Facebook PixelThe dynamic AI-powered playlist matching with your heartbeat
Brainstorming
Tour
Brainstorming
Create newCreate new
EverythingEverything
ChallengesChallenges
IdeasIdeas
Idea

The dynamic AI-powered playlist matching with your heartbeat

Image credit: https://unsplash.com/photos/3efwRzEHaoo

Loading...
jnikola
jnikola Jul 24, 2021
Please leave the feedback on this idea
Originality

Is it original or innovative?

Feasibility

Is it feasible?

Necessity

Is it targeting an unsolved problem?

Conciseness

Is it concisely described?

Bounty for the best solution

Provide a bounty for the best solution

Bounties attract serious brainpower to the challenge.

Currency *
Bitcoin
Who gets the Bounty *
Distribution
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.


Introduction
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).

The idea
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.

Benefits
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.

Implementation
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?

[1]https://medium.com/@Spotify/groove-is-in-the-heart-matching-beats-per-minute-to-heart-rate-271a79b7f96a

7
Creative contributions

Finding resonating music is way more complicated than just matching the BPM with listener's heart rate

Loading...
Povilas S
Povilas S Jul 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.




Please leave the feedback on this idea
Loading...
jnikola
jnikola3 years ago
Hi Povilas S! Thank you for your exhaustive feedback! I am glad this topic provoked your detailed contribution.

I agree that the heart rate does not define the music preference of an individual. You mentioned situations where a person could be stressed and want to relax, or it might be bored and want to do something active. Please take into account that I was talking about fine-tuning the playlists with the heart rate monitoring and matching, but a person still has the freedom to choose the desired type of playlist. If a person wants to relax, no matter how high the heart rate is, they will never choose to play upbeat music. The same works for the opposite. If a person wants to dance and start moving, it will not play slow music but will choose the faster beats. That is the moment when the algorithm starts to do the magic. It takes this heart rate as the starting point, tracks the change, and fine-tunes the playlist to keep up with the heart rate change dynamics. It would learn from your workout habits, chill routines and would act accordingly.

When I researched this topic, I found two papers backing up my theory. The first one reports that the gradual increase in music tempo (2%/minute) successfully increases the heart rate (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5339732/). They also noted that the heart rate elevation is mediated by sympatho-respiratory coupling with the played tempo. Another paper showed a similar thing to what you mentioned. Scientists showed that 93% of subjects experienced a decrease in heart rate following the slow song and 100% of subjects experienced an increase in heart rate following the fast song (https://www.emerginginvestigators.org/articles/the-effect-of-music-on-heart-rate/pdf). I agree that more studies are needed and I will do my best to find more proof of my theory.

So, the conclusion is that the music beat definitely affects the heart rate of an individual. In my opinion, that's not creating false positives, but a potential shortcut to a perfect atmosphere that is used every day. Yes, the point of the bpm tracking would be to find the right starting point. The next goal would be to lift the atmosphere to the desired level. In both segments, the fact that music affects the heart rate is beneficial to the success of the algorithm and leads to a mutual benefit. Yes, things are much different if we talk about DJ fine-tuning the party atmosphere or a person chilling on the porch. Some situations undoubtedly require more effort from the algorithm and that's where we could make a change.

This idea is still very far from the app that picks the right song for the right moment. It's a small step towards finding the right tune, but we will have to do a lot of research on the brain to be able to predict what could be the best song for the next moment. AI will definitely help here, I agree. But I didn't dare to go that far. Concerning the dangers of matching the bpm of the music with the heart rate, I didn't hear that! I will definitely check it, but what I think that could play a major role is the volume and the frequency of the music, rather than the tempo.

And finally, yes. The algorithm would work much better on wordless songs. That's why it is a clever idea to test it on songs like that. In the future, I think the lyrics will just be another filter that will help the algorithm to filter the vast amount of available songs much easier and choose the right one more precisely.
Please leave the feedback on this idea
Loading...
Povilas S
Povilas S3 years ago
Juran So the conclusion for me still seems to be that the attempt to match/approximate the BPM with listener's heartbeat is not the right approach. Since the tempo of the music can effectively influence a person's heartbeat, this might serve as means for an app to get you to the desired state, for example, the user might choose the mood they want to get to from their current state, whether they want to be energized, relaxed, focused, etc., the app would know what heart rate represents particular emotional states or at least is likely to be associated with them and with the help of music with specific BPM would try to increase or slow down user's heart rate accordingly.

So the approach would be to rather change the heart rate with the help of music than try to match it. Similar to the heart rate BPM might only be appropriate to pick if a person's heartbeat is already close to how it looks in their desired state. The music with a similar tempo would then act as a sort of reinforcement. I'm not sure if starting with the BPM similar to the person's heartbeat is better than starting with an entirely different BPM when trying to move the person into a different mood. The articles that you cite prove that music can effectively change the heart rate, but there's nothing about the benefits of approximating or matching it with the initial heart rate. I'm still wondering about that part.
Please leave the feedback on this idea
Loading...
jnikola
jnikola3 years ago
Povilas S Yes, yes, I completely agree with you! I am still skeptical about it, too, but the reason why I think the matching is needed is to change the heart beating effectively. One of the papers I cited described the gradual change in tempo (2%/min) to be more effective than significant change (>3%/min). Therefore, I presumed that first matching, and then slowly leading the heartbeat in the desired direction could be a more effective method. It's how marketing works, right? You first try to meet, make friends and understand ("resonate") and then you slowly and surely influence their thoughts :D
Please leave the feedback on this idea

Individualize music taste

Loading...
Darko Savic
Darko Savic Jul 24, 2021
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:
  • surprise me
  • 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
Please leave the feedback on this idea
Loading...
Povilas S
Povilas S3 years ago
Spotify already does at least some of the things you mentioned. It does everything you mentioned in the first paragraph. It probably does the rest too, but without the user's interactions, in a silent mode, I think it goes as part of their algorithm. I understand you mean that the app should do this in addition to what Juran is suggesting. Anyway, it would be easier to combine the two since some of these functions are already implemented into an existing music streaming platform.


Please leave the feedback on this idea

The music can be overlaid with binaural sound effects

Loading...
Samuel Bello
Samuel Bello Aug 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.
Please leave the feedback on this idea

Music beat synchronized walking/running

Loading...
Povilas S
Povilas S Oct 25, 2022
I came up with a related idea that is easier to realize. It's quite fun to synchronize your steps to the music beat as you walk and listen to music, so I thought why not make an app for this?
The app would have to register the frequency of your steps, I think most fitness trackers are able to do that. Then you'd either choose a song yourself and try to match your steps with the BPM of the song or the app would choose the music for you to guide your walking/running exercise that way. It could start with slow BPM songs and slowly increase it by choosing songs with faster and faster BPM until you had to run eventually. The app would have to take into account your listening history to make sure you'd like the songs it chooses for you at least to an extent.
You'd look at the screen of your phone and see whether your steps are synchronized with the beat or not. The interface could be similar to the musical instrument tuning apps:
The arrow leans right if you walk/run too fast and left if too slow, the more it deviates from the center the more you miss out. You have to keep the arrow within the green section in the middle.
Please leave the feedback on this idea
Loading...
Shubhankar Kulkarni
Shubhankar Kulkarnia year ago
This is a good way to learn rhythm while exercising.
Please leave the feedback on this idea

Music that gets you moving - a personalized workout motivator

Loading...
jnikola
jnikola Jul 24, 2021
We all love listening to music sometimes. What if we used music as a motivator to get moving?

The idea
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.

[1]https://www.scientificamerican.com/article/psychology-workout-music/

Please leave the feedback on this idea

Idea for an experiment: do people with higher average heart rate prefer music with higher BPM?

Loading...
Povilas S
Povilas S Nov 19, 2021
I've recently started checking BPMs of the songs that I like with online BPM finders and noticed that a few different dance songs that I find very moving have very similar BPM (122-124). Then I remembered this idea session and the notion that people prefer musical tempo which is closer to their heartbeat and it got me thinking could your preferences for music tempo be related to the specifics of your average heart rate. In other words - could your individual physiology influence your preferences for certain music.
Average resting heart rate differs greatly among different people. So does the preferences in music. There have been studies investigating how people's preferences for a certain music tempo relate to the heart rate, but they were about measuring the heart rate "in action", not the average resting heart rate which is specific to the individual.
Momentary heart rate is influenced by the activity you are engaged in. If your heart rate is measured while you listen to music, the music itself influences it, so such an experiment is not "clean" in that sense. I haven't found studies that would try to determine the relationship between the average resting heart rate of different people and their preferences for music with certain BPM.
The experiment: The average heart rate of a participant in a proposed experiment would have to be determined by measuring their pulse in a resting state on different days and times of the day and then averaging the results. Preferences for music with certain BPMs can be measured by both inclinations towards certain musical genres (each genre has its specific BPM range). As well as by preferences for specific, favorite songs. The correlations between average heart rates and inclinations towards music with a specific tempo would then be tested.
To make the latter part easier, a software which takes a favorite music collection (e.g. playlists on spotify, youtube, etc.) of a person, analyses all songs there and presents a distribution of a BPM range, can be used - it will show to which BPM range the majority of the songs fall into.

[1]https://songbpm.com

[2]https://play.soundplate.com/analyze/track

[3]https://pubmed.ncbi.nlm.nih.gov/8570336/

[4]https://pubmed.ncbi.nlm.nih.gov/8532483/

Please leave the feedback on this idea

The above-mentioned tool as a DJ-ing tool

Loading...
jnikola
jnikola Jul 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.
Please leave the feedback on this idea
Loading...
jnikola
jnikola3 years ago
I agree with you about music directly affecting the heart rate. Thankfully, there is always resistance, if people do not feel the music. Therefore, I hope the inner mechanisms of each individual can fight against a bad atmosphere and still send useful information to the AI DJ. Of course, the AI should be able to distinguish between the "background noise" you mentioned. Good notice!

The cameras and microphones could be a nice add-on feature and help to make the algorithm more efficient in recognizing the mood and fine-tune the playlist accordingly.
Please leave the feedback on this idea
Loading...
Darko Savic
Darko Savic3 years ago
I think music directly affects the heart rate so whatever the DJ does would bias the listeners' heart rate reporting devices.

Instead, could AI be able to tell how the listeners are feeling based on audio and video that it gets via cameras and microphones in the venue? That's basically what the DJ does based on people dancing and cheering/screaming. The AI could maybe pick more subtle cues. There would be some kind of optimum mood the DJ strives towards and the AI would help achieve this.
Please leave the feedback on this idea
Loading...
Darko Savic
Darko Savic3 years ago
I wonder how people would like something like this (middle) at the DJ table on the party https://images.techhive.com/images/article/2015/09/ibm-watson-100616825-large.jpg
Please leave the feedback on this idea

Add your creative contribution

0 / 200

Added via the text editor

Sign up or

or

Guest sign up

* Indicates a required field

By using this platform you agree to our terms of service and privacy policy.

General comments