In Task 8 you look at a picture of something genuinely strange, then call a friend and describe it so they can picture it too. You get 30 seconds to prepare and 60 seconds to answer. The person on the phone cannot see the image, and raters are not testing whether you identify it correctly, they are testing whether your words let someone else see it. The demo above uses the exact interface and timing of the real test, so everything you practice here transfers directly.
Trained raters score every CELPIP speaking task on the same four equally weighted dimensions. The picture is strange by design, and each dimension rewards a different part of how you talk your way through it.
The number and quality of your details and the order you deliver them in. An overall impression first, then specifics in a logical path beats jumping randomly around the picture.
Range and precision. Comparisons like about the size of a suitcase or shaped like a snail shell count as precision, you never need the exact technical name for anything.
Rhythm, pronunciation, intonation, and grammar working together. A short silent pause while you find a description beats a string of fillers every time.
Did you stay on the phone, describe the scene for your listener, and keep going for the full time? A sharp description with no greeting or closing still leaves points here.
Find the strangest thing in the picture and pick two or three details you can describe. When the 30 seconds run out you hear a loud "Start speaking now" and recording begins immediately, there is no button to press.
Greet your friend, say why you are calling, and give the scene in one sentence, something like you will not believe what I am looking at. It sets the phone tone and buys your brain time to organize.
Work through two or three details in a logical order and give each one a color, a size, or a position, then compare it to something familiar. When something has no obvious name, saying what it looks like is exactly what raters want to hear.
Say what you make of it all, offer a guess with it might be or I have a feeling, then close like a real call. In this demo that means inviting your friend to come see for himself, an answer that just stops reads as unfinished.
These phrases turn a missing word into a description and a guess, which is exactly the skill Task 8 measures. Practice them until they come out automatically.
Stopping dead because you do not know a word is the classic Task 8 failure. Describe around it instead, a long metal stick with a flat end says far more than silence.
Saying there is a truck and some people gives a list, not a picture. Attach a color, a size, or a comparison to each thing so the listener can actually see it.
Launching into description with no greeting and stopping with no goodbye sounds like a read report. Tone is scored, so talk to your friend, not at the picture.
Racing through ten details leaves all of them vague. Two or three elements described properly beat a complete inventory delivered in a panic.
You are describing something truly strange, so let your voice show it. A monotone run through kills the rhythm and intonation raters listen for.
Describe it instead. Say what it looks like, how big it is, or what it reminds you of, raters reward the workaround because describing around a missing word is a real communication skill, not a failure.
No, there is no correct answer to check against. Raters score how clearly you describe and how naturally you speculate, so a hedged guess like it might be some kind of costume scores just as well as a right one.
You are not supposed to cover it all. Pick the two or three strangest elements and describe them properly, a focused answer that fills the full minute scores higher than a rushed tour of everything.
