In a startling revelation this past April, fans of the iconic Britpop band Oasis were abuzz with the discovery of a “lost” album. This collection, capturing the essence of late-90s culture with its powerful choruses and Liam Gallagher’s distinctive Mancunian voice, seemed like a dream come true. However, the dream was short-lived. The tracks, intriguingly named AIsis, a form of AI music, were not genuine Oasis creations but artificial renditions of what might have been. Delving deeper, these weren’t entirely AI-generated either, challenging our conventional understanding of artificial intelligence in music.
Mophonics Weighs In: The AI audio challenge
Kristoffer Roggemann and Steph Altman, the creative minds behind music and sound studio Mophonics, provide insights into the complexities of generative AI in the audio realm. According to them, audio AI lags behind its counterparts in other creative fields. Kris points out a scientific reason: the human ear’s discerning nature. While 24 frames per second can fool the human eye into perceiving motion, over 40,000 audio samples per second are needed to convince the ear of a sound’s authenticity. He humorously adds, “I once heard an audiologist claim that if our eyes were as powerful as our ears, then we could look up from Earth and be able see a quarter sitting on the surface of the moon!”
Steph brings another perspective, highlighting the advanced tools available to the VFX and animation industries due to their long-standing history, a luxury audio AI hasn’t enjoyed.
The truth behind AIsis
So, how did AIsis come into being? The tracks, it turns out, were human compositions, styled after Oasis. The AI’s role was limited to mimicking Liam Gallagher’s vocals. While impressive, it’s a far cry from generating original content. Steph emphasizes that the idea of AI recreating a popular song convincingly, let alone crafting an original masterpiece, is currently a stretch.
The music industry is currently surfing a wave of AI enthusiasm. However, Kris and Steph urge caution. Kris sees AI’s potential in democratizing creativity, envisioning a future where even a novice can produce basic AI-generated music affordably. Steph, however, warns against premature expectations. While AI can produce rudimentary music, the demands of top-tier brands require a higher caliber of composition. Steph believes that the chasm between average and exceptional music is vast, and AI hasn’t even mastered the average yet.
Both experts concur on the irreplaceable value of human creativity in music. Kris envisions professional music-makers pushing boundaries, crafting original pieces that steer clear of the monotony of algorithmic music. Steph foresees the need for human prompt engineers to fine-tune AI compositions, potentially becoming tomorrow’s composers.
Kris also touches upon the intricate art of mixing and mastering, referencing the legendary Billie Jean which underwent 91 mixes before finalization. He doubts a machine’s capability to replicate such nuanced decisions.
The essence of music
Nick Cave’s 2019 reflection on AI’s potential to craft a great song resonates deeply in this context. He opined that while AI might produce a good song, it lacks the human experience that elevates a song to greatness. Steph echoes this sentiment, emphasizing that standout content requires a level of craft and soul that AI currently cannot replicate.
Kris acknowledges the strides in music innovation, even before AI’s advent. However, he believes that while AI has made significant progress in text and visuals, its magic in music remains to be seen.
The intersection of AI and music is a captivating blend of potential and reality. While AI’s capabilities are expanding, the soul of music, for now, remains undeniably human. For the team at Mophonics, the essence of great music-making, rooted in human creativity, stands timeless.