I can't say anything about the app as I've not used it. However, the concept is (in my opinion) flawed for a large number of potential users - the music my friends listen to is absolutely no indication of what I might like.
I'm too old. I've spent the past 30 or so years (I'm 36) developing a taste of my own. I no longer meet or socialise with people around music (eg go to gigs with friends). Similarly, my "favourites" are people I admire and respect for things entirely different to what their music tastes are. Most of the new music I find these days is based on what I listen to already or in curated theme-based playlists from sites like http://ShareMyPlaylists.com. Automating something like music discovery from a social graph is skewed (in my opinion again, and based on instinct rather than science) towards young people who still socialise in what I would call "taste networks" - groups based around shared tastes. Older people don't do that in my experience.
Good luck with it though. There's still a giant market out there. And I'll still try it just to see what it thinks I might enjoy.
I agree. It's very difficult to estimate someones taste.
For example I like a hardcore band but only because they are different from all other hardcore bands. Most programs will estimate I like hardcore but that isn't the case.
I'm not sure but I think I like the music I like because they have this special 'mood' that fits me. And the lyrics are also important for me. Question is: how can we turn this into meta data suitable for matching.
Interesting question ohwp. I do think that recommendation systems are only going to get better. Lyrics and sentimental analysis are definitely going to play a part in that.
One of the main reasons we built Soundwave though was to take a more human based approach to music discovery. Personally, I'd take three friends who have similar hardcore tastes than me over any system (currently built) because I too was annoyed at how poor the recommendations were. Our app will hopefully reduce the friction in finding those people with similar tastes.
Thanks for the comment onion2k - fair point. We definitely have a target market in mind with the app which is a younger demographic whose musical taste may not be completely developed just yet. Having said that, I'm sure you have a couple of friends you still turn to for tips on the latest and greatest?! As for favorites, again I'd agree with you but we've had a huge amount of interest from people trying to discover what sports stars are listening to before a big game etc. Social graphs are so subjective that this really is a binary type of idea. Some people will love it and use it a lot (hopefully), others will use other channels for discovery and stay away from the social curation side.
Thanks for raising this point so amicably though - I'll be sure to update the post with the links when we're live!
Actually, the social graph is directed - like, say, twitter. So I can explicitly follow people I know have a penchant for picking up decent new music first.
This might be the "muso" in your real-world social group or it might be a respected journalist outside that group. It might even be Bono himself.
I'm in a similar age bracket, and I've had "so what are you listening to these days" conversations with friends. I find them quite rewarding, even though my broader tastes have settled down.
I'm biased (obviously) but I think we've built something along the lines of those conversations - and more (check-out maps!).
Comments
I can't say anything about the app as I've not used it. However, the concept is (in my opinion) flawed for a large number of potential users - the music my friends listen to is absolutely no indication of what I might like.
I'm too old. I've spent the past 30 or so years (I'm 36) developing a taste of my own. I no longer meet or socialise with people around music (eg go to gigs with friends). Similarly, my "favourites" are people I admire and respect for things entirely different to what their music tastes are. Most of the new music I find these days is based on what I listen to already or in curated theme-based playlists from sites like http://ShareMyPlaylists.com. Automating something like music discovery from a social graph is skewed (in my opinion again, and based on instinct rather than science) towards young people who still socialise in what I would call "taste networks" - groups based around shared tastes. Older people don't do that in my experience.
Good luck with it though. There's still a giant market out there. And I'll still try it just to see what it thinks I might enjoy.
I agree. It's very difficult to estimate someones taste.
For example I like a hardcore band but only because they are different from all other hardcore bands. Most programs will estimate I like hardcore but that isn't the case.
I'm not sure but I think I like the music I like because they have this special 'mood' that fits me. And the lyrics are also important for me. Question is: how can we turn this into meta data suitable for matching.
Interesting question ohwp. I do think that recommendation systems are only going to get better. Lyrics and sentimental analysis are definitely going to play a part in that.
One of the main reasons we built Soundwave though was to take a more human based approach to music discovery. Personally, I'd take three friends who have similar hardcore tastes than me over any system (currently built) because I too was annoyed at how poor the recommendations were. Our app will hopefully reduce the friction in finding those people with similar tastes.
Thanks for the comment onion2k - fair point. We definitely have a target market in mind with the app which is a younger demographic whose musical taste may not be completely developed just yet. Having said that, I'm sure you have a couple of friends you still turn to for tips on the latest and greatest?! As for favorites, again I'd agree with you but we've had a huge amount of interest from people trying to discover what sports stars are listening to before a big game etc. Social graphs are so subjective that this really is a binary type of idea. Some people will love it and use it a lot (hopefully), others will use other channels for discovery and stay away from the social curation side.
Thanks for raising this point so amicably though - I'll be sure to update the post with the links when we're live!
Actually, the social graph is directed - like, say, twitter. So I can explicitly follow people I know have a penchant for picking up decent new music first.
This might be the "muso" in your real-world social group or it might be a respected journalist outside that group. It might even be Bono himself.
I'm in a similar age bracket, and I've had "so what are you listening to these days" conversations with friends. I find them quite rewarding, even though my broader tastes have settled down.
I'm biased (obviously) but I think we've built something along the lines of those conversations - and more (check-out maps!).