Frequency Loom

Change Desk compares two intact sealed acoustic runs at the same planned position. Rank band-level changes and share a report linked to the exact stored evidence.

See what changed at a position

Change Desk compares two intact sealed acoustic runs at the same planned position. Rank band-level changes and share a report linked to the exact stored evidence.

• Open a commissioning project → Verify → Change Desk. • Choose baseline and later runs at the same planned position. Confirm comparable stimulus level, gain, microphone position and room conditions. • Compare broadband and matching-band changes, then share the frozen evidence references.

Both runs must be accepted, nonvoided and seal-valid, with matching input chain, route, sample rate, scale, stimulus and band set. Matching metadata cannot prove matching physical conditions. Differences are not a transfer function, EQ recommendation or proof of improvement.

Frequency Loom is an acoustic commissioning fieldbook for installed sound systems. It turns a listener-area plan into a traceable record of positions, witnessed microphone runs, acceptance results, and handoff evidence.

Start with the job, not a generic meter. Record the site, room, work order, operator, client witness, installed system, microphone, interface, channel, and field calibration context. Lay out one or more listener areas with named measurement positions, microphone heights, plan coordinates, notes, and required repeat counts.

At each position, capture a live evidence run while the installed sound system or another external source plays pink noise, speech-shaped noise, program material, or a sweep; use Ambient only for background capture. Frequency Loom records broadband RMS, one-third-octave band values, active input route, sample rate, captured duration, clipping, and level stability. Short, clipped, unstable, incomplete, interrupted, or route-changing runs are held out of verification instead of being silently treated as valid.

The acceptance matrix evaluates project-defined targets for:

• Planned position coverage
• Listener-area level spread
• Worst cross-position band spread
• Repeat-run agreement
• Optional referenced ambient and commissioned-level limits

Every retained run receives a deterministic SHA-256 evidence seal. Later changes cause an integrity mismatch and exclude the run. Voided runs stay visible as ledger events. Export the complete job as structured JSON, CSV, or a readable commissioning report through the iOS share sheet.

The commissioning workflow requires no account or developer network service and works offline; Methodology links open external reference websites only when you tap them. Microphone samples are analyzed on device and are not saved as audio. Projects remain in private local storage until you deliberately export them. There are no analytics, ads, tracking SDKs, remote flags, or subscriptions.

Uncalibrated work is labeled relative dBFS. A documented reference match can produce referenced dB SPL values, but Frequency Loom does not claim that an iPhone, microphone, calibrator, or input chain conforms to IEC, ISO, ANSI, or AVIXA standards. The app is a commissioning evidence fieldbook, not a class-rated sound level meter or standards certificate.