PUZZLE WOOD Sound Contents by AI
Summary
Through this AI expedition, I set out to explore and experiment with sound-generating AI tools to discover their practical potential in creating multisensory, synesthetic content.
As the visual content creator for Puzzle Wood, I wanted to translate the brand’s fragrance into sound — to capture the atmosphere of its scent through AI-generated audio.
I was deeply curious to see how AI would interpret the human sensory experience of inhaling a fragrance, and how closely it could reflect that emotional nuance through sound.
Theme
‘In the languid early afternoon,
somewhere in an unknown part of nature, closed eyes without thinking.
Then a fresh breeze blew in from the distance and touched moment’
somewhere in an unknown part of nature, closed eyes without thinking.
Then a fresh breeze blew in from the distance and touched moment’
Puzzlewood, with its green accord, opens with a refreshing burst of citrus, followed by soft, verdant notes of neroli and eucalyptus.
In the heart, neroli and citrus intertwine gracefully, revealing a modern freshness — one that is neither overly sharp nor reminiscent of intense phytoncide greenery.
As the scent unfolds, a gentle woody base sustains the harmony of the heart notes, eventually melting into a subtle musk trail that lingers with understated warmth.
To more vividly convey Puzzlewood’s message — “the scent of rest,” a fragrance that calms and centers the mind — I envisioned a tangible scene to translate this feeling into sound.
Scene Visualized Through Midjourney (Album Cover Image)
A man sitting in grass in front of clouds:: 3in the style of Jessica drossin, ren hang, camera high-angle, —ar 116:85 —style raw* – Upscaled (2x)
Process
While there are many AI generative tools capable of creating sound, I chose to experiment with MUSIA (Muzia) and ChatGPT for this project.
Through ChatGPT, I gathered insights on the musical atmosphere, genre, and chord progressions, and based on those tips, I used MUSIA to generate melodies.
These melodies were then refined through a simple mixing and production process to create the final sound.
Mainstream AI music tools such as SOUNDRAW and AIVA instantly generate complete sound samples once the user inputs parameters like mood, genre, tempo, and duration — allowing the user to simply select the version closest to their desired outcome. Having already experienced this approach, where high-quality samples are endlessly generated on demand, I wanted to explore a slightly more intricate, hands-on AI tool that required deeper creative engagement.
Mainstream AI music tools such as SOUNDRAW and AIVA instantly generate complete sound samples once the user inputs parameters like mood, genre, tempo, and duration — allowing the user to simply select the version closest to their desired outcome. Having already experienced this approach, where high-quality samples are endlessly generated on demand, I wanted to explore a slightly more intricate, hands-on AI tool that required deeper creative engagement.
ChatGPT “You’re the best composer.” — Role assignment.
I provided ChatGPT with detailed descriptions of the sensory impressions and fragrance notes I wanted to express, and in return, it offered musical atmosphere and chord progression guidelines that complemented them.
We had several engaging conversations, a few of which are transcribed below.
What I found most interesting was ChatGPT’s literary foundation — it was remarkably skilled at expressing emotion and nuance, almost like a storyteller.
Another notable aspect was its clear sense of subjectivity; when presenting multiple options, it always explained precisely why one was the most suitable.
Moreover, unless the change was as drastic as switching from jazz to heavy metal, its interpretation of the musical mood remained largely consistent with my vision.
Before producing the track, I already had a faint sense of the feeling I wanted to achieve.
The atmosphere proposed by ChatGPT leaned more toward a slower, chill-house style than I initially imagined, and unless I altered the genre parameters myself, the overall mood it suggested remained consistent throughout the process.
Me :
You are the best artistic composer to create music that expresses this scent.
Write a title and description for this music.
Write a title and description for this music.
Chat GPT :
Title: “Harmony of Serenity”
To create a refreshing and sensual atmosphere that reflects the pleasant comfort and sophisticated style of the A-flavor (A-flavor = puzzle
wood), the “Chillout” or “Smooth Jazz genres would be a good fit. These genres combine relaxed rhythms with sophisticated melodies to help
convey a pleasant sensation to the listener. Therefore, these genres will match the personality of the A-flavor while making the music richer
and more sophisticated.
“Harmony of Serenity” is an artistic composition that translates the alluring scent of A-flavor (Puzzle Wood) into sound.
Inspired by green relief, crisp air, and the serene essence of nature, each note feels like a gentle guide leading the listener into a fragrant world.
The soft, tranquil melodies invite a sense of calm, while each instrument conveys the subtle nuances of A-scent — its clarity, freshness, and depth.
This piece becomes an artistic journey where music and fragrance intertwine in harmony.
To capture an even more composed and refined mood, the piece was written in C minor, introducing a new chord progression that lends a polished sensibility and an overall smoother tone.
Compared to earlier versions, this rendition also carries a slightly darker and deeper emotional shade, adding a new layer of sophistication to the composition.
Me :
Tell me your favorite and why
Chat GPT :
We chose the title “Harmonious Serenity” because it captures the essence of both pieces of music and the emotions we want to convey to the
listener. “Harmonious” refers to the harmonious and coherent quality of the music, and “Serenity” refers to the peaceful and calm atmosphere,
making it the perfect way to encompass the essence of both pieces of music. This way, listeners can expect to feel the harmonious tranquility
of the music.
The word “Harmonious” reflects the music’s sense of balance and coherence, while “Serenity” evokes its peaceful and calming atmosphere.
Together, they form the perfect expression of the two pieces’ shared spirit — allowing listeners to experience a sense of tranquil harmony through the music.
Me :
Recommend overall music genres, mood and code guide.
Chat GPT :
New Music Information:
Genre:Ambient/Chillout
Cm (C minor)
Running:*2min
Features: Soft Synth Pads, Synth Arpeggio, Synth Leads, Full Synth Ensemble
〔Cm] 〔Gm] 〔Cm] 〔Cm] (Synth Arpeggio) ————————————————————————————— Remaining sections omitted ———————————————————
Remaining sections omitted
Genre:Ambient/Chillout
Cm (C minor)
Running:*2min
Features: Soft Synth Pads, Synth Arpeggio, Synth Leads, Full Synth Ensemble
〔Cm] 〔Gm] 〔Cm] 〔Cm] (Synth Arpeggio) ————————————————————————————— Remaining sections omitted ———————————————————
Remaining sections omitted
MUSIA Plugin and DAW (Logic Pro): “Instant Melodies with a Single Click”
Following the guidelines suggested by ChatGPT, users can input chords, mood, and tempo, and the system automatically generates melodies that harmonize with those elements.
Each melody can be exported in two formats — WAV and MIDI.
The MIDI file can be directly dragged and dropped into Logic Pro (DAW), allowing for more detailed mixing and customization to achieve the desired sound.
The most remarkable aspect of this process is that the melody is composed instantly by artificial intelligence.
Whereas a human composer might begin by sketching a few main chords and gradually developing the melody — a process that can take considerable time — MUSIA produces melodies spontaneously with a single click.
It’s an incredibly efficient method of ideation, though it also evokes a subtle sense of unease when compared to the organic spark of human inspiration.
Entering the MUSIA code allows infinite melody generation,
and inputting the chord progression suggested by ChatGPT into the plugin automatically produces melodies.
and inputting the chord progression suggested by ChatGPT into the plugin automatically produces melodies.
I performed a simple mix by combining the individual MIDI melody files generated by MUSIA in the DAW.
I combined and mixed the individual MIDI melody files suggested by MUSIA in the DAW through a simple mixing process.
The combination of the main chord progression proposed by ChatGPT and the melodies generated by MUSIA strongly reflected the consistent chill house style that ChatGPT had initially suggested.
Without changing the genre, I experimented with combining the melodies like puzzle pieces, one by one, exploring various arrangements to find harmony within the established flow.
During this process, ChatGPT recommended using the key of C major, but it felt too bright and lively, so I switched to G major to create a calmer and more relaxed atmosphere, completing the final composition.
“Harmony of Serenity”
PUZZLE WOOD SOUND CONTENTS by AI
PUZZLE WOOD SOUND CONTENTS by AI
Final_Chill House / BPM 114 / G Minor
- Amorepacific Creatives
- Visual & Sound
- 우시아