Why does dual camera video audio sound worse after you upload it?
The file DualCam wrote did not change. The audio sounds different by the time someone hears it on Instagram, TikTok or wherever you sent the video because the platform that carried it there did two things to that track on the way: it re-encoded the audio, the same way it re-encoded the picture, and on most platforms it also adjusted how loud the track plays back so it matches everything else in the app.
The picture side of this — why a dual-camera video looks softer once it has been re-uploaded somewhere — is its own question with its own answer. This one is about the track underneath it: a single mono AAC recording that starts smaller and more modest than the video next to it, and what happens when something re-compresses and re-levels it a second time.
What DualCam actually wrote, on the audio side
DualCam records one audio track per video, not one per camera — a single mono AAC stream at 44.1 kHz, roughly 96 kbps, taken straight from whatever iOS has set as the phone’s default audio input. There is no gain control to check and no separate mix to balance; the track is written at whatever level the room and the microphone actually produced, and it does not change after the fact. Saving to Photos or opening the share sheet hands over that exact file, audio included — neither step touches it.
So immediately after recording, in DualCam’s own library, that track is as good as it is ever going to get. Anything that sounds different later happened somewhere downstream, in whichever app carried the file the rest of the way.
Two separate things happen on the way out
A second, lossier compression pass
Nearly every messaging or social app re-encodes a video before it will send or accept it, and that re-encode is not a video-only operation — it processes the whole file, container and all, which means the audio track gets compressed again as part of the same pass. A lossy codec like AAC works by throwing away detail a listener is least likely to miss, and doing that twice to the same track is not the same as doing it once: fine detail the first pass kept can be exactly what the second pass decides it can live without.
What that loss sounds like on a voice track is fairly specific. It shows up first as a slight dulling of the top end — sibilant “s” and “t” sounds losing some of their edge, cymbals or clinking dishes in the background turning slightly softer or smeared — well before the words themselves become any harder to make out.
A loudness adjustment, not just a compression pass
Most platforms that host video also normalize loudness: they measure how loud a track plays on average and turn the whole thing up or down to match a target level, so that clips from wildly different sources do not jump in volume back to back in a feed. That is a leveling step, separate from compression, and it is a routine part of how these platforms handle every video that comes through them, not something specific to a dual-camera file.
It is also why the same clip can come out a little quieter on one platform and a little louder on another — each one is normalizing to its own target, not preserving whatever level DualCam actually wrote. A track that sounded right in DualCam’s own library can still get nudged either way once it is somewhere else.
Why a small mono track has less slack than it looks like it should
It is tempting to assume a track this modest has nowhere further to fall, but the opposite is closer to true. DualCam’s audio is already a deliberately restrained recording: mono rather than stereo, because a phone held at arm’s length is not producing a stereo image worth preserving, and encoded at a bitrate that is a small fraction of the roughly 10 Mbps the video track uses. That restraint is a reasonable choice for what a phone microphone can actually capture, but it also means there is less headroom sitting unused in the first pass — less of the kind of inaudible detail a second compression pass can quietly remove before you notice.
A video shot in 4K on a single camera can absorb a re-encode and still look clean, for the same reason: it started with detail to spare. A dual-camera video’s picture cannot do that as easily, because its canvas already tops out at 1080 pixels wide. The audio track is the same story in miniature — it was never carrying much more than it needed to in the first place.
What does not change
The words are still the words. A re-encode and a loudness pass reshape how a track sounds at the margins — its top end, its overall level — not what was said or when. If a dual-camera video’s audio and picture are drifting apart in time after you upload it, that is a separate, unrelated problem with its own cause, not something a re-encode or a loudness adjustment does on its own.
What actually keeps the original audio intact
The same three habits that protect the picture protect the track underneath it, because both are riding through the same pipeline. AirDrop moves the file itself between two Apple devices rather than routing it through a server built to re-encode everything that passes through, so it is the closest thing to handing someone the exact audio DualCam wrote. Sending over Wi-Fi rather than a slow or metered cellular connection matters here too, since some apps compress more aggressively when they assume the connection cannot take a bigger file. And sharing the original from DualCam’s library or from Photos, rather than passing along a copy someone already downloaded from somewhere else, keeps it to one re-encode and one loudness pass instead of several stacked on top of each other.
Common follow-up questions
Does DualCam compress or adjust the audio when I save or share a recording?
No. Saving writes an exact copy of the file into Photos, and sharing hands that same file to the iOS share sheet. Any change in how the audio sounds afterward happens in whichever app carries it from there, not in DualCam.
Why does the same video sound a little louder on one app than another?
Most platforms normalize loudness to their own target level as part of handling every video they host, so the same original track can get turned up or down by a different amount depending on where it lands.
Is the audio still mono after I upload it somewhere that supports stereo?
It will still sound mono. DualCam records a single mono track at the source, and a platform that technically supports stereo has no second channel of real information to add — there is nothing to recover that was not there to begin with.
Can I get the original audio quality back after uploading?
Not from the uploaded copy — the re-encode and loudness adjustment happen on the platform’s side and are not reversible from there. If audio quality matters for a specific clip, share the original file directly, ideally by AirDrop, rather than re-uploading a version someone already downloaded.
Want to just do this?
DualCam records the iPhone front and back cameras at the same time and writes one finished MP4 while you shoot. Free, no account, no ads, nothing leaves the phone.