“Stereo audio” in a delivery spec is ambiguous, and the ambiguity is not pedantic. It can mean one stereo track. It can mean two mono tracks that a system will pair. It can mean a stereo pair inside an eight-channel track where the other six are something else. All three are “stereo” to somebody, and only one of them is what the receiving system expects.
Tracks and channels are different counts
A container holds tracks. A track holds channels. Both numbers matter and they are routinely confused.
- One stereo track — 1 track, 2 channels.
- Two mono tracks — 2 tracks, 1 channel each, 2 channels total.
- Programme mix plus M&E, both stereo — 2 tracks, 4 channels total.
- A 5.1 mix — usually 1 track, 6 channels. Sometimes 6 mono tracks.
A tool that reports only “2 channels” cannot distinguish the first case from the second. A tool that reports only “2 tracks” cannot distinguish the second from the third. You need both counts, and for anything beyond a single stereo pair you need to know what is in each track.
What the extra tracks usually are
Deliveries carry more than the mix people hear because the material gets reused.
M&E — music and effects, the full mix with dialogue removed. This is what makes a programme dubbable into another language. A delivery that is supposed to carry M&E and does not is a delivery that cannot be localised, and it is often discovered months later by somebody in another country.
Stems — the mix split into its groups, typically dialogue, music and effects as separate tracks. More flexible than M&E and larger.
Audio description — a narration track describing what is on screen, either as its own track or pre-mixed.
Clean and dirty versions — with and without on-screen text or graphics audio.
A typical broadcast delivery might therefore be: track 1 stereo programme mix, track 2 stereo M&E, track 3 stereo audio description. Three tracks, six channels, and a spec that just said “stereo” would have been satisfied by one of them.
Channel order is part of the spec
Within a multi-channel track, the order is fixed by convention and getting it wrong is not subtle. The common film order for 5.1 is L, R, C, LFE, Ls, Rs. The SMPTE order used in some broadcast contexts is L, R, C, LFE, Ls, Rs as well, but you will meet L, C, R variants in older material and in some game audio.
Swap centre and LFE and dialogue comes out of the subwoofer. That is not a subtle degradation; it is a phone call. It is also invisible in a waveform view and inaudible on a stereo monitor, which is how it survives to delivery.
For multi-track deliveries the equivalent question is track assignment: which track is the programme mix. Systems that expect track 1 to be the mix will happily play the M&E if the tracks were written in the other order, and the result is a broadcast with music and effects and no dialogue.
Sample rate and bit depth
Two smaller fields that produce a disproportionate number of rejections.
48 kHz is the professional video standard, and it is not negotiable in most delivery specs. Material that came from a music workflow is often 44.1 kHz, and it will be resampled somewhere downstream by something you did not choose.
24-bit is the usual ask for masters, 16-bit for some distribution formats. The failure here is quieter than sample rate: a 16-bit file where 24 was specified plays perfectly and simply has less room before the noise floor, which matters if anyone processes it again.
Both are read straight from the container, which makes them cheap to check and therefore inexcusable to miss.
Reading what a file actually carries
The awkwardness is that most viewers summarise audio by its first track. Open a three-track delivery in a player and you will see “stereo, 48 kHz, 24-bit”, which describes track one and says nothing about the other two.
FileLister reports the audio layout per file rather than per first track: how many tracks the container holds, how many channels in total, and a per-track breakdown showing each track's channel count and codec, with the track language where the file declares one.
The total channel count is the field that catches the common error fastest. A batch where every file reports six channels total and one reports two is a file that was exported without its M&E, and that is visible from sorting a column rather than opening ten files.
Checks worth putting in a filter set
For a recurring delivery, four requirements cover most of it:
- Audio track count equals the number the spec asks for. Catches missing M&E and missing description.
- Total channel count equals the expected sum. Catches a track that is present but mono when it should be stereo.
- Sample rate equals 48000. Catches material that came out of a music session.
- Bit depth equals what the spec says.
One nuance on the first: a requirement of “no audio” is best expressed carefully. A file with no audio and a file whose audio was never examined can look the same to a naive check, and a spec that says “no audio” passing a full stereo mix because nobody read the tracks is worse than no check at all. A tool that distinguishes “measured, and there is none” from “not measured” is doing the right thing.
The part a check will not do
Everything above is structural. It confirms that three tracks exist, that they have the right channel counts, and that the sample rate is right. It does not confirm that track 2 is actually the M&E rather than a duplicate of track 1, which happens more often than anyone would like.
That still needs somebody to listen. What the structural check buys is that when somebody does listen, they are listening to a file that is at least shaped correctly — and they are listening to three files rather than thirty.