This article addresses questions about stability and variation in folk singing in relation to different learning methods such as oral transmission from recordings, learning in live situations or from musical notation, and in folk singing studies within higher music education (HME). A case study is presented where methods for audio analysis were used in order to examine the development of interpretation in recordings of contemporary folk singers. The singers learned the songs either by oral transmission from a living person, from a recording, or from a transcription. The study shows that singers learning by oral transmission proceeded from an initial mimicking that reproduced the source closely, to the presentation of versions holding both stable features and extensive variation of other features. Stable features included relative ‘phrase timing’, ‘voice timbre quality’, pronunciation and ‘rhythmising patterns’. Variation was found in features such as melody, micro-rhythm, micro-timing, intonation praxis, articulation, and ornamentation. The different learning methods are discussed in relation to teaching and learning in HME. By way of describing how musical material is both transferred as a certain singer’s version and further developed by the learner, the term ‘the singer’s imprint’ is introduced.