A peaking audio equalizer with three cascaded bands and a fixed Q of 1 changes audio by adding or subtracting gain in a broad region around each center frequency, not by adjusting every frequency below or above a hard boundary. Each of the three disclosed controls operates on a region rather than a hard cutoff, applies integer gain from -12 dB to +12 dB, and runs the same Q=1 peaking biquad regardless of the source material. Cascading all three in the same direction can push samples above digital full scale, which is why a uniform safety scale appears whenever the raw processed peak exceeds 0.99. Reading the numbers the page reports, instead of judging purely by louder playback, is the first move that separates a workable EQ from a noisy one. The rest of this article walks through the mistakes that show up most often when using an audio equalizer, and shows how the Audio Equalizer handles each one in the open rather than behind a black box.
Mistake: Boosting When You Actually Wanted Volume
The most common mistake is treating a peaking filter as a volume knob. Sliding Bass, Mid, and Treble upward feels like it should make the track louder, but a peaking filter changes the balance between the chosen region and the rest of the spectrum. If every band goes positive, the average level tends to stay near where it started while the low, mid, and high regions fight for the same headroom. The result often sounds harsher and thinner, not louder. When the goal is a louder file, the right move is a gain stage applied to the whole signal, not three positive peaking filters stacked together. The Audio Equalizer separates these two ideas on purpose: it exposes Bass, Mid, and Treble as -12 to +12 dB adjustments, never as a master volume slider, so the mistake becomes obvious the moment a listener asks why the file is not louder.
Mistake: Treating the Three Controls Like a Graphic 10-Band EQ
Hardware graphic equalizers and many software plugins offer a row of fixed-frequency sliders, each tuned to a specific octave. The three-band equalizer here is not that tool. Bass is centered at 100 Hz, Mid at 1,000 Hz, and Treble at 10,000 Hz when the decoded sample rate allows it. For sample rates below about 22,050 Hz, the Treble center is reduced to 45 percent of that rate so it stays below the Nyquist limit instead of asking an invalid filter to operate outside the available spectrum. That detail matters when loading older or low-rate audio. If you expect Treble to land at 10 kHz on a 16 kHz file, it will not, and the mistake is treating the readout as broken rather than reading what the page actually reports about the decoded sample rate.
| Band | Center frequency | Q value | Adjustment range |
|---|---|---|---|
| Bass | 100 Hz | 1 | -12 to +12 dB |
| Mid | 1,000 Hz | 1 | -12 to +12 dB |
| Treble (normal sample rate) | 10,000 Hz | 1 | -12 to +12 dB |
| Treble (low sample rate) | 45% of decoded sample rate | 1 | -12 to +12 dB |
Mistake: Forgetting That Q=1 Is a Broad, Disclosed Choice
Q=1 means each peaking filter shapes a relatively wide region around its center rather than a narrow spike. This is a product decision stated up front, not an inference from the source file, and it cannot be changed on the page. The mistake is expecting surgical cuts the way a parametric EQ with adjustable Q would deliver. A Q=1 Bass control at +6 dB does not isolate one note on a bass guitar; it raises a broad low-end shelf that also lifts rumble, room noise, and microphone proximity effect. Stacking three broad positive bands is the fastest way to lose clarity. Cuts are usually safer than boosts because they reduce competing energy rather than adding to it, which keeps the spectrum's overall shape closer to the source.
Mistake: Stacking Boosts Until the Safety Scale Fires
The three sections are cascaded in Bass, Mid, then Treble order, and that order matters because each stage's output feeds the next. Cascaded boosts can create samples above digital full scale even when the original input never clipped. The Audio Equalizer measures the absolute peak across every processed sample and every channel before encoding the WAV. If that raw peak is above 0.99, one uniform scale factor reduces the entire processed result so the largest magnitude becomes 0.99. The page reports raw peak, output peak, and the exact safety scale. The mistake is to ignore those numbers and assume the EQ "just worked" because a file downloaded. The safety scale protects the PCM encoder from independently chopping individual peaks, but it can also reduce overall level. When the safety scale drops below 1.0, the EQ is telling you that your combined boosts exceeded digital headroom and the result was scaled back. Pulling a band down by a few dB and rerunning usually produces a louder, cleaner file than letting the safety scale silently absorb the overshoot.
| Reading on the page | What it means |
|---|---|
| Raw peak below 0.99 | Cascaded boost did not exceed full scale; output peak equals raw peak and safety scale is 1.0. |
| Raw peak above 0.99 | Cascaded boost exceeded full scale; the entire processed signal was scaled uniformly so the output peak is 0.99. |
| Safety scale below 1.0 | Level was reduced to prevent clipping. The lower the scale, the more headroom was used by the boosts. |
Mistake: Judging EQ by Louder Playback
The human ear tends to prefer whatever sounds louder in the moment, so a boosted version can feel like an improvement during A/B testing even when it has masked detail, raised sibilance, or emphasized rumble. The mistake is to set Bass, Mid, and Treble until the new file feels better at default level, without matching perceived loudness to the original. The Audio Equalizer is not loudness normalization. It does not target LUFS, RMS, replay gain, or any streaming-platform specification, so any volume change is purely a side effect of how the filters reshaped the spectrum. A fairer check is to listen to both versions at a matched perceived level on speakers or headphones that matter to the use case. Comparing only the louder file to the quieter original will bias the result toward whatever was boosted, which makes the EQ look better than it is.
Use the Audio Equalizer Without These Mistakes
- Open the Audio Equalizer page and pick one browser-decodable audio file within the 50 MiB and five-minute decoded limits.
- Leave all three controls at 0 dB and apply the equalizer to confirm the flat path is genuinely flat before changing anything.
- Make one small change at a time. Start with a single band, set an integer value between -12 and +12 dB, and re-apply.
- Read the raw peak, output peak, and safety scale. If the safety scale drops below 1.0, pull a band down or switch from a boost to a cut on another band.
- Prefer cuts to stacked boosts when the source already feels balanced. Cuts reveal shape; stacked boosts add energy the source did not have.
- Preview the result and compare against the original at a matched perceived level, not at default playback volume.
- Download the PCM16 WAV only after the numbers and the preview both look right.
Mistake: Forgetting the Output Is a New PCM16 WAV
Another recurring mistake is to expect the downloaded file to keep the original codec, bitrate, tags, artwork, chapters, loop markers, and container metadata. It will not. The Audio Equalizer writes a new uncompressed PCM16 WAV, interleaved by frame, with a header that records the actual channel count, decoded sample rate, byte rate, block alignment, bit depth, and complete data length. Original MP3 or AAC compression, bitrate, ID3 tags, cover art, and container-specific metadata are not preserved. Web Audio's AudioBuffer may resample while decoding, so the result reports and writes the decoded sample rate rather than claiming the source container rate survived unchanged. Each channel keeps the same number of sample frames, and all accepted channels are filtered independently with identical settings. Processing happens in the current browser tab, and Lizely does not receive the audio bytes, decoded samples, EQ settings, filename, or result. If the job needs the original metadata to survive, this page is the wrong tool and a full audio editor should be used instead.
When the Three-Band Equalizer Is the Wrong Tool
The compact three-band design is honest about what it is not. It is not a graphic 10-band system, not a linear-phase mastering processor, and not an automatic room correction system. The page states plainly that one setting does not improve every recording. Use a full audio editor when the job needs precise center frequencies, adjustable Q, spectrum analysis, automation, linear-phase processing, loudness metering, dithering controls, codec selection, or metadata preservation. If you find yourself fighting the limits of three fixed Q=1 bands, the answer is a different tool, not a more aggressive boost. A peaked channel with two stacked positive bands can also reveal bugs in a recording that the EQ cannot fix: rumble that needs a high-pass filter, sibilance that needs de-essing, and distortion that needs a different source. Reading the page's reports and switching tools when they make sense is part of using an audio equalizer correctly. For a checklist that fits on one screen, the three-band checklist guide walks the same ideas in shorter form.
Limits Worth Checking Before You Apply
The page rejects an over-limit file before filtering and never silently truncates the end. The hard limits are 50 MiB for the source file, five minutes of decoded audio, one through eight channels, a sample rate from 8,000 through 192,000 Hz, and no more than 30 million channel samples. A decode failure produces an error and no stale download is presented. The mistake here is to assume the tool will trim a too-long file or stretch a low-rate file to a higher rate. It does neither. A familiar extension is also not a guarantee: containers can hold codecs that a browser cannot read, so an MP3 with an unusual codec inside the MP3 container can fail to decode even though the extension matches. The page reports the decoded sample rate and channel count, which are the numbers the WAV header will actually carry, so reading them is part of avoiding the mistakes that come from assuming the source survived unchanged.