How to read a recipe

The first parts of this book keep two promises: the object chapters say what each tool is for, and the machine chapters say why to trust it. This part makes a third kind of promise. A recipe puts several objects on one patch cord and chases a specific sound — a record you have heard, an instrument you have coveted — and tells you honestly how close the kit gets.

Recipes are held to the house rules, adapted:

• Every knob named exists, spelled the way the attribute is spelled. A recipe is checkable against the reference pages; if it says @decay 0.8, that attribute takes that value on the shipping object.
• Settings are starting points, not measurements. A recipe's numbers get you into the neighborhood; your ears walk the last block. Where a chapter number is a measurement (a decay time, an alias floor), it still cites the executed notebook or pinned test that carries it — the recipes borrow those numbers rather than re-deriving them.
• Provenance stays honest. When a recipe chases a record, it says what is documented about how that record was made and what is folklore. When it chases an instrument, it leans on the same published analyses the kernels were built from. What a recipe never does is claim to be the record — mix, room, tape, and hands are not in the box.
• Every recipe ranks its ingredients. The house habit from the Moog recipe in the oscillator chapter: list what each element buys, in order of importance, so you know what to cut first when CPU or taste says so.

Each recipe has the same skeleton: the sound and where it came from, the signal chain, the settings (tables for knobs, grids for patterns), what each ingredient buys, and — because every tool is sometimes the wrong tool — when to leave the recipe and cook something of your own.