<!--
{
  "availability" : [
    "iOS: 15.0.0 -",
    "iPadOS: 15.0.0 -",
    "macCatalyst: 15.0.0 -",
    "macOS: 12.0.0 -",
    "visionOS: 1.0.0 -"
  ],
  "documentType" : "symbol",
  "framework" : "CreateML",
  "identifier" : "/documentation/CreateML/MLSoundClassifier/ModelParameters-swift.struct/featureExtractionTimeWindowSize",
  "metadataVersion" : "0.1.0",
  "role" : "Instance Property",
  "symbol" : {
    "kind" : "Instance Property",
    "modules" : [
      "Create ML"
    ],
    "preciseIdentifier" : "s:8CreateML17MLSoundClassifierV15ModelParametersV31featureExtractionTimeWindowSizeSdvp"
  },
  "title" : "featureExtractionTimeWindowSize"
}
-->

# featureExtractionTimeWindowSize

A time duration, in seconds, the training session uses for each audio sample it reads from an audio file in
a dataset.

```
var featureExtractionTimeWindowSize: TimeInterval { get set }
```

## Discussion

The time-window size value defaults to `0.975` seconds and must be in the range `[0.5, 15.0]`.

Training sessions that use
[`MLSoundClassifier.ModelParameters.FeatureExtractorType.vggish(revision:)`](/documentation/CreateML/MLSoundClassifier/ModelParameters-swift.struct/FeatureExtractorType/vggish(revision:)) ignore this value
and always use a time-window size of `0.975` seconds.

---

Copyright &copy; 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)