RETAINING BINAURAL CUES WHEN MIXING MICROPHONE SIGNALS

Abstract

A method of mixing microphone signals. First and second microphone signals are obtained from respective first and second microphones. In at least one affected subband, the first and second microphone signals are mixed to produce first and second mixed signals. At least one reference subband of the first and second microphone signals is processed in order to identify a binaural cue between the first and second microphone signals, the reference subband being distinct from the or each affected subband. The affected subband in the first and second mixed signals is modified in order to re-emphasize the identified binaural cue.

Claims

1. A method of mixing microphone signals, the method comprising; obtaining first and second microphone signals from respective first and second microphones; in at least one affected subband, mixing the first and second microphone signals to produce first and second mixed signals; processing at least one reference subband of the first and second microphone signals in order to identify a binaural cue between the first and second microphone signals, the reference subband being distinct from the or each affected subband; and modifying the affected subband in the first and second mixed signals in order to re-emphasize the identified binaural cue.

2. The method of claim 1 wherein identifying the binaural cue comprises analysing the reference subband in the first and second signals in order to identify a level, magnitude or power difference between the first and second signals in the reference subband.

3. The method of claim 2 wherein modifying the affected subband in the first and second mixed signals comprises applying respective first and second emphasis gains to the first and second mixed signals in the or each affected subband.

4. The method of claim 3 wherein identifying the binaural cue comprises analysing the reference subband in the first and second signals in order to identify a time difference between the first and second microphone signals.

5. The method of claim 4 wherein modifying the affected subband in the first and second mixed signals comprises applying the time difference to the first and second mixed signals in the or each affected subband.

6. The method of claim 5 wherein the mixing comprises mixing the signals from at least two microphones, in low frequency subbands, so that the signal which is suffering from least wind noise in each of the low frequency subbands is preferentially used in that subband for further processing in both of the mixed signals.

7. The method of claim 6 wherein the mixing comprises mixing the signals from at least two microphones, in middle-to-high frequency subbands, so that the signal which is suffering from least lens focus motor noise in each of the affected subbands is preferentially used in that subband for further processing in both of the mixed signals.

8. A device for mixing microphone signals, the device comprising: first and second inputs for receiving respective first and second microphone signals from respective first and second microphones; and a digital signal processor configured to, in at least one affected subband, mix the first and second microphone signals to produce first and second mixed signals; the digital signal processor further configured to process at least one reference subband of the first and second microphone signals in order to identify a binaural cue between the first and second microphone signals, the reference subband being distinct from the or each affected subband; and the digital signal processor further configured to modify the affected subband in the first and second mixed signals in order to re-emphasize the identified binaural cue.

9. A non-transitory computer readable medium for mixing microphone signals, comprising instructions which, when executed by one or more processors, causes performance of the following: obtaining first and second microphone signals from respective first and second microphones; in at least one affected subband, mixing the first and second microphone signals to produce first and second mixed signals; processing at least one reference subband of the first and second microphone signals in order to identify a binaural cue between the first and second microphone signals, the reference subband being distinct from the or each affected subband; and modifying the affected subband in the first and second mixed signals in order to re-emphasize the identified binaural cue.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0028] An example of the invention will now be described with reference to the accompanying drawings, in which:

[0029] FIG. 1 is a schematic of a system for determining a mixing ratio in each of one or more affected subbands;

[0030] FIG. 2 is a schematic of a system for assessing inter-aural level differences in reference subbands in order to determine suitable emphasis gains to be applied to each of one or more affected subbands in accordance with a first embodiment of the invention;

[0031] FIG. 3 is a schematic of a system for applying emphasis gains to affected subbands in the embodiment of FIG. 2;

[0032] FIG. 4 is a schematic of a system for applying a time difference to affected subbands in accordance with another embodiment of the invention; and'

[0033] FIG. 5 is a schematic of a system for applying both emphasis gains and a time difference to affected subbands, in accordance with yet another embodiment of the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0034] Focus noise in video recording, being the noise of an auto focus motor of the lens of the video camera, is a situation where subband mixing between multiple microphone signals may be applied for example between about 4 kHz and 12 kHz. The following description uses subband signal mixing to ameliorate focus noise as an example, however it is to be appreciated that other embodiments of the present invention may be applied to low frequency subband mixing to address wind noise, for example.

[0035] FIG. 1 shows part of a system 100 for mixing 2 microphone signals. If it is supposed that the mic1 signal is more affected by focus noise than the mic 2 signal, then the system is configured to mix the microphone signals in affected subbands, and to use the mixed output as the new mic1 output, so that the mixed output suffers less noise as a result of the mixing. The inverse applies when the mic2 signal is more affected by noise. To achieve this, both microphone signals are analysed at 110, 112 using DFT or any other suitable subband analysis method, and the two selectors 120, 122 select which subbands are affected subbands that are to be mixed. The mixing ratio module 130 of FIG. 1 calculates the mixing ratio in each affected subband selected by the selectors. a.sub.j is the mixing ratio applied on mic1 and (1a.sub.j) is the mixing ratio applied on mic2, and j is the subband index. In this mixing procedure, stereo or binaural cues will be diminished or lost because the mixed signal and mic2 signal are being made more similar or even identical in each affected subband.

[0036] FIG. 2 is a schematic of a system 200 for assessing inter-aural level differences in reference subbands in order to determine suitable emphasis gains to be applied to each of one or more affected subbands in accordance with a first embodiment of the invention. The two selectors 220, 222 select which subbands are affected subbands that are to be mixed. The Interaural level differences (ILD) module 230 calculates the inter aural level differences D.sub.j (also referred to as ILD.sub.j). The emphasis gains module 240 uses the D.sub.j and .sub.j values to calculate emphasis gains G.sub.j using the equation:

G.sub.j=(1.sub.j)*(ILD.sub.j1)+1

[0037] The gain G.sub.j is one (0 dB gain) if the mixing ratio is 1 (no mixing), or if the ILD.sub.j is 1 (i.e. mic1 and mic2 signals are of the same level). The calculation of Gj in other embodiments can take different forms, such as:

G.sub.j=(1.sub.j).sup.2*(ILD.sub.j1)+1;

[0038] FIG. 3 shows the subband gains being applied on both microphones before mixing. The emphasis gains are applied to emphasize the difference between the mixed output and the mic2 output, and thereby re-emphasise binaural cues carried by such level differences. The total subband gains (including mixing, emphasis gain) applied by block 320 on mic1 are aj*Gj. The total subband gains applied by block 322 on mic2 are (1aj)*Gj.

[0039] FIG. 4 shows an embodiment in which a time difference is applied by block 440 on the mixed output, in order to re-emphasise binaural cues. A fixed delay is applied by block 442 on mic2 in case the time difference is a negative value, i.e. when sounds arrive at mic1 earlier than at mic2. In this embodiment, the time difference of arrival (TDOA) between the two microphones is calculated using a generalized correlation method (C. H. Knapp and G. C. Carter, The generalized correlation method for estimation of time delay, IEEE Trans. Acoust., Speech, Signal Processing vol. 24, pp. 320-327, August 1976). The time difference is then applied on the mixed output for those subbands affected by noise, so that after the mixing the mixed output and mic2 will have the same time difference as the original mic1 and mic2 signals, thus better preserving binaural cues. The fixed delay applied at 442 is the microphone spacing between mic1 and mic2 divided by the sampling rate.

[0040] In alternative embodiments similar to FIG. 4, the time difference of arrival could instead be calculated during the IDFT stage using the phase shift of reference subbands.

[0041] FIG. 5 illustrates yet another embodiment of the invention in which both a time delay 540 and emphasis gains G.sub.j are used to reemphasise binaural cues.

[0042] It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.