Troubleshoot Audio Effects Online by confirming that the source is one browser-decodable file no larger than 50 MiB, then checking the effect amount, frame counts, and peak before export.

Audio Effects Online applies one selected effect per run entirely in the current browser tab. File reading, decoding, sample processing, WAV encoding, and download creation happen locally; audio is not uploaded to Lizely, stored in an account, rendered on a server, or altered by a hidden AI transformation.

how do i troubleshoot a problem when i use audio effects online
How Do I Troubleshoot Audio Effects Online Problems?

Check file decoding before changing the effect

Start with the source because an invalid or undecodable file cannot reach the effect stage. Decoding uses the browser’s Web Audio implementation, and the W3C Web Audio API defines decoded buffer length in sample frames. MP3, WAV, M4A, AAC, Ogg, WebM, and FLAC filenames are accepted, but actual codec and profile support varies by browser and operating system; a recognized extension does not prove that a particular file will decode.

Input check Supported requirement Troubleshooting response
File size One encoded file no larger than 50 MiB Choose a smaller source file and retain the untouched original separately.
Browser decoding The browser must decode the file’s actual codec and profile If decoding fails, no result is created. Try a compatible browser or operating system, or export a WAV copy from the source application.
Channel layout One to eight equal-length channels Use a source or re-export that meets the supported channel layout.
Sample rate 8,000–192,000 Hz Use a compatible export if the source falls outside the supported range.
Decoded size No more than 30 million channel samples Choose a shorter or compatible source. Multichannel audio can reach this budget sooner than its frame count alone suggests.

The 50 MiB check applies to the encoded file, while the 30-million-channel-sample check applies after decoding. They are separate gates. Test with a copy so that troubleshooting does not replace your source, and keep the original available for comparison and for starting another effect.

Run a clean Audio Effects Online test

  1. Preserve the source and select one file. Keep an untouched copy, then choose one browser-decodable encoded file no larger than 50 MiB.
  2. Select one operation. Choose echo, lightweight reverb, reverse, fade in, fade out, or normalization. The tool applies only one effect during each run.
  3. Set the amount. Use a whole-number setting from 0% through 100%. The amount blends or scales the selected operation.
  4. Apply the effect locally. Wait for the current job to finish. Changing the file, effect, or amount invalidates older work, and a stale job cannot replace newer state.
  5. Inspect the current measurements. Check channel count, sample rate, input and output frame counts, output peak, and the exact WAV byte size reported after processing.
  6. Interpret the frame counts correctly. Reverse, fades, and normalization retain the input frame count, while echo and reverb add tail frames. A complete output above 30 million channel samples fails without truncation or a downloadable file.
  7. Download and check the result. Save the complete PCM16 WAV, open it in your usual audio player, and compare it with the untouched original.

For a longer chain, download one result, inspect it, and use that WAV as the input to a later run. Repeated lossy decoding and PCM processing can reveal source artifacts, while repeated clipping or normalization can change dynamics, so isolate each stage before proceeding.

Match the result to the selected effect

Some apparent errors are actually the documented behavior of the selected operation. Use the exact behavior below to decide whether the file is faulty or the chosen setting simply does not create the result you expected.

Effect What it does Troubleshooting clue
Echo At 100%, each input sample is retained and a half-strength copy is added 200 milliseconds later. Lower amounts scale the delayed copy linearly. The added tail increases output frames. Strong dry and delayed overlap can also clip.
Lightweight reverb Adds reflections at 29.7, 37.1, and 41.1 milliseconds with weights of 0.50, 0.35, and 0.25 multiplied by the amount. It is an early-reflection design, not a room, convolution, or professional hardware simulation.
Reverse Reads sample frames in the opposite order and blends them with the original according to the amount. If the result still sounds partly forward, check the amount rather than the file format.
Fade in Scales from the selected dry proportion at the beginning to full level at the end. Confirm that the direction was correct; fade out performs the inverse.
Normalization Finds the largest absolute decoded sample and scales it toward a 0.95 peak. It does not target LUFS, perceived loudness, broadcast compliance, dynamics, or true peak between samples. Silent input remains silent.

Reducing the amount changes the echo or reverb contribution but does not shorten its delay or allocated tail. It therefore may reduce clipping without solving an output-budget failure. When the complete output fits the budget, the full tail is processed rather than cutting off the final echo.

Separate clipping and silence from other problems

If a result exists but sounds harsh or distorted, inspect the output peak and consider the selected amount. Strong echo and reverb settings can clip where dry and delayed samples overlap. Finite output samples are constrained to the legal -1 through +1 range before WAV quantization, so regenerate from the untouched original at a lower amount rather than treating the clipping as a download error.

A silent source remains silent after normalization because there is no nonzero sample peak to scale toward 0.95. Likewise, normalization is a sample-peak operation, not an equal-loudness process: two recordings with different dynamics or perceived loudness can still sound quite different after processing.

When effects are chained, remember that each run applies one transformation to the WAV supplied as its new input. Processing an already clipped or repeatedly normalized file cannot restore information that was lost in an earlier step, so troubleshoot the earliest stage where the result stopped meeting your expectations.

Verify the WAV before diagnosing export loss

The download is a new interleaved, little-endian PCM16 RIFF/WAVE file. It preserves the processed channel samples and sample rate. The Microsoft WAVEFORMATEX reference explains PCM block alignment: at 16 bits, a mono frame uses two bytes and a stereo frame uses four. The page reports the exact WAV byte size after processing, so use that displayed value rather than estimating from the source file size.

Source codec, bitrate, compression quality, tags, artwork, chapters, cue points, loop markers, and container metadata are not carried into the export. A missing tag or larger WAV is therefore not proof that the effect failed; the original encoding and metadata have been replaced by the processed PCM16 output.

Keep the downloaded result separate from the original source. This makes it possible to compare the dry recording with the processed file, reverse an unwanted choice, or begin again without accumulating another decode and transformation.

Separate input, processing, and export failures

  • Input layer: If no decoded output exists, check the actual codec and profile, the 50 MiB encoded-file limit, the supported channel count, sample rate, and decoded channel-sample budget.
  • Processing layer: If a result exists but sounds wrong, compare the effect and amount with the documented behavior, then inspect output frame count, peak, and the complete-output budget.
  • Export layer: If the downloaded file plays correctly but its codec or metadata differs from the source, verify that it is the expected new PCM16 WAV.

The cleanest isolation test is the untouched original, one effect, and one fixed amount. Record the browser and operating system, codec and profile, file size, channel count, sample rate, effect, amount, frame counts, and peak so the problem can be reproduced precisely. If decoding fails, the tool creates no result; if the complete effect output exceeds its budget, it creates no file rather than truncating it. Recognizing those hard limits prevents a limit-related failure from being mistaken for a broken download.

Related reading: Online Reverb: Real Room or Studio Processor Emulation?.

Related reading: Common Mistakes When Using Audio Effects Online.