Digital Art Music Series: Playing Hide and Seek with Myself—AI-generated vocals and lyrics
My first time trying AI-generated vocals and lyrics—
I composed and arranged the music, and even had the AI “learn” the sound of wind and rustling straw, to give her voice a slightly scratchy, textured feel 🌾
When I first saw this artwork by @chuchunxx, I asked her:
“Is this the present self looking at your future self?”
She replied,
“Nope! It’s just hide-and-seek 😆”
From that moment on, I couldn’t un-hear it.
The idea of playing hide and seek with myself stuck with me.
After all, you are always the hardest person to find.
Conflicted, layered, always shifting.
We spend so much of our lives just trying to “find ourselves.”
It’s often only when something truly happens that we realize which version of ourselves shows up.
Before that, we just think we know—and have endless “solutions” ready.
But the truth is, we’re still searching.
On the production side, this piece took a long time—not because the software was hard (it was actually very user-friendly), but because there were just too many possibilities.
Each version sounded reasonable, even good.
So in the end, I returned to the core motivation—simplifying the musical and expressive elements.
Other experimental directions, like lyrics that are just phonetic “spells” with no meaning, also worked well.
But the longer I listened, the more I felt that they didn’t really connect to the idea of “self.”
In the mixing process,
Jack @j_ack0831 brought out the footstep sounds more than I originally planned,
to make the feeling of quietly searching more deliberate.
He also carved out space in the mix so the ears could “hold their breath,”
and we intentionally blended the vocal into the atmosphere—
as if the voice were drifting in the air, not standing out too much.
It’s meant to be felt more than heard.
We also tried using AI for the mix for the first time.
It took quite a bit of time to explore and tweak,
but in the end, we didn’t find a version that felt quite right—
maybe it just needs more trial and error across future works. Lyrics: hide, seek,hide, seek,Playing hide and seek with myself,
In the shadows.
I laugh, I cry, I hide, I seek,
I‘m out of breath.
In the corners of my mind,
I search for clues.
a hint, a sign,
of me I hope to find.
Playing hide and seek with myself.
Playing hide and seek with myself. Digital Art Music 系列：Playing hide and seek with myself 跟自己捉迷藏｜首次嘗試讓AI 去唱歌和寫詞，我作曲和編曲，還給她學了一點稻草和風聲的音效，讓聲音稍微有點刺刺毛毛的🌾。記得第一次看到這幅圖的時候，我就問 @chuchunxx 小chu說：『這是現在的自己在看著未來的自己嗎？』她說：『沒有捏～這是捉迷藏😆！』自此之後就回不去了，我腦中一直揮之不去跟自己捉迷藏這個主題，自己總是最難捉模的，矛矛盾盾矛矛盾，光是找自己，可能就花了大半輩子，往往要遇到了事情，才知道自己會做出哪個選擇，在沒遇到事情之前，都屬於”很”知道該怎麼做，方法也很多，但等遇到了才發現，只是一直不斷地在找自己。在製作的時候花了很多時間嘗試，不是因為軟體很難學，相反的軟體很簡單使用，會猶豫和思考是因為太多種可能了，每個方式都很合理效果也不錯，所以最後還是回到簡化動機的部分，同時也簡化表現面，其他的可能性，留給下一個作品去表現，像是沒有意思只有發音如魔咒的歌詞，其實效果也很好，但是聽久了跟”自己”並無太多關聯性。在混音的部分， @j_ack0831 子傑將腳步音效聲調整的比我預設的更出來，讓偷偷往前進找尋的意圖更明確一點，然後整體也盡量騰出空間，讓耳朵有屏氣凝神的餘地，人聲也盡量融入在樂器和空間裡，像是在空氣中流動著，不要太突出，以氛圍為主。我們這次也初次嘗試了用AI混音，花了蠻多時間討論和調整，但最終還是沒有找到比較適合的樣貌，可能還需要更多作品去嘗試看看。歌詞：hide, seek,hide, seek,Playing hide and seek with myself,
In the shadows.
I laugh, I cry, I hide, I seek,
I‘m out of breath.
In the corners of my mind,
I search for clues.
a hint, a sign,
of me I hope to find.
Playing hide and seek with myself.
Playing hide and seek with myself. #digitalpainting #illustration #illustrator #digitalillustration #backgroundartist #underscore #BGM #backgroundmusic #artgallery #digitalart #digitaldrawing #night #nighttime #darkness #nightview #nightphotography #nightvibe #starrynight #捉迷藏 #hideandseek #AI #dreamtonica #synthesizerv #オリジナル曲 #サントラ #ピアノ #자작곡 #오에스티 #originalsong #피아노
Comments