Recording and mixing a French sound: The case of Beethoven’s Erard Frères piano

Project Reports

May 12

Recording and mixing a French Sound: The case of Beethoven’s Erard Frères Piano
Interdisciplinary timbre studies in music, organology, sound recording and acoustical analysis

Interactive Project Report

Authors

Martha de Francisco, Malte Kob, Tom Beghin, Timo Grothe, and Aybar Aydin

Researchers

Martha de Francisco (McGill University), Malte Kob (Detmold University of Music), Timo Grothe (Detmold University of Music), Tom Beghin (Orpheus Instituut), Chris Maene (Pianos Maene)

External Collaborators

Caspar Ernst (Detmold University of Music), Jithin Thilakan (Detmold University of Music), Aybar Aydin (McGill University), Marco Petrella (McGill University), Nicholas Bendzsa (McGill University)

Published: May 12, 2025

Abstract

In the crucial decades of the development of the piano at the start of the 19th century, essential musical questions centered on sonic and performative nuances embodied in the contrast between different school of piano building in Vienna and Paris. For this ACTOR funded research project realized at the end of May 2022 at the Hochschule für Musik Detmold, two replicas of historical keyboard instruments that Beethoven employed to compose some of his major piano works are being contrasted. Replicas of an 1803 French Erard Frères piano and a ca. 1800 Viennese Walter piano were alternately played by historical performer and researcher Tom Beghin. A multitrack immersive sound recording captured the most characteristic sounds of both instruments in the close and mid-range. Simultaneously acoustical recordings and measurements were realized to analyze the origin of the sound waves in different distances. Perceptual analysis of the microphone signals in comparison with the acoustical data collected as well as materials from a pilot project served to describe in detail the timbral characteristics. The evaluation of all the recorded audio takes as well as acoustical investigations of the timbre, directivity, and the effect of the second sound board of the Erard are ongoing. Educational audio materials have been developed.

Musical Parameters

A newly built replica of Beethoven’s 1803 Erard Frères piano (Orpheus Institute, 2016-present) has allowed for a study of “Frenchness” through the sonic and embodied explorations of a well-known and all-important Viennese pianist-composer. Relevant to the ACTOR mission, the collaborative research builds on questions that emerged from a recording on the instrument in 2019, which featured music by Beethoven as well as French contemporaries Louis Adam and Daniel Steibelt.

On a fundamental level, we ask what differentiated a “French” from a “Viennese” sound; but taking into account the skills of both builder and player, along with the complex affordances of a particular type of instrument, we also ask what made Beethoven a “fast learner” on a new instrument as opposed to his Paris-based rival Steibelt, who was an “expert” of the for him familiar French piano — one that lent itself well to what in France was known as son continu, or the skill of sustaining or “spinning” sound: Steibelt was a widely recognized champion of this technique of tremolo.

The French Erard piano was contrasted with a Viennese fortepiano by Anton Walter. To first establish a framework of two longstanding traditions, we recorded on each instrument fragments from Mozart (to represent an Austrian tradition) and Clementi (to represent an English/French tradition). We then zoomed in on the period of 1803–05 and a more focused comparison of “Viennese” and “French” with fragments of Beethoven (Op. 53, “Waldstein”) and Steibelt (Op. 64). Organological parameters in need of assessment included the presence of a second soundboard in the Erard (adding an additional layer of vibrations over its “real” soundboard, which is thinner than an English but thicker than a Viennese one) and the sound-reflecting behavior of the lid (which on a Viennese fortepiano was either left closed or taken off altogether).

Research Questions and Methodology

Musical questions are aimed at understanding the sonic differences between two historical pianos of the same era linked to Ludwig van Beethoven, the sonic possibilities of which helped shape his compositions and thus on a broader level influenced the development of the pianistic repertoire in the early 19th century. To facilitate the comparisons between these two pianos, an interdisciplinary research project was realized at the Hochschule für Musik Detmold. A multitrack immersive sound recording was designed to capture the most characteristic sounds of both instruments in the close and mid-range while circular arrays of microphones provided a capturing of sound radiation around and above each instrument. Simultaneously, acoustical measurements were realized to analyze the origin of the sound and directivity in different distances. Perceptual analysis of the microphone signals in comparison with the acoustical data collected as well as materials from the pilot project (a recording of the Erard piano from 2019) served to describe in detail the timbral characteristics, aiming at shedding light on how organological differences – strengthened by distinct manners of composing and playing – manifested themselves and how the early sound reflections contributed to the sonic character of two building and playing styles.

The pilot recording realized at the Akademiezaal in Sint-Truiden, Belgium, in 2019, released in 2020 (Evil Penguin Classic 0035), raised additional questions in connection with the sound radiation of the Erard to the furthest corners of the room.

The Immersive Sound Recording

*Microphone set-up diagram for an immersive sound recording*

The Digital Performance Room in the Audienzsaal of the Music University in Detmold was a very suitable laboratory for the research experiments with the Erard Frères and the Anton Walter pianos, both built by Chris Maene. Concentric metal frames of 7 and 4 meters in diameter allowed microphones to be placed around one piano at a time, and in different heights, to capture the sound radiated from the instrument in different directions. In addition to the circles of microphones (24 calibrated Røde microphones (type NT5) or Neumann (KM 84)) and two pairs of main microphones in omnidirectional pattern (DPA 4006) were positioned in relative closeness from the instrument in a small AB - large AB configuration. Additionally, four wide cardioid microphones (Schoeps MK 21) were placed high above the piano to capture the height information of the instrumental sound.

Tom Beghin performed a selection of musical segments with works by Mozart, Clementi, Beethoven and Steibelt, as well as chromatic scales or single tones. He alternated playing on the Walter, or on the Erard, the latter both without and with its second soundboard. Other experiments included differences in sound with or without the use of the piano lid. Additionally, a layer of acoustic absorbers was placed or removed from under the instrument to compare the effect of the bottom reflections of the sound with or without acoustic absorption.

Every one of 103 takes recorded in any of these conditions was carefully documented in a Master Table that contained all the research information for each condition. The following is an excerpt:

Acoustical Measurements

In parallel to these recordings, the room acoustic conditions (reverberation time, STI, C80, etc.) were captured using impulse response measurements and other standardized methods. Malte Kob and his team investigated the specific properties of the instruments using acoustic measurement techniques such as an acoustic camera (Visor from Head Acoustics) to visualize the sound radiation, along with laser scanning vibrometry (PDV-100 from Polytec) to analyze the vibration of the strings and soundboards. Animated polar spectrograms were generated and used to compare the acoustic response of the two instruments during the performance of a particular moment in the music.

*Malte Kob installs the laser scanning device assisted by Jithin Thilakan*

The Double Soundboard of the Erard Piano

The addition of an optional second soundboard, which would be suspended above the strings of the Erard, has not yet been studied in detail. Historically, it has not been clear which role the board would play on the sound of the piano, and opinions fluctuated between considering that there was a marked sonic difference all the way to questioning whether the soundboard had only a practical function of protecting the instrument against dust. The research team approached these questions in an interdisciplinary manner. After the recordings, detailed listening sessions were realized by the performer and the tonmeister to compare the takes using the second soundboard, with those without it, and in different conditions. Tom Beghin and Martha de Francisco realized a perceptual analysis that enabled them to understand the sonic effect on the performance and to predict the musical advantages for the use of the second soundboard under different conditions. Listening tests for a larger audience of interested performers, researchers and students have been considered. Malte Kob and his team were able to measure and demonstrate the vibrational behaviour of the soundboard using scanning-laser-Doppler vibrometry. Further analysis and evaluations will follow.

*Tom Beghin places the second soundboard in the Erard*

Evaluations (Work In Progress)

Different methods are being used to evaluate the materials collected in these recordings.

Perceptual analysis
Acoustic camera and vibrometric measurements
Animated polar spectrograms
Spectrograms for the time-frequency analysis
Directivity analysis
Auditory analysis and mixing

Perceptual Analysis

Here are examples of notes, observations and suggestions for action obtained through critical listening and perceptual analysis by Tom Beghin and Martha de Francisco to gain musical insight into the sonic characteristics of the two instruments, as well as to project compositional and performance considerations.

We hear this – can we confirm with measurements?

Topic 1: Signature sound - the melodic line vs the harmonic line:

The characteristic sound of the Viennese piano (Walter) is bright. It has a significant content of high frequencies that allow the performance of melodic lines to be easy to play and easy to follow by the listener. The sustain of the single notes is relatively short, chords do not linger after being played, the sound decays quickly.
The characteristic sound of the French instrument (Erard) centers on a longer decay time of single notes. Especially the lower frequency components of this sound ring on longer after being played, allowing the harmonic context to be easily followed. Melodic lines are clearly audible but lack a concentration of high harmonics on every note.
Is the stability of the tremolo different in the time domain and/or the timbre domain?
Tremolo effect appears easier with Erard + soundboard. Walter: slight impression of mechanical action.
Do we see these differences in the images of the acoustic camera?

Topic 2: Erard – single vs double soundboard

With a single soundboard, the melodic and harmonic lines are balanced.
With the addition of the second soundboard, the instrumental sound appears to be more concentrated, louder. It gains a certain poignance in the salient components that the pianist is choosing to highlight in his performance. It gives the impression of a certain frequency compression.
Is there a stabilizing function of the 2. Soundboard? Is this of advantage? More details audible without 2. Soundboard?
Sb stabilizing function used by son continu
Idea: analyse the projection/directivity effect using the 24-mic recording

Plan a listening test to confirm the differences of the double soundboard effect
Goal: would the sb effect be beneficial for the performance of the composition?
Hypothesis: not only stabilizing but also projecting in favour for the player

Topic 3: Sound propagation in the horizontal domain

Walter: A historical observation: “The best sound can be heard in the prolongation of the tail.” We observe that by soloing the relevant microphones on the ring the sound found in that area favours the high-frequency components.
In comparison, the sound heard in front of the open lid appears to contain more lower frequency components, which makes it sound more “complete” for the listener, but less distinct for following melodic lines (a characteristic of every piano, historical or modern).
Tail towards listener beneficial? Difference with/without lid? Different effects for Walter and Erard? Differences in spectral balance?

Topic 4: Sound propagation in the vertical domain

Walter and Erard: Does the characteristic sound remain consistent high above and around the instruments? Evaluate the mic signals on the upper ring using a high-resolution directivity and LTAS analysis.

Essential Sound Comparison

A selection of essential soundtracks was prepared that includes examples of the music excerpts in stereo played on the Walter, or the Erard without the second soundboard, or the Erard with the second soundboard.

Beethoven, Op. 27 No. 2, i, beginning, “Moonlight”

Sonic characteristics of Walter and Erard (with double soundboard)

Adam Op. 8 No. 2, ii, beginning

The pantalon effect (playing without dampers) on Walter and Erard (with double soundboard)

Adam Op. 8 No. 2, ii, Tremendo

Tremolo comparison on Walter and Erard (single and double soundboard)

Beethoven, Op. 53, i, beginning, “Waldstein”

Performance comparison on Erard with single soundboard and double soundboard

Other comparisons juxtapose music examples of the instruments with or without the open piano lid.

Mozart K 283, i, beginning:

Animated Polar Spectrograms

Gained from the signals of the 24 calibrated Røde microphones (type NT5) on the lower ring (at 1.13 meters), the researchers created visualisations of the sound radiation and directivity of the instruments with polar plots that allow a visual comparison of similar moments in the music on both instruments. The polar spectrogram displays spectra in a circular isobaric chart, where the colour represents the normalised sound pressure level radial dimension represents decreasing frequency, and the angular dimension indicates the recording angle.

Polar spectrogram and audio channel (L: Erard, R: Walter )

The main idea behind these visualizations is to look at the energy distribution in the sounds, and their main radiation directions. The heavily smoothed sound spectra are limited to display frequency range below 4 kHz, in form of spectral envelopes. The color shading from black to light yellow corresponds to the level in dB, as shown by the color bar to the right of the plot. The brighter colors indicate a larger sound pressure level. The frequencies are scaled along the radial dimension: 4 kHz in the center, 100 Hz at the outer diameter.

These so-called polar spectrograms are most meaningful in anechoic conditions, where the recorded microphone signal represents the radiated sound, exclusively, as demonstrated e.g. for the oboe (Grothe, 2023).

In the present project, being constrained to record in a small circular space with absorbent curtains within a concert hall, we used cardioid microphones to enhance the recording of the direct sound without the reflected portions.

Side-by-side comparison of the same music recorded on the different instruments visually depicts sound differences and allows for timbre explorations (Grothe, 2025), in particular in the present study:

The yellow "ring" in both plots shows a strong level around 500 Hz, approximately equally picked up by all mics, in an omnidirectional pattern. Lobes of more directional radiation are recognized for the same note in different areas around both instruments.

Spectrogram for the Time-Frequency Analysis

Aybar Aydin: “In this example we can see the fundamental and the harmonics of the notes of the chromatic scale performed on this take. The Y- Axis (vertical axis) represents frequencies in Hz from 0Hz to 20 kHz. The colours represent the intensity of those frequencies in dB: from blue/dark blue (lowest intensity) to bright yellow (most intensity). The X-Axis represents the time, from the beginning of the audio sample to the end”.

Conclusion

The ACTOR research project “Recording and mixing a French sound” represents a concentration of interdisciplinary activities to gain insight into questions around two of Ludwig van Beethoven’s keyboard instruments.

Several musical questions complement historical performer and researcher Tom Beghin’s explorations as presented in his various publications on Beethoven’s French piano (Beghin 2022, 2020). A thorough sonic evaluation of all the recorded audio takes as well as questions linked to the pilot project (CD recording by Martha de Francisco and Tom Beghin released in 2020), now including the juxtaposition of the French Erard with a Viennese Walter led to considerations about the specific choice of each of these instruments as a more suitable option for the performance of certain works. It also led to the production of 3D audio recording materials for educational purposes. Examinations of a variety of acoustical measurements are giving insight into the directivity of sound radiation of the instruments and they are verifying the observations of perceptual analysis. A particular focus has been dedicated to the second soundboard of the Erard and its musical use, centering on thorough acoustical investigations and critical listening analysis. The evaluations are ongoing. Publications will follow.

The immersive recording of essential music examples for the comparisons of the instruments in all research conditions, coming from 32 microphones (8 studio and 24 array microphones) has been prepared as a tool for students in the partner universities to study the sonic characteristics of the music recorded and documented under the Erard and Walter study. A complete table with precise information regarding each of the conditions will allow the students to perform a detailed auditory analysis of the sounds on each point, both for training of critical listening as well as for the sonic evaluation and mixing of the complex music recording in 3D sound.

*Tom Beghin performs on the Erard in the recording set-up with both rings at 220 cm*

References

Beghin, T. (2022). Beethoven’s French Piano: A Tale of Ambition and Frustration. Chapter: Mixing a French Beethoven Sound, 335-339. Chicago: University of Chicago Press.
Beghin, T. (2020). Beethoven and his French Piano. Sonatas Op. 53, Op. 54 and Op. 57 / Adam Sonata Op. 8 No. 2 / Steibelt Sonata Op. 64 CD recording. EPR Classic 0036. Pilot project, record production by de Francisco, M. Trailer: https://www.eprclassic.eu/items/beethoven-and-his-french-piano
Beghin, T. and S. Maes (2020). A Farewell to Paris: The Story of Beethoven and his French Piano. Documentary film. Orpheus Institute and Evil Penguin TV https://orpheusinstituut.be/en/beethovenerard/tv-documentary
De Francisco, M., Lazarov, V. (2022). Music collaborations, timbre, and recorded sound. Part One: Thoughts on Recording Pianists. Dialogues. Timbre and Orchestration Resource (TOR).
https://timbreandorchestration.org/writings/dialogues/musical-collaborations-timbre-and-recorded-sound/part-one
Grothe, T. (2023) “Oboe-In-Depth, Visualization of Oboe radiation in the Timbre and Orchestration’s (TOR)” Online Instrument Catalogue. https://timbreandorchestration.org/all-instruments/oboe/in-depth
Grothe, T. et al. (2025) Detmold interactive musical Instrument Timbre Explorer (DmITE) – a web application for experiental learning in musical acoustics. In: Proceedings of DAS|DAGA 2025 Copenhagen, pg. 416-417, DEGA e.V. Berlin, 2025. https://doi.org/10.71568/dasdaga2025.636