Compress Audio Online — Reduce File Size Without Losing Clarity

How Audio Compression Works

When people say "compress this audio file," they usually mean one of two different things. The first is codec compression — converting to a format like MP3 that mathematically discards inaudible frequencies. The second, which is what this tool does, is sample rate reduction: lowering how many times per second the audio waveform is measured. CD-quality audio is sampled 44,100 times a second (44.1 kHz). Drop that to 22.05 kHz or 8 kHz and you capture fewer data points, which means a smaller file — at the cost of the highest frequencies. For spoken word, those frequencies barely matter. For music, they're part of what makes a track sound full.

Podcasts, voice memos, lecture recordings, and customer support calls make up most of the audio people need to shrink — and it's mostly speech, not music. Human speech sits mostly between 300 Hz and 3,400 Hz, the same range telephone networks have used for a century. An 8 kHz sample rate comfortably covers that range and produces files a fraction of the size of the original. That's why this tool offers three presets instead of one: Low (8 kHz) for voice-only content where file size matters most, Medium (22 kHz) as a general-purpose middle ground, and High (44 kHz) for music or anything where you want a smaller file without an audible quality hit.

Most audio compression sites work by uploading your file to a server, processing it there, and sending back a download link. That adds a round trip, and it means a copy of your audio — which might be a private voice memo or a confidential recording — sits on someone else's infrastructure, even if only briefly. This tool uses the Web Audio API to decode, resample, and re-encode the file entirely within your browser tab — nothing leaves your device. For anyone compressing sensitive recordings, that distinction is the reason to use a browser-based tool at all.

Reducing sample rate is a lossy, one-way process — you can't get the discarded frequencies back by upsampling later. It's good practice to keep your original file until you've confirmed the compressed version is good enough for your purpose. The output downloads as WAV rather than MP3: keeping WAV at a lower sample rate avoids stacking a second lossy codec on top, which would degrade quality further. If you specifically need an MP3, run the compressed file through the MP3 Converter afterward.

Understanding Bitrate, Sample Rate, and Codec Compression

Audio file size is controlled by three separate factors, and it's easy to confuse them. Sample rate is how many times per second the audio waveform is measured — 44,100 Hz (44.1 kHz) for CD quality, 22,050 Hz (22 kHz) for medium quality, 8,000 Hz (8 kHz) for telephone quality. Bitrate is how much data per second is used to store the encoded audio — 128 kbps, 192 kbps, 320 kbps for MP3, for example. Codec compression is the algorithm used to encode audio into a specific format like MP3, AAC, OGG, or FLAC, each with different tradeoffs between file size and quality.

This tool adjusts sample rate only. It doesn't touch bitrate or codec — the output is WAV, which is uncompressed at whatever sample rate you choose. If you started with a 44.1 kHz WAV and compress to 22 kHz, you've halved the amount of raw data captured per second, so the file size roughly halves too. But WAV is still uncompressed, meaning it stores every sample without discarding any (within the new, lower sample rate limit). If you then convert that 22 kHz WAV to MP3 at 128 kbps, you're applying codec compression on top, which further reduces size by discarding psychoacoustically masked frequencies that humans are less likely to perceive.

The reason this tool outputs WAV rather than MP3 directly is to avoid double lossy compression. If your input is already an MP3, it's been through one lossy encoding pass. Decoding it, resampling to a lower rate, then re-encoding to MP3 again stacks two lossy steps, compounding artifacts. Keeping the intermediate output as WAV (which is lossless at whatever sample rate it's stored) means you can listen, confirm quality, and then apply codec compression if needed, rather than baking both steps into one irreversible operation.

For comparison: a 3-minute song at 44.1 kHz stereo WAV is about 30 MB. Resample to 22 kHz and it drops to 15 MB (still WAV). Convert that 22 kHz WAV to MP3 at 128 kbps and it shrinks to about 3 MB. Convert the original 44.1 kHz WAV directly to MP3 at 128 kbps and you get roughly the same 3 MB, but with slightly better quality since you didn't throw away high frequencies before encoding. The best workflow depends on your priority: if file size is critical and you're compressing speech (where high frequencies don't matter), resample aggressively first. If quality matters and you're compressing music, go straight to codec compression at a moderate bitrate and skip sample-rate reduction entirely.

Choosing the Right Preset for Your Content

Low (8 kHz) — Voice-Only, Maximum Compression

8 kHz sample rate is what old landline telephones used, and it's perfectly adequate for spoken word. Human speech contains most of its intelligibility between 300 Hz and 3,400 Hz; an 8 kHz sample rate captures up to 4 kHz (the Nyquist limit is half the sample rate), which covers that range with headroom. Voices sound slightly thin or tinny compared to full-quality recordings, but every word remains clear, and file sizes shrink dramatically — often by 80–85% compared to 44.1 kHz originals.

Use 8 kHz for voice memos, dictation, customer support call recordings, simple podcasts where the content matters far more than audio fidelity, and any scenario where you're hitting file size limits on email attachments or upload forms. Don't use it for music, sound effects, or multi-speaker recordings where you want to preserve tonal nuance — instruments and singing lose definition rapidly below 16 kHz, and 8 kHz makes them sound like they're coming through a telephone.

Medium (22 kHz) — Balanced, General-Purpose

22.05 kHz (often rounded to 22 kHz) is half of CD quality and the sweet spot for most non-professional audio. It captures frequencies up to 11 kHz, which includes the clarity and "air" that make voices sound natural and music sound reasonably full, while still cutting file size roughly in half compared to 44.1 kHz. Most podcast hosting platforms and streaming services consider 22 kHz acceptable for voice-focused content, and casual listeners on phone speakers or laptop built-ins won't notice a difference between 22 kHz and 44.1 kHz for speech.

Use 22 kHz for podcasts, interviews, audiobooks, lecture recordings, and any content where you want a smaller file without an obvious quality drop. It's also safe for background music or simple instrumental tracks where you're not aiming for audiophile fidelity. If you're unsure which preset to pick and the content is mostly voice with some incidental music or sound effects, start with Medium — it's forgiving enough for mixed content and compresses well enough to solve most file-size problems.

High (44 kHz) — Near-Original, Music-Safe

44.1 kHz is CD-quality sample rate and the standard for professional audio production. Keeping your file at 44 kHz means you're not discarding any frequency content (up to the 22.05 kHz Nyquist limit, which is above the range of human hearing for most people). File size won't shrink much — maybe 10–20% if you're also converting from a higher bit-depth or stereo to mono, but sample rate alone doesn't reduce size significantly at 44 kHz. The point of the High preset is to re-export at the same quality while ensuring the output is a clean WAV ready for further processing or archival.

Use 44 kHz when you're compressing music, when you plan to edit or mix the audio further and don't want to lose headroom for EQ or effects, or when file size isn't the bottleneck and you just want to normalize or clean up the file without degrading quality. If your goal is meaningful compression, you're better off converting to a codec like MP3 or AAC rather than staying at 44 kHz WAV, which doesn't compress much on its own.

Combining Sample Rate Reduction with Codec Conversion

For maximum size savings, apply both techniques in sequence: resample first (to discard high frequencies you don't need), then encode to a lossy codec like MP3 (to discard psychoacoustically redundant data). A 44.1 kHz stereo recording compressed to 22 kHz WAV, then encoded to MP3 at 128 kbps, can shrink to 5–10% of the original size while remaining perfectly intelligible for voice content. Just be mindful that each lossy step is permanent — if you over-compress and later realize you need better quality, you'll have to start over from the original source file.