Signal Processing and Speech Communication Laboratory
homestudent projects › Implementation of a modular audio analysis software in Python

Implementation of a modular audio analysis software in Python

Status
Finished
Type
Bachelor Project
Seminar Type
Audio Signal Processing, Seminar (3 SE)
Announcement date
02 Mar 2016
Student
Malte Merdes, Philipp Merz
Mentors
Research Areas

Abstract

SNARE is a modular audio analyze tool focused on handling long multichannel recordings. E.g. recording a workspace with several microphone positions for an entire workday to analyze noise pollution. This is possible due to its block structure file handling, which distinguishes it from other analyze tools. Furthermore SANRE does not depend on specialized hardware, any audio-interface/microphone set-up can be used. SNARE includes a standardized octave band analysis, sound pressure level over time and histogram (sound pressure level occurrence distribution) with the option to calibrate analysis and generate reports. Analysis can be extended by a plugin system.
It is implemented in Python, and in contrast to most analyze tools open source and cross-platform. Project files and binaries for Windows and OSX are available at: https://git.io/snare.
SNARE can use WAV files with 16bit or 24bit with 44.1kHz, 48kHz or 96kHz with an arbitrary number of channels as data source or directly record multiple channels from any audio interface. The tool supports a calibration source with 94dB at 1kHz or can perform quantitative analysis for non calibrated recordings.
The first part of this thesis features a short tutorial on how-to use SNARE, a detailed manual describing all aspects of the tool, as well as a guide on how to extend the tool’s functionality, while the second part covers the development of SNARE and technical documentation.