Documentation
Understand your audio.
From simple explanations to detailed methods: discover what each analysis observes, how to read its results and where its limits lie.
Clean means no detector raised a finding. A dash means a measurement is unavailable. Audition flagged passages and read the limits before drawing a conclusion.
Authenticity & spectrum
- Real bit depth
Unused bits and quantization grids beneath small residuals.
- Upsampling
Persistent empty high band in a high-rate container.
- AAC / MP3 transcoding
Statistical search for quantization grids left by lossy codecs.
- Spectral cutoff
Content boundary, cliff and upper-band level.
Stereo & phase
- Dual mono
Identical channels or a negligible stereo difference.
- Global polarity
Left/right correlation and strong whole-track opposition.
- Local & band phase
Brief or hidden opposition across four frequency bands.
- Stereo balance
RMS level difference between left and right channels.
- HF Stereo
Stereo narrowing in the high band. Experimental cue.
Levels & dynamics
- DC offset
Mean displacement from zero, measured per channel.
- Clipping
Runs of samples at or very near full scale.
- True peak
Estimated inter-sample peaks, in dBTP.
- Dynamics (DR)
Gap between sample peak and loud-passage RMS.
Loudness
- Integrated LUFS
Programme loudness with K-weighting and measurement gates.
- LUFS-M max
Loudest 400 ms window and its location.
- LUFS-S max
Loudest 3-second window and its location.
- Loudness Range (LRA)
Variation in short-term loudness, expressed in LU.
Integrity & discontinuities
- Impulses
Suspected isolated pulses, with locations to audition.
- Dropouts
Short exact-zero gaps within active audio.
- FLAC MD5 signature
Decoded PCM compared with the signature stored in FLAC.
- MD5 & CRC32 fingerprints
File-byte identification, including after a move.
DSD sources
- PCM source in DSD
Search for boundaries consistent with 44.1 / 48 kHz PCM.
- DSD heritage in PCM
Spectral valley and ultrasonic rise consistent with a DSD transfer.