Portrait of a Voice
I’m interested in the uniqueness of the human voice, and its role as a completely non-visual identifier. For example, I’ve been told that I speak like a New Yorker, which I think means that I speak relatively quickly and definitively; I look like my parents, but we have completely different accents; I use different voices if I’m chatting with friends vs. working at the Miller gallery. Voices are a synthesis between larger characteristics like culture and social roles as well as individual traits like introversion/extroversion and
Input: 1 minute of natural, relaxed human speaking.
- Users can be invited to introduce themselves briefly (something they’re accustomed to doing) or choose from a list of prompts to speak about for the duration of the recording.
Output: Visual composition of user’s voice
- Shapes and colors are manipulated by sound attributes
- Low level goal: 2 color gradient + 1 nested circle + overall opacity affected by volume
- pitch relates to a color: higher pitch = warmer color, deeper = cooler
- 2 colors representing the highest + lowest pitch recorded
- words spoken in 1 minute (speed of speaking) = # of times a circle is nested in the center of the composition
- More complexity:
- 3 color gradient, pitch level most detected (likely the average pitch) added
- speed of speaking also relates to space between concentric circles:
- Low level goal: 2 color gradient + 1 nested circle + overall opacity affected by volume
- After the recording, the user will see their portrait projected on the wall in front of them. The gallery of other users’ portraits will be projected on the opposite side, and theirs will populate to join them.
Inspiration:
library of bird call frequency visualizations https://experiments.withgoogle.com/bird-sounds
IG: @stefyloret
Kandinsky, Composition 8
library of portraits