Signal processing unit for bone conduction microphone and speaker An implementation of delay, pitch shifting and echo cancellation
| dc.contributor.author | Eliasson, Daniel | |
| dc.contributor.author | Stauffer-Kee, Lucien | |
| dc.contributor.department | Chalmers tekniska högskola / Institutionen för elektroteknik | sv |
| dc.contributor.examiner | Reinfeldt, Sabine | |
| dc.contributor.supervisor | Fredén Jansson, Karl-Johan | |
| dc.date.accessioned | 2024-06-20T07:54:22Z | |
| dc.date.available | 2024-06-20T07:54:22Z | |
| dc.date.issued | 2024 | |
| dc.date.submitted | ||
| dc.description.abstract | Abstract This paper introduces the development of a signal processing unit for bone conduction microphone and speaker generating real-time adjustable delayed auditory feedback (DAF) and frequency altered feedback (FAF). These types of feedbacks are known to effectively increase fluency in people who stutter. The incidence of stuttering is about 1% of the world population. The effects of DAF have been noted since the 1950s. The user’s voice is captured by a microphone, processed, and then relayed back through a speaker. DAF introduces a delay of 50 to 200 milliseconds, while FAF alters the pitch by a quarter to a full octave. A notable challenge for bone conduction devices is the potential for strong feedback paths between speaker and microphone, which can result in echo or oscillation. The signal processing algorithms, including DAF, FAF, and echo cancellation, were developed and tested using MATLAB®. Additionally, an analog chain was constructed and evaluated on a breadboard, featuring variable amplification and power amplifier to directly drive a passive speaker. To achieve a standalone device, the algorithms were ported to a microcontroller, which was further enhanced with a user-friendly interface, including a rotary encoder and LCD, allowing adjustments of the algorithms without programming expertise. The entire system was then integrated onto a custom-designed printed circuit board (PCB), combining both analog and digital circuitry. The MATLAB® script successfully implements all algorithms; DAF, FAF and echo cancellation. It can be used either with sound files like wav or mp3, or through the use of a audio interface the MATLAB® script can be used in real-time for live application such as a test with a real person. The hardware implemented design on PCB has a working and tested DAF, and a untested implementation of echo-cancellation. Due to limitation in floating point performance of the microcontroller a pitch shifting algorithm remains incomplete. The hardware device has sufficient audio quality with a total harmonic distortion of about 3% adhering to IEC 60645-1. This work lays the groundwork for future enhancements, particularly in refining the pitch-shifting capability. | |
| dc.identifier.coursecode | EENX30 | |
| dc.identifier.uri | http://hdl.handle.net/20.500.12380/307954 | |
| dc.language.iso | eng | |
| dc.setspec.uppsok | Technology | |
| dc.subject | Keywords: Signal processing, stuttering, DAF, FAF, echo-cancellation, Raspberry Pi Pico, audio. | |
| dc.title | Signal processing unit for bone conduction microphone and speaker An implementation of delay, pitch shifting and echo cancellation | |
| dc.type.degree | Examensarbete för masterexamen | sv |
| dc.type.degree | Master's Thesis | en |
| dc.type.uppsok | H | |
| local.programme | Embedded electronic system design (MPEES), MSc |