Cases/NVH/C05
NVHOEM sound-quality teams / Tier 1 interior trim

Psychoacoustic Sound Quality (SQ) Four-Metric Evaluation

01

Pain Points

  • !"Premium vs. cheap" relies on subjective scoring, lacking repeatable objective metrics.
  • !A-weighting is too far from the human ear; door-closing and fan sounds need psychoacoustic evaluation.
  • !Each of the four metrics is computed with a different tool, and their time series don't align.
02

The Tinia Approach

Compute the four psychoacoustic metrics in parallel, then align and merge them into a composite SQ score; with subjective ratings available, AutoML can distill a discriminant function.

Cut evaluation segmentsLoudness + Sharpness + Roughness + TonalityAlign and mergeComposite SQ score
03

Nodes Used

Audio Segment Splitaudio_segment_split
Cut out the evaluation segments
Loudnessloudness
Loudness (sone) time series
Sharpnesssharpness
Sharpness (acum)
Roughnessroughness
Roughness (asper)
Tonalitytonality
Tonality metric
Feature Mergefeature_merge
Align and merge the four metric time series
Chart Viewerchart_viewer
Four metric curves + composite score

Core nodes; add or remove steps as needed for your data and standards — the node graph is always editable.

04

Expected Output

  • Time series for all four metrics: Loudness / Sharpness / Roughness / Tonality.
  • A composite sound-quality score that replaces part of the subjective evaluation.
  • Optional: a subjective-objective regression discriminant function (AutoML distillation).
Reference Standards / Engineering PracticeISO 532 series · ECMA-418-2 · DIN 45631/45692

Want to run this workflow on your own data?

Request a Trial
NVH

Other cases in this domain