A parameter distribution setting method including performs learning based on a gradient learning method in advance such that a mutual information between a probabilistic distribution of an output of a reservoir device and an ideal probabilistic distribution of the output increases, and setting a parameter distribution of parameters defining element derivation in a plurality of elements constituting the reservoir device in a device model for the reservoir device.

A method of fabricating a micro-electrical-mechanical system (MEMS) transducer chip scale package. The method comprising: providing (101) a front side pre-fabricated semiconductor die wafer (1) comprising a plurality of individual die that each comprise at least a MEMS transducer. And back etching (104) the semiconductor die wafer (1) at the back side (4) of the semiconductor die wafer (1) by etching an acoustic die channel (5) through each respective die of the plurality of die and etching a die back volume (6) into each respective die of the plurality of die. The semiconductor die wafer (1) is capped with a cap wafer (16) such that a wafer level packaged MEMS transducer wafer is provided containing multiple MEMS transducer chip scale packages.

A MEMS transducer package (1) comprises a semiconductor die element (3) and a cap element (23). The semiconductor die element (3) and cap element (23) have mating surfaces (9, 21). The semiconductor die element (3) and cap element (23) are configured such that when the semiconductor die element (3) and cap element (4) are conjoined, a first volume (7, 27) is formed through the semiconductor die element (3) and into the semiconductor cap element (23), and an acoustic channel is formed to provide an opening between a non-mating surface (11) of the semiconductor die element (3) and either a side surface (10, 12) of the transducer package or a non-mating surface (29) of the cap element (23).

A MEMS transducer package (1) comprises a semiconductor die element (3) and a cap element (23). The semiconductor die element (3) and cap element (23) have mating surfaces (9, 21). The semiconductor die element (3) and cap element (23) are configured such that when the semiconductor die element (3) and cap element (4) are conjoined, a first volume (7, 27) is formed through the semiconductor die element (3) and into the semiconductor cap element (23), and an acoustic channel is formed to provide an opening between a non-mating surface (11) of the semiconductor die element (3) and either a side surface (10, 12) of the transducer package or a non-mating surface (29) of the cap element (23).

A MEMS transducer package (1) is provide having a semiconductor die portion (3) with a thickness bounded by a first surface (9) and an opposite second surface (11). The package further has a transducer element (13) incorporated in the second surface (11) and a die back volume (7) that extends through the thickness of the semiconductor die portion (3) between the first surface (9) and the transducer element (13). The package is completed by a cap portion (23) that abuts the semiconductor die portion (3) at the first surface (9).

A MEMS transducer package (1) comprises a semiconductor die element (3) and a cap element (23). The semiconductor die element (3) and cap element (23) have mating surfaces (9, 21). The semiconductor die element (3) and cap element (23) are configured such that when the semiconductor die element (3) and cap element (4) are conjoined, a first volume (7, 27) is formed through the semiconductor die element (3) and into the semiconductor cap element (23), and an acoustic channel is formed to provide an opening between a non-mating surface (11) of the semiconductor die element (3) and a side surface (10, 12) of the transducer package.

A MEMS transducer package (1) comprises a semiconductor die element (3) and a cap element (23). The semiconductor die element (3) and cap element (23) have mating surfaces (9, 21). The semiconductor die element (3) and cap element (23) are configured such that when the semiconductor die element (3) and cap element (4) are conjoined, a first volume (7, 27) is formed through the semiconductor die element (3) and into the semiconductor cap element (23), and an acoustic channel is formed to provide an opening between a non-mating surface (11) of the semiconductor die element (3) and a side surface (10, 12) of the transducer package.

A MEMS device comprises a housing with an interior volume, wherein the housing includes an access port to the interior volume; a MEMS sound transducer in the housing, and a mechanical barrier structure having a plate element that is fixed by elastic spacers to a carrier and overlaps the access port, and providing a ventilation path passing a boundary region of the plate element, wherein a clearance of the ventilation path is set by the distance of the boundary region of the plate element to the housing or by the distance of the boundary region of the plate element to a blocking structure that opposes the boundary region of the plate element.

The present invention provides a bone conduction microphone including a housing and a circuit board connected with the housing. The circuit board has an acoustic channel. The microphone further includes a vibration assembly forming a first conduction cavity and a second conduction cavity. The vibration assembly includes a vibration member and a frame. The frame, the vibration member and the circuit board form a first conduction cavity. The frame, the vibration member and the circuit board form a second conduction cavity. The vibration of the vibration member is conducted to one side of the vibration diaphragm, and is also conducted to the other side of the vibration diaphragm. Compared with the related art, the bone conduction microphone of the present invention can effectively improve the sensitivity and the signal to noise ratio.

Noise-canceling audio devices may include a first vibration member, a second vibration member, and a microphone supported by a housing. A feedback, noise-cancelation circuit may be operatively connected to the microphone, the feedback, noise-cancelation circuit configured to generate a first portion of a modified audio signal by combining an audio signal with a noise-canceling signal generated in response to a signal from the microphone to at least partially cancel at least a portion of an audible response of the second vibration member. A feed-forward, noise-cancelation circuit may be operatively connected to the microphone, the feed-forward, noise-cancelation circuit configured to compare the signal from the microphone to a predetermined SPL profile and generate a second portion of the modified audio signal configured to at least partially cancel environmental noise, the feedback, noise cancelation circuit configured to output the modified audio signal only to the first vibration member.