What Task 3 actually asks
An image appears. You describe it into a microphone.
That sounds like the easiest of the eight tasks. It produces some of the weakest answers on the test, and always for the same reason.
The failure: naming instead of describing
"In this picture I can see a man. There is a woman. There is a table and some chairs. There is a dog near the door. There are two children. There is a window and outside the window there are some trees."
Every word is correct. A listener has learned almost nothing.
That is an inventory, not a description. It tells you what is present. It does not tell you what is happening, where anything is in relation to anything else, or what kind of moment this is.
Here is the same scene, described:
"It looks like a family kitchen in the middle of a Saturday morning. A man is at the stove with his back to the room, and a woman is at the table with a laptop open, though she is looking over at two children on the floor rather than at the screen. The children have emptied a box of something — cereal, maybe — and one of them has clearly just realised it was a mistake. A dog is standing right at the edge of the mess, very interested."
Same objects. Completely different answer. And it is not longer because of vocabulary — it is longer because it says where things are, what they are doing, and what is going on between them.
The one instruction that fixes this task
Describe it to someone on the phone who cannot see it.
That single reframe fixes almost everything. When you are on the phone you never say "there is a man." You say "there's a guy at the stove with his back to us." Position, action and relationship arrive automatically, because otherwise the person on the other end has nothing.
This is not a trick — it is the actual demand. Task 8 makes it explicit by asking you to describe an unusual image to someone who cannot see it. Task 3 wants the same thing without saying so.
The shape that works
- One sentence for the whole scene. Where are we, roughly when, what kind of moment is this?
- The main action. Whoever is doing the most important thing, and what they are doing.
- Work outward from there — who else, where in relation to the first person, what they are doing.
- The background. Weather, time of day, what is out of the window, what is on the walls.
- One small detail nobody would list. Something at the edge. This is what makes a marker believe you were actually looking.
Notice that steps 1 and 5 are the two most people skip, and they are the two that separate a description from an inventory.
Language that carries the weight
Not vocabulary — position and action.
- at the back / in the foreground / off to one side / just behind her / at the edge of the table
- is leaning / is reaching for / has just put down / is turned away from
- it looks like / it seems to be / judging by the light
That last group matters more than it looks. You are allowed to guess. "It looks like a Saturday morning" is not a claim you can be wrong about — it is exactly what a fluent speaker says about a photograph, and it lets you describe a situation rather than only its contents.
What is not being marked
- Not your accent. Being easy to follow matters; sounding Canadian does not.
- Not whether you are right. Nobody knows whether it is Saturday. Commit anyway.
- Not how many objects you named. This is the trap. Naming more things makes the answer worse, not better, because it crowds out the description.
And the rule underneath everything:
A response giving insufficient information to assess scores 0 or 1.
An inventory can hit that. Sixty seconds of "there is… there is… there is…" has described nothing that could be assessed.
Two habits to practise
Practise on your own photographs. Open any photo on your phone and describe it aloud for forty seconds, to an imaginary person on the phone. You will hear yourself sliding into "there is" within about ten seconds. Catching that is the whole skill.
Start with the whole before the parts. If your first sentence is "it looks like…", the rest of the answer usually arranges itself. If your first sentence is "I can see a man", you are already making a list.
Try one
[Try Task 3: Describing a Scene →] — one full task, marked, no card required.