Olhos music fest _branding
Daniel Lopes 1, Pedro Martins 2 and Penousal Machado 3
CISUC, Department of Informatics and Engineering, University of Coimbra
— This project is about the creation of a music festival’s
dynamic brand, which reacts to music and customises itself to any
person participating in the event.
Keywords — brand, dynamic, live, music visualisation, face
detection, eye detection, festival.
Olhos (eyes) is a small village in Portugal where there is a
music festival called Olhos Music Fest (Eyes Music Fest). This
project is about the creation of that festival’s graphic identity,
which works dynamically and even lives, reacting to music
(sound) and customising itself to any person participating in
this event, either it is a musician, part of the staff or a visitor.
The main concept behind the festival’s branding was to pass
the energy and feeling of “what it is like to participate in the
festival” — an extrovertive mood, with people dancing free.
First of all, a logotype was created because it was important
to communicate the name of the festival (Fig. 1). However,
branding and graphic identity are much more than a logo.
Thus, it was not taken as the main feature of this brand.
We tried to understand how to translate our concept (what it
is like to participate in the festival”) into graphics. In other
words: how to visualise that experience. So, we started to
define the very elementary units that compose the festival —
music, space and people — and we used those as the starting
point to develop the brand. Just like Neue Design Studio
information visualisation techniques to create the Norwegian
Meteorological Institute’s logo  (to create graphics), which
reacts to weather statistics, we set up to use information
visualisation to create not only graphics but (and primarily)
Fig 1. Olhos Music Fest
’s Logotype (black) concept, explained (red draw)
There are at least two types of inputs to visualise music —
amplitude and frequency of the audio signal. From these two,
the amplitude is probably the easiest one for humans to
perceive. Thus, that was the one to go. However, because a
brand must be unpaired and it was meant to translate our
concept into the graphics, the solution for this visualisation
could never be a trivial sound visualiser. Instead, we decided to
visualise music by developing a dancing artefact, which reacts
to sound’s amplitude — the “dancing free” part of the concept.
This artefact mimics a boneless arm (chain) capable to spin
itself around infinitely.
Visualising space may seem redundant unless we can add a
new layer and turn to see things in a different way.
To visualise space, we opted for simplifying the
environment into four ranges of brightness and represent them
with geometric symbols. Using cameras to render the space
around, we were able to create a whole new way of visualising
it, and create an “extrovertive mood” (part of the concept),
almost mimicking a psychoactive experience. This layer was
used as a background pattern.
Because this is “Eyes Music Fest”, the brand could not be
completed without making the eyes, the main graphic motif.
Though, for people visualisation, we only use their eyes so we
may recognise a friend, although he/she remains anonymous
for the crowds. Those eyes were then used as the main unit to
build the dancing sounds, which works over the space
Changing the inputs will consequently change outputs. So, by
applying this visualisation techniques into the brand’s
graphics, we could create a highly dynamic and personalisable
To better understand this dynamic brand proposal, we refer
the reader to the following link, which contains some videos
about the project:
II. HOW IT IS MADE
The dancing eyes (music and people visualisation):
The technical process of the project started by using
library to detect eyes on photographs.
However, it is very likely for the algorithm to pick other
elements in the image that are not eyes. As such, our solution
consisted of detecting first the faces in the images/frames and
then looking for eyes inside the detected region of interest.
Once a good eye detection was achieved, the next step was
to create an array of eyes for each eye that has been found and
then place each one relating to the one before (its parent), with
a linear distance between them. In other words, we form a
chain of eyes for each eye that has been detected (Fig. 2).
Fig 2. Eye chains over Dalla Marta
1 firstname.lastname@example.org, 2 email@example.com, 3 firstname.lastname@example.org
After that, we set them up to move autonomously around each
respective “father eye”, with a linear speed (linear angle
increment), backward or forward depending on a probability.
Also depending on a probability, each element may change its
moving direction (backwards or forward).
Now that the eyes could move, they just needed o be
moving on rhythm — dancing. Thus, besides its own natural
movements, for each frame, the amplitude from a song playing
was added to each eye “moving angle”. That means that the
tips of each chain move faster than its origin and that each one
moves as faster as the song’s amplitude. However, it also
means they will never stop dancing even if there is a silence
(they will just slow down, waiting to be energetic again).
The environment patterns (space visualisation):
The goal of this space visualisation was to make it in a way
that it could tell something about the objects in space (you can
identify darker or lighter zones and so even identify some
silhouettes) but also that it could be abstract enough to let the
viewer to interpret it in his own way, and be away from the
real and objective world. So, this filter was made out of
First, we picked up the brightness of each pixel in a linearly
spaced grid. Then, we divided the brightness spectrum into
four ranges and for each of those ranges, we have associated a
symbol (or lack of), whose density reflects the brightness level.
Because the system may work with real-time video, it has a
great potential to be used live on VJ performances, interactive
posters, or any other kind of physical installations in the
It can be capturing the same image as the “eyes algorithm”
is (“background is the eye’s owner”) or it can be used to
capture some other different images. In both alternatives, eyes
and background can then be mashed up (eyes over
background). Figure 4 shows the first alternative.
Fig 3. Still from Olhos Music Fest’s
image filter on webcam.
Fig 4. Example of a still illustration to Olhos Music Fest’s posters (eyes
dancing over a dynamic background made out of symbols).
The generated artefacts can be exported and then used to
compose static or animated design objects such as videos,
posters or merchandising (post-production). However, they can
also be used in real-time applications such as interactive
posters, VJ performances or photo-spots in the festival’s
In terms of post-production, two principal graphic elements
were developed — an animated one and a non-animated one.
The non-animated one was a printed poster showing the
artists coming, side by the respective day and time. The poster
used the festival’s logotype, complemented by Fig. 4. Then,
all the typography and composition is properly in tune with the
brand’s rules created (the brand is not only about the logotype
of even only about this generated images; it is a whole
collection of pieces together that identify the festival) (Fig. 5).
Fig5. Olhos Music Fest’s
The animated element was a Facebook
banner where, as
well as in the poster, we introduced the invited artists.
However, here it was possible to use all the potential of the
sound, space and people visualisations. So, we used a picture
from each artist/band to pick their eyes and make them dance
at the rhythm of a song (from Whales
— one of the invited
bands) at the same time we were introducing them. Because we
were working with static images (photos from artists) in the
banner, we used a different input to the background filter
) — a video from someone dancing.
Inspired by works as “Oto Nové Swiss Poster
 or “EV NT interactive poster wall
” , the proposed real-time application was an
interactive poster. Figure 5 depicts a digital version of the
poster, where the static image was replaced by the output of
the real-time algorithm, working with the input of a webcam.
This way, the poster can customise itself to any person or
group of persons looking at it.
As already mentioned, this algorithm can have many other
applications such as VJ performances, by making it work with
the input of a camera on stage, pointing it at musicians or to
To watch the banner and poster animating, as well as the
algorithms working, we refer the reader to the same link
 Armin, “Where the Cold Wind Blows”, 29 September 2010. [Online].
 Studio Feixen and J. Giger, “Oto Nové Swiss Poster,” 9 March 2017.
[Online]. Available: www.otonoveswiss.co.uk. [15/03/2018].
 LUST, “Camera Postura,” LUST, 2014. [Online]. Available:
 Peder Eskild, Silje Barth and Aleksander Wassum, “EV NT interactive
poster wall concept”, 2012. [Online]. Available: https://vimeo.com/43598671