How to reduce mp3 bitrate — free, browser-based, no upload
- Step 1Open the Bitrate Changer — Go to the bitrate-changer tool. FFmpeg 8.1 loads into the page on first use; nothing installs on your computer.
- Step 2Drop your MP3 — Drag the MP3 onto the dropzone (MP3, WAV, FLAC, M4A, OGG, Opus, or video are all accepted as input — but the output is always MP3). Free tier accepts up to 50 MB / 30 minutes per file.
- Step 3Read the current bitrate first — Before reducing, note the source bitrate. Reducing a 320 kbps file to 128 kbps saves real space; 're-reducing' a file that is already 128 kbps to a higher number just enlarges it without adding quality.
- Step 4Pick a target bitrate — Use the Bitrate dropdown: 320 kbps (best), 256, 192 (default), 128 (podcast), or 64 (voice). For shrinking music, 128 kbps is the common floor; for spoken word, 64 kbps mono-source material stays clean.
- Step 5Run the re-encode — Click to process. FFmpeg decodes the MP3 and re-encodes it at the new constant bitrate locally — speed depends on duration and your CPU, not your connection.
- Step 6Download and verify the size — Download the new MP3. Confirm the file size dropped as expected (CBR size is predictable from the cheat sheet below) and spot-check that the ID3 tags and cover art survived.
MP3 size by CBR bitrate (per minute and per hour)
Constant-bitrate MP3 size is bitrate times duration plus a tiny header overhead. Figures are approximate (1 kbps = 1000 bits/s; 1 minute of CBR audio ≈ bitrate/8 × 60 bytes).
| Target (dropdown label) | Approx. per minute | Approx. per hour | Typical use |
|---|---|---|---|
| 320 kbps · best | ~2.4 MB | ~144 MB | Archival music, no further shrink wanted |
| 256 kbps | ~1.9 MB | ~115 MB | High-quality music, slightly smaller |
| 192 kbps · default | ~1.4 MB | ~86 MB | Good music quality / size balance |
| 128 kbps · podcast | ~0.96 MB | ~58 MB | Streaming, music at the common floor |
| 64 kbps · voice | ~0.48 MB | ~29 MB | Spoken word, lectures, voice memos |
What this tool does vs. what it does not
Grounded in the bitrate-changer UI and the runBitrateChange / runConversion path. Verify before assuming a control exists.
| Capability | In this tool? | Detail / where to go instead |
|---|---|---|
| Constant bitrate (CBR) | Yes | Encoded with libmp3lame -b:a <target> |
| Variable bitrate (VBR) | No | The dropdown sets a single CBR value; there is no VBR/quality slider |
| Output format choice | No | Output is always MP3; there is no format selector for this tool |
| Preserve ID3 tags + cover art | Yes | -map_metadata 0, cover re-attached, -id3v2_version 3 |
| Change sample rate | No | Use sample-rate-converter |
| Hit an exact file-size target | No | Use audio-compressor (MB target) |
Cookbook
Concrete before/after re-encodes. Sizes are computed from the source bitrate, target bitrate, and duration — your file may vary by a few percent due to headers and ID3 frames.
320 kbps album track down to 128 kbps for phone storage
A favourite 4-minute track ripped at 320 kbps is ~9.6 MB. Dropping it to 128 kbps for a phone with limited space cuts it to ~3.8 MB while staying acceptable for earbuds on the go.
Input: track.mp3 320 kbps CBR 4:00 ~9.6 MB Target: 128 kbps (dropdown: '128 kbps · podcast') FFmpeg: libmp3lame -b:a 128k Output: track.mp3 128 kbps CBR 4:00 ~3.8 MB (~40% of original)
Spoken-word recording to 64 kbps
A 45-minute interview exported at 192 kbps is ~77 MB. For a voice-only file, 64 kbps is plenty and brings it down to ~22 MB.
Input: interview.mp3 192 kbps 45:00 ~77 MB Target: 64 kbps (dropdown: '64 kbps · voice') Output: interview.mp3 64 kbps 45:00 ~22 MB
What NOT to do — bumping a low-bitrate file 'up'
Setting a higher target than the source does not restore quality. The file grows but the discarded detail is gone for good.
Input: voicmemo.mp3 64 kbps 10:00 ~4.8 MB Target: 320 kbps (mistaken 'upgrade') Output: voicemo.mp3 320 kbps 10:00 ~24 MB Result: 5x larger, same (or slightly worse) audible quality
Batch the same target across an album (Pro)
Free tier is one file at a time. On Pro (up to 10 files), drop a whole album and re-encode each track to the same bitrate in one session.
Free tier: 1 file / run, 50 MB, 30 min cap Pro tier: 10 files / run, 200 MB, 120 min cap Workflow: drop 10 tracks -> set 192 kbps -> process -> download each
Tags and art survive the shrink
Reducing bitrate keeps the ID3 metadata and embedded artwork, so the smaller file still shows correctly in a music library.
Input: song.mp3 320 kbps [title/artist/album + 600x600 cover] Target: 192 kbps Output: song.mp3 192 kbps [same title/artist/album + same cover] (FFmpeg: -map_metadata 0, cover re-attached, -id3v2_version 3)
Edge cases and what actually happens
Target bitrate higher than the source
By designIf the source is 128 kbps and you pick 320 kbps, FFmpeg re-encodes at 320 kbps. The file gets larger but no quality returns — the missing data was discarded by the first encode. Only reduce when size is the goal.
Input is not an MP3 (WAV/FLAC/M4A/OGG/Opus/video)
SupportedAny accepted audio (or a video's audio track) is decoded and re-encoded to MP3 at your chosen bitrate. The output extension is always .mp3 — there is no option to keep the original container here.
File exceeds 50 MB on the free tier
Tier limitFree tier caps input at 50 MB and 30 minutes per file. A long, high-bitrate MP3 can exceed this. Trim it first with audio-trimmer, or upgrade for the 200 MB / 120 min Pro limits.
Recording longer than 30 minutes (free)
Tier limitThe free duration cap is 30 minutes per file, separate from the size cap. A 45-minute show under 50 MB still hits the duration limit on free; Pro raises it to 120 minutes.
Variable-bitrate output expected
Not supportedThis tool only writes CBR via -b:a. There is no VBR or quality-target mode. If a publisher requires VBR specifically, this tool is not the right fit.
Re-reducing an already-low file
By designRe-encoding 128 kbps to 96/64 kbps each adds a fresh lossy generation, so quality degrades a little more each pass. Always re-encode from the highest-quality source you have, not from a previously shrunk copy.
ID3 tags not appearing after shrink
PreservedTags and cover art are carried with -map_metadata 0 and -id3v2_version 3. If a player shows nothing, it may not read ID3v2.3, or the source simply had no tags to copy.
Output sounds 'thin' at 64 kbps for music
Expected64 kbps is tuned for voice. Music at 64 kbps shows audible artefacts in cymbals and reverb. Use 128 kbps or higher for music; reserve 64 kbps for spoken-word sources.
Browser tab closed mid-encode
ExpectedAll processing is in-tab. Closing the tab cancels the job and nothing is saved server-side (because nothing was uploaded). Re-open and re-run.
Frequently asked questions
Does the file get uploaded to a server?
No. FFmpeg 8.1 runs as WebAssembly inside your browser tab, so the MP3 is decoded and re-encoded locally. Nothing is uploaded and nothing is stored on a server.
Can I increase the bitrate to improve quality?
No. Re-encoding at a higher bitrate makes the file larger but cannot recover detail the original lossy encode discarded. Bitrate reduction is a one-way trade — only reduce when size is the constraint.
What bitrates can I choose?
Five constant-bitrate targets: 320, 256, 192 (default), 128, and 64 kbps. They are selected from a single dropdown.
Is the output CBR or VBR?
Constant bitrate (CBR). The tool encodes with libmp3lame -b:a <target>. There is no variable-bitrate or quality-slider mode in this tool.
What format is the output?
Always MP3, no matter what you drop in. There is no output-format selector for the bitrate changer. To convert containers, use a dedicated converter like wav-to-mp3.
How much smaller will my file get?
At CBR, size is bitrate times duration. Going 320 to 128 kbps yields roughly 40% of the original size; 320 to 64 kbps yields roughly 20%. See the cheat sheet above.
Will my tags and album art survive?
Yes. ID3 metadata and embedded cover art are carried across the re-encode (-map_metadata 0, cover re-attached, -id3v2_version 3).
What is the lowest bitrate I should use for music?
128 kbps is the common floor for music. Below that, artefacts in cymbals and reverb become noticeable. 64 kbps is best reserved for voice.
What is the largest file I can process?
Free tier: 50 MB and 30 minutes per file, one file at a time. Pro: 200 MB / 120 minutes / 10 files. Pro-media and Developer: up to 100 GB with no duration cap.
Can I batch multiple MP3s?
The free tier processes one file per run. Pro allows up to 10 files per run; higher tiers allow more.
Does reducing bitrate change the duration or pitch?
No. Bitrate only affects how many bits per second represent the audio. Duration, pitch, and tempo are unchanged. For those, use the pitch or tempo tools.
How do I hit an exact target like 'under 8 MB'?
The bitrate changer sets a bitrate, not a file size. For an exact MB target use audio-compressor, which sizes the encode to a megabyte goal.
Privacy first
Every JAD Audio tool runs entirely in your browser via FFmpeg (WebAssembly) and RNNoise. Your audio files never leave your device — verified by zero outbound network requests during processing.