Source-linked AI summary
Rhythm Is a Dancer: Designing Interactive Rhythm Feedback for Beginner Dancers
Bettina Eska, Annika Kilian, Paweł W. Woźniak, Jakob Karolus
TL;DR
Beginner dance practice can be difficult, while existing interactive systems do not always support learners’ progress. SkeletonDance uses teacher-informed rhythm feedback, and participants reported regaining rhythm and feeling more confident, though objective accuracy did not show a significant feedback effect.
Problem
Dance is complex to learn, and existing interactive systems may not sufficiently support learners’ progress during practice.
Source: §1 · p. 2Method
Results
Participants reported that SkeletonDance helped them regain rhythm and feel more secure, while objective rhythm accuracy showed no significant effect of feedback condition.
Source: §1 · p. 2Takeaways & Limitations
The findings indicate that rhythm feedback can support novice dancers’ subjective practice experiences, with prior dancing experience potentially affecting objective effectiveness.
Source: §1 · p. 2, §8 · p. 25The limited study duration prevented the authors from ascertaining quantitative improvement; evaluating changes over time would require longitudinal study.
Source: §8.4 · p. 26
Abstract
from arXiv · showhide
Learning how to dance can readily overwhelm beginners, especially without effective guidance from a dance teacher. Existing interactive systems often do not sufficiently support the learner's progress. We investigated how targeted feedback on rhythm keeping interactively supports dance practice for novice dancers by introducing SkeletonDance. Our design is grounded in motor learning theory and conceptualized through interviews with dance teachers, following established teaching strategies. SkeletonDance automatically detects rhythm flaws and provides assistance through mimicking clapping feedback, a common instructional technique in dance lessons. In our study, participants reported that SkeletonDance helped them to re-establish lost rhythm and increased confidence during practice, especially among novices. Though objective performance metrics did not consistently confirm these effects during controlled test sessions. Our work highlights that feedback can support novice dancers' subjective practicing experiences and demonstrates how prior dancing experience moderates the objective effectiveness of such minimal, teacher-inspired interventions.
1 Introduction
Dance’s complexity makes effective learning support difficult, while score-based systems and demanding setups leave gaps for learners seeking to practice. SkeletonDance uses teacher-informed, rhythm-focused clapping feedback; participants reported feeling more secure and regaining rhythm, although objective accuracy showed no significant feedback effect.
1 Introduction
- Dance’s complexity makes it challenging to design interactive systems that effectively support learning.Dance requires coordinated motion, extensive training, and physiological development for most users.Source: §1 · p. 2
- Score-based dance-game feedback often does not explain mistakes, making it harder for learners to reason about errors and improvements.Such feedback may work in party-game settings but has limited application for dancing as a hobby or outside competition.Source: §1 · p. 2
- Existing dance feedback systems can require complex setups, so this work targets minimal-setup feedback for novice dancers.Prior systems have used full-body motion tracking.Source: §1 · p. 2
- Participants reported that clapping feedback helped them regain rhythm and feel more secure, but objective rhythm accuracy showed no significant effect of feedback condition.Differences in prior dancing experience may help explain the objective results.Source: §1 · p. 2
- SkeletonDance combines motor-learning theory and dance-teacher interviews to design a system supporting rhythm practice through detection and feedback.Its camera-based system overlays a clapping sound when the dancer is out of sync, mimicking a common teaching technique.Source: §1 · pp. 2–3
2 Related Work
Dance-support systems span entertainment games, education tools, and sensor-based feedback, but supporting reflection, bodily experience, and accessible practice remains challenging.
2 Related Work
- Dance is expressive and complex, making it difficult to identify features that capture its expressive qualities.Personal interpretations of music and intertwined movements add to the challenge for interactive support.Source: §2 · p. 3
2.1 Interactive Dance Systems and Feedback
- Dance-learning systems differ in their feedback strategies and learning goals, spanning commercial products and research prototypes.The broader design challenge is balancing performance insight with clear feedback that supports constructive reflection and bodily understanding.Source: §2.1 · p. 3, §2.3 · p. 4
2.1.1 Dance Games.
- Commercial dance games commonly score timing or movement accuracy but offer coarse feedback that does not explain errors or support reflection.Learners must infer whether mistakes arise from timing or movement inaccuracies.Source: §2.1.1 · p. 3
2.1.2 Dance Education Systems.
- Dance-education systems address technique through tools such as color-coded positions and movement-execution evaluations, while teachers use demonstrations, verbal or gestural explanations, and touch.Experts often provide augmented feedback on small chunks of movement.Source: §2.1.2 · p. 3
2.2 Sensing Technology and Dancing
- Dance-support research uses motion tracking and varied technologies, including IMUs, wearables, VR, AR, and smartphone applications, to generate feedback from deviations from expert performance.Rhythm systems range from retrospective analyses to real-time beat emphasis, though sensing limitations can reduce accuracy.Source: §2.2 · pp. 3–4
2.3 Towards a Deeper Understanding of Dancing and Bodily Experiences
- Dance feedback should support proprioception and bodily understanding as well as technical proficiency, while delivering appropriate stimuli at the right time.The paper frames teacher-like feedback as a design goal.Source: §2.3 · p. 4
3 Methodology: From Theory to Practice
The methodology uses motor-learning theory as a design lens, not as an outcome to validate, and centers evaluation on dancers’ subjective experiences alongside rhythm measures.
3 Methodology: From Theory to Practice
- The study uses motor-learning theory to guide design rather than empirically validate the theories, focusing evaluation on dancers’ subjective experiences.The authors note that short-term quantitative metrics can be difficult to use for capturing gradual or implicit changes in dance learning.Source: §3 · p. 4
- The research questions address design requirements for beginner-dancer feedback systems and whether SkeletonDance supports rhythm keeping and body perception.The proposed methods include theory review, teacher interviews, rhythmic-accuracy analysis, questionnaires, and post-study interviews.Source: §3 · pp. 4–5
- The methodology combines motor-learning theory with interviews of experienced dance teachers to derive design requirements and inform SkeletonDance.Figure 2 presents this theory-and-practice basis for the system design.Source: Figure 2 · p. 5, §3 · p. 5
- The evaluation compares video-only practice with video plus clapping feedback within each participant, measuring rhythmic accuracy and gathering questionnaires and interview data.Questionnaires and post-hoc interviews also target feedback and its effect on proprioception.Source: §3 · p. 5
4 Theory: Motor Learning and Dancing
Motor-learning theory informs the paper’s design choices: practice and feedback support skill development, while schema theory, attention, and task decomposition guide feedback for dance.
4.1 Phases of Motor Learning
- Motor skills develop through cognitive, associative, and autonomous phases, and feedback is a critical component of skill improvement.Learners move from an error-prone general understanding toward refinement and more automatic performance.Source: §4.1 · p. 6
4.2 The Schema Theory - Generalized Motor Programs
- Recall and recognition schemas connect movement specifications with outcomes and sensory consequences, and develop through repetition and variability.The schemas govern familiar movements and help facilitate learning novel ones.Source: §4.2 · p. 6
4.3 Conceptualizing Feedback
- Feedback may communicate knowledge of results about movement outcomes or knowledge of performance about movement patterns.The two forms can coexist and may be difficult to distinguish when correct movement is the goal.Source: §4.3 · p. 6
4.4 Attentional Focus Impacts Motor Learning
- An external attentional focus emphasizes a movement’s intended effect, and cited research reports stronger learning and movement effectiveness than with internal focus.Internal focus directs attention to one’s body movements.Source: §4.4 · pp. 6–7
4.5 Improving Subskills Contributes to Overall Skill
- Breaking dance into smaller subskills lets novices focus on one aspect at a time before combining it with more complex tasks.Practiced subskills can transfer to similar tasks after being internalized.Source: §4.5 · p. 7
4.6 Summary
- The paper’s theory synthesis motivates repetition and appropriate feedback for schemas, external focus, and identifying dance subskills suitable for interactive systems.It treats motor-learning theory as guidance for design decisions.Source: §4 · p. 6, §4.6 · p. 7
5 Practice: Exploring Opportunities for Interactive Systems to Support Dancing Practice
Interviews with dance professionals explored how teaching practice and motor learning theory could guide interactive support for dance practice. Teachers emphasized rhythm and coordination, practical teaching strategies, and feedback that adapts to learners while remaining supplementary to real lessons.
5 Practice: Exploring Opportunities for Interactive Systems to Support Dancing Practice
- The inquiry focused on instructors’ approaches to timing and rhythm, connecting interview insights to motor learning theory and the social context of dance practice.The authors conducted interviews with five dance professionals to study pedagogical strategies and complementary tools.Source: §5 · p. 7
5.1 Procedure
- Five dance professionals were interviewed about teaching strategies, classroom challenges, digital tools, sensors, and useful feedback.The semi-structured online interviews had five phases and covered four dance styles; four researchers used iterative thematic analysis to develop five themes.Source: §5.1 · pp. 7–8, §5.2 · p. 8, §5.3 · p. 8
5.3 Findings
- Teachers linked dance performance to both movement technique and timing, and commonly teach rhythm by breaking sequences down and counting beats.They also use simple movements such as bouncing, walking, or stomping to help students embody the beat.Source: §5.3 · pp. 8–9
- Teachers attributed timing problems more to motor-skill and coordination difficulties than to a lack of rhythm sense, and favored immediate, positive feedback.They said early correction prevents error consolidation and supports motivation.Source: §5.3 · p. 9
- Teachers proposed visual, haptic, and auditory rhythm cues, including lights for synchronization, beat-pulsing wearables, and clapping or counting.One teacher cautioned that a vibrating floor could distract, especially for older students.Source: §5.3 · p. 9
- Teachers viewed technological systems as supplementary to real lessons, warning that unsupervised practice may risk injury or overreliance on the system.They noted that users might skip warm-up and stretching routines.Source: §5.3 · p. 9
5.4 Implications and Connection to Motor Learning Theory
- The authors connected teachers’ immediate feedback to schema consolidation, clapping or counting to external focus, and feedback frequency to learners’ motor-learning phases.They also framed rhythm as a subskill that teachers build through whole-sequence demonstrations followed by slower, music-free practice of parts.Source: §5.4 · p. 10
6 SkeletonDance - Supporting Rhythm Keeping Through Interactive Feedback
SkeletonDance combines dance tutorials with camera-based rhythm detection and teacher-inspired clapping feedback. Its design targets beginner practice with a lightweight setup and feedback that activates when rhythm irregularities are detected.
6 SkeletonDance - Supporting Rhythm Keeping Through Interactive Feedback
- SkeletonDance combines video instruction with clapping feedback, using auditory cues to support rhythm practice without adding visual-channel clutter.Its components resemble real-world dance classes or video tutorials and draw on teacher interviews and motor learning theory.Source: §6 · p. 10, §6.1 · p. 10
6.1 Design of SkeletonDance
- The design requirements call for immediate, teacher-like feedback focused on rhythm, adapted to learner progress, and supported by a simple webcam-and-speaker setup.Learners can practice at their own pace and repeat exercises; feedback activates only when needed.Source: §6.1 · pp. 10–11
6.2 Practicing with SkeletonDance
- Practice uses a video tutorial that teaches movements gradually, counts beats without music, and adds music at the end, followed by a 15-minute practice session and an unsupported rhythm test.Users can revisit tutorial sections during practice.Source: §6.2 · p. 11
6.3 System Implementation
- The system connects webcam pose detection, feedback assessment, and a graphical interface over a local network to support solo practice.It uses an off-the-shelf webcam and streams skeleton data with time synchronization; no webcam images are saved or further processed.Source: §6.3 · p. 11, §6.3.1 · pp. 11–12
6.3.1 User Pose Detection.
- For beginner-focused sequences, the system uses lightweight movement detection rather than an elaborate algorithm for many dance movements.A webcam and pose estimator track a 33-landmark skeleton at 20 frames per second.Source: §6.3.1 · p. 11, §6.3.2 · p. 12
6.3.2 Algorithm for In-Beat Detection.
- For each bar, the detector sums hip-to-knee and knee-to-ankle distances into Δleg and checks whether its highest peak aligns with the bar’s start or end.The processing window is shifted by half a bar to elicit one peak response per window.Source: §6.3.2 · pp. 12–13, Figure 5 · p. 12
6.3.3 Technical Feasibility Assessment.
- Preliminary trials with three users performing on-beat, early, late, and off-beat movements checked whether detections matched intended timing deviations.The authors state that perfect or exhaustive detection accuracy was not their goal for a supportive system.Source: §6.3.3 · p. 13
6.3.4 Integrating Feedback for the Dancer.
- Clapping activates after three consecutive out-of-rhythm bars and deactivates analogously; the pipeline combines pose estimation, Δleg calculation, and feedback classification.A voting algorithm was used to smooth the feedback experience.Source: §6.3.4 · p. 13, Figure 6 · p. 13
7 Evaluation
The study compared video-led dance practice with and without SkeletonDance’s clapping feedback, then tested performance without instruction or feedback. A mirror, screen, music playback, and webcam supported the practice setup.
7.1 Study Design
- Participants practiced two comparable, simple leg-movement sequences, one with video and clapping feedback and one with video alone, in counterbalanced order.Each sequence included three similar moves; participants tested each sequence without feedback after practice.Source: §7.1 · p. 14
- The test removed both video instruction and feedback, measuring performance after supported or baseline practice rather than during assisted dancing.Participants performed the practiced sequence through one playback of the song.Source: Figure 7 · p. 14, §7.1 · p. 14
7.2 Apparatus
- Participants followed a video tutorial using a mirror for self-observation, while music played and a webcam recorded their movement.The camera recording was converted to stick figures for movement analysis; images and videos were not stored.Source: §7.2 · pp. 14–15, Figure 8 · p. 15
7.3 Procedure
- Before practice, participants completed a rhythm test; each then had 15 minutes to practice each sequence before the unsupported test and questionnaires.The study counterbalanced conditions and sequences, collected questionnaires after each round, and concluded with a semi-structured interview.Source: §7.3 · pp. 15–16
- Custom questions rated perceived timing and rhythm on a 0–100 visual analog scale from disagreement to agreement.The questions were administered after each control modality.Source: Table 1 · p. 16, §7.3 · p. 16
7.4 Metrics and Analysis
The evaluation combined camera-derived dance-performance measures with questionnaires and interviews. Analyses accounted for dancing experience as a moderator and used mixed-effects models for the quantitative measures.
7.4.1 Objective Metrics of Dance Performance.
- Rhythmic accuracy, rhythm-loss count, and mean out-of-rhythm duration captured performance during unsupported tests after practice.Rhythmic accuracy was the share of in-rhythm bars; rhythm-loss count excluded consecutive out-of-rhythm bars, and duration measured bars until recovery.Source: §7.4.1 · p. 16
- Feedback duration measured how long clapping was active during the feedback practice session, assessing system functionality rather than dance performance.It was available for exactly one training session per participant.Source: §7.4.1 · p. 17
- Each objective metric was modeled with condition and dancing experience as fixed effects and participant as a random effect; significant effects were reported.The authors checked model assumptions and used Tukey-adjusted p-values for multiple comparisons.Source: §7.4.1 · p. 17
7.4.2 Subjective Metrics.
- Questionnaires assessed perceived rhythmic support and standardized constructs, while semi-structured interviews elicited detailed experiences for thematic analysis.Measures included workload, flow, self-appraised performance, and perceived creepiness.Source: §7.4.2 · p. 17
7.5 Participants
The analysis used data from 24 participants recruited for little or no dancing experience, after excluding two who did not complete the study.
7.5 Participants
- Twenty-four participants were analyzed; their reported previous dancing experience averaged 24.42 (SD 16.77), and they were not familiar with the dance sequence beforehand (familiarity mean 25.2, SD 32.3).The group comprised 14 men, 9 women, and 1 diverse participant; mean age was 25.58 years (SD 2.65).Source: §7.5 · p. 17
7.6 Results
Objective test results varied with dancing experience: feedback favored lower-experience participants on rhythmic accuracy, while condition effects were not consistently significant. Participants nevertheless described rhythm recovery and support, alongside requests for more explicit guidance.
7.6.1 Objective Metrics.
- 60.7% versus 59.8% mean rhythmic accuracy favored feedback over baseline descriptively, but condition had no significant main effect and the experience interaction was significant.The feedback advantage appeared among participants with lower dancing experience; the predicted condition lines crossed as experience increased.Source: §7.6.1 · p. 18, Figure 9 · p. 18
- Higher dancing experience was associated with fewer rhythm losses, while no other rhythm-loss effects were significant; the low-experience feedback pattern reversed as experience increased.The experience effect was significant, but the condition-by-experience pattern was not.Source: §7.6.1 · pp. 18–19, Figure 10 · p. 19
- Mean out-of-rhythm duration was 3.53 bars in baseline and 2.42 bars with feedback, but condition, experience, and their interaction were not significant.Duration decreased slightly with experience in both conditions, consistent with quicker rhythm recovery.Source: §7.6.1 · p. 19, Figure 11 · p. 20
- Clapping feedback averaged 33.5 seconds and tended to run longer for less-experienced dancers, though no effects were statistically significant.The duration measure assessed feedback activation during practice, not test performance.Source: §7.6.1 · p. 20, Figure 12 · p. 21
7.6.2 Questionnaires.
- Participants rated feedback higher for noticing the system and using it to regain rhythm, with significant condition effects on both questions.Feedback also scored higher descriptively on most custom items, and participants generally rated it positively.Source: §7.6.2 · pp. 20–21, Figure 13 · p. 22
- Workload ratings did not differ by condition, and decreased with experience; self-appraisal rose with experience, while creepiness and flow showed no significant effects.Self-appraisal was descriptively higher under feedback; creepiness was slightly higher but nonsignificant.Source: §7.6.2 · p. 21
7.6.3 Interviews.
- Participants described clapping as helping them hear the beat, regain rhythm, and feel more confident, although some found its link to movement unclear.Some reported that it subtly shifted attention toward the beat or influenced timing without conscious use.Source: §7.6.3 · p. 23
- Interviewees also described novice challenges and requested clearer, richer guidance, including movement cues and explicit confirmation of whether they were on beat.They cited difficulty combining movement with music, remembering steps while keeping rhythm, and suggested remote practice as a useful context.Source: §7.6.3 · p. 24
8 Discussion
Participants reported that clapping feedback helped them recover rhythm and feel safer, while objective usefulness varied with dancing experience. The discussion frames SkeletonDance as a novice-oriented, motor-learning-informed design, with findings bounded by the study’s duration and two-sequence evaluation.
8.1 SkeletonDance’s Clapping Feedback Supports Rhythm During Practice
- Beginners reported developing a better feeling for rhythm, using clapping to recover after losing the beat, and feeling safer during practice.Participants said the beat became more prominent; the objective recovery-time analysis showed no significant effects, although descriptive analysis suggested faster recovery with feedback.Source: §8.1 · p. 25, §8 · p. 25
8.2 Dancing Experience Moderates the Usefulness of Feedback
- Feedback benefited less-experienced dancers most, while greater experience aligned with stronger baseline rhythm accuracy, lower workload, and higher self-appraisal.The feedback advantage diminished as experience increased; the authors note it may have been redundant or interfering for more experienced participants.Source: §8.2 · p. 25
8.3 Motor Learning Theory as a Design Lens
- The authors use motor learning theory as an interpretive lens, linking immediate, rhythm-focused, momentarily adaptive feedback and lightweight access to supportive early-practice conditions.They did not aim to empirically evaluate motor learning theories.Source: §8.3 · pp. 25–26
8.4 Limitations
- The evaluation used only two dance sequences, so its findings are bounded by a small sequence sample.The sequences were teacher-approved and considered exemplary; transfer of established rhythm to other dances is presented as a theoretical expectation.Source: §8.4 · p. 26
- The limited study duration prevented the authors from ascertaining quantitative improvement; they identify longitudinal evaluation as necessary to track expertise and learning trajectories.Participants nevertheless reported actively using the clapping feedback and feeling safer and more confident.Source: §8.4 · p. 26
- The minimal, rhythm-focused design prioritizes novice access over long-term mastery, although some participants wanted more explicit encouragement and usefulness declines with experience.The authors chose minimal feedback because it reflects established dance-class approaches and prior research indicates excessive feedback can impede practice.Source: §8.4 · p. 26
9 Conclusion
SkeletonDance was developed from dance-teacher interviews and motor learning theory to provide rhythm feedback for novice practice. Participants reported that its clapping cues helped them regain rhythm and feel better supported, while the paper highlights lightweight beat-focused feedback for beginner dancers.
9 Conclusion
- Teacher interviews identified beginner difficulties and informed SkeletonDance’s motor-learning-grounded approach to rhythm feedback.The system generated human-like clapping sounds synchronized with the music to accentuate its beats.Source: §9 · p. 26
- Participants reported that clapping helped them regain rhythm after losing it and made practice feel safer, more confident, and less strained.They also found the system user-friendly and suitable for remote, independent practice.Source: §9 · p. 26
- The authors emphasize interactive beat-enhancing feedback as a support for beginner dancers’ motivation and practice.They identify customizable feedback and basic-webcam data collection as potential directions for dance education.Source: §9 · p. 26
A Semi-Structured Interview: Dance teacher
The interview guide covers teachers’ backgrounds and teaching practices, then asks how technology might support dance learning and timing. It probes how instructors address timing difficulties and what feedback or tools could help learners.
A Semi-Structured Interview: Dance teacher
- The guide asks about teachers’ demographics, dance and teaching histories, styles, and views on key components of dance and good dancing.It asks teachers to rank components such as timing, technique, enjoyment, and body position.Source: A Semi-Structured Interview: Dance… · p. 29
- Teachers are asked how they teach new sequences and maintain timing, including how they respond when students move well but struggle to keep time.The guide probes methods, tricks, and resources for helping pupils recognize and maintain rhythm.Source: A Semi-Structured Interview: Dance… · pp. 29–30
- The guide explores technology for dance and timing support at home or in class, including sensor choices, feedback modes, and when feedback should be given.It also asks about digital-class tools, potential training effects, and system considerations.Source: A Semi-Structured Interview: Dance… · pp. 30–31