This application is directed, in general, to microphones arrays and, more specifically, to a microphone array for capturing multiple audio channels.
Audio recording using one-dimensional (1-D) (i.e. linear) or two-dimensional (2-D) (i.e. planar) microphone arrays to capture stereo or surround ambience is a well-established practice (see, e.g., Rayburn, “Eargle's The Microphone Book: From Mono to Stereo to Surround: A Guide to Microphone Design and Application,” Focal Press, 2011; Rumsey, “Spatial Audio,” Focal Press, 2001; Gerzon, “The Design of Precisely Coincident Microphone Arrays for Stereo and Surround Sound,” 50th Audio Engineering Society Convention, London, March 1975; Williams, “Migration of 5.0 Multichannel Microphone Array Design to Higher Order MMAD (6.0, 7.0 & 8.0) With or Without the Inter-format Compatibility Criteria,” Paper 7480, 124th Audio Engineering Society (AES) Convention, May 2008; and Yong, et al., “Sound Source Localization for Circular Arrays of Directional Microphones,” Proc. IEEE ICASSP, pp. 93-96, March 2005. Commercially available 2-D microphone arrays include the Soundfield SPS200 SW controlled microphone from TSL Professional Products Ltd. of Marlow, UK, and the Zoom H2N surround/stereo audio recorder from Samson Technologies of Hauppauge, N.Y., USA. These conventional arrays consist of a few closely spaced bidirectional or unidirectional (e.g., cardioid) microphones and have proven relatively effective in generating multiple audio channels.
One aspect provides a system for generating multiple audio channels. In one embodiment, the system includes: (1) an array of omnidirectional microphones and (2) a beamformer coupled to the array and operable to transform signals produced by the array into multiple directional audio channels.
Another aspect provides a method of generating multiple audio channels. In one embodiment, the method includes: (1) producing signals from each of an array of omnidirectional microphones and (2) employing a beamforming technique to transform at least some of the signals into multiple directional audio channels.
Yet another aspect provides a module for an audio recorder/transmitter. In one embodiment, the module includes: (1) a shell, (2) an array of omnidirectional microphones coupled to the shell, (3) a beamformer coupled to the array and operable to transform signals produced by the array into multiple directional audio channels and (4) an interface coupled to the beamformer and operable to convey the multiple directional audio channels into the audio recorder/transmitter.
Reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:
As stated above, conventional microphone arrays generally consist of a few closely spaced directional (e.g., bidirectional or unidirectional) microphones. Unfortunately, directional microphones suffer from known limitations, including proximity effect and heightened wind noise sensitivity. Low-cost directional microphones are particularly susceptible to off-axis coloration. These shortcomings require compensation, typically taking the form of baffles and high-pass filters. Directional microphones also require proper placement within the array and carefully designed acoustic packaging that allows the directivity to be preserved. All of these contribute to the cost of any product that includes such an array.
It is realized herein that omnidirectional microphones have several advantages over directional microphones, at least in terms of proximity effect and wind noise sensitivity. Further, they do not exhibit off-axis coloration. Some conventional stereo/surround microphone arrays do use omnidirectional microphones. However, these require large spatial dimensions, such as the “Polyhymnia Pentagon” described in Kamekawa, “An Explanation of Various Surround Microphone Techniques,” http://www.sanken-mic.com/en/qanda/index.cfm, or intricately designed acoustic isolation (via baffles and acoustic tubes) between the microphones as in the DPA5100 system described in Nymand, “Developing the 5100 Mobile Surround Mic,” Resolution, April 2009, http://www.dpamicrophones.com/en/Microphone-University/Surround Techniques/-/media/PDF/MicUni/Resolution—5100.pdf. It is thus realized herein that a microphone array employing omnidirectional microphones may have substantial advantages over one employing directional microphones.
Beamforming can be used to provide directivity using as few as two omnidirectional microphones arranged in a closely spaced, end-fire linear array (which may be a “sub-array,” defined as a portion of a larger array). For example, the U.S. patent application on which priority hereof is claimed and incorporated herein by reference teaches various beamforming techniques applicable to omnidirectional microphones. A beamformer to be described below may apply the techniques taught therein or other conventional or later-developed techniques for processing signals produced by various microphone arrays introduced herein.
Microphone beamforming is conventionally employed to suppress directional interference or ambient noise. In the present context however, it is realized herein that microphone beamforming may also be employed to obtain directional response along desired directions.
For XY stereo, desired directivity may be achieved with a pair of cardioid beams having a specified angular separation. For mid-side (MS) stereo, desired directivity may be achieved with a forward-looking cardioid beam and a side-looking bi-directional beam and appropriate mixing. Surround sound acquisition may involve a mix of cardioid, hyper-cardioid or even more directional beams, depending on the microphone and computational resources that are available.
For example, microphone beamforming may be employed to form two cardioid beams in opposite directions along the axis of a dual-microphone end-fire array consisting of a first microphone M1 and a second microphone M2. The array can thus be considered as two virtual arrays: a first sub-array formed by M1-M2, and a second sub-array formed by M2-M1. (This notation will be used throughout this disclosure to denote an ordering of microphones in a sub-array.) Microphone beamforming may be carried out to form these two beams simultaneously; thus they are not mutually exclusive. For applications in which acquisition of stereo or surround audio is required, end-fire arrays leveraging microphone beamforming can be exploited across a variety of 2-D microphone array geometries having multiple microphones. Accordingly, introduced herein are various embodiments of microphone arrays appropriate for yielding multiple audio channels, e.g., stereo or surround audio channels. Various embodiments employ microelectromechanical systems (MEMS) omnidirectional microphones, electret microphones or combinations thereof.
It is realized herein that the inter-microphone spacing between microphones M1 and M2 sets a limit on the highest audio signal frequency beyond which spatial aliasing can occur. The spacing should be less than half the wavelength of the highest frequency signal to be processed. For example, for wideband voice applications where the sampling rate is about 16 KHz, the microphone spacing should be within about 21.2 mm. For full-band (20 KHz) audio applications with sampling rates of 44.1 KHz or 48 KHz, the spacing should be within about 8.5 mm.
Some microphone array embodiments described herein employ a spacing of about 8 mm for at least some microphones in the array embodiments. As a result, certain of the microphone array embodiments may have a relatively compact footprint. In one microphone array embodiment having seven microphones, the footprint is less than 4 cm2.
Certain embodiments of the microphone array described herein use different combinations of microphones to capture different frequency bands. As described above, microphone spacing should be within about 8.5 mm to support a full audio bandwidth, nominally defined as about 20 KHz. However, it is realized herein that microphone spacing and directionality bear a direct relationship. Consequently, a relatively close microphone spacing has the effect of reducing the directional performance at low frequencies. Directivity can be improved by increasing the microphone spacing, but higher frequencies will begin to alias, preventing the desired full 20 KHz audio bandwidth from being supported.
It is realized herein that directivity can be preserved, and aliasing resisted by using different combinations of microphones located along the same line. Thus, in various embodiments described herein, two or more microphones having a closer spacing are employed to generate a band of higher frequencies, and two or more microphones having a wider spacing are employed to generate a band of lower frequencies. The two bands can then be combined to yield the desired audio bandwidth. Allocating of microphones between or among multiple bands may be termed “nesting” herein.
The array has a first microphone M1, second microphone M2 and a third microphone M3 located at vertices of a triangle. The first microphone M1 and the third microphone M3 are operable to provide signals transformable into a first left channel L1 and/or a third right channel R3 (i.e., sub-arrays M1-M3 and M3-M1). The second microphone M2 and the third microphone M3 are operable to provide signals transformable into a first right channel R1 and/or a third left channel L3 (i.e., sub-arrays M2-M3 and M3-M2). The first microphone M1 and the second microphone M2 are operable to provide signals transformable into at least one of a second left channel L2 and/or a second right channel R2.
As
For front surround, the three-microphone embodiment of
For mid-surround, two XY stereo pairs with different width may be generated. Assuming 2θ equals 90°, the first left and right channels L1, R1 may be obtained using beamforming (sub-arrays M1-M3 and M2-M3, respectively). The second left and right channels L2, R2 having a 180° angular separation may be obtained using microphones (sub-arrays M1-M2 and M2-M1, respectively).
As
Assuming 2θ equals 90°, a center channel C may be formed using M2-M4. XY stereo operation is possible with 90° or 180° separation. XY-90° stereo may be achieved using microphones M1-M4 for the first left channel L1 and microphones M3-M4 for the first right channel R1. XY-180° stereo may be achieved using sub-array M1-M3 for the second left channel L2 and sub-array M3-M1 for the second right channel R2. MS stereo may be achieved using the center channel (sub-array M2-M4) and a bi-directional response from microphones M1 and M3 (e.g., sub-array M1-M3).
The array of
In the illustrated embodiment, the spacing between microphones M1 and M4, M2 and M4 and M3 and M4 is uniform. In one specific embodiment, the spacing is about 21 mm, which (at a KHz sampling rate) makes the array suitable for wideband voice applications.
As
As
Before describing further microphone array embodiments, it should be noted that the channel nomenclature employed herein (e.g., left, right, center) are relative to an implicit primary axis (often pointing to the subject of a recording, such as an orator or a musical band, or parallel with the optical axis of a lens). Those skilled in the pertinent art should understand that such arrays may be rotated with respect to such primary axis. In such case, while the relative positions of the microphones remains constant, the nomenclature given to the audio channels produced by beamforming would change (e.g., a right channel might become a center channel; a center channel might become a left channel; and so forth). Array rotations will therefore not be further discussed.
In
XY-180° stereo is obtained using (M1-M2) and (M2-M1) sub-arrays. The other stereo separation angles are determined by θ and φ. For example, these angles can be selected such that array provides L, R stereo pairs at 60°, 120° and 180°.
No further reference will be made to angular separation between channels or linear separation between microphones for the following illustrated embodiments. Those skilled in the pertinent art will understand that wide variations are possible in both without departing from the scope of the invention.
The first microphone M1 and the sixth microphone M6 are operable to provide signals transformable into a first left channel L1 and/or a thirteenth left channel L13. The second microphone M2 and the seventh microphone M7 are operable to provide signals transformable into a first right channel R1 and/or a thirteenth right channel R13. At least two of the first microphone M1, the fourth microphone M4 and the seventh microphone M7 are operable to provide signals transformable into a second left channel L2 and/or a twelfth right channel R12. At least two of the second microphone M2, the fourth microphone M4 and the sixth microphones M6 are operable to provide signals transformable into a second right channel R2 and/or a twelfth left channel L12. The first microphone M1 and the fifth microphone M5 are operable to provide signals transformable into a third left channel L3 and/or an eighth right channel R8. The second microphone M2 and the third microphone M3 are operable to provide signals transformable into a third right channel R3 and/or an eighth left channel L8. The first microphone M1 and the second microphone M2 are operable to provide signals transformable into at least one of a fourth left channel L4 and a fourth right channel R4. The third microphone M3 and the sixth microphone M6 are operable to provide signals transformable into a fifth left channel L5 and/or a fifteenth right channel R15. The fifth microphone M5 and the seventh microphone M7 are operable to provide signals transformable into a fifth right channel R5 and/or a fifteenth left channel L15. The third microphone M3 and the seventh microphone M7 are operable to provide signals transformable into a sixth left channel L6 and/or an eleventh right channel R11.
The fifth microphone M5 and the sixth microphone M6 are operable to provide signals transformable into a sixth right channel R6 and/or an eleventh left channel L11. At least two of the third microphone M3, the fourth microphone M4 and the fifth microphone M5 are operable to provide signals transformable into at least one of a seventh left channel L7 and a seventh right channel R7. The sixth microphone M6 and the seventh microphone M7 are operable to provide signals transformable into at least one of a tenth left channel L10 and a tenth right channel R10. The first microphone M1 and the third microphone M3 are operable to provide signals transformable into a ninth left channel L9 and/or a fourteenth right channel R14. The second microphone M2 and the fifth microphone M5 are operable to provide signals transformable into a ninth right channel R9 and/or a fourteenth left channel L14.
Nesting may be employed to yield different frequency bands. For example, the second left channel L2 may be formed with a combination of an upper frequency band and a lower frequency band in which the first, fourth and seventh microphones M1, M4, M7 form a three-microphone sub-array for the upper frequency band and the first and seventh microphones M1, M7 form a two-microphone sub-array for the lower frequency band. Likewise, the seventh left channel L7 may be formed with a combination of an upper frequency band and a lower frequency band in which the third, fourth and fifth microphones M3, M4, M5 form a three-microphone sub-array for the upper frequency band and the third and fifth microphones M3, M5 form a two-microphone sub-array for the lower frequency band.
The second microphone M2 and the sixth microphone M6 are operable to provide signals transformable into a first left channel L1 and/or an eighth right channel R8. The second microphone M2 and the fourth microphone M4 are operable to provide signals transformable into a first right channel R1 and/or an eighth left channel L8. At least two of the first microphone M1, the fourth microphone M4 and the seventh microphone M7 are operable to provide signals transformable into a second left channel L2 and/or a tenth right channel R10. At least two of the third microphone M3, the sixth microphone M6 and the seventh microphone M7 are operable to provide signals transformable into a second right channel R2 and/or a tenth left channel L10. The first microphone M1 and the fifth microphone M5 are operable to provide signals transformable into a third left channel L3 and/or a ninth right channel R9. The third microphone M3 and the fifth microphone M5 are operable to provide signals transformable into a third right channel R3 and/or a ninth left channel L9. The first microphone M1 and the sixth microphone M6 are operable to provide signals transformable into a fourth left channel L4 and/or a seventh right channel R7. The third microphone M3 and the fourth microphone M4 are operable to provide signals transformable into a fourth right channel R4 and/or a seventh left channel L7. At least two of the first microphone M1, the second microphone M2 and the third microphone M3 are operable to provide signals transformable into at least one of a fifth left channel L5 and a fifth right channel R5. At least two of the fourth microphone M4, the fifth microphone M5 and the sixth microphone M6 are operable to provide signals transformable into at least one of a sixth left channel L6 and a sixth right channel R6. At least two of the second microphone M2, the fifth microphone M5 and the seventh microphone M7 are operable to provide signals transformable into a center channel C and/or a back center channel B.
As with
At least two of the first microphone M1, the fourth microphone M4 and the seventh microphone M7 are operable to provide signals transformable into at least a first left channel L1 and/or a fourth right channel R4. At least two of the third microphone M3, the fourth microphone M4 and the fifth microphone M5 are operable to provide signals transformable into a first right channel R1 and/or a fourth left channel L4. At least two of the first microphone M1, the second microphone M2 and the third microphone M3 are operable to provide signals transformable into at least one of a second left channel L2 and a second right channel R2. At least two of the fifth microphone M5, the sixth microphone M6 and the seventh microphone M7 are operable to provide signals transformable into at least one of a third left channel L3 and a third right channel R3. At least two of the second microphone M2, the fourth microphone M4 and the sixth microphone M6 are operable to provide signals transformable into a center channel C and/or a back center channel B.
The “snowflake” configuration of
The module includes a shell 821. Coupled to the shell 821 is an array of omnidirectional microphones 822. A beamformer 823 is coupled to the array 822. The beamformer 823 is operable to transform signals produced by the array 822 into multiple directional audio channels. An interface 824 is coupled to the beamformer 823. The interface 824 is operable to convey the multiple directional audio channels into the audio recorder/transmitter 810. The audio recorder/transmitter 810 may then record or transmit the multiple directional audio channels. In an alternative embodiment, the audio recorder/transmitter 810 may record the “raw” signals produced by the array for subsequent beamforming, as will now be explained. In various alternative embodiments, the interface 824 is or includes a Universal Serial Bus (USB) interface, an IEEE 1394 High Speed Serial Bus interface, a Bluetooth/WiFi wireless link interface, a proprietary (e.g., iPhone) bus interface, a read/write buffer memory interface (e.g., Advanced Microcontroller Bus Architecture (AMBA) High Performance Bus (AHB) or Advanced eXtensible Interface (AXI)) or a system-on-a-chip (SoC) interconnect. The interface 824 may be of any conventional or later-developed type.
Those skilled in the art to which this application relates will appreciate that other and further additions, deletions, substitutions and modifications may be made to the described embodiments.
This application is a continuation-in-part of U.S. patent application Ser. No. 13/531,211, filed by Rayala, et al., on Jun. 22, 2012, entitled “Real-Time Microphone Array With Robust Beamformer and Postfilter for Speech Enhancement and Method of Operation Thereof,” commonly assigned with this application and incorporated herein by reference.
Number | Date | Country | |
---|---|---|---|
Parent | 13531211 | Jun 2012 | US |
Child | 13932805 | US |