In the new streaming app under development, one of the current menu items is “discovery” which could immediately remind listeners of “discover weekly” by you know who.
As soon as one hits a thesaurus, different ontological questions begin to arise. For me, there’s also an itch that can’t quite be scratched… if we’re trying to evolve the economics behind streaming, should we not also evolve the very concepts of music discovery as well?
So lets start a discussion on this… @auggod would you feel like posting a screenshot of the menu, or even the discovery page itself? (Totally understand if you’d like to keep it under wraps for now!)
I can’t remember if I mentioned it before or not, but I always thought it could be interesting to have a map (which may open up an avenue with the fun, more human illustration style shown in one of the other threads). Then it might differentiate from both the corporate streaming way of displaying new music and the location tag system in use elsewhere, but in a more imaginative, community-focused way.
I guess I always imagine scenes of people when I think of different styles of music and how place influences that. Realistically, it’s probably a really hard thing to put into practice, but I wondered if it might help with the wording, then you would feel like you’re walking into spaces defined by those sounds, rather than scrolling through a platform. I’ve read about a few mapping experiments with you know who, but they only seem to be short-lived, publicity-driven projects rather than an ongoing archive or attempt at building a community. Happy to email some examples if relevant.
Of course, never in a million years with the Evil Beast that shall not be named. However, this would certainly work:
@auggod I can’t imagine it being too hard to add geo points based on artists stated location?
Would be amazing to do a little tour around the world, where you have to manually explore different regions, zooming in, moving the map around until you find a new cluster… or an isolated artist. Imagining this makes me realize how much streaming feels virtual, as if coming from some imaginary realm, rather than a studio or a house in specific locations. This map idea of yours @RobertaFidora solves that for me!
I guess we can just use opencage (https://opencagedata.com/) to geocode locations (cities) then we can use leaflet with openstreetmap and carto (https://carto.com/) for the tiles. i have good experience using leaflet. we can have clusters, heatmaps.
Well, this is all really excellent and positive to read! I thought of a couple of further considerations to this idea, extending the sharedness aspect:
Obviously, there may be musicians who might not feel comfortable with opting in and perhaps that could be covered with an “off the grid”, “random” or “lucky dip” button (maybe illustrated by a little planet/periscope/question mark or something so it’s still exciting to find something new and people don’t feel excluded if they don’t want to share their location).
The other benefit is that a map isn’t a fleeting selection that disappears quickly or becomes old news and it isn’t so hierarchical in that way, so as an artist in that system, where you can move side to side rather than between older and newer, there’s always the chance someone could find you (which reminds me of some of the older internet projects aggregating videos or songs that have “never been seen”, etc. - they’re kind of fun). However, you could also build in sub groups if you really wanted to, so if the area is London, you can have subcategories of “new sounds”, “all sounds” or later on “Resonate editorial” - which might encourage musicians in different music scenes to contribute pieces about their local ecosystem and pave the way for other commentaries on local cultures from music writers.
If artists tag themselves with multiple locations (so if they started making music in one place and then moved elsewhere), you can see how music scenes cross-pollinate and which places intersect, which might help provide more context about their work. To add to this, if anyone wanted to find “related music” after listening to one artist, this could potentially make it easier (and be freeing for anyone reluctant to classify their sound with strict genre labels).
I vaguely recollect seeing something about a radio show or podcast from Resonate to highlight new releases and saw another thread on the forum about Last.fm, which allows users to tag by place and listen to radio stations based on those tags. Don’t know if appropriate or not, but could be interesting to be able to listen to different regions and cities. Also kind of inspired by this project, which maps the areas served by different stations: https://www.pirateradiomap.com
I’m conscious of the whole attention economy thing, but feel that maybe this is more aligned with the ritual of going to a record shop and flicking through the categories until something catches your eye or visiting a new city and looking at the gig listings rather than the aforementioned “stats and rankings” maps that other companies have used.
And absolutely, I love the idea that someone who’s making music in an extremely remote location might be able to find collaborators they didn’t know existed, or that a listener somewhere else in the world could visit that remote place through the sounds produced there.
I love the idea that someone who’s making music in an extremely remote location might be able to find collaborators they didn’t know existed, or that a listener somewhere else in the world could visit that remote place through the sounds produced there.
On that theme, there’s a radio station I found via Radio Garden up in Longyearbyen, essentially the furthest north town on the planet! Arctic Outpost Radio – It’s existence feels like a legend in itself! It exclusively plays old tunes from the ~1930s-50s which makes it even more enigmatic to me.
Definitely different from hearing music written in Longyearbyen, but still gave me a fascinating feeling of connecting to a place I’d never connected to.
I think we give too much importance to whatever has been done before in the short existence of the streaming era.
For me, the best findings in music are often very fortuit. Whatever tools can help, ok but there’s only a very small chance it will find anything worthy for me. That does not happen every day or every month.
The discovery page on Resonate is not some fancy tool you should use to find the next sound you’ll want to listen all day long but rather, at least give some idea of what up with Resonate at the moment.
I guess, it’s just badly named so let’s find another term if you want but I never heard of that discover weekly thing. Discovery is the title of an album by Daft Punk too.
Radio-ish, i’m fine with this.
@hakanto I like that radio.garden site, too bad they messed up ssl support.
As a listener on you know which platform … I value the links between artists more than the discover playlist itself.
So over there in that world you can jump between “similar artists” that are more nuanced than it might first appear. Sometimes it’s driven by collabs. So in my exploration of electronic dance music I may end up at say an acoustic cover of Dreams by Lissie because of Morgan Page or Kommode from Royksopp and those two styles of music are radically different but linked by a collaboration.
It’s definitely something evident where musicians collaborate with vocalists.
There are local scenes in real life but more often than not you’re going to see leaps in geography of thousands of miles that can link a random Welsh town to a beach in California.
Music scenes themselves are very incestuous. So if you’re trying to explore a genre often the navigation of that scene is who worked with who.
The best way I find these links in a genre? Wikipedia. To that end you don’t need an extensive catalogue because that database is universal, it’s wikipedia. Think about the experience in a record store and crate digging, more often than not you’re engaged in a conversation with a “linker” someone with some knowledge who can say oh ok so you should listen to Snowmine, and you never would have got there without that interaction.
Yes +1 for Wikipedia! This is also topical because in the works for some time we have had a design for a ‘feeds’ page: As a user (listener / artist / label… anyone) can ‘subscribe’ to others’ profiles and their music in different ways: follow / recommend so they can share their interests and also get a feed of ‘events’ - new public playlists, artists they recommend. See this issue on our Github.
We also have a generic way of capturing ‘member of’ / ‘works with’ relationships and of course links to other platforms, or wiki, where users can put down the ‘breadcrumbs’ to be found by visitors to their profile page.
This could become something of a ‘social graph’, where privacy becomes a big concern, so we are thinking of doing this in a privacy-respecting way, with privacy permissions each way between the followed and the follower, and control over what is shared in ‘public’.
We have to be realistic though… wiki is a huge (and ethical) source to tap into and there is no point in recreating it, or expecting users to maintain it when they have already done a good job on their wiki page or their own website, which we would simply link to. We can perhaps supplement that experience with tools to navigate our own co-operative community in a decent, ethical way.
I wonder if we could/should look at Wikipedia as a preferred ‘discovery partner’ in some way?
Discovery could also be powered by the creative passport. The thing I like about wikipedia is that it’s more organic and in-depth. With biographies there’s an enrichment also in getting to know artists.
What about some tensor flow thing that could pick names out of an artists biography? With a possibility to date an era of an artist also … Take Ani Difranco, going to the start of the discography there are different intensities compared to later era’s of that material.
Something my wife said to me, she likes Pandora vs Spotify because Pandora seemed more organic and less trapped in a bubble.
I wonder if that kind of enrichment of artist information could be possible here.
Yes … Exactly the sort of thing that @afuturemodern was describing … harvest the external context of the music from many sources as well as the music tempo/mood/‘genre’ and use this to support an AI-augmented discovery that YOU control. This is very different from a discovery experience that is the product of an AI weapon that is pointed at YOU, the listener, and uses all the things it knows about YOU and your social graph (friends that it knows about) to make suggestions for music that it thinks PEOPLE LIKE YOU like, or that nudge buying behaviours for YOUR SEGMENT that marketers want.
So many cool ideas and concepts being thrown around here!
Someone mentionned at some point in the previous posts older websites with features to browse “never seen” or barely viewed/heard content. That could be a cool filter to add to a “discovery”/exploratory search tool, or simply a cool idea to build a feature around, giving a chance for some fresh or obscure material to have a turn in the spotlight.
Actually, to go a bit further, I’d wish for a distinction to be made between new music with a low play count, and music thay’s been sitting there for a while, also with a low play count.
As a musician who makes pretty niche music that I think takes from too many genres/sounds to be effectively promoted, that’s a feature that would help my music reach potential listeners looking for new experiences.
Interested in exploring this idea further too, seeing as a comprehensive wiki usually features on discovery sites like last.fm and similar. I do remember reading though that Wikipedia had a bias problem (and this article is a good example). Do you know if they have a better handle on that now and whether it would potentially affect using them as a discovery partner?
