Explore the future of sound identity with AI

Rather than suffer the artificial intelligence revolution, why not harness it to create unique sound identities?

Since 2022, the Audiospheric studio has been exploring the creative opportunities offered by generative AI applied to sound. From vocal design to sound design, via generative music, we test and hijack these tools to better adapt them to the requirements of brands and creatives.

Today, sound identity is a strategic lever for capturing attention and making a lasting impression. How can AI serve this challenge, while retaining a human and emotional dimension? Discover our experiments.

Voice AI: redefining the voice of brands

Voice AI demo :

Corporate motion design video(see the film)

Tools used: ElevenLabs, Speechelo, Wellsaidlabs, Listnr, Murf, Play.ht

Challenges:

• Uneven audio quality between different platforms.
• Several generations are needed to achieve a satisfactory result.
• Difficulties linked to languages with little training (e.g. French).
• Risk of auditory hallucinations.

Solutions provided:

• Benchmark the best solutions for specific uses.
• Methodology specific to AI voice-over to maintain a natural flow.
• Careful checking and correction during mixing.

What we remember:

Voice AI can be a powerful asset for creating voices, but still requires human supervision to ensure consistent sound quality. English voice models are better trained than French ones, which can affect the quality of results.

​

IA sound design: creating unique sound worlds

Demo IA sound design :

Tools used: Stable Audio, ElevenLabs, Optimizerai, MMaudio

Challenges:

• The systems don't understand certain technical terms (e.g. "skidding").
• Presence of spurious frequencies in generated sounds.

Solutions provided:

• Rephrasing of prompts in descriptive language ("rapid displacement of earth" instead of "skidding").
• Multi-layer approach to enrich sound textures.
• Rigorous post-production to optimize every element.

What we remember:

Every AI has its strengths and weaknesses. AI can be a formidable ally in the design of unique, immersive soundscapes. The challenge lies in adapting the tools and manually enriching the results for optimum quality.

​

IA music: generating and revisiting musical identities

Generative music, composing with AI

Demo IA music soundtrack :

Demo IA sung music :

Tools used: Suno, Udio, Riffusion, Stable Audio, Google Magenta, AIVA, Jammable, Kits

Process :

1. ChatGPT or MistralAI-assisted speech generation.
2. Texto-to-Music with Suno, Udio.
3. Remix and post-production in DAW software (Ableton Live, ProTools, ...)

Challenges:

• Find the right prompt for a specific musical style.
• Understand the characteristics of different genres and instruments.
• Audio quality still below industry standards.

Solutions provided:

• Standardization of prompts for better control of results.
• Post-production work to improve sound quality.

What we remember:

AI offers great creativity and unexpected results, although audio quality can sometimes fall short of industry standards.

​

Musique Interpolée, a fusion of musical worlds

Musical interpolation makes it possible to hybridize different styles to offer original compositions, while respecting a brand's sound identity. It pays tribute to influences while creating a sense of nostalgia or familiarity.

IA interpolated music demo :

Challenges encountered:
- Finding compatible pieces to interpolate.
- Ensuring consistency between merged elements.
- Copyright risks.

Solutions provided:
- Creation of new compositions to avoid legal disputes.
- Use of recognizable elements with a unique transformation.

What we remember:
Interpolation opens up fascinating possibilities for modernizing a sound identity or bringing different musical cultures into dialogue. Interpolation also makes it possible to link different communities through the same work, thus enriching the sound identity.

​

Audio AI tools

AI Vocal Tools: ElevenLabs, Speechelo, Wellsaidlabs, Listnr, Murf, Play.ht

IA Music Tools: Suno, Udio, Riffusion, Stable Audio, Google Magenta, AIVA, Jammable, Kits

IA SFX Tools: Stable Audio, ElevenLabs, Optimizerai, MMaudio

Towards new sound horizons

AI tools are revolutionizing audio creation and opening up unprecedented possibilities. However, their profusion and technical limitations require an expert approach to make the most of them. At Audiospheric, we believe that AI is not an end in itself, but a catalyst for creativity. It enables us to push back the boundaries of sound design, provided it is integrated with intelligence and sensitivity.
These tools can work wonders when properly mastered, but you may feel limited. Co-creation with an expert in audio AI then becomes a suitable solution for pushing the boundaries of sound creation.
Interested in exploring these new perspectives for brands? Let's talk about it.