Audio Effects Online is a browser-based tool that applies exactly one of six bounded effects — a 200 millisecond echo, a lightweight three-tap reverb, reverse, fade in, fade out, or peak normalization — to a local audio file and exports the result as a PCM16 WAV. Each effect runs with a whole-number amount from 0% through 100%, which blends or scales that operation. All decoding, sample processing, and WAV encoding happen in the current browser tab, so the original file never leaves the device. The target reader has a single audio clip — a voice take, a music stem, a sound design sketch, or a downloaded podcast segment — and wants one specific change applied quickly. They do not want to install a digital audio workstation, learn a multitrack workflow, or sign up for a cloud render service, and they want predictable results: the same input, effect, and amount must produce the same output every time. That contract — six disclosed effects, one effect per run, fixed amounts, local processing, and a standards-based WAV — is what this article walks through, starting with what each of the six operations actually does to the audio samples.
The Six Effects and What Each One Does
Audio Effects Online does not give you a rack of seventeen plugins. It gives you six operations with disclosed math. Knowing what each one actually changes helps you pick the right tool for the task instead of guessing or stacking effects on top of each other by trial and error.
| Effect | What it does to the samples | When to reach for it |
|---|---|---|
| Echo | Adds a copy of the input delayed by 200 ms, scaled by half the selected amount, on top of the original | Slap-back vocals, doubled sources, call-and-response dubs |
| Lightweight reverb | Adds three discrete early reflections at 29.7, 37.1, and 41.1 ms with disclosed weights of 0.50, 0.35, and 0.25 multiplied by the amount | Quick space or "air" on a dry recording for a sound-design draft |
| Reverse | Reads sample frames in opposite order and blends with the original according to amount | Reverse cymbals, hidden messages, experimental transitions |
| Fade in | Scales from the selected dry proportion at the start to full level at the end | Smooth intros, soft openers, removing a hard click at the head |
| Fade out | Scales from full level at the start to the dry proportion at the end | Clean endings, smooth cutaways, removing a hard click at the tail |
| Normalization | Finds the largest absolute sample and rescales the whole file toward a 0.95 peak | Files that were recorded too quietly or too hot, before downstream processing |
The reverb and echo rows are deterministic by design. The reverb does not simulate a room, plate, spring, or branded hardware processor, and it does not model diffusion, damping, frequency-dependent decay, stereo diffusion, or a professional tail. Treat it as a quick draft for space, not as a mixing-stage decision. The same is true for normalization: it scales the sample peak to a fixed 0.95 target and does not measure LUFS, perceived loudness, dynamics, true peak between samples, or broadcast compliance. Silent input stays silent.
Apply an Audio Effect Online in Your Browser
The page is built around three short actions. The full run takes longer to describe than to perform.
- Open Audio Effects Online and choose one browser-decodable audio file no larger than 50 MiB. The page accepts filenames ending in MP3, WAV, M4A, AAC, Ogg, WebM, or FLAC, but the actual decoder is your browser's Web Audio implementation — a recognized filename does not guarantee that a particular encoded file will decode.
- Select one effect from the six available: echo, lightweight reverb, reverse, fade in, fade out, or normalization. Set the amount as a whole number from 0% through 100%. The amount scales or blends that specific operation only.
- Apply the effect locally. The page reports channel count, sample rate, input and output frame counts, output peak, and the exact WAV byte size. Click download to save a new interleaved little-endian PCM16 RIFF/WAVE file. Your original file is left untouched on disk.
The W3C Web Audio specification defines the decoded AudioBuffer length in sample frames, which is the unit the tool reports back to you. Microsoft documents the WAVE PCM block alignment as channels multiplied by bits per sample divided by eight; at sixteen bits that gives two bytes per mono frame and four bytes per stereo frame. Knowing this is what lets the tool quote an exact byte size for the output instead of rounding.
What the Download Preserves and What It Drops
The export is always a fresh PCM16 WAV. Floating-point samples are clipped to the legal -1 through +1 range before quantization and then converted to signed sixteen-bit PCM. The channel layout and sample rate of the decoded input are preserved.
What the export does not preserve: the source codec, bitrate, compression quality, tags, artwork, chapters, cue points, loop markers, or any other container metadata. The download is not a re-encode of the input — it is a freshly written WAV that contains only the processed samples. If you need to keep tags or artwork, store them elsewhere before exporting.
Limits on File Size, Channels, and Sample Rate
The tool enforces several interacting limits. The input file must be at most 50 MiB and must decode with 1 to 8 equal-length channels at 8,000 through 192,000 Hz. The decoded input and the complete processed output must each stay within 30 million channel samples. Echo and reverb can pass the input budget yet fail the output budget because their tails add frames — in that case nothing is truncated and no file is created, so a shorter input or a smaller amount is the fix.
If decoding fails, the tool creates no result. Changing the file, effect, or amount invalidates earlier work, and stale jobs cannot replace newer state. Decode contexts are closed and temporary object URLs are revoked on replacement and unmount, which keeps the browser tab from leaking memory across runs.
Chain Effects by Running the Tool Again
Only one effect runs per pass, which keeps each transformation testable and predictable. If a fade-in followed by normalization, or a reverb layer on top of a reverse pass, is the goal, the workflow is sequential. Step one: run the first effect and download the WAV. Step two: inspect the result — listen to it, or check the reported output peak and frame count. Step three: use that downloaded WAV as the input to a new run with the next effect and amount.
Repeated lossy decode and PCM processing can reveal source artifacts, and repeated clipping or normalization shifts the dynamics. Keep the untouched original on disk so a clean start is always one click away. A guide that walks through the same six effects from a different angle is available in Add Reverb and Echo to Audio Online: Six Local Effects.
How to Read the On-Screen Numbers
After applying an effect, the page reports several values that tell you whether the run worked as intended.
- Channel count: matches the decoded input, between 1 and 8.
- Sample rate: matches the decoded input, between 8 kHz and 192 kHz.
- Input frame count: the number of sample frames that were decoded before processing.
- Output frame count: the number of frames after the effect, which can be larger for echo and reverb because their tails are fully preserved.
- Output peak: the largest absolute sample value, clipped to the -1 through +1 range.
- WAV byte size: the exact size of the downloaded file, derived from the PCM block alignment and the output frame count.
Two numbers deserve special attention. A higher output frame count than input frame count means a tail was added (echo or reverb) — the file got longer, which is expected. An output peak near 0.95 after normalization is the target; values that sit exactly at 1.0 or -1.0 indicate the safety scale engaged because the original recording was already hot.
The math behind the reverb weights is direct. At amount 100%, the three wet taps are added with weights of 0.50, 0.35, and 0.25. The amount scales those weights linearly: at amount 50% the weights become 0.50 × 0.5 = 0.25, 0.35 × 0.5 = 0.175, and 0.25 × 0.5 = 0.125. Lower amounts mean a lighter blend with the dry signal; higher amounts push the reflections closer to the upper bound where clipping protection is most likely to engage.
That transparency — disclosed delays, disclosed weights, disclosed targets, disclosed limits — is the core value of the tool. There is no server render, no hidden AI transformation, and no account. The reliable contract is narrow: decode one supported file, apply the exact disclosed operation to every frame, preserve the complete calculated tail within the sample budget, encode a standards-based PCM16 WAV, and report exactly what changed.
If you're weighing options, Start Reversing Audio Locally in Your Browser covers this in detail.