What is delayed auditory feedback?

DAF plays your own voice back into your ears a fraction of a second after you speak, usually between 50 and 200 milliseconds. The mismatch between what you are saying and what you are hearing changes how you speak, and for many adults who stutter it changes it immediately and noticeably.

Does DAF work for stuttering?

For a substantial number of adults it produces an obvious change in the room, on the first try. What is far less clear is whether the effect lasts once the device comes off, and the research on carry-over is thin. Use it as a practice tool rather than a solution.

What delay setting should I use?

Start at 50 ms and work up. Shorter delays around 50 to 75 ms feel closest to natural speech; longer ones near 150 to 200 ms have a stronger effect but slow you down and are harder to talk through. The right setting is personal and it is worth ten minutes of trying rather than accepting a default.

Why does hearing my voice late change how I speak?

Nobody can say with certainty. The most common explanation is that the delay makes you attend to speech as an ongoing action rather than launching each word and moving on, which slows the rate and softens the starts. A second explanation is simpler — it is a distraction, and novel distractions have long been known to change speech temporarily.

Is a DAF device better than a DAF app?

A dedicated device is discreet and always ready, and it costs several hundred to a few thousand US dollars. An app on a phone with wired earbuds does the same signal processing for a fraction of that. Latency matters more than brand — Bluetooth adds its own delay on top of yours, so use wired headphones.

Delayed auditory feedback (DAF) · Stuttering Recovery

DAF plays your own voice back to you a fraction of a second late, and for many adults who stutter that changes speech immediately. What it does, what the research supports, and how to choose a delay.

Speak into a microphone, hear your own voice back a tenth of a second later, and something odd happens: your speech changes. For most people it becomes halting and strange. For a large share of adults who stutter, it goes the other way, and it goes there immediately.

That effect is the whole of delayed auditory feedback. It is one of the oldest findings in the field, it is genuinely striking the first time, and it is surrounded by more overclaiming than almost anything else in this subject. This page is about what it does, what is actually known, and how to use it without building a dependency.

What DAF does to your voice, in one paragraph

Your ears are part of how you speak. You monitor your own voice continuously and adjust in real time, far below the level of conscious attention. DAF breaks that loop by inserting a delay, usually between 50 and 200 milliseconds, between what you produce and what you hear. Your speech system now receives feedback about a sound you finished making a moment ago. The usual result is that you slow down, stretch vowels, and start words more gently — which happen to be the same three things fluency shaping spends months teaching.

Try it now

If you want to know whether this does anything for you, the answer takes about ninety seconds and costs nothing.

Use wired headphones or earbuds. This matters more than anything else on this page: Bluetooth adds 100 to 300 ms of its own latency, on top of the delay you set, and you have no way to know how much. A free browser tool with wired earbuds is a more accurate test than an expensive wireless setup.

Then read a paragraph aloud at your normal pace and notice three things: whether your rate changed without you deciding to change it, whether the starts of words felt different, and whether it was comfortable or maddening. All three answers are useful, and the third one is not a failure — a meaningful minority of people find DAF actively unpleasant and get nothing from it.

Why the delay changes speech, as far as anyone knows

Two explanations compete and both are probably partly right.

The feedback account. Speech is a closed loop, and you normally run it fast and automatically. Delay the returning signal and the loop cannot run automatically any more, so speech shifts towards a slower, more monitored mode. Slow, monitored speech has fewer disruptions in almost anyone, which is why the effect looks so dramatic.

The distraction account. Novelty changes speech temporarily. Speaking in time with a metronome, in chorus with another person, in an unfamiliar accent, while listening to noise — all of them produce a similar effect, and all of them fade as the novelty does. On this account DAF is a very good member of a large family rather than something unique.

The reason this matters practically: if the feedback account dominates, the change should transfer with practice. If distraction dominates, it will fade. The honest position is that both are happening, in proportions nobody has pinned down, and that the transfer question is the one the research has not answered well.

What the research shows, and where it is weak

What is reasonably solid:

The immediate effect is real, repeatable, and observed across decades of studies going back to the 1950s.
A substantial proportion of adults who stutter show a clear change in the room while DAF is running.
Combining DAF with frequency-altered feedback — shifting the pitch of the returning voice as well as delaying it — tends to work better than delay alone, which is why commercial devices do both.

What is weak, and what sellers of devices tend to skip:

Carry-over is barely studied. The question everyone wants answered is what happens after you take it out. The follow-up data is thin, the sample sizes are small, and the studies rarely run long enough to matter.
Adaptation happens. Some users report the effect diminishing over weeks of continuous use, which is exactly what the distraction account predicts.
Non-responders are common. A meaningful share of adults get nothing, or find it worse. There is no reliable way to predict who in advance, which is a good argument for trying it free before paying for anything.
Real-world settings are harder than labs. A noisy room, a phone call, several people talking — these are the situations that matter and the ones least represented in the literature.

None of that makes DAF useless. It makes it a tool with a known immediate effect and an unknown durable one, which is a perfectly reasonable thing to use as long as you know that is what you have.

Choosing a delay: 50 ms, 75 ms, 120 ms

There is no correct setting and the defaults are arbitrary. Spend ten minutes finding yours.

50 ms. The lightest. Closest to natural, easiest to hold a real conversation through, weakest effect. Start here.
75 ms. A common middle. Noticeable, still usable for talking to another person.
120 to 150 ms. Strong. Slows you down considerably. Good for deliberate practice, difficult for actual conversation.
200 ms and beyond. Most people find this disruptive rather than helpful.

Two rules of thumb. Shorter delays suit faster speakers. And if you find yourself fighting the delay rather than settling into it, the number is too high.

Devices compared with apps

A dedicated in-ear device is discreet, always available, and built for the job. It also costs several hundred to a few thousand US dollars, and the price is largely about the hardware engineering rather than about the signal processing, which is not complicated.

A phone app with wired earbuds does the same processing. The trade is that you have a phone in your hand and a cable running to your ear, which is fine in some situations and not in others.

If you are deciding, the sequence that wastes the least money is: try a free browser version with wired earbuds, then a cheap app for a few weeks in real situations, and only then consider hardware — and only if you have established that you are a responder and that the effect has not faded.

Who it helps and who it does not

It tends to help if you experienced a clear change on the first try, if your difficulty is mostly blocks and repetitions rather than avoidance, and if you want a practice aid rather than something to wear all day.

It tends not to help if you felt nothing, if it was actively unpleasant, or if the hardest part of your day is the calls you do not make. DAF changes the mechanics of speaking. It does nothing at all about avoidance, and avoidance is what most adults who stutter would name as the real cost if you asked them at the end of a week rather than in the middle of a block.

Using DAF as practice, not as a crutch

The distinction worth holding onto: DAF used as equipment is something you depend on and lose when the battery dies. DAF used as practice is a way of feeling what slower, gentler speech is like from the inside, so you can find it again without help.

That second use is the durable one. Read aloud with DAF running for five minutes, then read the same passage with it off and try to keep the shape. That is the exercise. The goal is not to keep the device in — it is to be able to reproduce, unaided, the thing the device showed you was available.

Every technique that shows up under DAF has its own page in the techniques section, and easy onset is the closest match to what you will feel happening. For the field’s own current summary of where altered feedback sits among the options, the ASHA practice portal is the reference to read.

If you want to try DAF, look for a dedicated DAF app or an in-ear device, and always start with wired headphones. Stuttering Recovery does not include DAF — our focus is rehearsal for real conversations.