Bumblebee

A small desktop app that converts XDF recordings (Lab Streaming Layer) into BrainVision files (.vhdr / .vmrk / .eeg) for analysis in BrainVision Analyzer and other EEG tools.

Point it at an .xdf file, pick which stream holds your signal, optionally pick the marker streams to keep, and Bumblebee writes a BrainVision dataset next to the original file.

What it does

• Reads any .xdf / .xdfz file and lists every stream it contains.
• Lets you choose one signal stream (EEG, ECG, EMG, …) to export.
• Lets you choose any number of marker/trigger streams whose events become BrainVision annotations, time-aligned to the signal.
• Streams that self-identify as EEG are scaled from µV to V (BrainVision's expected unit); every other stream type is exported in its raw LSL units with no scaling.
• Channel names are taken from the stream metadata when present, otherwise labelled CH1, CH2, …
• Runs the conversion on a background thread with a live log, so the window stays responsive.

Output

The result is written to the same folder as the source file, with the same base name and a .vhdr extension (plus its companion .vmrk and .eeg). For session1.xdf you get session1.vhdr. Any existing files of that name are overwritten.

Using the app

1. Open XDF … and select your recording. Bumblebee loads it and fills two tables — signal streams on top, marker/trigger streams below. (It sorts streams automatically: anything typed marker, trigger, events, etc., or with a sample rate of 0, goes to the marker table; everything else is a signal.)
2. Select the signal stream to export in the top table (exactly one). If there is only one signal stream it is selected for you.
3. Optionally select marker streams in the bottom table. Use Ctrl-click or Shift-click to pick several; leave it empty to export with no annotations.
4. Click Convert to BrainVision. Watch the log; a dialog confirms the output path when it finishes.

Running from source

The project uses uv for dependency management.

git clone https://github.com/markspan/Bumblebee.git
cd Bumblebee
uv sync            # creates .venv and installs mne, pyxdf, pybv, PySide6
uv run python Bumblebee.py

uv run uses the project's virtual environment automatically, so there's no need to activate anything. If you prefer to activate it:

# Windows PowerShell
.venv\Scripts\Activate.ps1
python Bumblebee.py

Building a standalone Windows executable

Two build routes are provided; either produces a single self-contained bumblebee.exe that needs no Python install on the target machine.

PyInstaller (fast builds, bundles the interpreter):

uv add --dev pyinstaller     # once
uv run pyinstaller bumblebee.spec
# → dist\bumblebee.exe

Nuitka (slower builds, compiles to machine code for faster startup):

nuitka_compile.bat
# → bumblebee.exe

Both embed bumblebee.ico as the application icon.

Requirements

• Python ≥ 3.13
• mne, pyxdf, pybv, PySide6

All are installed automatically by uv sync from the pinned uv.lock.

Licence

See LICENSE.