Speech Recognition & TTS

    AI · Intermediate

    Definition

    Speech recognition converts audio to text; text-to-speech generates audio from text.

    Voice assistants, captions and accessibility tools rely on them.

    Diagram

    waveform→ spectrogram→ model→ text

    Flowchart

    Record audioConvert to spectrogramNeural model predicts tokensDecode text

    Example: Word error rate

    errors, words = 3, 50
print(errors / words * 100, '%')

    Output: 6.0 %

    Key points

    • WER measures ASR quality
    • Whisper is a popular open model

    Interview questions

    What is a spectrogram?

    A picture of how audio frequencies change over time.

    More AI topics

    What is Artificial IntelligenceNeural NetworksGradient DescentCNNs & Computer Vision