Skip to main content
20% off your first payment$5.99 for your first week, then $7.49 · Ends November 15, 2026See plans

Free tool · no account · sourced

Dictation drill for TOEFL® Listening practice

Written by the Reach120 research teamReviewed
The TOEFL has no dictation task, and this does not simulate one. It is a diagnostic: hear a sentence, type what you heard, and see word by word what went missing. That tells you something a comprehension question cannot — whether you failed to hear a word or heard it and did not know it.

Dictation drill

SentenceGeology

Your browser does not offer speech synthesis, so the audio button is unavailable here. The sentences are printed below and the drill still works if someone reads one to you — but the listening half needs a browser that supports it, and most current desktop and mobile browsers do.

Why dictation, when the test never asks for it

The TOEFL has no dictation task and this tool does not simulate one. Dictation is a diagnostic: it separates words you did not HEAR from words you heard but did not understand, which a comprehension question cannot tell apart. If your transcription is accurate but the meaning was lost, the constraint is vocabulary. If the transcription itself breaks down, the constraint is listening.

The practical consequence is that the two problems need opposite responses. A hearing problem is fixed by more listening at speed — the words are known, and the bottleneck is catching them in connected speech, where they are shortened, run together and stripped of stress. A vocabulary problem is not fixed by more listening at all, and the hours spent on it are the most commonly wasted hours in TOEFL preparation. Transcribe a sentence and the difference is visible in a way no score can show you: the words you typed correctly but could not explain are the vocabulary list.

The sentences, and what each one is testing

Printed here in full on purpose. This is practice material rather than a test, so nothing is gained by hiding it — and if your browser has no speech synthesiser, someone can read one to you from this list.

1. Geology
The sediment at the bottom of the lake accumulates in annual layers, so counting them gives a record of the climate going back several thousand years.
Long noun phrases with several unstressed function words in a row. "at the bottom of the" is four syllables of almost no volume, and learners who are listening for content words alone lose the structure holding them together.
2. Biology
Some birds navigate by the position of the stars, and a bird raised indoors under an artificial sky will orient itself to whatever pattern it was shown.
A clause boundary with no pause after it. The comma is written but barely spoken, and the second half changes subject, so a listener who is still parsing the first clause misses the start of the second.
3. Economics
When a central bank raises interest rates, borrowing becomes more expensive, and the resulting fall in demand is usually intended rather than accidental.
Three multi-syllable Latinate words close together. "resulting", "intended" and "accidental" share a rhythm, and the -ed endings on the last two are almost inaudible at speed.
4. Art history
The blue pigment used in these paintings was ground from a stone quarried in only one region, which is why its presence can date a work quite precisely.
A passive verb chain, "was ground from a stone quarried in". Two past participles in one breath, and "ground" is a homophone that sends the listener to the wrong meaning until the sentence resolves it.
5. Archaeology
Archaeologists had assumed the settlement was abandoned suddenly, but the pottery found in the upper layers suggests it emptied over two or three generations.
A contrast built on one word. Everything turns on "but", which is short, unstressed and easy to lose — and a listener who loses it hears two agreeing statements instead of a correction.
6. Psychology
Recalling an event appears to rewrite it, which means a memory retrieved often is not necessarily the memory that has been preserved best.
A negation buried mid-sentence. "not necessarily" reverses the whole claim and arrives after twelve words of setup, by which point many listeners have already committed to the opposite reading.

These sentences were written by Reach120 for this tool, on general academic subjects of the kind the Listening section uses. They are not real TOEFL material and no ETS content is reproduced here.

How the marking works

The sentence is read by your own browser’s built-in speech synthesiser — no audio is downloaded and nothing you type leaves the page. Marking compares your words to the sentence’s words after lowercasing and removing punctuation, aligning the two so that a missed word shifts nothing after it. Capitalisation and punctuation are never marked wrong, because the drill is about hearing rather than about typing.

The result is a count of words matched in one transcription. It is not a band, not a Listening score and not a prediction of either.

Common questions

Is there a dictation section on the TOEFL?
The TOEFL has no dictation task and this tool does not simulate one. Dictation is a diagnostic: it separates words you did not HEAR from words you heard but did not understand, which a comprehension question cannot tell apart. If your transcription is accurate but the meaning was lost, the constraint is vocabulary. If the transcription itself breaks down, the constraint is listening.
Are these real TOEFL sentences?
These sentences were written by Reach120 for this tool, on general academic subjects of the kind the Listening section uses. They are not real TOEFL material and no ETS content is reproduced here.
Where does the audio come from?
Your own browser reads the sentence using the speech synthesiser built into your operating system. No audio file is downloaded and no text-to-speech service is called, which is why the drill is free and unmetered. The voice you hear is whichever English voice your device provides, so it will differ between a Mac, a Windows PC and a phone.
The voice does not sound like the TOEFL recordings. Does that matter?
For this drill, not much, and the difference is worth being honest about: a synthesised voice is more evenly paced and more clearly articulated than the recorded speakers on a real test. That makes it easier, so a strong result here is not evidence you would transcribe the real thing as well. What transfers is the diagnosis — which words you miss, and whether you miss them because you did not hear them or did not know them.
How is my transcription marked?
The sentence is read by your own browser’s built-in speech synthesiser — no audio is downloaded and nothing you type leaves the page. Marking compares your words to the sentence’s words after lowercasing and removing punctuation, aligning the two so that a missed word shifts nothing after it. Capitalisation and punctuation are never marked wrong, because the drill is about hearing rather than about typing.
How many sentences are there?
6, each chosen for a different reason it is hard to hear — unstressed function words, a clause boundary with no pause, participle endings that disappear at speed, a homophone, a contrast resting on one short word, and a negation that arrives late. Each one names its own difficulty after you check it, because the reason a word was missed is the part that transfers.

The next step

Free accounts get a metered number of AI-scored recordings; past that it needs a plan.

Say the sentence back and have it checked
All free tools

Cite or link this tool

Free to reference anywhere, with no permission needed. Paste this where you need the link:

<a href="https://www.reach120.com/tools/toefl-dictation-drill">TOEFL dictation drill</a> — Reach120

Or write the attribution in plain text:

TOEFL dictation drill — Reach120. https://www.reach120.com/tools/toefl-dictation-drill

No link is required. The embed carries its attribution as a rel="nofollow" link by default, and adding ?link=off to the embed URL removes the link entirely — the calculator works exactly the same either way.

Every tool here is free and needs no account. Each is reviewed 90 days after it ships: a tool that earns no measurable use is retired rather than left to rot, and its page says so on the way out. This one shipped on 2026-08-23, so its kill-rule review falls due on 2026-11-21.

Nothing you type here is stored or sent anywhere — the tool runs in your browser. This tool does not issue or predict an official score; official TOEFL® scores are issued only by ETS. Reach120 is not affiliated with or endorsed by ETS; TOEFL is a registered trademark of ETS, used here only to name the test this tool is about.

Sourced from the ETS 2026 test specifications, read (link confirmed ). Reach120 is an independent practice tool, not affiliated with or endorsed by ETS; the document is named here only to say what this page practices.

Reach120 is an independent practice tool. It is not affiliated with, endorsed by, or approved by ETS, and it does not provide official TOEFL® test scores.

TOEFL® is a registered trademark of ETS. This product is not endorsed or approved by ETS.