Audio Bitrate Explained: Why 128kbps and 320kbps Sound Different

Quick answer: Bitrate is how much data an encoder spends representing each second of audio, measured in kilobits per second (kbps). The higher the number, the more of the original sound survives; the lower it drops, the more the encoder has to discard. The catch is that the relationship isn't a straight line: differences at the high end (256-320kbps) are hard for most listeners to hear at all, while the same-sized drop at the low end (say, 128kbps down to 64kbps) is immediately obvious.

"128kbps" and "320kbps" get thrown around as a shorthand for audio quality, and the numbers do matter - but the gap between them isn't a straight line, and where it actually becomes audible has more to do with which frequencies get compressed away than with the number itself.

What bitrate actually measures

Bitrate is how much data is used to encode each second of audio - measured in kilobits per second (kbps). A higher bitrate gives the encoder more data budget to represent the original waveform faithfully; a lower bitrate forces it to discard more, the same lossy trade-off a compressed image makes with visual detail. For lossy audio formats, the encoder specifically targets frequencies and details human hearing is least sensitive to, trying to spend the available bitrate where it's least likely to be missed.

Where the number actually comes from

For uncompressed audio (WAV), bitrate isn't a choice at all, it's just sample rate multiplied by bit depth multiplied by channel count. A standard 44.1kHz, 16-bit stereo file works out to a fixed, fairly high bitrate with no encoder decisions involved. Lossy formats break that link entirely: the encoder targets a bitrate you (or the tool) choose, then decides moment to moment what to keep and what to throw away to hit it. That's the fundamental difference between "bitrate as a mathematical fact about the file" and "bitrate as a quality setting."

CBR, VBR, and why the same average number can sound different

Encoders can hit a target bitrate two different ways. Constant bitrate (CBR) spends the exact same amount of data on every second of audio, regardless of what's actually happening in it, a few seconds of near-silence get the same budget as a dense, busy passage. Variable bitrate (VBR) instead adjusts on the fly, spending more data where the audio is complex and less where it's simple, aiming for a target quality or average rate rather than a fixed rate per second. Two files with an identical average bitrate can sound noticeably different if one was encoded CBR and the other VBR, because VBR is spending its budget more intelligently rather than evenly.

Why the gap is more audible at the low end than the high end

Most listeners have real difficulty reliably telling apart a well-encoded file in the 256-320kbps range from an uncompressed original, especially on typical consumer playback equipment - the encoder has enough budget at that range to preserve nearly everything that matters perceptually. Drop well below that, though, and the discarded detail starts showing up as audible artifacts: a dulled high end, a flattened stereo image, or a faint "underwater" quality especially in complex passages with a lot happening at once. The relationship isn't linear - the difference between 320 and 256 is far subtler than the difference between 128 and 64.

What actually gets discarded first

Encoders don't cut randomly, they follow a rough priority order based on what human hearing is least likely to miss. Very high frequencies go first, since sensitivity to the top of the audible range drops off naturally with age and typical speaker or headphone response. Stereo width narrows next: at low bitrates, the left and right channels get encoded as more similar to each other than they really are, since preserving a wide stereo image costs data the encoder would rather spend elsewhere. Only once those cheaper cuts are exhausted does an aggressive encoder start visibly smearing the mid-range detail that's actually easy to notice, which is why very low bitrates sound distinctly "squashed" rather than just quieter.

Audio Format Converter converts audio to WAV (fully uncompressed, no bitrate trade-off at all) or WebM using the Opus codec (a modern, efficient lossy option) - entirely in the browser.

Why WAV or WebM instead of MP3

This is a deliberate, browser-driven choice rather than an oversight: MP3 encoding has historically involved licensed patent technology, which is part of why browsers haven't universally built in free, native MP3 encoding the way they've built in playback. WAV sidesteps the question entirely by not compressing at all. WebM's Opus codec sidesteps it by being an open, royalty-free codec built for exactly this kind of use - and by most technical measures, it's more efficient than MP3 at an equivalent bitrate, despite MP3 having far more name recognition with the general public.

Matching bitrate to what you're actually encoding

The "right" bitrate depends heavily on the source material, not just a fixed number to always reach for. Speech uses a much narrower frequency range and a simpler waveform than music, so a voice recording commonly sounds clear at a bitrate that would make a full band mix sound noticeably thin. Dense, wide-frequency music (orchestral pieces, anything with heavy cymbals or layered synths) tends to need more headroom than a simple acoustic recording before artifacts show up, since there's more going on for the encoder to have to compromise on.

Archival vs distribution

A file you're keeping as a long-term master benefits from staying uncompressed or lossless, since a lossy encode is a one-way decision, the discarded detail is gone for good, and any future re-encode of that lossy file just compounds the loss further. A file meant for casual listening, a podcast, a voice memo, background music, doesn't need that ceiling, and a moderate-to-high lossy bitrate is a reasonable, space-conscious choice specifically because the losses at that range are hard to perceive in normal listening conditions.

Common mistakes worth avoiding

Assuming higher is always meaningfully better. Past roughly 256kbps, most people can't reliably tell a lossy file apart from the original in casual listening, so pushing to the highest available setting mostly buys file size, not audible quality, for typical playback situations.

Re-encoding an already-lossy file at a higher bitrate. Once detail is discarded, a higher bitrate on the next pass can't bring it back, it just re-encodes the already-degraded audio more generously. The quality ceiling is set by the lowest-bitrate pass in a file's history, not the most recent one.

Using a very low bitrate for dense or wide-dynamic-range music. The same low bitrate that sounds acceptable on a simple voice recording can sound obviously compressed on a busy mix, because there's more detail competing for the same limited data budget.

Real-world size and quality tradeoffs

Bitrate and file size scale together in a fairly mechanical way, since bitrate is data per second and duration is just time: doubling the bitrate roughly doubles the file size for the same length of audio, all else equal. A three-minute track encoded at 320kbps typically lands at roughly double the size of the same track at 160kbps, not because the encoder is doing anything different structurally, just because it's been given twice the data budget to spend per second. That's a more predictable relationship than image compression, where size depends heavily on the content itself (a flat-color image compresses far better than a noisy photo at the same settings); audio's size math is comparatively straightforward once the bitrate is fixed.

The short version

Bitrate controls how much data an encoder gets to represent each second of audio, and the resulting quality gap is most audible at the low end, where an encoder is forced into real compromises, not at the high end, where most listeners can't reliably tell the difference from an original anyway. Where a tool lands on WAV or WebM instead of MP3 is a licensing-driven choice about what a browser can freely encode, not a quality downgrade - Opus in particular holds up well against MP3 at the same bitrate. Audio Format Converter handles the actual conversion, so the only real decision left is picking a bitrate that matches what you're encoding rather than defaulting to the highest number available.