Dry sound and ambient sound separation

An audio module configured to separate an audio input signal into a dry signal component and an ambient signal component, the audio module comprising a memory and at least one processing unit, the memory comprising instructions executable by the at least one processing unit, to cause the audio module to:

• generate a transferred input signal including transferring the audio input signal into a frequency space,
• apply a smoothing filter to the transferred input signal to generate an estimated ambient signal component, wherein the smoothing filter suppresses the dry signal component of the audio input signal to generate the estimated ambient signal component,
• determine the dry signal component based on the estimated ambient signal component, and
• determine the ambient signal component based on the determined dry signal component and the audio input signal,
• determine magnitudes of the transferred audio input signal,
• attenuate the determined magnitudes of the transferred audio input signal before applying a smoothing filter,
• extract fluctuations from attenuated magnitudes of the transferred input signal, and
• add the extracted fluctuation to an output of the smoothing filter to determine the estimated ambient signal component.

A method for separating an audio input signal into a dry signal component and an ambient signal component, the method comprising:

• transferring the audio input signal into a frequency space to generate a transferred input signal,
• applying a smoothing filter to the transferred input signal to generate an estimated ambient signal component,
• determining the dry signal component based on the estimated ambient signal component, and
• determining the ambient signal component based on the determined dry signal component and the audio input signal,
• wherein applying the smoothing filter to the transferred input signal to generate the estimated ambient signal component includes suppressing the dry signal component of the audio input signal with the smoothing filter to generate the estimated ambient signal component,
• wherein the audio input signal comprises a combined signal period, where the ambient signal component and the dry signal component are present and a reverberation signal period following the combined signal period, and where substantially only the ambient signal is present, the method further comprising:
• detecting the reverberation signal period in the audio input signal,
• determining an estimated reverberation signal period component in the estimated ambient signal component in the reverberation signal period, and
• comparing the estimated reverberation signal component to the audio input signal in the reverberation signal period to determine a smoothing coefficient used in the smoothing filter for generating the estimated ambient signal component,
• wherein the method further comprises:
• determining whether the reverberation signal period is currently present in the audio input signal,
• utilizing the audio input signal in the reverberation signal period for the estimated ambient signal component if the reverberation signal period is present in the audio input signal, and
• utilizing an output of the smoothing filter for the estimated ambient signal component if the reverberation signal period is not present in the audio input signal.

A method for separating an audio input signal into a dry signal component and an ambient signal component, the method comprising:

• generating a transferred input signal including transferring the audio input signal into a frequency space,
• applying a smoothing filter to the transferred input signal to generate an estimated ambient signal component,
• determining the dry signal component based on the estimated ambient signal component, and
• determining the ambient signal component based on the determined dry signal component and the audio input signal,
• wherein applying the smoothing filter to the transferred input signal to generate the estimated ambient signal component includes suppressing the dry signal component of the audio input signal with the smoothing filter to generate the estimated ambient signal component,
• determine magnitudes of the transferred audio input signal,
• attenuating the determined magnitudes of the transferred audio input signal before applying a smoothing filter,
• extracting fluctuations from attenuated magnitudes of the transferred input signal, and
• adding the extracted fluctuation to an output of the smoothing filter to determine the estimated ambient signal component.

A method for separating an audio input signal into a dry signal component and an ambient signal component, the method comprising:

• generating a transferred input signal including transferring the audio input signal into a frequency space,
• applying a smoothing filter to the transferred input signal to generate an estimated ambient signal component,
• determining the dry signal component based on the estimated ambient signal component, and
• determining the ambient signal component based on the determined dry signal component and the audio input signal,
• wherein applying the smoothing filter to the transferred input signal to generate the estimated ambient signal component includes suppressing the dry signal component of the audio input signal with the smoothing filter to generate the estimated ambient signal component,
• determining an onset in the audio input signal where at least a portion of the dry signal component is present in the audio input signal, and
• determining an offset in the audio input signal where at least the portion of the dry signal component is no longer present in the audio input signal, p1 wherein a reverberation signal period is determined as starting after the offset and ending when an output of the smoothing filter meets a predefined criterion, and wherein comparing the estimated reverberation signal component to the audio input signal in the reverberation signal period is carried out in a part of the audio inout signal after the determined offset.

A method for separating an audio input signal into a dry signal component and an ambient signal component, the method comprising:

• generating a transferred input signal including transferring the audio input signal into a frequency space,
• applying a smoothing filter to the transferred input signal to generate an estimated ambient signal component,
• determining the dry signal component based on the estimated ambient signal component, and
• determining the ambient signal component based on the determined dry signal component and the audio input signal,
• wherein applying the smoothing filter to the transferred input signal to generate the estimated ambient signal component includes suppressing the dry signal component of the audio input signal with the smoothing filter to generate the estimated ambient signal component,
• determining an onset in the audio input signal where at least a portion of the dry signal component is present in the audio input signal,
• determining an offset in the audio input signal where at least the portion of the dry signal component is no longer present in the audio input signal, and p1 determining a peak follower signal with exponential decay from the transferred input signal which is then attenuated by a factor smaller than 1, wherein the offset is detected when the audio inout signal crosses the peak follower signal at a point where the audio input signal decreases.

A method for separating an audio input signal into a dry signal component and an ambient signal component, the method comprising:

• generating a transferred input signal including transferring the audio input signal into a frequency space,
• applying a smoothing filter to the transferred input signal to generate an estimated ambient signal component,
• determining the dry signal component based on the estimated ambient signal component, and
• determining the ambient signal component based on the determined dry signal component and the audio input signal,
• wherein applying the smoothing filter to the transferred input signal to generate the estimated ambient signal component includes suppressing the dry signal component of the audio input signal with the smoothing filter to generate the estimated ambient signal component,
• determining an onset in the audio input signal where at least a portion of the dry signal component is present in the audio input signal,
• determining an offset in the audio input signal where at least the portion of the dry signal component is no longer present in the audio input signal, and
• determining a ratio of the ambient signal component to the audio input signal, wherein determining the ration includes:
• p1 determining an amplitude of the audio inout signal in a last sample of the audio input signal before the offset, p1 determining an amplitude of the audio inout signal in a first sample of the audio input signal before the offset, and p1 determining the ratio based on the amplitude in the last sample and the amplitude in the first sample.