) performing a resonance analysis step for the first buffer for the corresponding frequency band, thereby producing a filter response for the corresponding frequency band,
S4) convolving the sub-sampled signal or a first complete block of the first buffer with the filter response for the corresponding frequency band,
S5) adding the convolution result to a second buffer for the corresponding frequency band,
S6) upsampling a first complete block of said second buffer for the corresponding frequency band, and
S7) adding the upsampled result to a result buffer for the corresponding frequency band; and
summing first complete blocks of said result buffers for each frequency band to form an output block of an output digital sound signal.
10 . The computer-readable storage media according to claim 9 , wherein the storage media further comprises instructions stored thereon that, when executed by a processor, cause the processor to perform the following steps for at least a second sub-band signal of said at least two sub-band signals:
adding the sub-band signal to a first buffer for the corresponding frequency band,
performing a resonance analysis step for the first buffer for the corresponding frequency band, thereby producing a filter response for the corresponding frequency band,
convolving the sub-band signal or the first complete block of the first buffer with the filter response for the corresponding frequency band, and
adding the convolution result to a result buffer for the corresponding frequency band.
11 . The computer-readable storage media according to claim 9 , wherein the storage media further comprises instructions stored thereon that, when executed by a processor, cause the processor to perform in said step of performing resonance analysis on a buffer at least the steps of:
performing a fast Fourier transform on a plurality of latest blocks of said buffer,
converting the resulting complex response into a real response,
overpass filtering of the real response to retain only peaks of said real response,
forming a resonance response based on said filtered real response, and
forming the filter response based on the resonance response.