Rethinking Music Discovery
Listening to music is one of my favorite pastimes. It is a great way to escape from the hectic world around us and just relax for a little, or maybe get pumped up for a workout. How do we continually find new songs though?
Music is constantly released and leaves us with a wealth of songs to listen to. Part of the challenge in finding new music is wading through all these songs to find the handful of them that you like. This process can be highly time consuming, especially if you try out each song for 1–2 minutes before making a decision on whether you like it or not. Let’s say you listened to 25 songs for 2 minutes each and only found 5 songs you enjoyed. In total, you spent 50 minutes listening to music and only liked 20% of the songs. Therefore, if you would have known ahead of time which songs you liked, you would have only spent 10 minutes listening to music. To aid in the music discovery process, many music streaming services such as Spotify, Apple Music, Pandora, and Google Play Music all have some sort of recommendation system that suggests music based on your past interests and listening behavior. These recommendations save people time and allow them to find music that suits their interests more frequently. For a deeper dive into how recommendations are generated, check out this blog post by Sophia Ciocca!
My main music streaming service is Spotify and they have a “Discover Weekly” playlist that is updated every week with new song recommendations. While I enjoy having a curated list of songs that I would probably like, over time, I have found that the recommended songs are often very similar in genre. For me, I like to listen to many different genres even if some are more popular than others. Because of this, I tend to listen to another Spotify-created playlist: “New Music Friday”. Every week, the playlist is updated with around 100 of the latest song releases. I view this as a great way to for me to discover new music across many genres and save time on searching.
Over the countless weeks of New Music Friday that I have listened to, I have come to realize that I am biased to songs based on factors such as who the artist is, song name, or what the album cover looks like. One week, I was listening to a song that I was enjoying, but once I saw it was ‘House Party’ by Sam Hunt, I thought twice about adding it. This was mostly because I automatically think “country” when I hear Sam Hunt, and I’m not the biggest country fan, so my initial reaction was more negative than it should have been. From moments like this, I learned that if I like a song, then I should add it to my playlist and not let who sings it affect my decision. I know you are probably thinking, duh, why would you let it be any other way?! And my answer to this would be that it isn’t as simple as you think. Subconsciously, biases play a big role in determining the songs we like or dislike.
Spotify Voice Emerges
One day, after watching The Voice, I was inspired to create an app that provided an unbiased, music listening experience. For anyone unfamiliar with The Voice, it is a TV show where singers perform with the judges’ backs to them. It is only when a judge decides that they like the singer’s voice that they turn around and see what the singer looks like. This is done to ensure that a singer is picked based on their voice and not any other external factors such as how they look. So, based on this simple idea, I decided to create Spotify Voice (I know, the name makes it sound like it accepts voice commands for Spotify).
To deliver an unbiased listening experience, Spotify Voice will automatically hide the song name, artist, and album cover from view when listening to songs. When using the app, I wanted a simple and clean user interface that mirrors the simplicity of the Spotify app itself. The result can be seen below:
So how does it work?
When you click the Play New Music Friday button, a playlist, called spotify_voice, will be created under your account which duplicates the songs in the New Music Friday playlist. As you listen to music, liking and disliking a song can be done by clicking the heart and prohibition sign respectively.
Liking a song will add it to the designated playlist as shown by the dropdown menu. In this case, a liked song would be added to my ‘Add Me Maybe’ playlist. You can add songs to any of your public playlists. Liking a song will also reveal the song name, artist, and album cover and continue to play the song.
Disliking a song will skip the currently playing song and remove it from the spotify_voice playlist. This is one of the most integral features of the app. By removing songs as they are disliked/skipped, the spotify_voice playlist will automatically track the listening progress you have made. Let’s say there are 100 songs in this week’s New Music Friday and you listen to 20 of them before going to bed. The next morning, you would only have 80 songs left and wouldn’t have to track the 20 songs you listened to already. If songs weren’t removed as you went along, then you would be listening to the same songs you already liked/disliked before. Disliking a song also shows the song name, artist, and album cover.
New Music Friday, on average, brings about 100 new songs each week. This is obviously a lot of songs to sample in one sitting. The good news is that you can pick up where you left off by clicking the Play New Music Friday button later on and it will look for spotify_voice playlist and play it if it exists. So, all your listening progress is tracked and you can pick up where you left off over multiple listening sessions.
To Infinity and On To the Next One
So what’s next for Spotify Voice? Well, for starters, there are still a multitude of features I want to add to the app because it really is very simple in its current state. It provides an unbiased listening experience well enough, but has some flaws. The most pressing issue is that only I can use the app currently. This is because of the code frameworks I have downloaded that are not widely available in most computers. I want to figure out a way to host my app online and have it accessible by anyone because we all want an unbiased music listening experience right? This and other issues will be covered in the blog posts to come as I look to tackle a new feature, bugs, and usability each week for the next 10 weeks.
This blog is mainly for my composition class, so I will try to explain the technical aspects of app development intuitively. This first post was meant to break the ice and will go into more of the coding side in the next posts. I would also like to mention that if you have any suggestions for features you would like to see, definitely drop a comment below (and like, subscribe, and share this with your friends, family, dogs, cats, ya-da ya-da)!
PS: I think it would be fun to share a song I like with each blog post, so here’s this week’s: Thunderclouds by LSD ft. Sia, Diplo, Labrinth