Audio Info Inspector

How the Audio Info Inspector works

Audio info inspector: getting a good result

Audio files are good at hiding what they are. A "CD quality" download may have been made from a low bitrate source; a podcast episode may be stereo when the host only ever recorded in mono, doubling its size for nothing; a file that sounds fine may be sitting a hair below clipping with no headroom left for the next stage. The inspector decodes the file and reports the properties that matter when you are deciding whether it is fit for a job.

Format. The container is identified from the file extension and the MIME type the browser reports — WAV, MP3, AAC in MP4, FLAC, Ogg Vorbis, Opus, WebM, AIFF and so on. This is a label, not a deep parse: if a file has been renamed, the decoder will still open it, but the format line will say what the name claims. The bit depth line is the give-away in that case, since a renamed MP3 called .wav will report an impossible bit depth.

Sample rate and channels come from the decoder. The sample rate sets the highest frequency the file can contain — half the sample rate, so 22.05 kHz for a 44.1 kHz file — and mismatched rates are a common reason a clip sounds wrong when dropped into a project. The channel line also checks whether a stereo file is really stereo: if every sample in the left channel equals the right, the tool reports the file as effectively mono, because a "dual mono" file can be halved in size with no loss at all.

Duration and bitrate. Duration is shown as a clock reading and as a sample count, which is the number that matters when you are lining files up in an editor. Bitrate is worked out from the file size and the duration rather than read from a header, so it is the true average across the file — for a variable bitrate MP3 that is the number that determines quality and bandwidth, and it will not match the nominal figure a player shows.

Bit depth. Lossy formats do not have a bit depth; they store frequency coefficients and the decoder produces 32-bit float samples. For WAV and AIFF the tool estimates the depth from the size, duration, sample rate and channel count, which lands on 16, 24 or 32 for a normal PCM file and on something odd for a file that is not what its extension says.

Levels. Peak is the loudest single sample in decibels relative to full scale, and a value within 0.1 dB of the ceiling is flagged as at or near clipping. RMS is the average energy, a rough guide to how loud the file will sound. The crest factor is the gap between the two: a large crest factor means a dynamic recording with quiet passages and sharp transients, a small one means a dense, compressed master. Each channel is listed separately with its own peak, RMS and DC offset, so a left-right imbalance or a channel with a stuck offset is obvious at a glance. For a loudness reading against streaming targets use the Loudness Meter; for a sample-by-sample clipping check use the Clipping Detector.

Export. Plain text gives a two-column report suitable for pasting into an email or a ticket. JSON gives the same fields as a machine-readable object, useful if you are documenting a batch of files or feeding another script.

Related tools