Workflow: Mastering
Specula judges a master against where it is going. Under Music streaming, the delivery card says what Spotify, Apple Music, YouTube and Tidal do to the file, in a sentence, beside the true peak, the mono fold-down and what Apple Music's AAC encode does to the peaks. Where a fix would change the verdict, the card offers it by name.

A finished master still has to survive Spotify, Apple Music and the rest.
Four things go wrong between the bounce and the upload, and none of them is audible on a quiet pass through the monitors. The integrated loudness overshoots and Spotify turns the track down. A quiet master stays quiet on Apple Music, which only turns loud tracks down. The true peak clears −1 dBTP in the WAV and lands over it after the AAC encode. And a stereo image that thins out in mono loses its vocal on a phone speaker. Each of these is a number the file already contains, and the delivery card reads them out as soon as the file has loaded.
Deliver to Music streaming, read the card, check the encode and the mono fold-down, hear it at the platform's level, fix, export.
Apple Music's Sound Check, YouTube and Tidal only turn loud tracks down. A master at −18 LUFS plays at −18 on Apple Music, not at its −16 reference, and the card says so: "Apple Music plays this as recorded, 2.0 dB under its −16 LUFS reference." It reads green and offers no level fix.
Spotify (in its default mode), Amazon Music, Deezer and SoundCloud raise quiet tracks, so their "turns this up by" is a real gain, and Hear at their level plays it.
Music streaming holds Spotify (−14 LUFS), Apple Music (−16), YouTube (−14), Tidal (−14), Amazon Music (−14, with a −2 dBTP ceiling), Deezer (−15), SoundCloud (−14) and the vendor-neutral AES TD1008 (−18). Every target's reference, tolerance and ceiling is in the user guide, and a custom target joins the category beside them.

On macOS 27, Apple's on-device music analysis reads a song's sections, bars and stems, and Specula measures against them with its own BS.1770-4: each section's loudness on the timeline, shaded from quieter to louder, the vocal's loudness against the bed, and the sections as chapters for a table of per-section readings. Songs



Specula reads the finished master: what each platform does to its level, whether the true peak clears the ceiling before and after the AAC encode, whether the stereo image holds in mono, and the fix where one would change the verdict.