Définition
Audio classification helps you automatically understand what is happening in an audio signal. The model can recognize applause, alarm, genre of music, machine noise, dog barking, type of call or emotional tone of speech. This is different from speech recognition, where the main goal is to retrieve the text.
Exemple
The smart home system recognizes the sound of breaking glass and sends a warning to the owner.
Pourquoi c'est important
The term is important for finding tools that work not with text, but with real sound: security, monitoring, media and voice products.
Fonctionnement
The sound is converted into features or a spectrogram, then the model classifies the fragment into predefined categories or probabilities.
Où c'est utilisé
- smart home
- audio moderation
- equipment monitoring
Limites
The quality depends on the noise, microphone, fragment duration and set of classes. The model may confuse similar sounds and perform poorly outside the training domain.
FAQ
Why is “Audio Classification” useful to know?
The term is important for finding tools that work not with text, but with real sound: security, monitoring, media and voice products.