UMIK-1 Spectrum Analyzer

UMIK-1 Spectrum Analyzer is a high-performance, real-time audio spectrum analysis software developed in Python. It is specifically tailored for miniDSP UMIK-1 calibrated measurement microphone, bringing high-level acoustic engineering capabilities directly to your desktop.

Warning

Disclaimer: This software is strictly intended for learning, educational purposes, and hobbyist exploration only. It is not a certified acoustic measurement device. The DSP implementations are not calibrated against national or international metrology standards. Do not use this application for any formal, legal, or commercial real-world acoustic compliance measurements.

---

📸 Screenshots

Real-Time Spectrum Analyzer — FFT line graph with Max-Hold trace, Peak detection, Waterfall Spectrogram, and SPL readouts:

Offline Data Viewer — SPL time-series playback with spectrum snapshot at cursor and statistical summary table:

---

⚡ Key Features

• High-RES FFT Line Analysis: Adjustable FFT block sizes (practical limit ~65536 given the UMIK-1's 48kHz sample rate) to achieve sub-1Hz resolution. Note that extremely high FFT sizes will reduce UI responsiveness.
• IEC Standard 1/3 Octave Analysis: Single-click toggle from raw continuous FFT to coarse professional 1/3 Octave Band Bar Graphs.
• Dynamic Peak & Max Hold: Automatically detects dominant peaks and sticks a Max-Hold trace (both smoothly mapping to Continuous FFT tracks or snapping to Octave Bands).
• Multi-Weighting SPL Calculation: Native math implementation of Z-Weighting (Unweighted), A-Weighting (Human Ear Response), and C-Weighting (Loud Machinery Response).
• Target Band SPL Selector: Fully draggable LinearRegionItem boundaries that calculates the energetic SPL isolated strictly within your selected frequency region. Smooth snapping support for Octave bands.
• Interactive Waterfall Spectrogram: Time-series decimation rendering. Translates raw volume dynamics into color history to observe wave decays and resonance.
• Overload, Clipping & SPL Alarms: Native OS-level thread beep capability alongside flashing UI visual alerts in case the acoustic energy overloads safe levels.
• Native miniDSP UMIK-1 Calibration Support: Direct parsing of .txt calibration files to correct frequency responses dynamically in software.
• Continuous Measurement Recording: Output full .npz binary array datasets or standard .csv files for deep-dive offline acoustics research.
• Standalone Data Playback Viewer: Comes bundled with umik1_data_viewer.py to graph past recordings, calculate acoustic statistics (Leq, Lmin, Lmax), and seamlessly plot offline Time/Freq curves without re-activating the microphone!

---

🛠️ Installation & Setup

Pre-compiled Standalone Executable (Windows)

The easiest way to use the analyzer without installing Python or dependencies is to download the bundled application:

1. Navigate to the Releases tab on this GitHub repository.
2. Download the latest UMIK_1_Spectrum_Analyzer.zip from the dist/ release assets.
3. Extract the folder and run UMIK_1_Spectrum_Analyzer.exe. No installation required!

Python Environment (For Developers)

We recommend using conda to sandbox the GUI toolkit dependencies and reduce unnecessary package bloating:

# 1. Create a specialized environment
conda create -n umik_pack python=3.12
conda activate umik_pack

# 2. Install High-Performance Acoustic & GUI Packages
pip install numpy scipy sounddevice pyqt5 pyqtgraph

# 3. Clone this repository
git clone https://github.com/RealSeaberry/umik-1-analyzer.git
cd umik-1-analyzer

(Optional) If you want to bundle it into a standalone Windows .exe executable without forcing users to install Python, PyInstaller is natively supported:

$env:PYTHONNOUSERSITE = "1"; python -m PyInstaller --noconfirm --clean --onedir --windowed --name "UMIK_1_Spectrum_Analyzer" "umik1_analyzer.py"

---

🚀 Usage

After activating your Python environment, execute the main GUI entry code:

python umik1_analyzer.py

GUI Tool interactions:

1. Device Selection: Ensure your target measurement mic (e.g. UMIK-1) is selected in the bottom right corner.
2. Axis Rescaling:
   • Dragging the bottom X-axis manually repositions the frequency rendering view.
   • Right-Click-Dragging the Y-axis zooms the Amplitude scale. Note: Zooming the Spectrum Amplitude directly controls the Color Gain Levels of the Waterfall Plot!
3. Crosshair Tracking: Clicking anywhere on the Main Spectrum will snap a green crosshair line precisely to measure frequency and dBSPL.
   • Pro Tip: If you are in 1/3 Octave mode, the crosshair automatically smartly snaps directly to the nearest exact IEC center-frequency (e.g 1000.0Hz)!
4. Recording: Click the [⏺ Start Recording] to drop .npz packages into your active directory containing exhaustive full spectrum history data slices.

To review your exported recordings in a playback view:

python umik1_data_viewer.py

---

📐 Scientific DSP Principles

Unlike typical colorful "Music Player Audio Visualizers", UMIK-1 Spectrum Analyzer maps exact Sound Pressure Levels targeting scientific-grade measurement applications.

1. Real-Time Math Pipeline

• Audio frames are extracted directly via C-bindings using PortAudio (sounddevice).
• Depending on your "Windowing" function (Hanning, Blackman-Harris), the raw chunk undergoes structural attenuation. We mathematically calculate a coherent gain compensation factor (window.sum()) to guarantee amplitude magnitudes remain 100% truthful irrespective to the window style altering the time-domain leakage.
• The absolute dBFS is mapped algorithmically using 20 * log10(magnitude / normalization).

2. Calibration Integration

• Without a calibration file, the app defines 0.0 dBFS against internal constants and performs basic Flat-Z curve rendering.
• By attaching your UMIK-1 calibration .txt, the script conducts a 1D linear interpolation mapped directly upon the dynamic FFT Bins. Every FFT discrete sample is amplified or attenuated by its exact offset before arriving at your screen.

3. IEC 1/3 Octave Calculation

When switching onto Professional Octave Analysis, looking at 2000 individual FFT lines is dizzying. We resolve the standard mathematical bins base 10 algorithm (f_center * 10^(1/20) boundaries) and compute energetic summative logs 10 * log10(sum( 10^(powers/10) )) from raw FFT blocks to ensure true band-limited decibel representation.

4. ISO A / C Weighting Calculation

• Z-Weighting (Flat) reports the true acoustic pressure displacement in the air. This often results in "noisy" visually tilted pink-noise floors due to the immense scale of atmospheric infrasound.
• The script implements exact $s$-domain transfer function approximations mapped into identical Z-plane counterparts to accurately apply the A-Weighting (Deaf to extreme Lows/Highs) or C-Weighting curves across the calculated spectral energy field.

---