CELPIP SPEAKING · TASK 3
CELPIP Speaking Task 3: describe the scene so a listener can follow it.
The biggest mistake in Describe a Scene is trying to name everything. A stronger response gives the listener an organized mental picture: establish the setting, move through the image in a predictable order and describe meaningful actions with precise language.
What to do in the 30-second preparation time
You do not have time to script a response. Use the image to decide the route your description will follow.
- Name the setting. Park, station, market, office, street, event or another overall location.
- Choose a visual route. Left-to-right, foreground-to-background or centre-then-sides.
- Pick action clusters. Identify several people or groups doing something worth describing.
- Notice one or two precise details. Clothing, objects, relative position, expression or interaction can make the description more vivid.
Write only a few cue words if notes help you. Your preparation should guide the order of the response, not become a script you try to read.
A 60-second scene structure
| Part | Purpose | Useful language job |
|---|---|---|
| Opening | Orient the listener to the overall scene. | “This appears to be…” / “The scene shows…” |
| Route | Move through the picture in a consistent direction. | “On the left…” / “In the foreground…” / “Behind them…” |
| Actions | Describe what people are doing and how they relate. | Precise action verbs and present-progressive forms where natural. |
| Finish | Use the remaining time for a final meaningful detail instead of repeating the setting. | Add a visible interaction, atmosphere or detail you have not yet described. |
Organize before you add vocabulary
Vocabulary becomes useful when it helps the listener locate and imagine details. Spatial language—such as beside, behind, in front of, near the entrance, at the far end—often improves coherence because it tells the listener where to look mentally.
HARD TO FOLLOW
Object hopping
A man is walking. There is a bench. A child has a balloon. There is a tree. A woman is talking. The details may be correct, but the listener has no map.
EASIER TO FOLLOW
Spatial sequence
Begin with the setting, move through one side of the image at a time and connect nearby people through their actions and positions.
What CELPIP evaluates in Speaking
CELPIP publishes four Speaking performance dimensions: Content/Coherence, Vocabulary, Listenability and Task Fulfillment. For Task 3, that means:
- Content/Coherence: include enough relevant details and organize them so the listener can follow.
- Vocabulary: use varied, precise words and phrases naturally.
- Listenability: keep pronunciation, rhythm, grammar, sentence structure and self-correction understandable.
- Task Fulfillment: actually describe the visible scene with relevant detail for the full task purpose.
A pause or self-correction does not automatically ruin a response. The practical question is whether the listener can still understand an organized, relevant description.
Common Task 3 mistakes
- Trying to describe every object. Select meaningful details and organize them.
- Jumping randomly around the picture. Decide the route before speaking.
- Repeating “there is” for every sentence. Mix existence, action and location language.
- Inventing a long story. Focus first on what can actually be seen.
- Stopping early after the obvious details. Train yourself to notice smaller interactions and relationships.
A focused practice drill
Use a fresh scene for each attempt. Give yourself 30 seconds to choose a route, then record for 60 seconds. When you replay it, do not start by counting grammar mistakes. First ask whether somebody who never saw the image could reconstruct the main layout from your description. Then review vocabulary and listenability.
- Did the opening establish the setting?
- Could you hear a consistent visual route?
- Were people described through actions rather than labels only?
- Did spatial phrases make relationships clear?
- Did the response stay understandable even when you paused or corrected yourself?