Acoustic Overlays

Visualize pitch, formants, intensity, and more

Overview

Acoustic overlays display computed measurements directly on the spectrogram, enabling visual analysis of prosody, voice quality, and spectral properties. Acoustic measurements are computed via WebAssembly.

Available Overlays

Pitch (F0)

Display: Blue line with dots

Fundamental frequency (F0) tracking shows vocal fold vibration rate, essential for analyzing intonation, tone, and stress patterns.

Parameters: - Range: 75-600 Hz (default, adjustable for speaker) - Time step: 10 ms - Algorithm: Praat autocorrelation method

Interpretation: - Higher pitch = higher blue line - Dots mark detected pitch points (unvoiced regions have no dots) - Missing segments indicate unvoiced sounds (like /s/, /t/, /k/)

Pitch Range Settings

Adjust pitch range for different speakers: - Adult male: 75-300 Hz - Adult female: 100-500 Hz - Children: 200-600 Hz

Configure via config.yaml or UI settings.

Formants (F1-F4)

Display: Red dots

Formant frequencies represent vocal tract resonances, critical for vowel identification and acoustic phonetics.

Parameters: - Number of formants: 5 (tracks F1-F4, ignores F5) - Max formant: 5500 Hz (female), 5000 Hz (male) - Time step: 10 ms - Pre-emphasis: 50 Hz

Display: - F1 (lowest red dots) - Vowel height (high vowels = low F1) - F2 (second row) - Vowel frontness/backness - F3, F4 (upper rows) - Rhotic/retroflex indicators

Formant Presets:

Preset	Max Formant	Num Formants	Use Case
Male	5000 Hz	5	Adult male speech
Female	5500 Hz	5	Adult female speech
Child	6500 Hz	5	Children’s voices

Formant Tracking Accuracy

Formant tracking may be inaccurate for: - Very high-pitched voices - Noisy recordings - Non-modal phonation (creaky voice, breathy voice) - Nasalized vowels

Always verify formant values against auditory perception.

Intensity

Display: Green line

Intensity (loudness) measured in dB SPL (sound pressure level), useful for identifying stressed syllables and amplitude modulation.

Parameters: - Minimum pitch: 75 Hz (for period detection) - Time step: 10 ms - Smoothing: Yes (Praat default)

Interpretation: - Higher green line = louder sound - Peaks often correspond to vowels - Valleys correspond to consonants or silence

Harmonics-to-Noise Ratio (HNR)

Display: Cyan line

HNR measures voice quality by comparing periodic (harmonic) energy to aperiodic (noise) energy.

Parameters: - Minimum pitch: 75 Hz - Time step: 10 ms - Silence threshold: 0.1

Interpretation: - High HNR (>20 dB): Clear, modal voice - Medium HNR (10-20 dB): Normal voice with slight breathiness - Low HNR (<10 dB): Breathy, creaky, or pathological voice

Voice Quality Research

HNR is useful for: - Distinguishing modal vs. non-modal phonation - Detecting voice disorders - Analyzing consonant voicing - Comparing voice quality across speakers

Center of Gravity (CoG)

Display: Orange line

Spectral center of gravity indicates where spectral energy is concentrated, useful for fricative and sibilant analysis.

Interpretation: - High CoG (>6000 Hz): Alveolar fricatives (/s/, /z/) - Medium CoG (3000-6000 Hz): Postalveolar fricatives (/ʃ/, /ʒ/) - Low CoG (<3000 Hz): Dental/interdental fricatives (/θ/, /ð/)

Spectral Tilt

Display: Purple line

Spectral tilt measures the slope of the spectrum, indicating voice quality and phonation type.

Interpretation: - Negative tilt: More energy in high frequencies (breathy voice, /h/) - Near-zero tilt: Balanced spectrum (modal voice) - Positive tilt: More energy in low frequencies (creaky voice)

A1-P0

Display: Pink line

A1-P0 (amplitude of first harmonic minus amplitude of first formant) measures nasalization and voice source characteristics.

Interpretation: - Negative values indicate nasal coupling - Useful for detecting nasal consonants and nasalized vowels

Toggling Overlays

Via URL Parameters

Pre-configure overlays when embedding the mobile viewer:

<iframe src="https://ucpresearch.github.io/ozen-web/viewer?overlays=pitch,formants,intensity">
</iframe>

See Mobile Viewer Embedding and URL Parameters for details.

Values Panel

Hover over the spectrogram to see overlay values at the cursor position:

Time: 1.234 s
Freq: 523 Hz

Pitch: 245 Hz
Intensity: 68 dB
F1: 720 Hz  B1: 80 Hz
F2: 1240 Hz B2: 110 Hz
F3: 2650 Hz B3: 150 Hz
F4: 3500 Hz B4: 200 Hz
HNR: 15.3 dB
CoG: 5420 Hz

See Values Panel for details.

Performance

Note

Overlays are computed on-demand when toggled. Once computed, values are cached for redrawing.

Configuration

Customize overlay parameters via config.yaml:

pitch:
  floor: 75        # Minimum pitch (Hz)
  ceiling: 600     # Maximum pitch (Hz)
  timeStep: 0.01   # Analysis step (seconds)

formantPresets:
  male:
    maxFormant: 5000
    numFormants: 5
  female:
    maxFormant: 5500
    numFormants: 5
  child:
    maxFormant: 6500
    numFormants: 5

intensity:
  minimumPitch: 75
  timeStep: 0.01

See Configuration Reference for all options.

Troubleshooting

Pitch Track Missing

Problem: No blue dots visible

Possible causes: - Audio is unvoiced (normal for fricatives, stops) - Pitch range set incorrectly for speaker - Very noisy recording

Solution: - Adjust pitch floor/ceiling for speaker - Check if audio is actually voiced - Try different time region

Formants Look Wrong

Problem: Red dots appear in unexpected locations

Possible causes: - Wrong formant preset (male/female/child) - Nasalized vowels - Non-modal phonation - Noise or artifacts

Solution: - Switch formant preset (Settings → Formant Preset) - Check audio quality - Verify with auditory analysis - Manually verify data point values

Overlays Not Computing

Problem: Overlay checkbox enabled but no visualization

Possible causes: - WASM not initialized - No audio loaded - Audio too long (>60s) and zoomed out

Solution: - Wait for WASM to load (status indicator) - Load audio file - Zoom in if file is long

Research Applications

Prosody Analysis

Use pitch and intensity overlays to study: - Intonation patterns - Lexical tone - Stress and emphasis - Question vs. statement contours

Vowel Formants

Use formant overlay to: - Create vowel plots (F1 vs. F2) - Track vowel formant trajectories - Identify vowel targets - Detect vowel coarticulation

Voice Quality

Use HNR, spectral tilt, and A1-P0 to: - Compare modal, breathy, and creaky voice - Detect pathological voice - Analyze phonation types across languages - Study laryngealization

Fricative Analysis

Use CoG and spectral tilt to: - Distinguish /s/ vs. /ʃ/ - Measure acoustic contrast - Study lenition processes - Compare speaker-specific productions

--- title: "Acoustic Overlays" subtitle: "Visualize pitch, formants, intensity, and more" --- ## Overview Acoustic overlays display computed measurements directly on the spectrogram, enabling visual analysis of prosody, voice quality, and spectral properties. Acoustic measurements are computed via WebAssembly. ![All acoustic overlays enabled](../screenshots/acoustic-overlays-all.png) ## Available Overlays ### Pitch (F0) **Display:** Blue line with dots Fundamental frequency (F0) tracking shows vocal fold vibration rate, essential for analyzing intonation, tone, and stress patterns. **Parameters:** - Range: 75-600 Hz (default, adjustable for speaker) - Time step: 10 ms - Algorithm: Praat autocorrelation method **Interpretation:** - Higher pitch = higher blue line - Dots mark detected pitch points (unvoiced regions have no dots) - Missing segments indicate unvoiced sounds (like /s/, /t/, /k/) ::: {.callout-tip} ## Pitch Range Settings Adjust pitch range for different speakers: - **Adult male**: 75-300 Hz - **Adult female**: 100-500 Hz - **Children**: 200-600 Hz Configure via `config.yaml` or UI settings. ::: ### Formants (F1-F4) **Display:** Red dots Formant frequencies represent vocal tract resonances, critical for vowel identification and acoustic phonetics. **Parameters:** - Number of formants: 5 (tracks F1-F4, ignores F5) - Max formant: 5500 Hz (female), 5000 Hz (male) - Time step: 10 ms - Pre-emphasis: 50 Hz **Display:** - **F1** (lowest red dots) - Vowel height (high vowels = low F1) - **F2** (second row) - Vowel frontness/backness - **F3, F4** (upper rows) - Rhotic/retroflex indicators **Formant Presets:** | Preset | Max Formant | Num Formants | Use Case | |--------|-------------|--------------|----------| | Male | 5000 Hz | 5 | Adult male speech | | Female | 5500 Hz | 5 | Adult female speech | | Child | 6500 Hz | 5 | Children's voices | ::: {.callout-note} ## Formant Tracking Accuracy Formant tracking may be inaccurate for: - Very high-pitched voices - Noisy recordings - Non-modal phonation (creaky voice, breathy voice) - Nasalized vowels Always verify formant values against auditory perception. ::: ### Intensity **Display:** Green line Intensity (loudness) measured in dB SPL (sound pressure level), useful for identifying stressed syllables and amplitude modulation. **Parameters:** - Minimum pitch: 75 Hz (for period detection) - Time step: 10 ms - Smoothing: Yes (Praat default) **Interpretation:** - Higher green line = louder sound - Peaks often correspond to vowels - Valleys correspond to consonants or silence ### Harmonics-to-Noise Ratio (HNR) **Display:** Cyan line HNR measures voice quality by comparing periodic (harmonic) energy to aperiodic (noise) energy. **Parameters:** - Minimum pitch: 75 Hz - Time step: 10 ms - Silence threshold: 0.1 **Interpretation:** - **High HNR (>20 dB)**: Clear, modal voice - **Medium HNR (10-20 dB)**: Normal voice with slight breathiness - **Low HNR (<10 dB)**: Breathy, creaky, or pathological voice ::: {.callout-tip} ## Voice Quality Research HNR is useful for: - Distinguishing modal vs. non-modal phonation - Detecting voice disorders - Analyzing consonant voicing - Comparing voice quality across speakers ::: ### Center of Gravity (CoG) **Display:** Orange line Spectral center of gravity indicates where spectral energy is concentrated, useful for fricative and sibilant analysis. **Interpretation:** - **High CoG (>6000 Hz)**: Alveolar fricatives (/s/, /z/) - **Medium CoG (3000-6000 Hz)**: Postalveolar fricatives (/ʃ/, /ʒ/) - **Low CoG (<3000 Hz)**: Dental/interdental fricatives (/θ/, /ð/) ### Spectral Tilt **Display:** Purple line Spectral tilt measures the slope of the spectrum, indicating voice quality and phonation type. **Interpretation:** - **Negative tilt**: More energy in high frequencies (breathy voice, /h/) - **Near-zero tilt**: Balanced spectrum (modal voice) - **Positive tilt**: More energy in low frequencies (creaky voice) ### A1-P0 **Display:** Pink line A1-P0 (amplitude of first harmonic minus amplitude of first formant) measures nasalization and voice source characteristics. **Interpretation:** - Negative values indicate nasal coupling - Useful for detecting nasal consonants and nasalized vowels ## Toggling Overlays ### Via UI Checkboxes Click checkboxes in the overlay panel: - ☐ Pitch - ☐ Formants - ☐ Intensity - ☐ HNR - ☐ CoG - ☐ Spectral Tilt - ☐ A1-P0 Multiple overlays can be enabled simultaneously. ### Via URL Parameters Pre-configure overlays when embedding the mobile viewer: ```html <iframe src="https://ucpresearch.github.io/ozen-web/viewer?overlays=pitch,formants,intensity"> </iframe> ``` See [Mobile Viewer Embedding](mobile-viewer.html#embedding-in-web-pages) and [URL Parameters](../embedding/url-parameters.html) for details. ## Values Panel Hover over the spectrogram to see overlay values at the cursor position: ``` Time: 1.234 s Freq: 523 Hz Pitch: 245 Hz Intensity: 68 dB F1: 720 Hz B1: 80 Hz F2: 1240 Hz B2: 110 Hz F3: 2650 Hz B3: 150 Hz F4: 3500 Hz B4: 200 Hz HNR: 15.3 dB CoG: 5420 Hz ``` See [Values Panel](../tutorial/05-data-collection.html) for details. ## Performance ::: {.callout-note} Overlays are computed on-demand when toggled. Once computed, values are cached for redrawing. ::: ## Configuration Customize overlay parameters via `config.yaml`: ```yaml pitch: floor: 75 # Minimum pitch (Hz) ceiling: 600 # Maximum pitch (Hz) timeStep: 0.01 # Analysis step (seconds) formantPresets: male: maxFormant: 5000 numFormants: 5 female: maxFormant: 5500 numFormants: 5 child: maxFormant: 6500 numFormants: 5 intensity: minimumPitch: 75 timeStep: 0.01 ``` See [Configuration Reference](../reference/configuration.html) for all options. ## Troubleshooting ### Pitch Track Missing **Problem**: No blue dots visible **Possible causes**: - Audio is unvoiced (normal for fricatives, stops) - Pitch range set incorrectly for speaker - Very noisy recording **Solution**: - Adjust pitch floor/ceiling for speaker - Check if audio is actually voiced - Try different time region ### Formants Look Wrong **Problem**: Red dots appear in unexpected locations **Possible causes**: - Wrong formant preset (male/female/child) - Nasalized vowels - Non-modal phonation - Noise or artifacts **Solution**: - Switch formant preset (Settings → Formant Preset) - Check audio quality - Verify with auditory analysis - Manually verify data point values ### Overlays Not Computing **Problem**: Overlay checkbox enabled but no visualization **Possible causes**: - WASM not initialized - No audio loaded - Audio too long (>60s) and zoomed out **Solution**: - Wait for WASM to load (status indicator) - Load audio file - Zoom in if file is long ## Research Applications ### Prosody Analysis Use pitch and intensity overlays to study: - Intonation patterns - Lexical tone - Stress and emphasis - Question vs. statement contours ### Vowel Formants Use formant overlay to: - Create vowel plots (F1 vs. F2) - Track vowel formant trajectories - Identify vowel targets - Detect vowel coarticulation ### Voice Quality Use HNR, spectral tilt, and A1-P0 to: - Compare modal, breathy, and creaky voice - Detect pathological voice - Analyze phonation types across languages - Study laryngealization ### Fricative Analysis Use CoG and spectral tilt to: - Distinguish /s/ vs. /ʃ/ - Measure acoustic contrast - Study lenition processes - Compare speaker-specific productions ## See Also - [Spectrogram](spectrogram.html) - Base visualization - [Data Points](data-points.html) - Collecting measurements - [Tutorial: Acoustic Analysis](../tutorial/03-acoustic-analysis.html) - Step-by-step guide - [Configuration](../reference/configuration.html) - Customizing parameters