Quiz sur les spectrogrammes de Mel

Les réponses et les explications apparaissent au fur et à mesure que vous répondez au quiz.

1. Que représente un spectrogramme Mel ?
   • Energy across time and perceptually scaled frequency bands
   • Words recognized from the audio
   • The compressed file size
   • A list of note names
2. What is the mel scale designed to represent?
   • Uniform spacing of all frequencies in hertz
   • A nonlinear frequency spacing related to how listeners perceive pitch differences
   • Only signal loudness
   • Removal of all high-frequency content
3. What transform is commonly used to obtain short-frame frequency content before mel filtering?
   • A text tokenizer
   • Une jointure à une base de données
   • The short-time Fourier transform (STFT)
   • A color histogram
4. Why may a power spectrogram be converted to decibel units?
   • To remove the frequency axis
   • Pour traduire l'audio en texte
   • To reconstruct the original waveform exactly
   • To express power on a logarithmic scale relative to a reference
5. Which parameter controls the number of mel-frequency bands in librosa?
   • n_mels
   • hop_length
   • center
   • sample-format
6. In the Tacotron 2 pipeline, what does the vocoder do after mel prediction?
   • Turns audio into subtitles
   • Converts the predicted mel spectrogram into waveform audio
   • Sets a video frame rate
   • Chooses the number of mel bands
7. What does hop_length specify in a mel-spectrogram computation?
   • The number of microphones in the recording
   • The count of mel bands
   • The number of audio samples between successive analysis frames
   • The length of the generated transcript
8. What does fmax set in a mel filter bank?
   • The number of video frames
   • The length of the output transcript
   • The number of microphones in the recording
   • The highest frequency included in the mel bands