The Audio Equalizer output is always a brand-new uncompressed PCM16 WAV file, never the original MP3, AAC, OGG, FLAC, M4A, or WebM container you uploaded. Codec choice, bitrate, ID3 tags, embedded artwork, chapter markers, loop points, and any other container-specific fields are intentionally not written back into the file you download. The tool decodes the selected audio through the browser's Web Audio API, applies the three disclosed peaking biquads to the decoded samples, and then encodes a fresh interleaved little-endian 16-bit WAV with a header built from the actual decoded channel count, sample rate, and total sample count. Because the decoded sample rate is what gets written into the WAV header and reported on screen, a source that Web Audio resamples on decode will produce a WAV at the new rate rather than the rate advertised in the original container. The processed audio stays in the current browser tab and Lizely does not receive the bytes, the EQ settings, the filename, or the result.
What the Audio Equalizer Downloads
Every download produced by the Audio Equalizer is the same kind of file: an uncompressed WAV written from the decoded channel samples after the three-band peaking filter runs. The page accepts MP3, AAC, M4A, OGG, WebM, WAV, or FLAC as input, but the only output is a new WAV.
The WAV header records the channel count, sample rate, byte rate, block alignment, 16-bit depth, and the full data length of the equalized result. Each accepted channel is filtered independently with the same bass, mid, and treble settings, then the samples are converted from float to signed 16-bit little-endian PCM and interleaved by frame. The header layout follows the standard RIFF/WAVE structure documented by Microsoft in the WAVEFORMATEX reference, with byte fields sized to match the actual decoded stream.
The header values come from the decoded stream, not the source container. The on-screen report shows the actual sample rate and frame count that were written, the raw peak across every processed sample, the output peak after any safety scale, and the scale factor itself. Those numbers are the contract: if the report says 48,000 Hz, six channels, and 0.99 output peak, the WAV header will match.
Because the output is a new WAV, there is no copy of the source container next to it. The browser saves a single file with the extension you choose, and the original MP3, M4A, or FLAC is left untouched wherever it was loaded from.
What Stays With the File and What Does Not
The downloaded WAV carries the audio samples, the channel layout, and the decoded sample rate. It does not carry any of the metadata or compression characteristics of the original container. A side-by-side summary is more useful than a long list of caveats:
| Property | In the Downloaded WAV |
|---|---|
| Original codec (MP3, AAC, FLAC, OGG, etc.) | No — always PCM16 |
| Original bitrate and compression mode | No — uncompressed |
| ID3, Vorbis, or MP4 tags | No — not copied |
| Embedded artwork | No — not copied |
| Chapter markers and loop points | No — not copied |
| Container extension and sidecar metadata | No — output is .wav |
| Decoded sample rate (Hz) | Yes — written to the header |
| Channel count | Yes — up to eight channels |
| Sample frame count per channel | Yes — full length preserved |
| Channel ordering | Yes — same order as decoded |
| Audio sample values | Modified — run through three biquads |
The decoded buffer the page works from is defined by the Web Audio AudioBuffer specification: a typed-array-per-channel representation of the audio at one fixed sample rate and channel count. That buffer is the only source of data the WAV writer sees, so anything that did not survive decoding also does not survive encoding.
Limits on the Source File
Several hard limits exist to keep the filter and the encoder from running on unexpectedly large data in the browser tab. They are checked before any EQ is applied, so an over-limit file is rejected with an error rather than silently truncated.
- Input file size: up to 50 MiB.
- Decoded duration: five minutes or less.
- Decoded channels: one through eight.
- Decoded sample rate: 8,000 Hz through 192,000 Hz.
- Total channel samples: 30 million or fewer.
A familiar extension does not guarantee a successful decode. MP3, M4A, and other containers can hold codecs the current browser cannot read. A decode failure shows an error and no download button. Files that pass every limit get the full three-band EQ plus the reported numbers; files that fail any limit are rejected cleanly.
Equalize a File and Download the WAV
Once the codec question is settled, the workflow is short and repeatable:
- Open the Audio Equalizer page and pick a single browser-decodable audio file that fits the 50 MiB and five-minute decoded limits.
- Set Bass, Mid, and Treble to whole numbers between −12 and +12 dB. Zero on all three is a genuine flat path that copies decoded samples without running a nonzero filter section.
- Apply the equalizer locally. Bass is centered at 100 Hz, Mid at 1,000 Hz, and Treble at 10,000 Hz when the decoded sample rate allows it; otherwise Treble drops to 45 percent of the sample rate so it stays below Nyquist.
- Read the decoded sample rate, frame count, raw peak, output peak, and uniform safety scale on the report line. If the raw peak was above 0.99, the safety scale brings the largest magnitude to 0.99 across every channel.
- Use the preview to compare against the original at a similar perceived level, then download the new interleaved little-endian PCM16 WAV. The saved file is a single .wav — not the original container — with the values shown in the report.
For a deeper look at exporting the equalized audio specifically as a WAV, the three-band WAV walkthrough covers the same workflow in more detail.
When You Need More Than a Three-Band EQ
The three-band peaking design is a compact tonal tool, not a mastering processor. It is the right choice for a quick bass lift, a treble tame, or a mid cut, and it stays out of the way for everything else. It is not the right tool when the project depends on the source container, the tags, the codec, or anything else the WAV format does not encode.
Reach for a full audio editor — Audacity, Reaper, Logic, Ableton, or similar — when the workflow needs dithering controls, codec choice on export, ID3 or Vorbis tag editing, artwork embedding, chapter or cue markers, linear-phase processing, loudness metering to LUFS or replay gain, automation lanes, adjustable Q, or precise center-frequency control. None of those features live in the three peaking biquads or the PCM16 encoder that produce the WAV download.
If the goal is only to tweak the sound and end up with a clean uncompressed file, the Audio Equalizer produces exactly that. If the goal is to keep MP3 tags, AAC artwork, or any other container-level metadata, plan to re-tag the WAV manually or run the export through a desktop tool that supports the original format. Treating the WAV as a clean intermediate and re-tagging in a full DAW is the usual pattern for projects where the metadata matters as much as the audio.