What happens when you combine audio files with different sample rates?
Short answer: the joiner converts every clip to 44.1 kHz before joining, so clips with different sample rates play at the right speed and pitch. We joined 8, 22.05, 44.1, 48 and 96 kHz clips in one go and measured the file that came out:
- Speed and pitch: unchanged. Every test tone came back at its exact frequency (441.0 Hz stayed 441.0 Hz), and every clip kept its length to within 1 sample (0.02 ms).
- Mono with stereo: the result is stereo, and each mono clip plays on both sides at its own level (-6.0 dBFS in, -6.0 dBFS out on each side). If every clip is mono, the result stays mono.
- Volume: unchanged. The joiner does not even out loud and quiet clips.
- 5.1 files: folded to stereo the usual way, with the centre speaker on both sides and the LFE (bass) speaker left out.
To combine audio files like this, open the audio joiner, add all the files, and pick WAV under Export format if you want the joined clips exactly as converted. No sample rate setting is needed.
Why sample rate matters when you combine audio
A sample rate is how many values per second a file stores: 44,100 for CD audio (the Red Book standard), 48,000, or as little as 8,000 in some voice recordings. One output file can only have one sample rate. If a tool copied the values of a 48 kHz clip into a 44.1 kHz file without converting them, that clip would play 8.8% longer and lower in pitch, because its 48,000 values per second would be read at 44,100 per second.
So a joiner has to resample: work out new values at the output rate that describe the same sound. The audio joiner on this site asks the browser to decode every file at 44.1 kHz, so the resampling is done by the browser's own decoder, on your device. We tested it in Chromium; we did not test Firefox or Safari.
How we tested
- Clips: short files we generated ourselves, each made of pure sine tones so pitch, level and length can be checked exactly: a 22.05 kHz mono clip (441 Hz), a 48 kHz stereo clip (1,000 Hz left, 2,500 Hz right), a 44.1 kHz stereo clip, an 8 kHz mono clip (300 Hz, like a phone recording), a 96 kHz mono clip with a 30 kHz tone above hearing, a 48 kHz stereo clip with a 23 kHz tone, a 48 kHz 5.1 surround WAV with a different tone for each of its six speakers, and a 48 kHz stereo MP3 made with FFmpeg's LAME encoder at 192 kbps.
- Tool: this site's audio joiner, run in Chromium 153 through Playwright. All files were added with one file pick, crossfade 0, no fades, exported as WAV, and once as MP3 at 128 kbps.
- Measured: samples per clip in the joined file, the frequency and level of each tone, whether the left and right sides are identical, and how far each clip is from ideal sine tones (in dB; higher is cleaner).
Results: length, pitch and level after joining
Six clips at five sample rates went in; one 44.1 kHz stereo WAV of 529,199 samples came out. The page showed 12.000 s total, and the joined file held exactly the sum of the six clips.
| Clip | Samples in the join (ideal at 44.1 kHz) | Tone in, tone out | Level in, level out | Clean (dB) |
|---|---|---|---|---|
| 22.05 kHz mono, 2 s | 88,200 (88,200) | 441 Hz, 441.0 Hz | -6.0, -6.0 dBFS both sides | 86.1 |
| 48 kHz stereo, 3 s | 132,300 (132,300) | 1,000 / 2,500 Hz, 1,000.0 / 2,500.0 Hz | -6.0 / -12.0, -6.0 / -12.0 dBFS | 86.3 / 82.3 |
| 44.1 kHz stereo, 1 s | 44,100 (44,100) | 600 / 700 Hz, 600.0 / 700.0 Hz | -6.0, -6.0 dBFS | 86.6 / 87.0 |
| 8 kHz mono, 2.5 s | 110,250 (110,250) | 300 Hz, 300.0 Hz | -6.0, -6.0 dBFS both sides | 87.8 |
| 96 kHz mono, 1.5 s | 66,150 (66,150) | 5,000 Hz, 5,000.0 Hz | -6.0, -6.0 dBFS both sides | 86.2 |
| 48 kHz stereo MP3, 2 s | 88,199 (88,200) | 800 Hz, 800.0 Hz | -6.3, -6.3 dBFS | 80.1 |
What the numbers show:
- No speed change. Four clips kept their length to the sample. The 48 kHz MP3 lost one sample out of 88,200, and a separate 1-second 48 kHz WAV also came out at 44,099 samples. One sample is 0.02 ms.
- No level change. The -6.3 dBFS of the MP3 clip was already in the MP3: decoded on its own by FFmpeg at 48 kHz, it read -6.3 dBFS too.
- Mono clips had bit-identical left and right sides in the stereo result. When we joined only the two mono clips, the result was a mono WAV of 198,450 samples.
- Clean conversion. After resampling, the WAV clips were 82 to 88 dB away from ideal tones, including the 8 kHz clip that had to be stretched to more than five times as many samples.
Sound above 22.05 kHz: removed, but not always completely
A 44.1 kHz file cannot hold anything above 22.05 kHz. A resampler should filter it out; if it does not, it folds back as a lower tone (an alias). We put a tone above that limit into two clips:
| Clip | Tone above 22.05 kHz | Where an alias would land | Level found there |
|---|---|---|---|
| 96 kHz mono | 30,000 Hz at -12.0 dBFS | 14,100 Hz | -97.4 dBFS |
| 48 kHz stereo | 23,000 Hz at -6.0 dBFS | 21,100 Hz | -48.9 dBFS |
The 30 kHz tone was removed almost completely. The 23 kHz tone, just above the limit, came through as a 21.1 kHz tone 42.9 dB below the original. We do not know how common strong content between 22.05 and 24 kHz is in real 48 kHz files, so we only report what the test tone did.
5.1 surround files are folded to stereo
The output is mono or stereo only. Our 48 kHz 5.1 WAV (stored in the order left, right, centre, LFE, left surround, right surround, each speaker a different tone at -10.5 dBFS) came out like this:
- Left = left at -18.1 dBFS, plus centre and left surround at -21.1 dBFS each.
- Right = right at -18.1 dBFS, plus centre and right surround at -21.1 dBFS each.
- LFE: not in the result (below -105 dBFS on both sides).
That is the usual fold-down: centre and surrounds 3 dB below the front speakers, the LFE left out, and the whole mix turned down so the sum cannot clip. We measured it on 2026-09-29, after a fix. The version live earlier that day put the centre speaker only on the left and the LFE only on the right; the same fixed fold-down now runs in the cutter and the format converters on this site. We tested with a 5.1 WAV; the order inside other surround formats may differ. If the bass from the LFE matters, join a stereo version made from the original instead.
Exporting the combined file as MP3
The same six-clip join exported as MP3 at 128 kbps was a 44.1 kHz stereo file of 192,679 bytes. Decoded with FFmpeg 4.4.2, it started 1,105 samples (25.1 ms) later than the WAV export. That short silence at the start is added by the MP3 encoder; we measured it in Can you trim an MP3 to the exact millisecond?. MP3 export also re-compresses the audio; see Does cutting an MP3 lose quality? for the loss per bitrate.
How to combine files with different sample rates
- Open the audio joiner and add all your files. Each row shows the clip length and whether it is mono or stereo.
- Put them in order. You do not need to convert any file to one sample rate first.
- A 5.1 surround file is folded to stereo without its LFE (bass) speaker; if you have a stereo version made from the original, you can use that instead.
- If the clips are at different volumes, fix that in the source files; the joiner keeps each clip's level as it is.
- Pick WAV to keep the joined clips exactly as converted, or MP3 at 256 or 320 kbps for a smaller file, then press Merge and download.
Questions
Can I choose 48 kHz for the combined file?
No. The joiner always writes 44.1 kHz. If you need 48 kHz, convert the joined WAV afterwards in your editing program.
Will a phone recording sound better after joining?
No. Our 8 kHz clip became 44.1 kHz, but resampling cannot add sound that was never recorded. A file recorded at 8 kHz holds nothing above 4 kHz, and it still holds nothing above 4 kHz after the join.
Does mixing sample rates slow the joiner down?
We did not time it separately. All files are decoded and resampled when you add them, before you press Merge and download.