How to set the audio bitrate independently from video
- Step 1Drop your video — Add the file (
.mp4,.mov,.mkv,.webm,.avi,.m4v,.ts). It loads into FFmpeg.wasm in your browser — no upload. - Step 2Set Audio kbps for your content — Type 128 for speech, 192 for general, 256–320 for music. The field allows 48–320 →
-c:a aac -b:a <value>k. - Step 3Set Video kbps too (it re-encodes video as well) — Because video is always re-encoded, set Video kbps near the source's existing video bitrate if you only meant to change audio — that minimises the picture change. Default is 2500.
- Step 4Encode — The tool runs
libx264 -preset medium -b:v Xk -maxrate (1.2X)k -bufsize (2X)k -c:a aac -b:a Yk -movflags +faststart. Single pass. - Step 5Download the faststart MP4 — The output has AAC audio at your chosen rate and a re-encoded H.264 video track.
- Step 6If you only wanted the audio file, use a different tool — To pull the audio track out of the video, use the audio-track extractor instead — this tool always returns a full video.
Choosing an AAC audio bitrate
Practical AAC bitrate guidance by content type. The Audio kbps field accepts 48–320.
| Content | Audio kbps | Why |
|---|---|---|
| Voice / podcast / narration | 96–128 | Speech is fully transparent here; saves space |
| General mixed video | 128–192 | Safe all-round default |
| Music-forward content | 256 | Near-transparent for most listeners |
| Critical music / concert | 320 | Top of AAC's practical range (field max) |
| Tiny/voice memo clips | 48–64 | Lowest the field allows; expect audible loss |
Audio bytes saved per minute
Approximate audio-track size per minute at each bitrate. Lowering audio is a cheap way to trim a long clip.
| Audio kbps | ~Per minute | vs 320 |
|---|---|---|
| 320 | 2.4 MB | baseline |
| 256 | 1.9 MB | -21% |
| 192 | 1.4 MB | -40% |
| 128 | 0.96 MB | -60% |
| 96 | 0.72 MB | -70% |
What this tool does to audio vs video
Critical to understand: setting audio here also re-encodes the video. There is no audio-only or passthrough mode.
| Track | Action | Control |
|---|---|---|
| Audio | Re-encode to AAC | Audio kbps (48–320) |
| Video | Re-encode to H.264 (always) | Video kbps (100–50000) |
| Audio passthrough | Not available | Use a container tool instead |
| Output | Faststart MP4 (fixed) | — |
Cookbook
Audio-bitrate scenarios, each showing both numbers you type (video is always re-encoded too) and the command the tool runs.
Shrink a podcast video's audio to 96 kbps
Speech is transparent at 96; this trims the file. Keep video near the source rate.
Source: 1080p talking head, ~5 Mbps video, 256k audio Video kbps: 5000 Audio kbps: 96 ffmpeg ... -b:v 5000k -maxrate 6000k -bufsize 10000k \ -c:a aac -b:a 96k -movflags +faststart Saves ~1.6 MB/min on audio alone
Bump a music video to 320 kbps audio
Give music headroom while keeping a sensible video target.
Video kbps: 8000 Audio kbps: 320 ffmpeg ... -b:v 8000k ... -c:a aac -b:a 320k ... Audio near-transparent; video has ample bits
Minimise the picture change when you only want audio tweaked
Video is always re-encoded, so match its target to the source to keep it as close as possible.
Source video bitrate (from ffprobe): ~6200 kbps Goal: only change audio 256 -> 128 Set: Video kbps 6200, Audio kbps 128 (Picture re-encodes but at ~the same rate)
Voice memo clip at the 64 kbps floor-ish
For a tiny voice clip, push audio low to save every byte.
Video kbps: 800 Audio kbps: 64 Small file; audio is lossy but fine for a voice memo (48 is the absolute minimum the field allows)
When you actually want the audio extracted
This tool always returns a video. To get a standalone audio file, use the extractor.
Goal: pull the soundtrack out as a separate file -> wrong tool. Use audio-track-extractor (/video-tools/audio-track-extractor) This tool re-muxes audio back into an MP4 video
Edge cases and what actually happens
Video gets re-encoded even though you only changed audio
By designEvery run applies -c:v libx264 -b:v <Video kbps>k, so the picture is always re-encoded — there is no audio-only or stream-copy path. To minimise the visible change, set Video kbps near the source's existing video bitrate (check it with ffprobe).
You wanted to keep the original video untouched
Not supportedThis tool can't passthrough the video stream. If preserving the exact video is essential, change audio in a workflow that allows stream-copy, or extract/replace audio outside this tool. The bitrate setter always transcodes both tracks.
Audio quality didn't improve going from 128k source to 320k
No gainRe-encoding an already-lossy 128 kbps track at 320 can't recover detail the original discarded — it just makes a bigger AAC file. Higher Audio kbps only helps when the source audio still holds that detail.
Output codec is AAC even if your source was something else
By designAudio is always re-encoded to AAC, whatever the source codec (MP3, Opus, AC-3, PCM). If your delivery needs a non-AAC audio codec, this tool can't produce it.
Source has no audio track
PreservedWith no audio stream there's nothing to encode — the output is video-only and your Audio kbps value is simply ignored.
Multi-track / surround audio
CautionThe encode targets a single AAC audio output; multiple language tracks or discrete surround channels may be collapsed or not preserved as separate tracks. For multi-track work, handle audio with the audio-track extractor and a dedicated audio pipeline.
You actually want the audio as a separate file
Wrong toolThis always returns an MP4 video with the audio muxed back in. To export a standalone audio file, use the audio-track extractor.
Audio kbps below 48 rejected
LimitThe field floors at 48 and caps at 320. Values outside that range aren't accepted. 48–64 is usable for voice but expect audible artefacts on music.
Source over the tier size cap
RejectedFree 1 GB / 1 file, Pro 10 GB / 5, Pro-media 100 GB / 50, Developer 100 GB / unlimited. Files above your cap are rejected before encode. No duration limit applies.
Frequently asked questions
Can I change only the audio bitrate without touching the video?
No. This tool always re-encodes the video too (libx264 at your Video kbps). There's no audio-only or passthrough mode. To keep the picture as close as possible, set Video kbps to the source's existing video bitrate so the re-encode barely changes it.
What audio bitrate should I use?
128 kbps AAC for speech, 192 for general video, 256–320 for music. The field allows 48–320. See the content-type table above.
Is 320 kbps the maximum?
Yes — 320 kbps is the top of the Audio kbps field (and the practical top of AAC). The minimum is 48 kbps.
Does raising audio from 128 to 320 improve a lossy source?
No. Re-encoding an already-lossy 128 kbps track at 320 can't restore detail the original codec threw away — it only makes a larger AAC file. Higher bitrate helps only when the source still has the detail.
What audio codec does the output use?
Always AAC, regardless of the source audio codec. If you need PCM/WAV, Opus, or AC-3 output, this tool can't produce it.
How much space does lowering audio actually save?
Audio size scales with bitrate: 320 kbps is ~2.4 MB/min, 128 kbps is ~0.96 MB/min — about 60% less. On a long clip with music dropped to speech-grade audio, that adds up. See the per-minute table.
Can I extract the audio as a separate file here?
No — this always returns an MP4 video with audio re-muxed in. Use the audio-track extractor to export a standalone audio file.
What happens to multiple or surround audio tracks?
The encode produces a single AAC audio output, so multi-language or discrete surround tracks may not be preserved separately. For multi-track work, handle audio with the audio-track extractor.
What if my video has no audio?
Then there's nothing to encode and the output is video-only — the Audio kbps value is ignored.
Does it support 5.1 surround?
There's no surround/channel control in this tool, and the output targets a single AAC track, so a 5.1 source may be downmixed or not preserved as discrete channels. Use a dedicated audio pipeline for surround delivery.
Is my audio uploaded anywhere?
No. Both audio and video are processed in your browser via FFmpeg.wasm — nothing leaves the tab. Free allows up to a 1 GB source.
Why does the output stream so quickly?
Every output is a faststart MP4 (-movflags +faststart), so the moov atom sits at the front and the file plays and seeks immediately.
Privacy first
Every JAD Video tool runs entirely in your browser via WebCodecs and FFmpeg (WebAssembly). Your video files never leave your device — verified by zero outbound network requests during processing.