Intensity

Intensity is the acoustic measure of loudness — energy per unit area reaching the microphone. Vocal Analyzer computes intensity internally from the uploaded audio and uses it in the pitch-tracker voicing decision and the SNR estimate, but the dashboard does not surface intensity as a headline metric.

Why intensity is not banded

The key word is uncalibrated. The analyzer has no knowledge of:

• Microphone sensitivity. Different mics output different voltages for the same sound-pressure level.
• Preamp gain. Digital recording chains vary by tens of dB.
• Input-level settings. Software gain sliders, automatic gain control, and file-format normalisation all shift the numeric intensity value without changing the perceived loudness.
• Distance from mouth to mic. Doubling the mic distance reduces received intensity by ~6 dB.

Because of all of that, two recordings of the same speaker at the same perceived loudness routinely differ by 10–20 dB in raw file intensity. Comparing intensity across sessions — even with the same hardware — is not meaningful without a calibrated reference, which Vocal Analyzer does not ship.

What intensity is used for internally

• Voicing decisions. The pitch tracker uses intensity threshold crossings as one signal for which frames contain voicing.
• SNR estimation. The SNR warning compares an estimated speech- intensity median against a noise-floor estimate from the same recording, so the SNR value is meaningful within one recording even though the underlying intensities are not absolute.

These uses are self-consistent within a single analysis, which is the only scope intensity is reliable in.

When to trust relative intensity

Intensity values within a single recording are meaningful:

• Louder-than-average moments are really louder than quieter moments in the same file.
• Pauses and unvoiced segments drop the running intensity.
• Clipping (intensity pinned against the ceiling) is real and informative.

Intensity values across recordings are not meaningful without calibration:

• A "higher" intensity in one session vs another does not mean a louder voice.
• Loudness-based progress tracking is not supported.

If you want loudness tracking

There is no acoustic-only shortcut — real loudness tracking needs a calibrated microphone with a known sensitivity, a fixed recording chain, and a reference tone. For voice-training purposes, subjective loudness tracking (your own sense of volume, or a voice teacher's ear) is usually more useful than an uncalibrated acoustic number.

Caveats

• Do not compare across sessions. See above.
• Do not compare across apps. Different recording apps and OS audio stacks apply different default gains and filters.
• Do not treat as dB SPL. The value shown, if rendered, is not calibrated sound-pressure level.