How CELPIP Speaking Works: 8 Tasks, About 20 Minutes
Eight tasks, fifteen minutes of timed clock, and a recorder that starts on its own when preparation runs out. Nothing in this section can be paused, replayed or recorded twice.
The section in one page
CELPIP-General Speaking is eight tasks long. Each task gives you a fixed block of preparation time, then a fixed block of recording time, and both blocks reset when the next task loads. Added up, the preparation comes to 300 seconds and the recording to 540 seconds. Task 5 adds a further 60 seconds for choosing an option, which brings the timed part of the section to exactly 900 seconds.
That is fifteen minutes of clock that belongs to you. The section as a whole is usually quoted at around twenty minutes, and that is also true: the instruction screens, the microphone check and the moment between tasks all take real time, but none of it is time you can spend on an answer. Plan for twenty minutes in the chair and fifteen minutes of actual work.
The eight tasks are not eight versions of one task. Two of them hand you an illustration. One hands you photographs of options and makes you pick. One hands you an awkward situation and two different people you could take it to. Three hand you nothing but a sentence of text. The only thing they all share is the shape of the clock.
The eight tasks and their timings
Task 1, Giving Advice. 30 seconds prep, 90 seconds speaking. Somebody has a problem and has asked you what to do. Ninety seconds is one of the two longest answers on the test, which means two or three developed pieces of advice rather than a list.
Task 2, Talking about a Personal Experience. 30 seconds prep, 60 seconds speaking. One event from your own life, narrated and then reflected on. Sixty seconds is one story told properly.
Task 3, Describing a Scene. 30 seconds prep, 60 seconds speaking. An illustration of a busy everyday place, described to somebody who cannot see it. The picture stays on screen while you talk.
Task 4, Making Predictions. 30 seconds prep, 60 seconds speaking. The same illustration returns and you say what will most probably happen next. Describing it again is the commonest way this task is failed.
Task 5, Comparing and Persuading. 60 seconds to choose, then 60 seconds prep, then 60 seconds speaking. Three screens rather than two. You choose one option, and then argue for it against a different option that somebody else wants.
Task 6, Dealing with a Difficult Situation. 60 seconds prep, 60 seconds speaking. An uncomfortable situation and two people you could address. You pick one and speak to them directly.
Task 7, Expressing Opinions. 30 seconds prep, 90 seconds speaking. A question about an everyday issue with nothing on screen to lean on. The other ninety second answer, and the hardest one to fill.
Task 8, Describing an Unusual Situation. 30 seconds prep, 60 seconds speaking. Something odd on screen, described down the phone to a named person, ending in a request.
Where the fifteen minutes actually go
Six of the eight tasks give you 60 seconds to speak. Only Tasks 1 and 7 give you 90, and those two are the argumentative ones with no picture attached. That is not an accident: an argument needs room to concede and answer, and a description does not. If you are going to practise anything to a stopwatch, practise those two, because 90 seconds is considerably longer than it sounds when you have run out of reasons at the fifty second mark.
Preparation is not evenly distributed either. Tasks 5 and 6 get a full 60 seconds of it while everything else gets 30. In both cases you are being given the extra half minute because there is a decision to make before there is anything to plan. Spending that minute the way you would spend thirty seconds on Task 2 wastes the part of it that was the point.
The recording clock is the one that matters most to your score, because how much of it you use is measured directly. An answer that stops at twenty seconds on a sixty second task has not been developed, and no amount of accuracy in those twenty seconds repairs that.
Task 5 has a third screen, and it surprises people
Every other task is prepare, then speak. Task 5 is choose, then prepare, then speak. The first screen shows you two options and gives you 60 seconds to pick one, with no recording running at all. The second screen restarts the clock for 60 seconds of preparation. The third gives you 60 seconds to talk.
The detail that catches candidates is what happens between screens. The option you rejected disappears. On the speaking screen you see your own choice and a different option that somebody else is pushing for, which you have not seen before. So the comparison you owe is not between the two things you were shown at the start. It is between the thing you chose and the thing the other person wants.
If you do not choose in time, the system picks for you and moves on. That costs you nothing directly, since nothing on the choosing screen is recorded, but you then spend your minute arguing for something you did not select. Choosing quickly beats choosing well here. Neither option is better, only easier for you to talk about.
Write your first sentence out in full during preparation and nothing else in full. The opening is where dead air happens, and having the first ten words ready means the recorder catches speech from the first second rather than a pause.
No pausing, no going back, no second take
Preparation time cannot be shortened and cannot be extended. When it hits zero, recording begins on its own, whether or not you have finished planning. The most common consequence is a candidate still writing when the microphone opens, which produces four or five seconds of dead air at the front of the answer. That silence is part of your recording and part of the duration you were given.
Once you leave a task there is no way back to it. There is no pause control, no replay and no option to record the answer again. This is why a plan that depends on getting a good run at it is not a plan. Whatever comes out in those sixty or ninety seconds is the answer.
The practical consequence is that recovery has to happen inside the answer rather than after it. If you mangle a sentence at second fifteen, the move is to close the sentence, say the point again cleanly and continue. Stopping to restart from the top burns the time you needed for your third idea, and a self-correction that is handled in three words costs almost nothing.
What this section is not testing
It is not testing your accent. There is no criterion for sounding Canadian and no advantage in trying to. What is assessed under Listenability is how much work a listener has to do: your pace, your pausing and whether the words come out clearly enough to be caught the first time. A strong regional accent that is clearly articulated at a steady pace is not a problem. Unclear, rushed speech in any accent is.
It is also not testing whether your opinions are correct, whether your advice is wise or whether your story actually happened. None of that is checkable and none of it is scored. What is scored is the language you build the answer out of, and how completely you did the thing the prompt asked for.
Finally, it is not a memory test. The prompt stays on screen while you speak, and on the picture tasks so does the image. Nothing needs to be held in your head. Your thirty seconds are for deciding what to say, not for memorising what you can already see.
Learn the two 90 second tasks by number, not by name. Tasks 1 and 7 need a third idea in reserve, and the moment to work out what that third idea is during preparation, not at second fifty five with the clock still running.
Practise it
Five questions on the shape of the section, the timings and the rules that cannot be worked around.