Sound Effects: Finland
What 16 audio logos tell us about the future of audio branding.
We commissioned this study and built the Finnish case for it.
TUNEMENT brought the third report in the Sound Effects series to Finland: we selected the sample, sourced the audio material, briefed the AI comparison, and set the questions the research had to answer for brand owners rather than for researchers.
EmotiSphere Insights ran the testing and the analysis. The result is independent research with a practitioner's brief behind it.
Rather than argue about sound, we measured it.
We showed people a brand's visual logo on its own, then the same logo with an audio logo attached, and measured what changed. Eight brands had bespoke audio logos made by composers. Eight had audio logos generated by AI, tested against the same no-sound baseline.
Our position going in was that this is not a referendum on AI. It is a question of where human judgment still does work a model cannot, and that is what the report sets out to locate.
Finnish brands tested, eight bespoke and eight AI-generated.
Outcomes measured for every brand.
Method, applied identically across the sample.
Can AI make an audio logo that works?
Not a question of whether the output sounds finished. Whether it moves the brand, and how often it fails to.
What does a composer still do that a model cannot?
The report isolates the outcome where bespoke and AI diverge most sharply, and explains why.
Is a memorable sound the same as a valuable one?
Sound is unusually good at being remembered. The report tests whether that memory does anything for the brand.
What should a brand do about it now?
Where each route is worth using, and the appendix on generating short-form sound with AI.
Most audio testing asks people to explain reactions they don't have access to. Implicit Association Testing measures the response instead, a rapid sorting task that captures what a sound does before anyone can rationalise it.
Every brand was tested twice: visual logo alone, then visual and audio logo together. The difference is what the sound adds.
- Methodology
- Implicit Association Testing
- Brands
- 16 — 8 bespoke, 8 AI-generated
- Conditions
- Baseline vs. visual + audio lockup
- Outcomes
- Emotional appeal, memory encoding, human vs. synthetic brand feel
- Significance
- 95% confidence interval
- AI platform
- ElevenLabs, licensed training data
- Series
- Third report, after Saudi Arabia and Italy
“Anyone can generate a sound now. The work is deciding which one is yours.”
“The numbers say memorability is the easy part. Meaning is what a brand still has to choose, and that choice is not something you can prompt.”
A study by TUNEMENT & EmotiSphere Insights, with contributions from Herwig Kusatz of Sound Strategy))) and Dr. Caitlyn Trevor of SonicUXR.