Opening Finito the first time

macOS blocks apps from developers it does not recognise. Finito is one, so the first launch takes one extra step. This is expected, and it happens once.

1. Open the disk image and drag Finito into your Applications folder. Do not run it from the disk image.
2. In Applications, right-click (or Control-click) Finito and choose Open.
3. A dialog will warn that the developer cannot be verified. Click Open.

After that, Finito opens normally from the Dock, Launchpad, or Spotlight.

Common questions

"Finito can't be opened because it is from an unidentified developer"

Follow the right-click ▸ Open steps above. If you have already tried and it still refuses, make sure you moved Finito to Applications first, because macOS treats apps run from a mounted disk image more strictly.

Nothing happens when I double-click an audio file

Double-click, "Open With ▸ Finito", and dropping a file on the Dock icon all work from version 1.5.0 onward. On earlier versions they did nothing; update, or use File ▸ Open… (⌘O) inside the app.

"This file is not in an audio format Finito can read"

The file is damaged, is not actually audio despite its extension, or is a protected purchase such as a DRM-locked iTunes track. Try playing it in QuickTime or Music: if they cannot open it either, the file is the problem. Converting it to WAV in another app will usually settle it.

The app is slow, or playback stutters

Finito keeps the waveform, analysis, and live master in memory. On an 8 GB Mac with a long track, close other large apps first. If playback stutters, try a different output device from the picker above the waveform. Editing processor settings and switching Studio sections should remain smooth while audio plays; if a particular control causes an interruption, include that control and your output device in a bug report.

My export sounds different from what I heard

Live playback and export use the same mastering controls and processor choices. Export may use a longer Linear Phase EQ filter and adds final format conversion and dither where selected. Headphone Reference, Gain Match, and monitor volume affect listening only. Open Delivery Review from the Export menu to audition the rendered master, compare codec previews, and inspect the finished file. If the difference remains large, please report it with the settings and export format.

Where are the detailed controls?

Switch to Studio at the top of the window. The Direction, Tone, Stereo, Character, Color, Dynamics, and Loudness tabs sit below the waveform and stay within reach as you scroll. Choose Expand beside the waveform to see larger meters, fades, and trim. Tone's routing, phase, High Pass, and Low Pass controls sit beside its EQ spectrum; Dynamics has separate Low-End, Compression, and Impact pages. The in-app Help guide explains every section.

How do I start in Guided Mode?

Analyze the song, then choose its Style, Strength, and Loudness target. Natural, Warm, and Open change broad tonal character; Gentle, Balanced, and Bold rebuild the Smart recommendation with different amounts of processing. The Listen card compares Original and Mastered with Gain Match on, while Compare strengths auditions complete rendered alternatives. Fine Tune adds optional Tone, Punch, and Width trims, and What Changed explains the active processors. Cancelling a strength analysis keeps the existing processing and style. Guided and Studio share the same settings, playback, undo history, and export.

How do I compare without a volume bias?

Use Original and Mastered in the waveform area, then turn on Gain Match before judging tone or punch. Until you Verify or Export the current settings, Gain Match estimates the adjustment from makeup gain. A current render supplies a measured LUFS difference; processing edits discard that measurement until you verify again. The louder source is turned down, including Original when the master is quieter, and a measured zero leaves both levels unchanged. Gain Match, Headphone Reference, and monitor volume affect listening only. The live meters below the waveform measure playback; Verify or Export provides the final whole-track figures.

How do I compare Smart directions?

Open Compare Smart in Studio's Direction page, then choose Render comparisons. Transparent, Balanced, and Competitive appear in three aligned cards with queued, preparing, and prepared states. Once all three complete renders are ready, the cards show measured loudness, true peak, crest factor, and limiting, with separate Listen and Use controls. Playback matches the quietest render and keeps the same song position when you switch directions; Pause, Stop, and the position slider control the audition. Your current settings stay intact until you choose Use. Smart Boundaries and processor bypasses remain respected, and closing or cancelling applies no direction.

What does Full-song Review do?

After analysis, choose Review full song in Guided Mode or Studio ▸ Direction. Finito renders the complete current master and compares its envelopes with the source throughout the trimmed song. It marks reduced attack contrast, introduced level movement, and low-band gain movement as timestamped listening prompts. Deliberate fades and silence are excluded from flags; these measurements are not a judgement of musical quality.

When the evidence warrants it, the review tests one small reduction to active compression, Low-End Leveler, or positive Limiter Drive within your Smart Boundaries. A refinement is offered only if it improves the measured tradeoff, stays within 0.75 LU of the current master, respects the peak ceiling, and avoids a material regression in any checked passage. Listen to Current master and Refinement at matched loudness, or choose a timestamp to compare that moment. Use refinement applies only the listed changes; Undo restores the previous settings. Closing or cancelling keeps your settings and removes the comparison renders.

Changing settings, source audio, or Smart Boundaries makes the result stale and requires another review. These auditions use the source-rate mastering render without Headphone Reference, delivery format conversion, codec loss, or dither. Use Delivery Review and finished-file verification for those delivery checks.

How should I use Headphone Reference?

Open Headphones beside the waveform. Listening selects Measured studio, a modeled translation room, or an imported response, plus crossfeed, room amount, Trim, and Subwoofer Boost. Subwoofer Boost starts Off and adds a 60 Hz low shelf with up to +6 dB of boost for checking bass-heavy systems. Calibration selects a headphone profile, correction amount and range, maximum combined boost, and separate bass and treble preferences. Its graph shows the applied EQ response before preamp and matching; it is not a measurement of your headphones' final response. All headphone adjustments affect listening only.

Measured studio uses the University of Surrey IoSR listening-room recording: Genelec 8330A speakers at ±30°, captured with a forward-facing Cortex MKII head and torso. Each of its four speaker-to-ear paths comes from the recording, with relative timing and level differences preserved. The captured speaker placement and head orientation stay fixed. Preparation applies one common safety gain and fades the quiet tail to fit 683 ms. Attribution: Jon Francombe, University of Surrey, 2017, CC BY 4.0.

Dry, Nearfield, Living room, Car, and Club remain modeled translation checks. Speaker Presentation changes these models; measured and imported responses keep their captured presentation. Room amount blends the complete response with dry calibrated headphones rather than adjusting reflections alone. The measured capture uses a generic listener and does not track head motion. Its realism can vary with hearing anatomy and headphone fit, and controlled listening comparisons are still pending. Check familiar finished recordings at 0% and 100% room amount, and confirm tonal decisions on another trusted playback system. A missing or damaged measured capture shows an error rather than substituting a model.

Headphone Gain Match has a separate power control and Playing music or Estimate methods. Estimate provides an initial match from a stereo probe. Playing music captures three seconds of audible Original or Mastered playback, then holds the adjustment steady; silence does not count. The status distinguishes preparation, waiting, paused capture, measurement, a completed match, Estimate, and bypass. Recheck clears the previous capture even while paused; resume live song playback to measure another passage. Changing the listening setup also starts a new match. The displayed match offset excludes Trim and the fixed headroom reserve, which applies equally to the processed and dry comparison. This match is separate from Original/Mastered Gain Match near the waveform.

Monitor Peak shows the headphone chain's protected sample peak in dBFS, separately from the master's true-peak meter. Recent protection reports automatic level reduction; it is not a clipping detection result. Off, paused, and preparing states show no live peak reading. Reference and separate preview players do not supply the live headphone-matching capture.

Save preset… retains the complete listening setup, including Subwoofer Boost; recall preserves the current Headphone Reference On/Off state. Imported AutoEq profiles persist locally. Import response… accepts a four-channel WAV ordered LL, LR, RL, RR: left speaker to left ear, left to right, right to left, and right to right. Responses must be 8–192 kHz, at most 683 ms and 16 MB. Use responses you own or are licensed to use; ordinary stereo room responses and proprietary vendor presets are incompatible. The response is copied into the local library so moving its original file does not break saved presets.

Can I recover a session after the app is interrupted?

Finito saves local recovery checkpoints in the background after edits and periodically, waiting for analysis, repairs, and stem summing to finish. After an interrupted run, choose Restore at startup or File ▸ Recover Interrupted Session. The last complete checkpoint opens as unsaved work, including settings, applied repair audio and recipes, learned noise profiles, stem settings, and A–D versions. Original audio and reference files are still required and can be relocated if they moved.

Later keeps the checkpoint; Discard Checkpoint removes it. Normal session close or quit clears that session's recovery copy. Recovery never overwrites a saved session and may not contain the very last edit before an interruption. Use Save Session for a permanent copy.

Why does Delivery Check ask me to verify again?

Changing processing settings makes the previous render out of date. Preflight guidance can still identify settings to review, but a loudness target is not a measurement of the finished master. Choose Verify or Export to measure the current settings. Delivery Review labels an older result as Last render; each completed export is independently reopened and checked, including its format, duration, loudness, and true peak.

How should I use repairs on a finished mix?

Open Repair beside the waveform. Choose a short waveform selection, or explicitly choose Whole Track for persistent hum or hiss. Start with a modest Amount, then compare Original, Repaired, and Removed in Preview. Removed should contain the unwanted defect rather than drums, cymbals, or intentional musical tails. Preview and Apply use the same repair scope; starting Preview stops competing playback. New De-Noise settings start with a 6 dB reduction ceiling and strong music preservation. Advanced settings open from the sliders button in each card header.

Does Scan for Problems repair the track automatically?

No. It lists suspected clicks, clipping, hum, short digital dropouts, and noise-like passages. Each finding needs listening review: Select chooses the passage, Preview auditions the suggested repair, and Skip dismisses it. A musical tone can resemble hum, and intentional texture can resemble noise. Spectral Selection opens the spectrogram, where dragging a rectangle defines both the time passage and frequency range. Reconstruct uses nearby sound for patches up to 0.5 seconds; longer areas use attenuation. Dropout reconstruction is limited to digital-zero gaps up to 25 milliseconds with reliable surrounding sound.

Where are Mouth De-click and De-reverb?

They are under Specialist restoration, below the everyday repair cards. Use them for exposed defects in a sparse passage. They cannot isolate a vocal from a dense mix or recover a completely dry source. Main-source De-reverb uses linked stereo processing, tonal and attack protection, and a 6 dB spectral attenuation limit at full Amount; always check Removed for lost ambience or sustain. Per-stem De-reverb now also offers linked stereo processing, an adjustable reduction limit, and sustain protection in Options. Audition it on a short passage before applying.

Can I change a repair after applying it?

Yes. Session Repairs keeps editable settings and bypass controls for each new repair. Changing an earlier step rebuilds later steps in order and replaces the working audio only when every step succeeds. Cancel leaves it unchanged. Saved sessions retain the recipes, baseline audio, and learned noise profiles. Older sessions still open using their baked repair audio.

What is available in per-stem Repair?

Open a stem editor and choose Repair. Ten tools address clicks, crackle, plosives, hum, clipping, noise, sibilance, unwanted room tails, spectral patches, and short digital dropouts. Options exposes detection, reconstruction, frequency, and protection controls. Scan Stem suggests candidates for listening review; it never applies a repair automatically. Spectral Selection lets you draw a time-and-frequency patch. The spectrum displays up to 20 seconds at a time: select a passage on the waveform and refresh to inspect it.

Preview processes the same complete scope as Apply, even when you listen to a shorter loop. Preparing a long stem can take longer; Cancel stops a pending render. Floating-point working files preserve reconstructed peaks for the mastering chain. Original, Repaired, and Removed let you judge the change in the mix.

To learn a noise floor, select a noise-only passage and choose Learn Noise from Selection in De-noise Options, then choose the passage to repair. Profiles travel with saved sessions. Applied per-stem passes can be edited, bypassed, enabled, or removed; later steps rebuild together, and cancellation or failure keeps the current audio intact. Older recipes retain their original processing unless you change an advanced option.

Where is Export?

Use the Export button at the top of the window or File ▸ Export. Both open the same format, sample-rate, dither, metadata, and preview choices; ⌘E opens the 24-bit WAV save panel. Finito warns before closing a session or quitting while an export is running.

Where did my temporary files go?

Finito writes working renders to your Mac's temporary folder and removes each one as it is replaced. macOS clears anything left behind on restart. Exports go only where you choose to save them.

How do updates work?

Finito checks for a new version about once a day and tells you when one is available. Updates are cryptographically signed and verified before they install. You can turn automatic checking off in the app's settings, and check by hand from Finito ▸ Check for Updates… at any time.

Is my music sent anywhere?

No. All processing happens on your Mac, including the Smart Analysis suggestions. See the privacy policy for the specifics.

Will there be an Intel or Windows version?

No. See system requirements for why.

Reporting a bug

Email frankiefinito@gmail.com. The more of this you can include, the faster it gets fixed:

• Your Finito version, shown in Finito ▸ About Finito, e.g. v1.17.0 · 20261006.1
• Your macOS version and Mac model, from Apple menu ▸ About This Mac
• What you did, what you expected, and what happened instead
• The exact wording of any error message
• The format, length, and sample rate of the file involved

Please do not send audio unless I ask for it. If I do, it is used only to reproduce your bug and deleted once it is fixed.

Requests and feedback

Finito is built by one person, and what gets worked on next is shaped largely by what people ask for. If something is missing, awkward, or gets in your way, say so: frankiefinito@gmail.com. There is no ticket system and no bot; the reply comes from me.