Audio or Story: Mystic to Mathematical: Spotify Engineering

December 17, 2021
Published by District Taino, Engineer

TL; DR For 2021 Wrapped, we were challenged this year to visually release a user’s audio aura based on how they listened. I like to think of it this way: if your music listening data becomes a person and the aura in the neighborhood goes to the reader, what would that person look like? This is the 2021 Rapid Audio Ara story.

A new wrapping experience

Spotify Wrapped is a company-wide effort with 300+ Spotifiers across 20+ teams. My team in particular is a design / engineering team and our main focus is to design, build and introduce the Rapid Personalized user experience to millions of Spotify users around the world. These personalized stories bring users into the habit of listening annually by telling creative stories.

Personalized experience shows listeners data stories such as their best songs, best artists and best genres of the year. In addition to creating these stories, our brand and creative team think outside the box and come up with new ways to illuminate user insights into audio streams.

Imagine. A meeting room full of professionals – of course a virtual meeting room. Eyebrows are raised, pens are clicked, and fingers are scratching the head. We were thinking about the answer to the question: how do we do it? Define Audio of a Spotify user? To help with this question, and to guarantee that the Ara we create is legitimate, we recommend an Ara reader, Mystic Michela.

According to Mystic Michaela, Aras is “your personal power signature. Everyone has one, and Ora readers see them as a combination of colors, representing every feature that makes you, you” (see Michaela’s full Aras here). For the purpose of our audio aura story, We can say by definition that a 2021 Spotify Rapid Audio Aura is a color combination of two colors that helps our listeners understand the vibe or mood of the music streamed this year.

Let’s come to the whiteboard

Challenge # 1

Return the top two mood sections and narrators associated with their 2021 listening history for each user.

Challenge # 2

Find out how much a user’s music is represented by the weight of each mood aka mood.

To address the challenges above, we used a dataset that describes the track mood, compiled each user’s listening history, and created six broad mood categories (such as “happy”, “calm”, “optimistic”) that would be compressed into two. . Aura of audio.

  1. Calculate the number of streams for a given track and retrieve its top mood descriptor as determined by the mood descriptor dataset.
  2. Put those moods in the bucket in one of the six mood sections.
  3. Perform an aggregate to find the total number of streams for each mood segment and get the mood descriptor with the maximum number of streams.
  4. Make a final calculation to find the percentage of a mood stream over the total number of streams.

In the example above, “happy” or “quiet” tracks are the most streamed, creating the two best moods found in a particular user’s music. 47% of “happy” users listen and the mood describes “happy”. “Silent” makes up 29% of user’s listening and granular narrator “Chill”.

Now, let’s turn these moods into colors

Click Dry-Erase Pen Cap … We have identified the best two audio aura moods of a user and the weight of their mood. Now let’s turn this data into color. Mysterious Michela took steps to lend her skills to Aras and their color. Along with his guidelines, we’ve assigned six key colors to the six mood-describing sections mentioned above (including “happy”, “calm”, and “optimistic”). To provide secondary color to the aura visuals, our designers present the weight of the mood through different levels of contrast. The more mood the user hears, the darker the color. The lower the user’s mood, the lighter the color. In the graphics above, the weight of the user’s mood is 0.47; Therefore, the secondary color can be found in rows labeled Encore 100. Their second mood weight was 0.29; Therefore, the secondary color can be found in rows labeled Encore 180.

The limitations of creativity and embrace

Now that we have a user’s audio aura color, the last question that remains for design and engineering to consider is: How can we make an ethereal aura visual engineer?

Our deadline was two weeks away. In our milestone plan, we’ve allocated a few days of engineering to each story as we’re committed to many data stories and several new features. We had a few days to explore an elegant aura visual that is possible for both iOS and Android. Part of the design challenge is to know what can be done within a given time for engineering. And part of the challenge for engineering is to accurately estimate how long it will take to create a given design. We had to come up with a solution that made good use of our time and achieved a confident engineering sign-off. This kind of situation wrapping is very unique to the design / engineering experience and in such situations, I have learned that creativity comes from embracing limitations and using what we have in a fancy way. Our designer used three elements:

  1. An animated gradient was created for last year’s 2020 wrapping campaign to create a “lava lamp” effect.
  2. Wrap ribbons made for 2021 to create organic and unique shapes.
  3. A vague effect of mixing everything together in a dream-like fashion

Since we’ve already created animated gradients and ribbons, we’ve been able to create timely aura visuals and accurately portray an ethereal aura with high confidence!

Wrap it up

The 2021 Rapid Audio Ora story is a peek into the unique experience of engineering and design. We took a secret idea and broke it down into its basic elements to create an exciting feature for the wrapping experience. Our ability to work with teams across Spotify helps us find creative solutions, enabling us to create stories that, hopefully, bring joy to our users. If you are interested in joining our efforts to bring new experiences to Spotify audiences, check out our open introduction!

Source link

Related Articles

Back to top button