Including the ones with awkward answers.
No. WavRead never writes to your files and never processes your mix. It measures and explains; your DAW does the editing. The capture plugin passes audio through untouched — the buffer that arrives is the buffer that leaves.
No audio, no analysis, no identifier, ever. Two requests exist in the whole app: a once-a-day check for a newer version, which you can switch off, and a one-time 80 MB model download if you use the built-in stem separation. Everything else runs offline, and the app works with the network cable out.
No account, no sign-in, no licence server. WavRead is free today. It will become a paid app — a one-time purchase — and the plan is that people using it before then keep what they have.
Because it is not code-signed with an Apple Developer certificate yet. Right-click the app and choose Open the first time. Signing and notarisation are on the list; until then that prompt is the honest consequence.
No. The build is Apple Silicon only. See requirements for the full picture.
Integrated loudness follows ITU-R BS.1770, loudness range follows EBU Tech 3342, and true peak is estimated by 4× oversampling as BS.1770 Annex 2 describes. Peak and clipping are measured across the real channels — a mono fold-down understates a peak living in one channel by up to 6 dB, which is a bug WavRead shipped once and now has a regression test against.
They are estimates and the documents label them as such. Chords and key come from pitch-class analysis, melody from a note-detection model, lyrics from speech recognition on the vocal part. Treat them as a good first draft, not a transcription you would print.
No, and it will not pretend to. What it can do — when you capture your master alongside the individual parts — is measure the difference between them: delay times in milliseconds, how long the added tail takes to decay, and the tone change through the bus. That is measurement, not identification. It reports "a repeat at 375 ms", never "this is a Space Designer".
Only if you want every track measured separately. One instance on your master gives you the finished mix. Busses are the sensible middle ground — five instances instead of forty. And if you would rather not use the plugin at all, export your tracks as audio files and drop the folder in: WavRead takes a folder of stems directly.
In one folder you choose, with everything in it: analyses, documents, captures and copies of your audio. Settings shows you where it is, how big it is, and moves it to another drive if you want — copying everything first and removing the original only once the copy is complete.
They stay readable. Documents carry the version and measurement schema that produced them, and when a measurement changes in a way that makes old and new numbers incomparable, WavRead says so instead of mixing them silently. You can reanalyze anything whose original audio is still where it was.
Because it could not be measured — the analysis predates the measurement, the file is too short for that window, or the take has no quiet gap to measure a decay in. A dash never means zero. A number that isn't there is more useful than one that's wrong.
That is what they are written for. The Track Document is plain Markdown containing measurements and no audio. Sharing it is your decision — WavRead never sends it anywhere itself.
Only when you press the button. Update now downloads the release, checks it against the checksum published in the release notes, confirms the app inside is the version it claims to be, and swaps it in — moving your current app aside first and putting it back if anything fails.