How to Practice Singing

Most singers practice in a way that feels productive but isn't. They run through songs, hit the hard part, back up, run through it again. This article is about why that approach limits improvement and what to do instead — grounded in voice science and applied vocal pedagogy for contemporary commercial music (CCM): pop, rock, R&B, and musical theater.

A foundational insight from K. Anders Ericsson's research on expert performance is that not all practice is equal. His 1993 study of violinists and pianists — still widely cited, though its "10,000 hours" oversimplification became notorious — found that what distinguishes top performers is deliberate practice: work that is (1) targeted at a specific skill gap, (2) at the edge of current ability, and (3) paired with immediate, accurate feedback.

Running through a song for the third time in a row fails on all three counts. You're repeating things you can already do, coasting on musical memory rather than exercising underlying technique, and your feedback (how you feel while singing) is unreliable — the sound you hear inside your head is significantly filtered by bone conduction and rarely matches what a listener hears.

This doesn't mean abandoning repertoire. It means most of your improvement time should come from exercises that target specific vocal coordination, not from song run-throughs.

There is no single right answer, but some useful constraints from the research:

Ericsson's research on musicians found that the best performers practiced in focused blocks of no more than about 90 minutes before taking breaks, and that benefits declined significantly above roughly four hours per day of deliberate practice. For concentration-heavy domains more broadly, effective focused work tends to be closer to one to two hours before quality degrades — though this is a rough pattern across studies, not a precise prescription from any single experiment. For voice specifically, the cap matters more than in most other instruments — the vocal folds are mucous-membrane tissue, not muscle, and overuse causes swelling that degrades the very coordination you're trying to train.

A more practical model for singers at most levels: 15 to 30 minutes of focused vocal exercise, four to five days per week, supplemented by song work. Consistent shorter sessions accumulate more learning than occasional marathon sessions. The cognitive science literature on distributed ("spaced") practice is consistent across many motor and cognitive skills on this point; voice-specific RCT evidence is thin, but the general principle is well-supported and widely applied in applied pedagogy.

One caveat: voice science doesn't have a clean RCT on daily singing duration for CCM singers specifically, so these numbers are informed estimates from applied pedagogy rather than controlled study results.

A well-organized practice session has three phases, each doing different physiological work.

The goal is to lower phonation threshold — the minimum air pressure and muscular effort needed to sustain vibration — before demanding anything of your voice. Research supports semi-occluded vocal tract exercises (SOVTs) as the most efficient warm-up tool.

SOVT means any exercise that partially closes the front of your mouth: lip trills, tongue trills, humming, straw phonation, or "ng" hums. Ingo Titze's 2006 paper in the Journal of Speech, Language, and Hearing Research provided the physical rationale: a partial occlusion at the lips or teeth raises the air pressure just above the vocal folds, which allows them to vibrate with less mechanical stress and at lower effort levels. A 2021 single-subject longitudinal case study by Castillo-Allendes, Cantor-Cutiva, and Hunter found measurable acoustic improvements (decreased jitter and shimmer, improved harmonics-to-noise ratio) after water-resistance straw phonation used as a one-minute warm-up — improvements that persisted over seven weeks of daily measurement. Because this was an n=1 study, it should be read as preliminary evidence that SOVT warm-ups produce real acoustic changes in individual voice users, not as a large-scale population finding.

The practical recipe: two to three minutes of lip trills or "ng" hums ascending and descending through your comfortable range. Keep everything light. Save your louder, more driven sounds for after the voice is warm.

This is the deliberate-practice core. Pick one or two specific coordination problems to work on — not "get better in general." Examples:

Different methods approach the passaggio differently: classical pedagogy and the Speech Level Singing (SLS) lineage (developed by Seth Riggs, focused on seamless register connection without laryngeal lifting) tend to emphasize vowel modification as you ascend — narrowing /iː/ toward /ɪ/, /ɑː/ toward /ʌ/. Complete Vocal Technique (CVT, Cathrine Sadolin's system, which organizes production by vocal mode rather than register) and Somatic Voicework (a sensation-based approach that de-emphasizes anatomical vocabulary in favor of kinesthetic feedback) focus more on register blending through resonance quality. Estill Voice Training teaches specific laryngeal configurations as separable figures — thyroid tilt, aryepiglottic narrowing, and so on — and builds voice qualities from combinations of those figures. All of these are real approaches with their own evidence bases, not competing superstitions — they're different leverage points for the same coordination challenge. If one doesn't click, try another.

One technique goal per session is a useful discipline. More than that and attention fragments, feedback gets murky, and the session devolves into the same kind of unfocused repetition you started with.

End with repertoire — but with one specific intention carried over from Phase 2. If you worked on breath support, apply it in one difficult phrase. If you worked on the passaggio, pick a chorus that crosses it. This transfers the isolated coordination into context without abandoning everything you just built.

Your internal experience of your singing is consistently inaccurate — not because you have a bad ear, but because bone conduction (vibration transmitted through your skull) adds significant coloration to what you hear from the inside. What you hear from outside — what everyone else hears — comes only through air.

Recording yourself, even with a phone voice memo, gives you access to the external signal. Play it back immediately after a take. The gap between what you thought you sang and what you hear on playback is informative diagnostic data, not cause for shame. (Research on self-voice perception confirms this discrepancy is real and consistent — bone conduction is part of a genuinely multi-modal experience of your own voice that recordings strip away.)

This practice is consistent across methods and applies at every level. What matters is reviewing the recording with a specific question ("Did the phrase end flat? Did the vowel shift between chest and head?") rather than listening generically for "how it sounds."

This is worth stating plainly, because it affects how you use online resources and who you study with.

Vowel modification: Classical pedagogy and much of the SLS lineage teaches that vowels should narrow as pitch ascends — /æ/ toward /ʌ/, /i/ toward /ɪ/. Some CCM belt-focused approaches prefer keeping vowels brighter and more open higher up, managing acoustic resonance differently. Neither position is wrong; they work with different aesthetic targets and resonance strategies.

Laryngeal height: Classical training often explicitly trains a low larynx. In CCM, the more common goal is a neutral larynx — not deliberately lowered, not pulled up from tension. Singers imitating operatic sounds into a pop context often carry a low larynx that muddies their upper range.

How much anatomy to teach: Estill Voice Training and CVT include significant biomechanical vocabulary (thyroid tilt, aryepiglottic narrowing, etc.). SLS and some Somatic Voicework approaches emphasize sensation cues and acoustic targets without the anatomy. Both can produce results; the question is what kind of learner you are.

Knowing that methods disagree — and why — helps you assess advice you encounter rather than taking any one coach's approach as the only possible truth.

The Five-Note Scale (Mee May Mah) is a standard entry point in CCM vocal pedagogy. It steps up and back across scale degrees 1–2–3–4–5–4–3–2–1 on the Italian vowel series: mee–may–mah–moh–moo–moh–mah–may–mee. The exercise runs at 88 bpm in eighth notes, with the piano doubling the melody as a reference.

What it trains: This is classified under resonance — the goal is a consistent tonal core that doesn't shift dramatically between vowels. The "m" onset at each syllable encourages cord closure before the vowel opens, which produces a cleaner, more centered tone than starting directly on a vowel. Rotating through five vowels on a single phrase trains you to maintain resonance consistency across different tongue and lip positions.

What to listen for: Does the "mah" at the top of the scale feel and sound as resonant as the "mee" at the bottom? Does the "moh" drift flat or breathy? Does the phrase shape remain even going back down? Any vowel that consistently drops out is a specific coordination gap to bring into Phase 2 of your next session.

The exercise moves up a half step per iteration, covering your voice's range systematically. Vocal Habit tracks your pitch accuracy after each key.

<!-- EMBEDDED EXERCISE: five-note-scale-mee-may-mah -->

---

Consistent short sessions beat occasional long ones. For most adult learners, four to five focused sessions per week of 20–30 minutes each builds skill more efficiently than one or two long sessions. The voice responds to regular stimulus followed by rest — the vocal fold tissue needs recovery time, not just accumulated volume.

It depends on how you practice. Light warm-ups and moderate technique work every day are generally fine and build habit. Loud, high-effort singing — belt rehearsals, sustained high notes, heavy chest — every day without rest days creates cumulative fatigue. Build in at least one or two lighter days per week, and listen to your voice. A voice that sounds rough or feels effortful at the start of a session likely needs recovery, not more pushing.

Yes, with some caveats. Exercises with built-in constraints — SOVT exercises, patterns that step through the passaggio in small intervals — teach coordination without requiring a teacher to be in the room, because they regulate pressure and register transitions automatically to some extent. What's harder to do without a teacher is diagnose why something isn't working; a single session with an experienced teacher can often identify a pattern that a singer missed for months of self-study.

Start with semi-occluded exercises: lip trills, an "ng" hum, or humming on a scale. These exercises lower phonation threshold and warm the folds with less mechanical impact than open-vowel singing. Two to five minutes is usually enough. Save louder, more driven sounds for after.

Bone conduction — vibration through your skull — adds coloration to the sound you hear internally. External listeners (and recordings) only hear the air-conducted signal, which sounds different. This gap is normal and not a sign of a bad voice. Recording yourself regularly and listening critically is one of the most efficient ways to close the gap between your internal experience and your actual output.

Yes. Vocal fold tissue responds to hydration, sleep quality, allergies, reflux, hormone changes, and how much you spoke or sang the day before. Day-to-day variation is not a meaningful measure of whether you're improving. Trend over weeks and months is the right unit of measurement.

If hoarseness or significant vocal changes persist for four weeks or more, the 2018 AAO-HNS Clinical Practice Guideline recommends seeing a laryngologist — an ENT specialist who can scope the vocal folds — rather than waiting it out or working around the symptom. See a doctor sooner if you have throat pain, difficulty swallowing, a neck mass, or a history of tobacco use.

---