Violinist Eszter Osztrosits and pianist Imre Dani in conversation

Recordings-Informed Performance

For us—present-day artists raised on modernist ideals—early sound recordings throw light on so many differences, or even contradictions, between the notation and the interpretation of the Romantic repertoire which today’s performers must reckon with. We think that the instrumental level of the performer, their individuality, the creative ideas arising from their personality, and intimate knowledge of the musical language of the era in question, are all indispensable for an authentic interpretation of Romantic works. In our project at the Liszt Academy of Music in Budapest, guided by Dr László Stachó, we aimed to realise a historically informed performance of Edward Grieg’s Violin Sonata in C minor (Op. 45). Fritz Kreisler’s and Sergei Rachmaninov’s legendary recording of this sonata seemed to us an ideal starting point: in their performance, the two personalities shine through – and isn’t this a key aspect of romantic performing practice? But even though Kreisler and Rachmaninov gave us a lot of inspiration, during the process of preparing for the performance we realised that their recording is far from being a faithful representation of late 19th-century and early 20th-century performing aesthetics. The discrepancy is most conspicuous in the playing of Kreisler, who used vibrato and portamento much more frequently than his predecessors and most of his contemporaries. From contemporary, or near-contemporary, violin playing, we studied recordings by Joseph Joachim, Jelly d’Arányi, Jenő Hubay, Ede Zathureczky, and Carl Flesch, and strove to embrace features of their performing styles. In turn, regarding the piano part, we considered recordings by Fanny Davies, Benno Moiseiwitsch, Alfred Cortot, and Edward Grieg himself as exemplary. We are excited to share a few thoughts about how we worked on the second movement of the sonata. But first, here’s the opening movement which already shows many of our findings in action.

Grieg, Violin Sonata No. 3, Op. 45/i, Allegro molto ed appassionato. Eszter Osztrosits (violin) and Imre Dani (piano)


– The most exciting experiment was to discover elements such as vibrato, portamento, and tempo rubato which are not marked in the score but are essential features of late 19th-century interpretations. In Grieg’s time, there existed no signs to directly notate vibrato, portamento, or nuances of rubato; but the form, the character of the thematic materials, the harmonic structure of the movement, alongside the notated performance instructions (especially Grieg’s own dynamics), clearly showed me the way. But contemporary, or near-contemporary, recordings offered most inspiration. It was especially inspiring to closely listen to Edward Grieg’s Norwegian Dance, Op. 35 No. 2, played by Carl Flesch, where the great Hungarian-born violinist recreates the piece with unique, and at first hearing surprising (but clearly rule-based and in the end logical) microtiming patterns. Based on the recordings studied, I aimed to achieve a well-balanced, rich sound with minimal use of vibrato. This non-vibrato manner proved to be most powerfully expressive during the recapitulation of the main melody (4:08–4:17). Furthermore, in those parts that include marcato, tenuto, and long melody notes, I chose an intensive and dense vibrato. More precisely, to maintain the energy of the emotional charge of the musical process, I aimed to use a vibrato that mimics natural resonance similar to the vocal sound, which in some cases took a variable speed even during a sound. This is especially relevant for the section between 2:09 and 2:16, where pianissimo and marcato hint at such an interpretation. Regarding the marcati, I tried to strengthen the rhythmical play between the piano and the violin with a dense, low-amplitude vibrato focused at the beginning of the sound. In turn, Grieg’s tenuto signs seemed to refer to the production of a horizontal melody line, in which highlighting and vibrating the tenuto notes played a prominent role in connecting melody notes together. Here, the vibrato served the melodic process: it helped drawing the structure of the melodic motif, and producing a singing tone to connect the individual notes. Crescendo and decrescendo are also tightly linked to tempo, resulting in a natural, constant rubato during the entire movement. Portamenti seemed to me less prominent as most of the time the main melody moves in a narrow range, the only exceptions being in the middle part where the melody seeks to reach its highest note (2:16–2:24, 2:39–2:53). Here is this second movement:

Grieg, Violin Sonata No. 3, Op. 45/ii, Allegretto espressivo alla Romanza. Eszter Osztrosits (violin) and Imre Dani (piano)


– The recorded piano part is not my recreation of one specific artist’s style from the early 20th century; rather, it is intended to capture key general performance elements from that era. But of course, I can mention several pianists whose recordings had a major impact on my interpretation during this “detective project”: Gabriel Fauré and Alexander Scriabin (both from piano rolls), Sergei Rachmaninoff, Alfred Cortot, Benno Moiseiwitsch, Mieczyslaw Horszowski – and, needless to say, Grieg. László Stachó, whose doctoral course at the Liszt Academy in Budapest inspired our project, showed us an annotated score by Daniel Leech-Wilkinson of a Fanny Davies recording of Robert Schumann’s Davidsbündlertänze, which includes the timings of each beat and the length of time by which the bass anticipates the beat to which it belongs, in one of the movements – I found this very instructive. Another insightful example by Leech-Wilkinson, Cortot’s recording of Chopin’s Prelude No. 4 in E minor, shows not just the timing of each note but their loudness as well. (Both analyses can be found in ch. 6 of Leech-Wilkinson’s online book, The Changing Sound of Music.) The Grieg project by Sigurd Slåttebrekk and Tony Harrison, documented at, is a noteworthy experiment as well: it allowed us to look into the process of emulating the Norwegian composer’s performing style.
In the second movement of the Grieg sonata, the most interesting part to work on was the piano introduction: its material is perfect to discover such late 19th-century musical aspects as rich (and appropriate!) tempo rubato and anticipated bass-lines. I also sought to expand the dynamic range and emphasise specific notes and inner melody lines. For example, between 0:28 and 0:46, I separated the two voices in the right hand and sought to give them their unique freedom (especially that of the middle-voice right-hand appoggiaturas: G#–F#, F#–E, E–D#, D#–D, C#–C, E–Eb, then between 0:46 and 1:01, where the left hand’s inner motifs and the right-hand melody bear similarity to the previously mentioned appoggiaturas). I also paid attention to anticipating bass notes differently, gently emphasising the various degrees of harmonies/dissonances (0:17–0:27). In the middle part, I experimented with the syncopations (2:24–2:58) to give the accompaniment a more flexible “swing”. In the recapitulation (4:08–5:06), I was greatly inspired by Grieg’s recording of his own piano piece, To Spring (Op. 43 No. 5). The way he plays triplets during the whole piece is full of life and instantly captivating; this is an interpretation to listen to and to learn from!
And now the final movement:

Grieg, Violin Sonata No. 3, Op. 45/iii, Allegro animato. Eszter Osztrosits (violin) and Imre Dani (piano)

A lockdown version

In addition to these recordings, we would like to share our earliest experiment to create a historically informed performance of the sonata’s second movement. This was made during the first lockdown period. Needless to say, we were separated, therefore – after studying the literature and the historical recordings we’ve already listed above – we started to send “karaoke” versions of the violin and the piano part to each other. This method seemed to us quite unusual at first yet we immediately found it really entertaining to follow the musical ideas we shaped through these recordings. Here is our final result of this fun process from the first quarantine period back in 2020.

Text by Eszter Osztrosits and Imre Dani, 2021

Leave a reply