What it is: Listen and write exactly what you hear. What it’s for: Fix weak points: missing sounds, linking, endings (-s/-ed). Example: Write 2–3 sentences, then compare with transcript.

• Pick 10–20 seconds audio.
• Listen 3 times and write.
• Compare with transcript and mark mistakes.
• Repeat the same clip until you reach 90–100% accuracy.

• Word boundaries (where words start/end)
• Reduced forms (gonna, wanna)
• Grammar endings (works, played)

• *dictation* /dɪkˈteɪʃən/: nghe-chép
• *boundary* /ˈbaʊndri/: ranh giới
• *ending* /ˈendɪŋ/: đuôi (từ)