Listening at 1.5x: what speed does to comprehension
Speeding up audiobooks and podcasts saves genuine hours. What it costs depends heavily on the material, and rather less on the speed than people assume.
Playback speed controls are now standard, and the arguments about them are unusually polarised — either you're wasting time at 1x or you're not really absorbing anything above 1.2x. The research is more interesting than either position.
What happens to comprehension
Studies on time-compressed speech go back decades, and the broad finding is consistent: comprehension holds up reasonably well up to somewhere around 1.5x for most listeners on straightforward material, then declines more sharply beyond that. The decline isn't a cliff, and where it starts varies a lot between people.
Two factors move that threshold more than anything else:
Familiarity with the subject. If you already know the territory, your brain is filling gaps from existing knowledge rather than building understanding from scratch. Familiar material tolerates much higher speeds. Genuinely new, dense material does not.
Information density. A conversational podcast with three people repeating each other carries very little new information per minute. A technical audiobook carries a great deal. The same speed produces quite different loads.
Why 2x feels fine and often isn't
The subjective experience of comprehension is unreliable. At high speed you follow the thread, recognise the words and feel as though you're keeping up. What tends to degrade first isn't following along — it's retention and the ability to reconstruct an argument afterwards.
The practical test isn't "can I follow this", it's "can I explain what the last chapter argued". Try it once at your usual speed. If you can't, slow down.
The narration variable
Speed increase is relative to the original recording, which varies enormously. A deliberate, slow narrator at 1.75x may land at a more comfortable words-per-minute than a fast one at 1.25x. Average conversational speech sits somewhere around 150 words per minute; professional narration is often deliberately slower.
Which means a single speed setting across your whole library is the wrong approach. Adjust per book, or at least per narrator.
What speed does to the experience
Beyond comprehension there's a separate question, which is that pacing is part of the work. In fiction, pauses carry meaning — a beat before a line of dialogue does something. Compress everything uniformly and that flattens. Good narration is a performance, and speeding it up is closer to skimming a painting than reading a book faster.
For instructional or informational material this matters much less. For a novel you chose partly for the prose, it matters more than the time saved.
A reasonable approach
- Move in small steps. Go up by 0.1x and stay there for a session. Jumping straight to 2x guarantees a bad experience and tells you nothing.
- Vary by material. Fast for familiar and conversational, slower for dense or unfamiliar, slowest for fiction you care about.
- Slow down when you stop being able to rewind mentally. If you can't recall the last two minutes on demand, the speed is too high for that material.
- Use silence-trimming separately. Many apps can shorten gaps between words without compressing the speech itself. On conversational podcasts this recovers a lot of time with no comprehension cost at all, and it's the first setting to reach for.
- Don't optimise for finishing. Getting through more books isn't the goal unless you retain them.
The genuinely defensible position is somewhere in the middle: speed is a useful tool applied selectively, and a bad default applied uniformly.