Key and tempo in a working session: 91 situations

When the file will not say

1. A loop pack arrives and the key is nowhere in the filenames.

Play a representative loop or drop its audio file on Key. Check the result against a pitched instrument before using it to set the project.

2. A filename claims a key and the audio disagrees.

Compare the exact file and its current pitch with the tag. Either the tag or the detector can be wrong; listen before replacing metadata.

3. A sample was pitched down after it was bought, so its label is now stale.

Analyze the processed audio rather than the original label. Check whether playback speed changed the tempo as well as the pitch.

4. A single one-shot needs a note, and somebody wants a tempo too.

An isolated pitched hit can suggest a note, but one hit does not establish a repeating beat or a complete musical key. Use a tuner for the isolated note.

5. A collaborator sends a stem with no session notes attached.

Play or drop the stem. A sparse stem may not establish the full song key, so compare with another part when available.

6. An old bounce survived and the project file did not.

Analyze the bounce as evidence of the surviving audio. A detected tempo or key does not recreate the original project or its tempo map.

7. A voice memo was recorded on a phone at three in the morning.

Play the memo on the Mac or drop the supported file. Noise and an incomplete melody can leave the tonal context uncertain.

8. A vinyl rip is running slightly fast.

Check the sounding pitch and pulse. A speed change can affect both; an analysis label does not prove the record is at reference tuning.

9. A construction kit lists one key for the folder, and some files do not match.

Check the files separately. A folder label is not a guarantee that every loop establishes the same tonal center.

10. A pack was tagged by whoever uploaded it and nobody ever checked.

Use the audio to investigate a questionable tag, then verify the result by listening. Do not replace a whole library’s tags from one unverified reading.

11. Metadata sits inside the file, and the DAW has no field that shows it.

Use the readout to inspect the playing audio without relying on a metadata column. Detection and reading an existing file tag are different tasks.

12. A file plays fine and refuses to import into the DAW browser.

System-audio analysis can use a player that opens the file. File-drop support and the DAW’s import support are separate requirements.

13. A drum loop has a tuned ring buried under the hits.

A sustained ring may carry pitch without establishing a key. Compare the percussion with the bass or harmony before choosing a tonal label.

When the sound is not a file at all

14. A track is playing on a streaming service and cannot be downloaded.

Choose system audio and play a representative passage. The analysis is of the audible stream, not the service’s track database.

15. A video has an obvious hook and nothing useful in its description.

Analyze the passage of interest. A title or description does not establish the key of a processed edit.

16. A sample is being previewed in a store before a credit gets spent.

Play the preview and check the available musical context. A short preview may not contain enough information to settle key and tempo.

17. A sound is being auditioned inside a plugin's own browser, where no file exists on disk yet.

Analyze the audible preview through system audio. A one-note audition can suggest pitch without establishing a key.

18. A hardware synth is coming back through an interface.

Monitor the intended input through the Mac, or use microphone mode for sound in the room. Check what the selected input is actually receiving.

19. A turntable is running through the interface into the open session.

Analyze the monitored record and account for platter speed. Compare a longer steady passage when checking tempo.

20. A guitar is going through a pedal chain and into the room.

Choose an input that receives the guitar. Microphone mode can hear the room, while system audio follows the Mac’s playback route.

21. Somebody is holding up a phone across the room.

Microphone mode can hear a nearby phone. Place it clearly and reduce other sound; a noisy room changes the evidence.

22. A soundtrack is playing inside an app that has no export at all.

Use system audio when the soundtrack is audible on the Mac. Confirm the capture route and required macOS permissions.

23. Someone on a call plays a reference from their end.

The call’s received audio can be an input to analysis. Speech processing, interruptions and competing speakers can affect what reaches it.

24. A radio stream or a recorded set is running with no tracklist.

Check whether Key is listening or holding a result when the program changes tracks. Start another analysis for the new passage when needed.

When the room is waiting

25. A reference lands in a client session and someone asks the key out loud.

Start analysis while a representative passage plays. Check the result and confidence before giving the room a key.

26. An artist hums an idea at the mic before a project is even open.

Select microphone mode for the live voice. A short hum may identify a pitch without establishing the song’s mode or key.

27. A topline writer asks for the key at the top of the second verse.

Analyze the verse itself when its harmony differs. A held reading from an earlier section does not establish the current section’s key.

28. A session player is plugging in and wants the key before the count.

Use a representative demo phrase and confirm the tonal center with an instrument. Do not promise a result within a fixed number of bars.

29. Four people are working on four machines and the beat came from a fifth.

Each compatible Mac can analyze its playback. Use the same passage and settings when comparing results, and respect the three-Mac license allowance.

30. A vocalist wants to know whether the track sits in their range.

A key label is one part of the discussion. The melody’s actual highest and lowest notes determine the range the singer must cover.

31. A mix is playing back for a client and stopping it would break the spell.

Read the menu bar while listening. Test the setup before a client review rather than assuming a first-time permission prompt cannot interrupt it.

32. Someone asks a question in the middle of a take.

Prepare the input before recording and avoid changing project settings during the take. Read the result without sending until a safe point.

33. A band is about to be counted in and the tempo is still a guess.

Estimate the pulse from a steady demo section, then check it against a click. Agree whether the count-in uses half or double time.

34. An A&R plays a comparison record and asks what it would take to match.

Analyze a representative section of the reference. Key and tempo help describe it but do not specify its sound, arrangement or suitability.

35. A writer walks in with a reference and no idea what is in it.

Read the reference, inspect uncertainty and check it musically. One pass is not guaranteed to settle both values.

Getting the answer into the session

36. Pro Tools is open and the session is still at the default tempo.

Enable the optional Beta integration and grant Accessibility. Send tempo and key signature, then confirm the actual Pro Tools session values.

37. Logic Pro is open and the project needs both numbers before tracking starts.

Use the optional Beta integration with Accessibility. Check the project tempo and key signature after the send.

38. Ableton Live 12 is open and the key needs to land somewhere.

Configure the included control surface. Live 12 supports tempo and scale; Live 11 supports the tempo route.

39. FL Studio is open and the same question comes up.

Configure the included controller for the optional Beta tempo send. Use the displayed key separately when setting instruments or scales.

40. The DAW in use is not one of those.

Check the current integration list, which also includes Studio Pro, Bitwig, Cubase, Reason and REAPER. For a MIDI-file route, verify how the destination imports tempo and key data.

41. A tempo gets retyped and one digit lands wrong.

A supported send removes a manual transcription step. Confirm the destination values; removing typing does not guarantee the whole operation succeeded.

42. The DAW is not installed on this machine, so there is nothing to send to.

Use the displayed values and the documented file or clipboard options as appropriate. A missing destination should not be treated as a confirmed send.

43. A project has been running for an hour at the wrong tempo.

Save first and check how the DAW’s clips follow tempo changes. A correct new tempo can still affect existing placement or playback.

When it is the tempo that is the problem

44. A number could be right, or exactly double, or exactly half.

Compare the groove against a click at the candidate, half and double values. Both tapping and automatic analysis can encounter metrical ambiguity.

45. A loop does not land on a round number.

The current Key product lists decimal BPM as optional Beta. Check precision against a long passage rather than assuming more decimal places mean more accuracy.

46. A drummer speeds up through the chorus.

Check different sections and use the DAW’s tempo-mapping workflow if timing must follow the performance. One live value is not a sample-accurate tempo map.

47. Tapping it in gives a slightly different answer every time.

Analyze a longer steady section and verify with a click. Detection also estimates tempo; it is not immune to variation or ambiguity.

48. The groove is heavily swung and the count keeps slipping off.

Count the intended main beat rather than every swung subdivision. Check the click against the phrase before setting project tempo.

49. The intro has no drums in it yet.

Use a section with clearer rhythmic evidence. A drumless passage can still have pulse, but a sparse or free introduction may not support a useful BPM.

50. A track changes tempo partway through.

Analyze the sections separately. Check listening or held state and use a tempo map when the project needs to follow multiple tempos.

51. The tempo is very slow, and it could equally be half of a faster one.

Compare the slower and faster interpretations with the intended subdivision. Choose the pulse the musicians will count.

When the material is ambiguous

52. The audio is pure drums.

Unpitched percussion alone may support BPM without a useful key. Tuned percussion can carry pitch; inspect the actual signal instead of predicting a display color.

53. The audio is speech or a podcast.

Ordinary speech does not establish a musical key in the way a tonal song does. Do not use an analyzer’s key label as evidence that it does.

54. The audio is a riser, a sweep or plain noise.

A sweep can pass through pitches without establishing tonal harmony. Check musical context rather than treating the label as a key for the song.

55. A song modulates halfway through.

Read each section and listen for a change in tonal center. A single whole-file label can hide local changes.

56. Two parts are layered in different keys.

Inspect the parts separately and listen together. A single key label may be an inadequate description of the combination.

57. The part is sub bass and almost nothing else.

Isolate a useful passage and compare against a pitched instrument. A low fundamental with sparse harmonics can be hard to label reliably.

58. A master is loud, dense and fighting itself.

Try a passage or stem with clearer harmonic information. A different confidence display is evidence to investigate, not proof that the new reading is correct.

59. An ambient pad never moves off its root.

Use a tuner to identify the sustained pitch. A drone by itself may leave major or minor context unspecified.

60. Only half a second has played so far.

Let a representative phrase play. The amount of musical evidence matters more than a fixed elapsed time.

61. Somebody wants a number for how accurate the reading was.

A confidence indication is not a measured accuracy percentage. A meaningful accuracy claim needs a test set, reference labels and scoring rule.

When the machine is not on the internet

62. The session is happening on a plane.

Activate and check the license online before travel. The published offline window is up to 21 days after a successful online license check.

63. The studio has no working wifi.

Analysis is described as local, but license state still matters. Confirm the offline window before relying on a disconnected studio.

64. The material is unreleased and the client will not let it leave the room.

Review the current product privacy details and the client’s actual requirements. Local analysis is not a substitute for reviewing every tool in the signal path.

65. A machine is kept off the network on purpose.

Plan for the published 21-day offline license window. A permanently disconnected facility needs to account for the required license checks.

66. A facility blocks outbound traffic as policy.

Check whether the licensing and update routes can be used under the facility’s policy. A local analysis engine does not remove those operational requirements.

67. Service on the bus drops for an hour.

A valid offline license window can cover a temporary interruption. Confirm licensing before leaving reliable connectivity.

68. An agreement forbids sending the audio to a third party.

Read the actual agreement and the product’s current processing details. A website comparison cannot establish compliance with an agreement it has not reviewed.

69. Somebody in the room asks where the audio actually goes.

The product page describes analysis on the Mac. Network services for licensing, updates and support are separate from the analysis path.

When the material belongs to somebody else

70. Remix stems arrive under an embargo.

Analyze only within the permissions for the material. Consult the current product’s processing details rather than assuming any app meets an embargo’s conditions.

71. A beat is being auditioned before any money changes hands.

Analyze the preview as heard. A limited preview may not establish the whole song’s key or tempo.

72. A brief names a reference record and expects the pitch to match.

Check the reference’s actual version, pitch and pulse. A key and BPM pair does not describe every musical requirement in a brief.

73. A sample is going to be sent for clearance.

Analysis can inform a production decision, but it does not grant permission to use a sample. Keep the rights decision separate from the tonal reading.

74. The work is ghost production and cannot be posted anywhere.

Use a workflow consistent with the agreement and inspect product privacy details. Do not infer contract compliance from a key-detection feature.

75. A label needs everyone working from the same numbers.

Agree on the exact file, section and settings, then record the chosen values. Different inputs or versions can produce different readings.

Around the session

76. A whole folder needs sorting by key.

For sorting and tagging a collection, inspect library-oriented tools in the comparison guide. Individual file-drop analysis is a different workflow from batch library management.

77. An old catalog needs checking, track by track.

Key can inspect individual dropped files or playback. For a large collection, compare batch analysis, tagging and correction tools before committing to a manual process.

78. A solo is being transcribed by ear.

Use a candidate tonal center as a starting point, then transcribe the notes and rhythms. A key label is not a transcription.

79. A cover has to be learned before rehearsal.

Check key and tempo against the particular arrangement to be rehearsed. Another recording or live version may differ.

80. A kick is fighting the bass and nobody is sure of the root.

Use an isolated, sustained bass note and a tuner or instrument to identify pitch. A song-key label alone does not tell you the note of each kick or bass hit.

81. Toms need tuning to the song.

Choose the desired notes in the arrangement, then tune and listen to the drums. A key label does not prescribe one tuning for every tom.

82. A filter is self-oscillating and the pitch needs naming.

A tuner is the appropriate tool for the isolated oscillation. Identifying one frequency is different from detecting the key of a musical passage.

83. Two records are being checked before they get mashed together.

Compare candidate keys and tempos, then audition the transition. Compatible labels do not guarantee that melodies, chords and timing will fit together.

84. A cue has to match a temp track.

Analyze the relevant section of the temp and check the picture’s timing requirements. One BPM value does not establish synchronization to picture.

85. A guitar is tuned down a half step.

Compare sounding pitch with the chart and the instrument’s tuning. Analyze the actual performance rather than assuming the written fingering names the sounding key.

86. A vocal sample is going under a new chord progression.

Check the vocal melody against the proposed chords. A detected key is a useful candidate context, not a guarantee every note fits the progression.

87. A project cannot be opened because the DAW that made it is not installed.

Use the surviving bounce as an input. Analysis cannot restore missing edits, stems or project settings.

System requirements

88. The studio machine is an Intel Mac.

The published Key requirement is Apple silicon. Playback has different hardware requirements; do not infer compatibility between the products.

89. The Mac is on an older version of macOS.

The published Key requirement is macOS 14.4 or later. Check hardware and other software compatibility before updating a studio system.

90. The collaborator works on Windows.

Key’s current published desktop requirement is macOS on Apple silicon. Use the comparison guide to inspect tools for another operating system.

91. The work moves between a laptop at home and a machine at the studio.

One Key license covers up to three Macs under the current offer. Check activation and license state on the machines used for the session.

Compare the workflow

Key and BPM tool comparisons