Why can’t I hear my own accent?

The two reasons, separately

• Bone conduction. Your skull carries sound to your ears directly. Recorded, that path disappears, which is why your own voice sounds wrong on a recording — the recording is the accurate one.
• Expectation. You hear the word you meant. A listener hears the word you said. When those differ, only the listener notices.

What to do about it

Record one phrase. Play the original, then yours, back to back, twice. Write down one difference — not five. Then shadow the phrase again with that one difference in mind. This single loop does more than any amount of careful speaking.

What the research says

According to Whitworth, B. & Rose, H. (2025). A Systematic Review of Research on the use of Shadowing for Second Language Pronunciation Teaching. Research Synthesis in Applied Linguistics, 1(2). — a systematic review of 44 studies found that shadowing training can improve learners’ comprehensibility, intelligibility and accentedness, along with suprasegmental control such as fluency and prosody. The same review found the evidence on individual sounds (segmental control) inconclusive. Notably, the review criticises studies that rely on acoustic measurements without real listener judgements: what a listener hears is the measure that counts, which is as true for your own practice as it is for research.

Frequently asked

Why does my recorded voice sound so bad to me?

It does not sound bad to anyone else — it is simply the first time you are hearing yourself the way everyone else always has.

How often should I record myself?

Once a session is plenty. The point is one honest comparison, not a collection.

Source

Whitworth, B. & Rose, H. (2025). A Systematic Review of Research on the use of Shadowing for Second Language Pronunciation Teaching. Research Synthesis in Applied Linguistics, 1(2). DOI · open-access copy