Scale recognition training is practice at hearing a series of notes played in order (a scale) and identifying which kind of scale it is by ear alone. Where a chord is a stack of notes sounding at once, a scale is a line of notes moving in sequence — you hear the very arrangement of steps that gives a scale its bright, dark, or tense character.
This page is a manual for choosing the settings. Each setting is designed around established methods from music education. For each setting that calls for a decision, we give the recommended choice, what it is for, and what it trains — with the evidence behind it. If you arrived from a "?" in the settings screen, scroll to the option you need.
Training Mode
Normal is enough to start. Use Custom only to target specific scales, and reach for "Review When You Forget" and "Weak-Point Focus" once your results accumulate.
This setting decides the policy by which scales are chosen for you. There are four modes.
The other two — "Review When You Forget" and "Weak-Point Focus" — pay off once your performance data accumulates. They work the same way across every training, so they appear inline below.
"Review When You Forget" mode (spaced repetition)
Spaced repetition means an item you have already studied comes back just as you are about to forget it. The more reliably you answer something, the longer the gap before it returns; the items you miss come back sooner. It rests on the idea that recall is more durable when you space the reviews out rather than cramming the same item back-to-back.
This mode suits you once the pool of items has grown and you start noticing that things you once knew have slipped. The app handles the scheduling automatically, so you do not have to track what needs reviewing. It rewards short, daily sessions more than occasional long ones.
When in doubt, reach for it in the "maintenance" phase, after you have worked through the items once. Use the normal mode to build the foundation when everything is new, then let spaced repetition move that knowledge into long-term memory — a comfortable two-step approach. The benefit of spacing reviews for long-term retention was synthesized at scale by Cepeda et al. (2006), and the finding that the act of retrieval itself strengthens memory comes from Roediger & Karpicke (2006).
Weak-Point Focus Mode
Weak-point focus looks at your answer history, identifies the items with the lowest accuracy, and serves those to you more often. Instead of spending time on what you already answer reliably, it concentrates practice on exactly the places you keep missing. The app picks the targets automatically from your own results.
This mode suits the stage where your overall accuracy is climbing but a specific few items still trip you up every time. Because it keeps returning to your weak spots, it feels harder — and that controlled extra effort is precisely where the learning happens (the principle of desirable difficulties). It needs enough answer history before candidates appear, so play the normal mode for a while first.
When in doubt, drop in for a few concentrated sessions whenever you plateau at a given level and feel you "keep losing the same handful of items right before it would stabilize." Once the weak spots clear, return to the normal mode to keep your overall balance. Deliberately making yourself retrieve the items you tend to miss — the "testing effect" — was shown by Roediger & Karpicke (2006) to strongly benefit later retention.
The four modes exist to concentrate practice on the scales you currently need. The quiz format itself — recalling the answer rather than listening passively — strengthens memory more than re-listening (the testing effect). "Review When You Forget" and "Weak-Point Focus" automate this using your performance data.
Ref.: Roediger & Karpicke (2006) [1]
Difficulty
When unsure, start at Beginner. Step up one level once you reliably hold 80%+ at the current one.
Sets the range of scale types you'll be asked about (shown in Normal mode). Following an order also used at music conservatories, the pool widens step by step: Major/Minor → Modes (church modes) → Pentatonic/Blues → all scales.
The fewer scales you handle at once, the less your ear must process, and the more reliably you fix the cues for telling them apart. Securing a narrow pool before widening it gets you there faster than tackling everything at once. The tiers follow cognitive-load theory: keeping processing demands low early in a new skill aids learning.
Ref.: Sweller (1988) [2]
Modes (Church Modes)
Don't learn all seven modes at once — in Custom mode pick just two or three (say "Major + Dorian"), starting with combinations whose differences are clearly audible.
An explanation of the church modes asked when you choose Modes under difficulty (the seven scales formed by changing which note of the major scale you start counting from). The notes are the same, but starting elsewhere rearranges the steps (the whole/half-step pattern), giving each its own color. Ionian is the major scale, Aeolian is the natural minor, and Dorian, Phrygian, Lydian, and Mixolydian each have one easy "tell-tale note" to latch onto.
Telling modes apart is the skill of hearing the color produced by the arrangement of steps, not the notes themselves. Getting a few sharply different modes into your ear at a time makes each mode's tell-tale note easier to grasp. This follows an established approach in college-level ear training, where modal and scalar listening is built up in stages.
Ref.: Karpinski (2000) [3]
Melodic Minor Style
Leave it on the default, Jazz. Choose Classical only to match classical theory or exams, or to train your ear on the traditional motion where ascent and descent differ.
Decides how the melodic minor scale is treated on its descent (coming down from high to low). It appears only when melodic minor is in the question pool, and matters most when the playback mode includes a descent (Descending, or Both).
This setting lets you handle by ear the convention by which, in real music, the melodic minor changes shape between ascent and descent. Practicing the descent as distinct builds the ability to hear the motion correctly where notes swap between going up and coming down. Building melodic listening that includes ascending-versus-descending differences in stages is an established method in college-level ear training.
Ref.: Karpinski (2000) [3]
Playback Mode
Learn the shape in Ascending first, add Descending and Both as you get comfortable, and finish with Random.
Decides in which direction the scale is played. The same scale feels different to identify depending on the direction.
The aim is to hear the scale itself without leaning on one direction. Real music uses both ascending and descending lines. Varying the direction stops you depending on a single cue (like an ascending "song opening") and builds the ability to hear the step pattern itself. This follows the established ear-training principle of cultivating listening that includes melodic direction.
Ref.: Karpinski (2000) [3]
Playback Speed
Leave it on Normal. Drop to Slow only when a new scale trips you up; move to Fast once stable.
Decides how fast each note of the scale plays. Slow, Normal, or Fast: Slow lets you trace the half- and whole-step pattern one note at a time, while Fast approaches real-playing flow and trains snap judgment of a scale's shape.
Speed is adjustable so you can match the processing load to your stage of learning. While a new scale is unfamiliar, Slow lowers the load; once discrimination is stable, Fast trains instant judgment. Keeping the load lower early follows cognitive-load theory.
Ref.: Sweller (1988) [2]
Root Note (Key)
Use Random across all keys (the app's recommendation). An absolute beginner who wants to isolate the scale-type difference may temporarily Fix the root.
Decides how the scale's root (the reference note the scale starts from — that is, its key) is chosen each time. You pick "Random across all keys" or "Fixed root."
Varying the key exists so you can hear the scale's step pattern itself rather than leaning on one pitch height. Mixing different keys (interleaving) yields better long-term retention than blocking the same condition, as meta-analysis shows. A beginner fixing the root temporarily is the exception — leveling conditions only at the start to ease comparison.
Ref.: Brunmair & Richter (2019) [4]
- Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255.
- Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257–285.
- Karpinski, G. S. (2000). Aural Skills Acquisition: The Development of Listening, Reading, and Performing Skills in College-Level Musicians. Oxford University Press.
- Brunmair, M., & Richter, T. (2019). Similarity matters: A meta-analysis of interleaved learning and its moderators. Psychological Bulletin, 145(11), 1029–1052.