Sam Lapp sammlapp

BriteKit vs OpenSoundscape: A Comparison for Bioacoustic ML Development

Date: October 30, 2025 Purpose: Compare two Python packages for developing machine learning recognizers for bird species and other bioacoustic applications

Executive Summary

Both BriteKit and OpenSoundscape are Python-based tools for bioacoustic analysis using deep learning. BriteKit offers an end-to-end, configuration-driven workflow focused on model development from data collection through deployment. OpenSoundscape provides a flexible, modular library emphasizing composability and integration with pre-trained models from the Bioacoustics Model Zoo.

how to debug, and how to ask for help debugging

warnings vs errors: know the difference

breakpoints: interactive debugging in VS Code

print everything out: check intermediate values

reading error messages carefully

add contacts and choose recipients for "Check-In" messages at https://explore.garmin.com/Messages
on phone, install Garmin Explore app (The Garmin Messenger app allows you to send/receive messages when paired, but doesn't allow configuration)
on phone, turn on bluetooth and "allow new connections" if necessary
if not first time using InReach Mini, go to Menu -> Settings -> Phone -> pair (enters Pairning mode); then in Garmin Explore app, go to devices tab, click through screens to connect to your device
Messaging settings will now sync between website and device, can take a couple minutes
Check on handheld: menu, check in -> should see your check in presets. Selecting one shows contact list and message content.

Note: the Garmin Explore phone app is the only way to sync messages from the website to the handheld device, and the website is the only way to set up the message settings. Various instructions mention Garmin Messenger App, but at least for iPhone as

Use cases

Load, inspect, and manipulate audio files

``` from opensoundscape import Audio Audio.from_file('/path/audio.wav') ```

Head to the Audio tutorial notebook

Recommended steps for publicly sharing a model trained with OpenSoundscape

Uploading the saved model object to a public location such as Box, OneDrive, etc
- Save the model object in two different ways to improve usability:
(1) model.save() in OpenSoundscape to save the entire model object, which can be re-loaded only in the same version as OpenSoundscape

(2) To enable use with other OpenSoundscape versions, save the weights dictionary and additional model information as a dictionary with torch.save. For instance:

set up raspberry pi

use Raspberry Pi imager on mac to install Raspberry Pi OS on a microSD card

use the advanced setup option to pre-configure ssh access and wifi

remote access to the GUI:

From mac terminal:

ssh into pi (ssh [email protected], then enter password)
install tightvnc: sudo apt update (this might take a while) then sudo apt install tightvncserver
start the VNC server (will need to be restarted on each reboot, but you only specify the password the first time): vncserver then enter a password (password has 8 character limit)

	Model,Status,Error Type,Error Message,Full Traceback
	ConvNeXt,Success,,,
	EfficientNet,Failed,ConversionError,"Failed to convert the exported program to an ONNX model. [96mThis is step 3/3[0m of exporting the model to ONNX. Next steps:
	- If there is a missing ONNX function, implement it and register it to the registry.
	- If there is an internal error during ONNX conversion, debug the error and submit a PR to PyTorch.
	- Create an error report with `torch.onnx.export(..., report=True)`, and save the ExportedProgram as a pt2 file. Create an issue in the PyTorch GitHub repository against the [96monnx[0m component. Attach the error report and the pt2 model.
	Error report has been saved to 'onnx_export_2025-10-24_08-22-58-034161_conversion.md'.

	## Exception summary

	"""this script extracts a measurement of sound pressure level from an audio file

	specifically, it
	(1) calculates a spectrogram
	(2) subtracts a spectral noise profile based on the Nth percentile value of each spectrogram row
	(3) calibrates the spectrogram to sound pressure level using a calibration curve from a csv file
	(4) reconstructs a time series of rms values from the spectrogram
	(5) extracts the maximum value of rms and converts to dB
	"""

	1. Open Outlook _in a browser_ not in the desktop app
	2. Click the calendar icon on left, and find the calendar you want to sync on the left bar; click the menu (...)
	3. Click sharing and permissions
	4. Send invite to your Google email address
	5. In the outlook pane, right click on the calendar's link and choose copy url
	6. In Google calendar, in left pane next to "Other Calendars", click + and then "from URL"
	7. Paste the URL

	Building wheels for collected packages: antlr4-python3-runtime, lmdb, tokenizers
	Building wheel for antlr4-python3-runtime (setup.py) ... done
	Created wheel for antlr4-python3-runtime: filename=antlr4_python3_runtime-4.9.3-py3-none-any.whl size=144554 sha256=4592393736e0f23d96921e75beb495b156d5ef90144ed466ddfbfbc29d342b58
	Stored in directory: /Users/SML161/Library/Caches/pip/wheels/12/93/dd/1f6a127edc45659556564c5730f6d4e300888f4bca2d4c5a88
	Building wheel for lmdb (setup.py) ... done
	Created wheel for lmdb: filename=lmdb-1.4.1-cp310-cp310-macosx_11_0_arm64.whl size=93571 sha256=c35657cea497e13b71ac47d3f7cbc918c37c8f83acc739baff46ec3c418be390
	Stored in directory: /Users/SML161/Library/Caches/pip/wheels/e6/00/19/a038745f7b5c39772fad66e071042f5ecbefb215235c520e03
	Building wheel for tokenizers (pyproject.toml) ... error
	error: subprocess-exited-with-error