Microphone interface and IP core for always-on system

• US 9,478,231 B1
• Filed: 03/10/2015
• Issued: 10/25/2016
• Est. Priority Date: 03/10/2015
• Status: Active Grant

• Current Assignee: Cadence Design Systems Incorporated (Cadence Group)
• Original Assignee: Cadence Design Systems Incorporated (Cadence Group)
• Inventors: Soman, Manoj Shridhar
• Primary Examiner(s): Lao, Lun-See
• Application Number:
  US14/643,234
• Time in Patent Office: 595 Days
• Field of Search: 381 56- 58, 381/92, 381 941- 944, 381 711- 716, 381 7111- 7114, 700/94, 375/219
• US Class Current: 1/1
• CPC Class Codes: G06F 1/3206 Monitoring of events, devices or parameters that trigger a change in power modality
• CPC Class Codes: G10L 25/48 specially adapted for particular use
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10

Abstract

Methods and systems provide a partitioned IP core and hierarchical power management to reduce power consumption and footprint size of an “always-on” pulse density modulation (PDM) sensor system. The IP core may be partitioned into a register transfer level (RTL) block and a firmware block. The RTL may include a first stage decimation filter, storage, and, optionally, a sound energy detector. The firmware block may include subsequent decimation filter(s) and sensor processing logic, e.g., a sound trigger algorithm. In operation, the firmware block may conserve energy by being in a power-off or power-saving mode by default. Responsive to a trigger by the sound energy detector, the firmware block may wake up, receive data from the RTL block, and process the data. The sound energy detector may output the trigger based on characteristics of the received sample such as signal strength, noise strength, and type.

Claims

1. A system to process a sound sample, the system comprising:

a register transfer level (RTL) block, including;

a first filter to filter the sound sample; and

a firmware block, including;

a processor to parse the stored filtered sound sample and perform a function corresponding to the parsed sound sample; and

a second filter to receive and filter the stored filtered sound sample.

2. The system of claim 1, wherein the firmware block includes the second filter.

3. A system of claim 1, wherein the RTL block includes the second filter.

4. The system of claim 1, further comprising a sound energy detecting device communicatively coupled to the firmware block and at least one of the first filter and the second filter, to wake the processor responsive to a determination that a characteristic of the filtered sound sample is above a threshold.

5. The system of claim 4, wherein the characteristic is at least one of:

a signal strength, a noise strength, a strength of a combination of signal and noise signal.

6. The system of claim 4, wherein the threshold is determined by at least one of:

the processor and the RTL block.

7. The system of claim 6, wherein the threshold is determined by the processor.

8. The system of claim 4, wherein the sound energy detecting device wakes the processor such that the system monitors an environment of the sound sample.

9. The system of claim 1, further comprising:

an interface to output the stored sound sample in pre-defined bursts to at least one of the second filter and the processor; and

a storage apparatus to store the filtered sound sample, wherein the storage apparatus processes in a first-in-first-out (FIFO) order.

10. The system of claim 1, wherein the first filter includes a cascaded integrator-comb (CIC) decimator to decrease the sample rate of the sound sample.

11. The system of claim 1, wherein the second filter includes at least one of:

a half-band filter and a finite impulse response (FIR) filter, to decrease the sample rate of the stored sound sample.

12. The system of claim 1, wherein the firmware block implements a power-down mode that wakes in response to an output by the RTL block.

13. The system of claim 1, further comprising a low-power sound detecting device (LPSD) communicatively coupled to the second filter and the processor to classify the sound sample filtered by the second filter and to output the classification to the processor.

14. The system of claim 1, wherein the RTL block and the firmware block operate on an output of a pulse density modulation (PDM) microphone.

15. A computer-implemented method to process a sound sample using a partitioned register transfer level (RTL) block and firmware block, the method comprising:

filtering, by the RTL block, the sound sample;

waking, by the RTL block, the firmware block responsive to a characteristic of the filtered sound sample;

filtering, by the firmware block, the filtered sound sample;

parsing, by the firmware block, the sound sample filtered by the firmware block; and

performing, by the firmware block, a function corresponding to the parsed sound sample.

16. The method of claim 15, wherein the waking comprises waking the firmware block from a power-down mode.

17. The method of claim 15, wherein the filtering by the RTL block includes a first decimation and the filtering by the firmware block includes a second decimation.

18. The method of claim 15, further comprising:

measuring at least one of;

a signal strength and a noise strength of the filtered sound sample; and

determining, by the RTL block, the characteristic of the filtered sound sample based on the measuring; and

wherein the waking is performed responsive to a determination that the characteristic of the filtered sound sample is above a threshold.

19. The method of claim 15, wherein the firmware block enters a power-down mode responsive to a determination that at least one of:

a signal strength and a noise strength of the filtered sound sample is below the threshold during a predefined time period.

20. The method of claim 15, wherein the threshold is determined by the firmware block.

21. The method of claim 15, further comprising, prior to the parsing, classifying the sound sample filtered by the firmware block;

and

wherein the parsing is based on the classifying.

22. A non-transitory computer-readable medium storing program instructions that, when executed, cause an RTL block and a firmware block to perform a method to process audio, the method comprising:

filtering, by the RTL block, the sound sample;

waking, by the RTL block, the firmware block responsive a characteristic of the filtered sound sample;

filtering, by the firmware block, the filtered sound sample;

parsing, by the firmware block, the sound sample filtered by the firmware block; and

performing, by the firmware block, a function corresponding to the parsed sound sample.

23. The non-transitory computer-readable medium of claim 22, further comprising:

measuring, by the RTL block, at least one of;

a signal strength and a noise strength of the filtered sound sample; and

determining, by the RTL block, the characteristic of the filtered sound sample based on the measuring; and

wherein the waking is performed responsive to a determination that the characteristic of the filtered sound sample is above a threshold.

24. The non-transitory computer-readable medium of claim 22, wherein the waking comprises waking the firmware block from a power-down mode.

25. The non-transitory computer-readable medium of claim 22, wherein at least one of the RTL block and the firmware block is dynamically configurable.