Record Yourself Speaking: The Only Honest Mirror

Most people would rather do forty grammar exercises than open their phone recorder and talk for two minutes. The grammar exercises feel productive. The recorder feels like a test. But the recorder gives you something a worksheet never will: a clear record of what you actually sound like.

Hearing your own voice is uncomfortable in your first language. You hear a thinner, stranger version of yourself. In a second language, it feels worse. Every missing ending is on the file. Every pause is on the file. You cannot argue with it. That is exactly why it is the single highest-information practice act you can do alone.

I have a folder of recordings from my first year of learning another language. Some of them make me wince. That is the data.

Why people avoid it

Your voice sounds wrong. When you speak, part of the sound reaches your ears through your skull. A recording removes that. What is left sounds like a different person. That is true in your native language. In English, you also hear your accent, your slow spots, and your mistakes in real time. It feels too direct.

That feeling never fully disappears. You learn to listen through it.

But avoiding the recorder means you only hear the version of your speaking that your brain edited. That version is not real. You need the unedited one.

What your memory hides

While you speak, your brain is busy planning the next word. You do not hear yourself the way a listener does. You remember what you meant to say. You forget the long silence after "because." You do not notice that you said "em" nine times in ninety seconds. You think the final -s is there because you know it should be there.

A recording removes that protection. You might think you said: "I went to the store because I needed milk." The recording says: "I went to the store uh... because uh... I need... milk."

Three things show up within the first month.

• Fillers you never noticed: uh, em, like, a repeated so.
• Pauses that felt like half a second but lasted four seconds.
• The same three mistakes every week: missing third-person -s, using "since" with a period of time, saying "depend of" instead of "depend on."

You do not need a teacher to point these out. You need to listen once.

Two videos that fit this solo work:

• "No Partner? No Problem!" — three ways to practice alone
• "Daily English Speaking Practice" — built for repeating

The weekly protocol

Pick one prompt and keep it for at least four weeks. The prompt should be practical: describe your morning, explain how to get from your home to the nearest market, or summarize a short video you watched. Using the same prompt means you compare the same task. Changing the topic every week hides progress.

Do not write anything down. Do not prepare in your head for five minutes. Start the recorder, speak for two minutes, stop. Use the default voice recorder on your phone. Nothing else matters.

Do not listen the same day. Your first reaction will be emotional, not useful. Wait until the next morning. The distance makes it listenable.

When you listen, count. Do not just feel bad or good.

• Count long pauses: any silence over two seconds. Mark a line for each one.
• Count fillers: uh, em, like, you know.
• Count words per minute. Play the first 30 seconds, count the words, multiply by two. For target ranges, see how fast you should speak English.

The pause table in the fluency guide shows what different pause patterns mean. A pause before a known word often points to retrieval trouble. A pause after a grammar word often points to a grammar search.

For most B1-B2 learners, 110 to 140 words per minute is a clear range. Under 90 usually means you are stopping too often. Over 160 often means clarity drops.

Do not retake the recording. If you fall apart at minute one, that is the evidence. Next week, you will know where it breaks.

If you want the full daily plan, see the practice guide.

Compare yourself to a model

Recording alone shows what is wrong. Recording over a model shows what to fix. Pick a short clip from a speaker you understand. A clear teacher or a news anchor works better than a fast comedy scene.

Play one sentence. Start the recorder. Say the same sentence. Play your version, then play the original. Go back and forth. The difference is concrete: stress, rhythm, where the pitch rises, where a vowel is shortened.

The pronunciation guide calls this record-and-compare. It works best with five or six short sentences, not a long video.

In English Shadowing, tap the repeat button to loop one sentence, start your phone recorder while the loop plays, and then play both. The word-by-word translation keeps you from repeating sounds without knowing what you are saying.

How to score a recording

Use a table. Do not grade your accent. Grade whether a stranger could understand you.

Score the same dimensions for four weeks. One week is noise. Four weeks is a pattern.

The archive

Save each recording with the date. Do not delete the bad ones. They are the baseline.

Every month, open the first recording and the fourth. Listen to thirty seconds of each. Progress is invisible day to day. Across eight weeks, it becomes loud. You will hear shorter pauses, fewer fillers, clearer endings. If you do not hear that, you have a precise list of what to work on next. That beats a vague feeling of being stuck.

The guide on being stuck at B1 explains why progress hides in plain sight. The archive is how you pull it out.

In the app

English Shadowing makes the model comparison step easy. Open any YouTube video, loop one sentence with the repeat button, and record over it with your phone. The word-by-word subtitles and translation keep the meaning in front of you while you speak.

The app is free at englishshadowing.app. Nothing to install.