Bassoon2425

Microphone array project WiSe 24/25. Topic: Sound Radiation of a Bassoon

Abstract

The bassoon exhibits complex sound radiation characteristics, which can pose challenges during recording. In this project, a bassoon was analyzed in various tonal regions using a microphone array to identify the smaller sound sources that contribute to its overall sound radiation. By applying beamforming techniques to the array data, it was possible to generate videos of the bassoon with overlaid visual representations of its sound sources.

Key findings indicate that the bassoon’s sound radiation strongly depends on the configuration of opened and closed tone holes for each played note, as well as on the analyzed frequency range. The visualized sound source distribution can help audio engineers when recording a bassoon, either to capture an accurate representation of the instrument or to achieve a desired timbre.

The result Videos of the bassoon and its visualized sound sources can be seen here: https://tubcloud.tu-berlin.de/s/8i6G8Hsa4gQjEKa

This repository has two important folders: mic_positions and evaluation.
mic_positions contains all files needed to plan, build, and verify the microphone array.
evaluation contains all files needed to process the recorded array data.

Workflow Mic Positions

1. Vogel_plot.ipynb: Determines the appropriate microphone number per plane.
2. generate_ideal_mic_side_positions_as_excel.py: Calculates and saves ideal 2D microphone positions.
3. 2d_to_3d_coords.ipynb: Converts 2D positions on the microphone planes to 3D positions in the microphone array.
4. plot_and_measure_3d_coords.ipynb: Plots ideal 3D positions. Also used for measuring distances between microphones.
5. position_optimizer.ipynb: Optimizes ideal microphone positions with measured microphone distances.
6. calc_other_positions.ipynb: Calculates positions like cameras, bassoon, and reference from real-world distance measurements.
7. mic_calib_FileExport.ipynb: Takes microphone calibration data and formats it for further use with Acoular.

Workflow Evaluation Folder

There are 4 types of files:

• _[...] are custom functions.
• evaluation_[...] for calculating the beamforming results from raw array data.
• video_[...] for converting the beamforming result to a video (acoustic camera).
• SODIX_[...] for using the SODIX algorithm on raw array data.

evaluation_[...] files

1. Copy array measurements to evaluation\in\array_audio_data.
2. Set up via evaluation_config.py.
3. efficient_evaluation.py: Calculates a single beamforming result at a given grid, fps, band, etc. Can be used for testing.
   evaluation_wrapper.py: Calculates multiple beamforming results at a given grid, fps, band, etc. This can take a very long time (>10 hours).

video_[...] files

1. Find optimal video parameters with video_export.ipynb: It shows a single frame of a video. Could also be used to export a single video.
2. video_wrapper.py: Creates multiple video files at once.

The video processing is about five times faster than the beamforming.

SODIX_[...] files

--Not finished--
The structure is the same as with the evaluation_[...] files. The SODIX algorithm is used to describe the 3D directivity.

Set Up Environment

All non-standard Python packages should be listed in requirements.txt. Install the required packages using requirements.txt:

pip install -r requirements.txt
conda install -c conda-forge ffmpeg

The ffmpeg writer installation was only successful via conda. Therefore, not using conda could lead to problems.