Performer Interviews

Pianist 1: Performing this work was a highly memorable experience. Despite the challenge of collaborating with an AI pianist, it genuinely felt like playing with a real human partner. As in traditional two-piano settings, the process of listening and reacting to each other was naturally realized here. I was especially impressed by how the system reflected the musical material I played—sometimes reproducing it exactly, sometimes transforming it into an accompaniment, prompting me to improvise new melodic ideas over it. Also, the dynamic and timbral integration between my piano and the system was so coherent that at times it felt even smoother than playing with another person.

Pianist 2: What made the experience especially meaningful was that the team I had worked with—originally in a research context—stepped outside the typical framework of a technical showcase and engaged in artistic expression for the first time. At first, there was hesitation, but once the performance began, everyone truly enjoyed it, empathized with each other’s roles, and offered mutual artistic encouragement. That transformation was deeply moving and rewarding to witness.

Pianist 1: The interaction closely resembled performing with a real pianist. Just like in a standard duet, one performer initiates and the other responds. Interestingly, unlike human duos that may experience interpretive inconsistencies, this system offered more seamless integration and responsiveness.

Pianist 2: I’ve previously encountered contemporary works that manipulate sound in real time using live recordings—transforming audio or generating ambient textures from it. What stood out to me about this piece was how those ideas were extended and reinterpreted through the use of an acoustic piano, a Disklavier (automated piano), and real-time transcription. The fact that such concepts were realized not through electronic sound sources, but through the piano as a central medium, made the experience feel particularly fresh and compelling.

Pianist 1: Definitely. Like any two-piano performance, I had to listen carefully to the phrasing and timing of the other part and adjust mine accordingly. When the system played something more expressive or virtuosic, I naturally lowered my tension or adapted my flow to match its direction.

Pianist 2: This piece places a strong emphasis on timing-based interaction, which significantly shaped my approach to performance. When I played a phrase, the second piano would follow after a short delay, or respond with a transformed version in a canon-like manner, reminiscent of Baroque structures. As a result, I had to continuously listen to and anticipate the second piano’s responses and shape my playing accordingly. This required me to actively adjust tempo, phrasing, and articulation in real time, not merely to follow the score but to build the music collaboratively with the system. I felt that this ongoing process of temporal negotiation created an evolving musical form—one that added layers and variations beyond what was fixed in notation.

Pianist 1: I was constantly aware of the system’s responses. That awareness did not distract me—it enriched my playing. The interaction influenced my phrasing, and the system's responses sometimes even changed how I expressed certain passages. For a trained pianist, such mutual awareness is natural, and I was able to immerse myself deeply without difficulty.

Pianist 2: I was constantly aware of the second piano’s responses. The structure of the piece felt like a real-time musical dialogue, where listening to the system and reacting to it was essential. It was not just about reacting to preprogrammed cues—it was about hearing the system’s output and shaping my next phrase based on what I heard. For example, depending on when and how the second piano entered, I would adjust how I proceeded with the next phrase. This real-time awareness had a direct impact on the pacing and flow of the performance, making each moment dependent on mutual listening and response.

Pianist 1: Absolutely. The interactive nature of the piece directly affected my emotions. The system’s reactions shaped my interpretation in real time, making the experience deeply immersive.

Pianist 2: In a solo performance, there is always a certain pressure to fill the entire space with your own sound. In contrast, during this piece, I experienced my own musical phrases returning in transformed or distorted forms through the second piano. This gave me a strong sense that I was not filling the space alone—there was another presence contributing to the soundscape. It created a richer sensory immersion and a heightened awareness of real-time interaction, as if I were co-constructing the sonic environment with an invisible partner. Compared to traditional two-piano performances, which are typically based on some degree of mutual planning or agreed-upon improvisation, this piece allowed me to take the lead entirely, with the second piano responding in real time. That structure gave me a stronger sense of autonomy as a performer. Within that autonomy, I was constantly receiving and reacting to live musical feedback, which made the experience both uniquely immersive and dynamically responsive.

Pianist 1: During rehearsals, the latency in the system’s response made synchronization somewhat difficult. However, because the piece employs a modern, non-tonal idiom, such temporal flexibility felt musically acceptable—and even expressive.

Pianist 2: The most challenging aspect was the unpredictability of the system. While the second piano generally responds according to the notated material, in practice, unexpected variations often occur—such as subtle delays, added breaths, or shifts in emphasis. These deviations required me to remain constantly alert, monitoring the system’s behavior and adjusting my performance accordingly. That sense of uncertainty created a persistent tension, as I had to stay fully engaged with the second piano’s real-time responses throughout the performance.

Pianist 1: It required a continuous process of exploration. I experimented with different dynamics, rhythms, and articulations to understand how the system would react. This exploratory approach was essential.

Pianist 2: It took time to explore how the system responded to my playing through repeated rehearsals. It wasn’t just about learning fixed behaviors—it was about understanding the algorithm’s character: how it receives sound and how it returns it. That exploratory process was essential in gradually learning to interpret and guide the system’s responses in a musically meaningful way.

Pianist 1: Occasionally, there may have been glitches, but they were not easy to distinguish. I often interpreted those unexpected behaviors as aleatoric elements within the piece rather than disruptions.

Pianist 2: Yes, especially when the system began responding with a tempo, character, or nuance that significantly diverged from what I had initially established. As a performer, I have a certain expressive direction in mind, and when the response strays too far from that, it can feel disruptive. However, through rehearsals, we were able to identify those moments and adjust accordingly to find a more stable and coherent flow.

Pianist 1: Yes. There were times when the system reacted differently from what I anticipated. In such cases, I modified my interpretation to maintain musical continuity.

Pianist 2: Not particularly. Since the number of controllable parameters was limited in this performance, I didn’t feel the need to modify my interpretation. However, if I could suggest an improvement, I would have appreciated more flexibility in how the system responds—beyond just canonic imitation. If performers were given options to select or influence the type of response, I believe it would allow for richer interpretations and a more interactive experience.

Pianist 1: Absolutely. From the start, it felt like I was playing with another pianist. In some ways, it even felt like another version of myself—a musical mirror that listened and responded thoughtfully.

Pianist 2: Absolutely. The system didn’t feel like a passive tool merely reproducing sound—it actively filled the space with me and inspired artistic ideas alongside my own. I felt that we were shaping the music together, with the system taking on a significant role in carrying the performance forward. It gave me a genuine sense of musical partnership, as if we were part of a shared creative community.

Pianist 1: I definitely felt a strong connection. The basic mechanism—where the system listens to my playing and responds—creates a natural sense of dialogue. For example, when the system echoed my phrase and I layered a new melody on top, it felt like a true musical conversation. I didn’t experience any real sense of disconnection.

Pianist 2: Yes, I definitely felt that the system was responding to my musical expression. When my phrases returned through the second piano—often transformed or imitated—it felt like I was engaging in a dialogue with another version of myself. Although the responses were governed by certain rules, they aligned with the flow of my playing and contributed to shaping the musical space together. Rather than any sense of disconnection, I frequently felt a strong sense of connection throughout the performance.

Pianist 1: I didn’t feel alone on stage, which was somewhat reassuring, but my full concentration was on the sonic interaction with the second piano. I wasn’t consciously aware of the operator’s presence during the performance.

Pianist 2: I wasn’t particularly aware of the operator during the performance. My focus was entirely on the sound and behavior of the second piano. Still, the awareness that I wasn’t completely alone on stage provided a subtle but comforting sense of support.

Pianist 1: I hope for more nuanced timbral expression. If the system could more precisely reflect subtle articulations and tone color, it would deepen the sense of musical agency and interaction.

Pianist 2: It would be interesting if the pianist could have more agency in choosing the types of system responses. For example, if a wider range of feedback options were available—and the performer could select or influence them—it would enrich the interaction and allow for greater artistic individuality. I believe it would open up new expressive possibilities for both performance and composition.

Pianist 1: The most memorable part was the sense of genuine dialogue with an intelligent and responsive partner. In fact, it felt more meaningful than some collaborations I’ve had with human pianists. The system’s dynamic control and velocity responsiveness contributed greatly to its realism. The only frustration was the limited range of timbral expression—but overall, it was a profoundly engaging and rewarding experience.

Pianist 2: The most memorable aspect was the feeling of performing with a kind of shadow or alter ego. Although the responses were somewhat predetermined, I was initiating the musical flow, and the system responded accordingly. That interaction created a deep sense of connection and immersion.