Speech Technology

Module code: EE5C04 · Semester: Semester 1

Resources

• Module descriptor (TCD)

Questions by Year

0 of 8 done · tick papers and add notes as you go

• 2025/26
• 2024/25
• 2021/22
• 2020/21
• 2019/20
• 2018/19
• 2017/18
• 2016/17

Papers open on tcd.ie (sign in to TCD first).

Topics

• Speech production and acoustic phonetics LO1
• Short-time analysis and pitch LO2
• Formants and spectrograms LO3, LO6–7
• Hearing and perception; Mel scale LO4–5
• Machine learning for speech recognition, speaker ID and synthesis LO8, LO14
• Non-verbal cues in conversation LO9
• Bias, ethics, privacy and social impact LO10–13

Learning outcomes (exam covers LO1–14)

1. Relate speech production to the acoustic-phonetic properties of a language
2. Use short-time analysis to study time-domain properties such as pitch
3. Use formant analysis to compare groups of speakers
4. Explain the human auditory system
5. Relate non-uniform frequency bands (e.g. Mel) to human hearing
6. Interpret speech in a spectrogram
7. Design spectrogram analyses to reveal time-frequency properties
8. Analyse machine learning speech systems: recognition, speaker ID, synthesis
9. Evaluate verbal and non-verbal cues in conversation
10. Critique speech technology performance for minority users
11. Consider ethics and privacy in speech technology
12. Identify how speech technology helps society
13. Explore the state of the art: AI bias, multimodality, low-resource languages
14. Design a deep learning speech application

Summarised from the module descriptor, checked November 2026.