A musical cryptogram is a cryptogrammatic sequence of musical symbols which can be taken to refer to an extra-musical text by some 'logical' relationship, usually between note names and letters. The most common and best known examples result from composers using musically translated versions of their own or their friends' names (or initials) as themes or motifs in their compositions, with an early and commonly known example being the BACH sequence used by the composer. These are not really rigorous cipher algorithms in the formal sense, but more like musical monograms or labels using initials. The methods used historically by composers were either too incomplete (i.e., did not include all of the letters of the alphabet) or too simplistic to meaningfully encrypt long text messages. There is a separate history of music ciphers utilizing music notation to encode messages for reasons of espionage or personal security that involved encryption and/or steganography.

Because of the multitudinous ways in which various musical notes and textual letters and other elements can be related, or not, detecting hidden ciphers in music and proving accurate decipherment can be difficult.

From the initial assignment by Western music theorists of letter names to notes in the 9th century it became possible to reverse the procedure and assign notes to the letters of names. However, this does not seem to have become a widely recognized technique until the Romantic period. From the mid-19th century it has become quite common. Sporadic earlier encipherments used solmization syllables.

It is believed that this method was first used by Josquin des Prez in his Missa Hercules Dux Ferrarie. It was named Soggetto cavato by the later theorist Zarlino. Under this scheme the vowel sounds in the text are matched to the vowel sounds of the solmization syllables of Guido of Arezzo (where 'ut' is the root, which we now call 'do'). Thus the Latin name of the dedicatee 'Hercules Dux Ferrarie' (Ercole d'Este, Duke of Ferrara) becomes re-ut-re-ut-re-fa-mi-re, which translates as D-C-D-C-D-F-E-D in modern notation with C as 'ut'. This is used as the cantus firmus of the mass setting. Josquin's method was imitated by several of his contemporaries and successors, including Adrian Willaert and Costanzo Festa.

Since the note names only cover letters A to G (reflecting the octave repetition of these names), the problem arises as to how to cipher the rest of the alphabet. Historically there have been two main solutions, which may be labelled for convenience the 'German' and the 'French' methods.

Because the development of note names took place within the framework of modes, in the German-speaking world B♭ was named 'B' and B♮ was named 'H'. The most common musical cryptogram is the B-A-C-H motif, which was used by Johann Sebastian Bach himself, by his contemporaries and by many later composers. Other note names were derived by sound, for example E♭, 'Es' in German, could represent 'S' and A♭ the digraph 'As'.

Composers less fortunate than Bach usually seem to have chosen to ignore non-musical letters in generating their motifs. For example, Robert Schumann, an inveterate user of cryptograms, has just S-C-H-A (E♭, C, B♮, A) to represent himself in Carnaval. Sometimes phonetic substitution could be used, Schumann representing Bezeth by B-E-S-E-D-H. Johannes Brahms used B-A-H-S (B♭, A, B♮, E♭) for his surname in the A♭ minor organ fugue, and the mixed language Gis-E-La (G♯, E, A) for Gisela von Arnim, among many examples.

The 'French' method of generating cryptograms arose late in the 19th century and was more akin to normal encipherment. The most popular version involved writing out the letters H–N, O–U and V–Z in lines under the original diatonic notes A–G, as follows: