Add Echo to Audio

Echo is the simplest effect there is to describe and the easiest to get wrong: a copy of your sound, arriving later, quieter each time. This tool gives you the two controls that matter — how long until the first repeat, and how many repeats there are — plus the ones that separate a delay you would actually use from a novelty.

A repeat placed a fraction of a second behind a sound makes it feel bigger without changing what it is. Placed on the beat, it turns a flat part into something that moves. Echo is the oldest production effect for a reason: it costs nothing in arrangement and it puts a dry recording in a place.

• Vocals: a short repeat thickens a lead without being heard as an effect; a tempo-locked one adds movement between phrases.
• Guitars and synths: a dotted-eighth delay is the sound of most records you can name — it fills the space between notes and pushes the track forward.
• Drums and percussion: a bounced repeat on a snare or a shaker widens a mix without adding another part.
• Podcasts and voiceover: a subtle repeat under the voice adds presence; anything more hurts intelligibility.
• Sound design: long delays with high feedback and heavy damping turn a single hit into a receding tail.

Your audio is split in two. One copy passes through untouched. The other is delayed by the time you set, filtered if you asked for it, turned down by the feedback amount, and then sent round again — and again, until it falls below hearing. Those repeats are added back to your untouched original at the level you choose.

That loop is why feedback and echo level are different controls. The loop's gain decides how many repeats there are; the level decides how loud they sit against the original. It also means the tool knows in advance exactly how many repeats it will produce and exactly how much longer your file will be, and it shows you both before you render.

A delay that lands between beats fights the track. Enter your BPM and pick a note division, and the delay time is worked out from the tempo instead of being dialled in by ear — a quarter note at 120 BPM is 500 ms, an eighth is 250 ms, a dotted eighth is 375 ms. The resolved time is shown in milliseconds beside the control, so you can check it against your project rather than trusting a label.

• Dotted eighth: the most-used delay setting in modern production. It syncopates against the beat and fills the gaps between notes.
• Triplets: give a rolling, circular feel; useful under half-time drums where straight divisions sound stiff.
• Sixteenths: at fast tempos these blur into a single thickened sound rather than separate repeats.
• Out of range: very slow tempos on long divisions can ask for a delay longer than the tool offers. When that happens the control says so rather than silently using a different time.

With the spread control up, each repeat lands on the opposite side from the last: left, right, left, right, receding as it goes. It is the widest-sounding thing a delay can do and it costs nothing in level — the repeats are placed, not turned up. Keep it moderate on lead vocals, where a voice moving between speakers competes with the words, and take it as far as you like on guitars, synths and percussion. At zero the repeats stay where your source is.

Real echoes lose their high frequencies every time they bounce, which is why a tape delay sounds warm and a digital one can sound brittle. Damping applies a filter once per repeat rather than once to the whole mix, so the second repeat has been filtered twice and the tenth ten times. The result fades in tone as well as in level, which is what makes a long delay sit behind a track instead of cluttering it.

Thinning does the same from the other end, taking low frequencies out of each successive repeat. It is the fastest way to stop a delay on a bass-heavy source from turning the low end to mud. Both leave your original completely alone — they are applied inside the repeat loop, not across the finished mix.

After every render the tool sends a test signal through the exact chain your audio just travelled and reads back where the repeats actually landed. The delay time and repeat count it reports are measured from that, not copied from the settings — so if anything in the chain had moved them, you would see it rather than trusting a label.

• Delay time: the first repeat lands on the sample the setting asks for, with no drift between repeats.
• Your original: at the same sample rate, the audio before the first repeat arrives is identical to what you put in.
• Peak level: the finished file is measured and its true peak is shown, so the clip guard's ceiling is a fact rather than an intention.

Unlock the two sides and the left and right channels run at different delay times — a dotted eighth against a quarter, say. It is the oldest stereo delay trick there is, and it works because the ear places a sound by WHEN it arrives at each side rather than by how loud it is there, so the repeats open up across the image without any panning at all. It is a studio feature that browser-based echo tools do not offer.

Both sides are held to the same accuracy: each is an exact number of samples, and after every render the tool measures and reports them separately. Stereo bounce is unavailable while the two sides differ, because bouncing repeats between channels that already run at different times makes neither control mean what it says.

Bring in MP3, WAV, FLAC, OGG, M4A or AAC, or drop a video file and the audio track is taken from it. Write out MP3, WAV, FLAC, OGG, AAC or M4A. Anything the tool cannot open is refused at the file picker with a link to the converter, so you never wait on a spinner that then fails.

• Free: one file at a time, up to 50 MB and ten minutes, written at up to 44.1 kHz and 192 kbps.
• Pro: up to 25 files in one batch, downloaded as a single ZIP, written at up to 48 kHz, 320 kbps and 24-bit.
• Every part of the effect — all six presets, the full delay and feedback range, tempo lock, ping-pong, damping and the clip guard — is available on both plans.
• Tags are carried across when you ask for them. Embedded cover art is not, on any format.

• Start with a preset: Slapback for thickening, Rhythmic for movement, Dub for something you will notice. Then move one control.
• Set feedback before level: decide how many repeats you want, then decide how loud they are. Doing it the other way round means re-balancing twice.
• Listen before you render: the Listen button plays the effect live, so finding a setting costs no renders at all.
• Damp long delays: anything above 300 ms with real feedback wants damping around 4–6 kHz, or the repeats stay in front of the track.
• Use pre-delay on speech: 20–50 ms of gap between the voice and the first repeat keeps consonants readable.
• Leave the clip guard on: repeats add to the original, so it is easy to exceed full scale without noticing. It costs nothing when it is not needed.

Nothing is uploaded. Your audio is read, processed and written back on your own device, and no part of it travels over the network at any point — which is also why the result appears without a wait for an upload and a download.

Powered by optimised audio processing algorithms.

• Audio reverb
• Reverb + echo
• Audio equalizer
• Audio compressor
• Audio converter