Audio Effects Online lets you apply one bounded audio effect in your browser, set its amount from 0% to 100%, and download the complete processed recording as a PCM16 WAV without uploading the source file. To get started, choose a single browser-decodable file no larger than 50 MiB, select echo, lightweight reverb, reverse, fade in, fade out, or normalization, and enter a whole-number amount. Accepted filename types include MP3, WAV, M4A, AAC, Ogg, WebM, and FLAC, but actual codec and profile support depends on the browser and operating system. Before downloading, the page reports channel count, sample rate, input and output frame counts, output peak, and exact WAV byte size. Use Audio Effects Online for a focused local edit rather than a multitrack session: it performs one deterministic transformation per run and keeps your untouched original separate from the export. No account, server render, or hidden AI transformation is involved. Keep the source file, begin with a modest amount when an effect changes the waveform, and inspect the processed file before using it elsewhere in your project.
Check that the file can finish processing
Audio Effects Online relies on the browser's Web Audio implementation to decode the source. The W3C Web Audio API 1.1 defines decoded AudioBuffer length in sample frames, but an accepted filename extension does not guarantee that a particular encoded file will decode successfully. The browser and operating system determine the actual codec and profile support.
The tool validates both the encoded source and the decoded audio:
| Check | Supported value |
|---|---|
| Source file | One encoded audio file no larger than 50 MiB |
| Decoded channel layout | 1–8 equal-length channels |
| Decoded sample rate | 8,000–192,000 Hz |
| Decoded input budget | No more than 30 million channel samples |
| Complete output budget | No more than 30 million channel samples, including any calculated effect tail |
| Processing selection | One effect and one whole-number amount from 0% through 100% per run |
A decoding failure creates no result. Echo and reverb can also pass the input budget while pushing the complete output beyond 30 million channel samples because their delayed tails add frames. In that case, the tool does not truncate the tail or create a file.
Apply and download your first processed WAV
- Choose one source file. Open the Audio Effects Online page and select one browser-decodable audio file within the 50 MiB source limit. Keep the original unchanged so it remains available for comparison or a later restart.
- Select one effect and its amount. Choose echo, lightweight reverb, reverse, fade in, fade out, or normalization. Enter a whole number from 0% through 100%. The amount scales or blends the selected operation according to that effect's defined behavior.
- Apply the effect locally. Start the local transformation after the browser decodes the file and the input meets the channel, sample-rate, and channel-sample requirements. For echo and reverb, the complete delayed tail is allocated before processing so the last repeat is not cut off, provided the result still fits the output budget.
- Inspect the reported result. Check the channel count and sample rate, then compare the input and output frame counts. Review the output peak for clipping risk and confirm that the exact WAV byte size corresponds to the complete processed file.
- Download the PCM16 WAV. Save the complete interleaved, little-endian PCM16 RIFF/WAVE result under a new filename. Do not replace the source file; the WAV is a processed derivative containing newly encoded audio samples.
Changing the selected file, effect, or amount invalidates older work. Make one deliberate setting choice at a time, then inspect the new result before downloading it.
Choose one effect based on the change you need
The amount control blends or scales the chosen operation, but each effect has its own exact formula and purpose. Use the following definitions to select a starting point without expecting every amount control to represent the same wet-to-dry balance.
| Effect | Exact operation | Result and boundary |
|---|---|---|
| Echo | Every input sample is retained. At 100%, a half-strength copy is added 200 milliseconds later; lower amounts scale the delayed copy linearly. | Useful for a clear repeated sound. The complete delayed tail is retained when it remains within the output budget. |
| Lightweight reverb | Copies are added at 29.7, 37.1, and 41.1 milliseconds with weights of 0.50, 0.35, and 0.25, each multiplied by the selected amount. | Provides a small early-reflection design for a quick draft. It does not simulate a room, impulse response, or professional hardware reverb. |
| Reverse | Sample frames are read in the opposite order and blended with the original according to the amount. | Creates a controlled blend between the forward and reversed recordings. |
| Fade in | The beginning starts at the selected dry proportion and rises to full level by the end. | Shapes the opening of the recording rather than adding room reflections or repetition. |
| Fade out | The full-range fade performs the inverse of fade in. | Shapes the closing portion of the recording. |
| Normalization | The largest absolute decoded sample is used to scale the file toward a 0.95 sample peak. Silent input remains silent. | Targets sample peak, not LUFS, perceived loudness, dynamics, broadcast compliance, or equal loudness across recordings. |
The three reverb coefficients are product-defined settings, not measurements of a concert hall, plate, spring, studio, acoustic material, or branded processor. Strong echo and reverb amounts can clip where dry and delayed samples overlap. Finite output samples are therefore constrained to the legal −1 through +1 range before WAV quantization.
Read the output numbers before judging the file
Channel count and sample rate describe the decoded audio being processed. Input and output frame counts show whether the rendered length changed, which is especially relevant for echo and reverb because their complete tails add frames. The output peak provides a quick sample-level check, while the exact WAV byte size identifies the size of the new file rather than the original compressed source.
Do not treat output peak as a loudness or inter-sample true-peak measurement. For normalization, 0.95 is a sample-peak target; the tool does not calculate LUFS or equal perceived loudness. Listen to the downloaded copy as well as reviewing the numbers, since clipping, quantization, or source artifacts may only become obvious during playback. If the figures or sound do not match the intended edit, Fix a Result That Looks Wrong After Audio Effects Online provides a focused follow-up for checking the settings and result.
Know what the PCM16 WAV preserves
The download is a new interleaved, little-endian PCM16 RIFF/WAVE file based on the PCM layout described by Microsoft's WAVEFORMATEX structure. It carries the processed channel samples and sample rate into the new file. It does not preserve the source codec, bitrate, compression quality, tags, artwork, chapters, cue points, loop markers, or other container metadata.
Floating-point samples are quantized to signed 16-bit PCM during encoding, and finite samples are constrained before that conversion. Treat the WAV as a new editing master rather than as a lossless copy of the source file's encoding. The untouched original remains necessary for comparisons, alternate effects, and a clean restart.
Add another effect only after checking this one
The tool applies only one effect per run. To build a chain, download the first PCM16 WAV, inspect and listen to it, and then use that WAV as the input for the next run. This makes each transformation separate and keeps the source of any later change visible.
- Keep the original source outside the processing chain.
- Give each exported WAV a filename that identifies the effect and amount used.
- Confirm the frame counts, output peak, and playback before starting another run.
- Apply the next single effect to the downloaded WAV rather than returning to an older on-screen result.
Starting with a lossy encoded source can reveal existing artifacts during decoding and subsequent processing. Repeated clipping or normalization can also change dynamics, so avoid stacking operations automatically. A controlled first run followed by an inspection after every export is easier to understand and revise.