If you watched We Are Your Friends, you remember the scene where Zac Efron explains where biology and music converge. Although it was just an average movie, it got me thinking. What if they were right and you, as a DJ, can completely control the crowd by simple bpm adjustments? That's when I started digging.
Maybe you noticed that you love to chill, slow music when you lay in your deckchair on a Sunday morning. On the contrary, you enjoy faster dance music in a local bar on a Friday night. So, there must be some connection between the music tempo, heartbeat, and our mood. Data from Spotify reported the average beat of top 10,000 songs to range between 120 and 130 bpm (the above-mentioned movie also mentioned the 128 bpm to be the "golden" tempo).
What if we developed a tool/plug-in for Deezer, Spotify, and other music apps that would track our heartbeat recorded from the smartwatch, bracelet, or any other similar device and adjust the playlist/beat to match our current heartbeat? When you start running, it starts with slower tempo music of your choice (new random songs, favorites, etc), gradually speeding up with your pace increasing. If you slow down due to pain in the spleen, it would also slow down and imperceptibly cross-fade to a slower beat.
Combining the existing AI tools to track your favorite genres and artists with physiological signs of current body state, we could develop a strong algorithm able to read and learn from your mood, habits, and daily routines, and deliver the perfect song any time. There would be no need to stop your workout due to a high bpm playlist being interrupted by a slow love ballad.
It is not very easy to develop and implement this tool. The easier part is to add the feature of real-time heart rate data tracking to any music app. The harder part would be to dig the beats of all the available songs, check for the beat changes, bpm patterns, and display them as a novel parameter. The toughest part would be to match and combine the developed parameter with the user favorites and new discoveries in real-time.
What do you think?
What other physiological metrics could be used to find a perfect song for the moment?
Finding resonating music is way more complicated than just matching the BPM with listener's heart rate
Povilas SJul 24, 2021
Just because a person's heart is beating fast does not necessarily mean that they will enjoy upbeat music. They might, for example, be stressed and want to relax, for which slow beat music or even beatless music is way more suitable. The same is true for slow heart rate. The person might be bored and wanting to do something active, therefore enjoy the upbeat music way more than the slow one. And there are numerous combinations in between.
Maybe a general tendency of liking the music with the BPM which is closer to your heart rate does exist, but it has to be backed by studies. It would be quite easy to test this on your own by using a player that shows the song's BPM and at the same time monitoring your heartbeat with a smartwatch or so while listening to music to see if you prefer songs with BPM closer to your heart rate.
But then there's another complication of music affecting and changing your heartbeat accordingly, which Darko already mentioned in the comments. So it might simply be that it's not that people's heart rate affects their preferences in music but simply that it tends to slow down or speed up according to whatever the song is playing and this leads to creating a false hypothesis. Also, there's a health concern that approximating the BPM too closely to the listener's heart rate might create a resonance in the body and lead to heart problems. The latter might be another myth, but I've heard about that being a sort of safety rule in a setting of dance music.
The app picking the perfect song for a particular moment has been a major dream of mine for a long time. The key to that is most certainly personalization rather than applying some general principles. General principles that work universally are useful, but they don't solve the problem of personal inclinations towards particular music at particular times. Why do you enjoy a certain song at a certain moment while after an hour or a few minutes it might not resonate with you anymore and you'll need something different? How to know in advance what exactly will resonate with you next? This is the main problem to solve for enabling such an app to become a reality.
Essentially it might be as complicated as reading someone's thoughts, but I think the right approach would be to monitor the person's listening habits very precisely, with the help of AI-powered data analysis and search for any chronologically (or otherwise) defined patterns. It would be somewhat similar to the approach for mood forecasting. The monitoring of physiological parameters, especially brain activity would be immensely helpful, cause it would provide the AI with data to correlate to various musical parameters of different songs.
But apart from the resonance with your mood/emotions, there are many other features that are hard to take into account, for example, song lyrics - you might like the song because the lyrics resonate with your recent life circumstances or because it reminds you of a warm childhood memory, etc, etc. This is all on the level of thinking and associations that also significantly influence your musical preferences of the moment. Perhaps the right way to start would be to only play with songs that don't have any lyrics, that way there would be less mind and more feelings and physiology involved and it would be easier for the app to analyze the data and make predictions.
Music that gets you moving - a personalized workout motivator
JuranJul 24, 2021
We all love listening to music sometimes. What if we used music as a motivator to get moving?
I suggest using the above-mentioned algorithm to roam through the user's listening history and match it with the heart rate and the amount of movement recorded (steps, calories, etc). That way, the AI could understand what lifts you, what makes you move, and could help you in tough times. When there is nothing that can force you to get up out of the bed or start working out, the app would play a "lifter song", that would motivate you to start moving. The brain would relate the music with the body movements and hopefully, get you moving! It could work similar to the alarm or reminder feature, or you could allow the algorithm to track and plan everything by itself.
Based on each user's playlists and history the AI should learn their music preferences. Then pick from similar genres and try to surprise the user with something new they haven't heard before.
Also, the user could be given some options:
play only my kind of music
mix in my favorites X% of time (adjustable)
When an excellent new song comes on, the user should somehow be able to like it. The AI remembers it and serves it again later/tomorrow. The user could have a few more buttons for each song:
play more often
play less often
never play again
The music can be overlaid with binaural sound effects
Samuel BelloAug 15, 2021
Binaural beats are audio illusions that are achieved by playing sounds of different frequencies on either ear. Binaural beats are believed to have effects on a person's mood and concentration levels. They can be used to reinforce positive and emotions while steering the listener's mind from negative thoughts. They make it easier to take a person to their preferred (or chosen) state of mind.
This can be less restricting from an aesthetic point of view since the user can play any song they want but the song will be modified with audio effects to make it replicate the results that binaural beats give.
The above-mentioned tool as a DJ-ing tool
JuranJul 24, 2021
Have you ever been to a party where a resident DJ completely missed to impress you or the person playing songs at the party played only the songs of his/her choice? The songs did not fit the moment and you were not the only one to notice that? Well, maybe if they had this tool, it would never happen.
For whom would it be?
The idea is to implement the tool that was suggested by the session on every party where the amateur/professional DJ is not playing its music (in case they do, like concerts and themed parties, the crowd comes to enjoy the specific genre/artist).
How would it work?
The DJ would detect the smart devices in the area and read real-time heart rate data from the people enjoying the party. The data would be anonymous and would be used exclusively to display the current average heart rate of the partying crowd. The DJ would then adapt the playlist accordingly.
The mutual benefit
This approach would help many. On the one hand, DJs would have another parameter to adjust their playlists. If they want to raise to a completely new level of service, they could buy the pro version of the tool that would suggest them playlists or songs that would perfectly match the crowd's heart rate and a party type. The algorithm could even automatically play the music like already seen on Virtual DJ "auto" mode, where the software does all the transitions. The additional feature would be that you don't need to choose playlists and songs, just type the type of the party and preferred genres and the software would do the rest. That way, DJs would have more time to plan the strategy for the rest of the party, drink, and enjoy the night, while people would enjoy every moment of the party undisturbed.