I've been working on a project where I'm building my own 'Echo Nest' type of API to feed it. I didn't want to rely on their API since I felt I could offer something better/different.
It's extremely incomplete at the time, but I'm still building the API the hard way: scraping, analyzing the data, connecting the different sources, and making sense of how it will all go together.
The end result for Musicplexer will be something more than bop.fm, spotify, pandora, etc... I am planning some differentiating features that I don't want to discuss yet! :)
Btw, it's built with Meteor/MongoDB & Mojolicious + EC2/DigitalOcean
And... if enough people visit it, it'll probably die. ;)
Does Echo Nest have any patents in the area of analysis?
That would be my concern about this deal. Not just for you, but for the whole area. I hope this doesn't become more grist for our borked software patent system, wherein this deal mostly benefits only West Texas lawyers and their car dealers and realtors.
Wow, that's a generic list indeed. I may need to come up with some non-scientific way to tell which songs go well together... Maybe I'll have to take into account the current alignment of the stars.
Rdio's recently launched Stations feature (complete with "Powered by echonest" logo in the corner) is brilliant. It'll be a real shame if it goes away.
That would suck, but I think what makes Rdio the best of the bunch is it's people-powered discovery system. I don't use the Stations feature and it would probably do just fine without it.
Spotify has said they'll honour all standing business arrangements, so I imagine Rdio will have time to move to another provider or try and bring it in house.
Really happy for these guys. :) I consulted with Norwest for technical due diligence during the last fundraising round[1], and was super impressed with what this team has built. Audio classification at scale is a really difficult thing to get right.
They're also one of the few companies who have successfully navigated the minefield of the music business, which takes a painstaking amount of focus and determination. But above all, they created a remarkable service to powering the next-generation of music experiences. Spotify seems like an obvious fit (especially given the Spotify Platform[2]), and I'm really excited to see where they go next.
As an aside, The Echo Nest has been a huge supporter of the start/tech scene in Boston. For 5 years now they organized "Music Hack Day"[3] which brought together musicians and engineers to build fun projects. Here's hoping they keep it up!
I think this is a disaster. It's an awful, hideous, stupid (stupid for the wider industry) defensive acquisition on Spotify's part, paid for, presumably largely, by stock - I doubt they really have the cash to buy companies of the scale of The Echonest, who have taken close to $30m in investment.
It destroys value in The Echonest - and every other non-consumer-facing music service out there. The beauty of The Echonest was that they DIDN'T sell to punters. And now, they've been bought in a deal paid for (assumption on my part) with made up value, based on an exit some time soon. It's basically a huge signal that Spotify is preparing for an IPO, which will help the major labels (considerable stock holders - and stake holders) realise a return on their early 'in-faith' investment in Spotify - which has been massively rewarded by unbelievable ($100m+) catalogue advances and equity. Streaming doesn't make money. It will - one day - but right now, there is no certainty that Spotify will come out on top, or even as a player in this space in five years, ten years, whatever.
The whole streaming thing has been the biggest wholesale transfer of capital from investors to labels that the music industry has ever seen. Spotify is now "too big to fail". The advances the labels have taken, and the equity they have been given means that it is near impossible for them to let the streaming model of doing business fail. Even if it's currently cannibalistic of their other revenue, and even if there's a huge PR backlash from artists who don't understand that consumption models have changed. The Majors have been given enough actual cash that they can afford to stick it out.
The Echonest is one of the most exciting companies - if not THE most exciting company - in the music space today. They've taken approaching $30m of funding, and I have no idea what Spotify had to pay for a company who are presumably fighting off acquisition on a day by day basis from people who want to become players in the music industry. There are about ten buyers I can imagine would have wanted to beat the highest bidder.
Beats bought Topspin yesterday. A completely pointless acquisition, based, presumably, on Ian Rodgers' history with that company. Today, Spotify announce their acquisition of The Echonest. Neither company has ANYTHING exciting - on a day to day basis - to report (woo! new interface! woo! more shit headphones! woo! an artist partnership! which our competitors announce tomorrow! woo!) - and apart from the lacklustre launch of Beats earlier in the year, no one in that space has had anything exciting to report in ages.
Streaming music is BORING. It's now pretty much just a commodity service, and it's becoming a very dull market, where the major players have nothing to distinguish themselves. The land grab is over, it's about who can sign what partnerships (people talked about Metcalfe's law with Whatsapp, but there's something similar at play here now - how can we partner with to bring massive numbers of consumers to our service? The landscape changes after Nokia and BOinc fucked that one up, but it won't be long before operators/ISPs try to get back in that game as partners rather than providers...)
CURATION! DISCOVERY! ACCESS! It's boring as hell. How do you announce something and get some press attention? - buy something that IS exciting!
We're looking at a race to the bottom.
Spotify have two or three options with The Echonest; leave it as it is and let it power their competitors discovery engines; take it in house, shut it down for everyone else, and steal a brief competitive march on the other guys; absorb it, keep it as it is, but charge shitloads for the API to other services, generating a real (non-consumer) revenue stream for the company. Any of these, though, it's just sad.
Music streaming is not innovation - it's reached the point that it is just another commoditised service now, and very soon we'll be at the stage that consumers just don't care where their music comes from; it will just be another bundled offering from infrastructure services (mobile, TV, ISP), or businesses like Netflix and Amazon, who are destroying conventional content payment models from the outside in. iTunes Radio, anyone? iOS carplay launched ONLY with audio partners, and iTunes want to steal that market 100%.
Absolutely and without doubt, when Amazon Prime bundles in a streaming service, I will cancel my existing streaming provider contract - because, unlike Lovefilm vs Netflix, there is absolutely no catalogue differentiation. If you're streaming music, you stream ALL the music, not a bit of it - and that is the way services have been built, not least because of the way the whole digital supply chain in music works. You can't have a BIT of the catalogue - you have all of it. The labels (hello, TV & film studios - wake up!) saw this very early on (or, at least, very early on in this part of the emerging narrative) - the more people who supply their content, the better, and the more money they and the rights holders make.
There are a few really exciting companies for whom The Echonest would have been a perfect target - people with brilliantly complementary data sets that could dovetail with their business. To sell out to an uncertain-futured company like Spotify is just sad.
^^^Nice way to seize the moment.
Interesting website concept. Just a suggestion for the homepage: since the top header is so tall, how about moving the play song links from the bottom of each song-artwork to be a universal play symbol (arrow) overtop the artwork, maybe increasing from translucent to full opacity upon hovering, noting contrast issues with relevant artwork. NOTE: at my 1080p pc monitor and font size settings I dont even see the song title 'above the screen's fold', so you may even want to think about that as well, maybe the same solution as the playbutton using overlay and opacity settings. It is a cool header bg though so dont mess with that, except maybe moving the search box onto it since having it below eats vertical space from the song listings as well.
Last point/question: I'm assuming the $1600 headphone giveaway is for the musicians, not listeners. Good growth tactic but maybe make it more clear OR offer something to listeners as well. I don't know much about the song development process market but maybe implement or communicate something about whats expected/gained from listeners for improving the music (assuming you actually want to help musicians rather than eat their souls lol)
EDIT: minor ui change to song pages: move everything on the separate login page as an embedded form on song pages (above the comments). No need to ask me to click and hope you've integrated with my favorite 3rd party.
EDIT2; Startup based advice: you might want to look into the business model of that site that advertised on thepiratebay using popups autoplaying (fairly decent) music videos of indie artists competing for a prize. Cant remember the name or find it on google, maybe it shutdown? Had a one wordish name derivied from video, definitely started with a V
It's a nice demo. I struggle to see why anyone beyond you, the artist, and a few hardcore fans would care though.
Like I said above, streaming is boring, it's a commodity service. Recommendation and curation are part and parcel of it; that's all you're doing, with a twist. The Echonest drove that. Do something else in music. Something that actually matters. And ideally makes us all BETTER.
And quite apart from all of that, your $1600 headphones are about as good as the bargain basement DAC in your run-of-the-mill laptop, for the average punter. So don't give them away to someone who just doesn't care.
So, who will come along and provide an alternative to the likely hundreds of competing streaming stores/platforms that would likely prefer Spotify not to have their data? Would a crowd-funded music tagging system in the style of Pandora's Music Genome Project have any chance of success?
I, for one, am ecstatic. I started a project to catalogue every single live music event on the planet, no matter how small or big. That's one less ogre in the way. Enjoy your new leashes, Echo Nest.
sam, what do you think are the exciting area's in music and tech?
Is it the advent of HTML audio/canvas to get rid of Finale/Sibelius and move scorewriters and scoresharing to the cloud? Or maybe even live coding a la Ableton but on HTML5 audio?
Or web-based collaboration of musicians and mixing of different track?
Or more back-end AI and Big Data of acoustic analysis and AI will we can hopefully get some auto-generated Miles Davis/Charlie Parker comps like those CS paper generators?
The fact that so many of these online services are basically using The Echo Nest to build their "music discovery" engines (aka their radio option) is a bit unnerving. What's there to differentiate them if they all use the same methods to pick what song they send you...?
Also shameless plug, I got really tired of Spotify, Pandora, last.fm robotic stations and built (http://www.ssradio.me). 100% NOT powered by Echo Nest.
It depends. EchoNest has a similar track feature for any given track, which could trivially be used for playlisting. But it also provides an incredible amount of data. There are countless ways you could create different playlisting algorithms with their APIs.
I was a QA intern at Echo Nest in Somerville over the summer and was super surprised to see this today! They're a really a fantastic and fun group of people and doing incredible work, having a fun time while they do it.
(Speaking, naturally, only my own opinions here:) I'm excited for the possibilities of growth that it'll offer the company, but also a little nervous about how it'll affect the Nest's ability to serve their other (previous?) clients. I personally signed on with Rdio because I loved the deep Echo Nest integration—I hope that's not going anywhere.
Looking hopefully, but anxiously, towards the future.
Hi, I'm curious what are the other applications of Echo Nest API besides music recommendation? Can it assist composers or performance artists in composition? (e.g., input scores of Charlie Parker, output a solo on key of X based on previous known solo's). Can it identify harmonic structures in music? (e.g., which mode is used in song X and in a composition, when a key is changed). I also live in the greater Boston area and has always viewed EchoNest with curiosity, so very curious what you can do with their API other than their "acoustic fingerprinting" (e.g., genre/bpm/key/mood/color etc.)
The API has some excellent stuff regarding audio segmentation. It represents any track as a list of ~2,000 segments. Each segment has energy in 12 chromatic bins measured, too. There's also information on beat times, and a Machine-Learning-based estimate of verse/chorus/solo and even more great stuff like that.
Full on auto-transcription like you hint at is a hard AI problem, like speech recognition, which will never by fully solved (according to me), but I, for one, would LOVE to see a musical editor enhanced by EchoNest metadata. One of these days, i'll add to to Audacity hahaha.
BTW, if you are in Boston area, come to the Music-Tech meetup. It's a "gem" and I'll be happy to intro you to the EN API (I worked there 6 years ago as a consultant).
Nope, though you may be interested in David Cope and Experiments in Musical Intelligence. Input Bach MIDIs, receive 500 compositions in the style of Bach. Turing complete.
Echo Nest has a few other fun features, like danceability, mood, tempo, and others. An API playground, too, for now.
I hope they don't. Last.fm is an app inside Spotify right now, and scrobbling is integrated. All I wish they did is better integration (love tracks, recommendations coming from last.fm instead of whatever they have, etc.), all of which can be done without having to buy last.fm. Last.fm as a standalone service for discovery is good as it is.
Last.fm has a great neighbor-based recommendation engine. Combine that with Echo Nest's user-based machine learning approach and Spotify's user base, and you have a pretty robust product
Oh wow! These guys are in Davis Square in Somerville not far from me...I've been working on some idea about the same thing for 4 years now...back then people would laugh at the idea because Pandora was already there. But now?
I had just started to dabble with python and echo nest (the remix API if I remember correctly) and now I wonder if their system will still be available or if now it is going to be proprietary/closed.
This is very interesting, correct me if I'm wrong, but Spotify was one of the power users of echo nest's APIs. Maybe they had to use it so much that they had to acquire it.
Rdio also uses The Echo Nest for their radio feature, and it's great. I always chalked up Spotify's radio sucking vs. Rdio's radio being so good due to Spotify having shitty integration with The Echo Nest... maybe that will change. One thing that always struck me about Spotify is that they never seemed to put a priority on anything around actual music discovery.
I think it's crazy that echonest still doesn't have any album information in their api, for building anything to do with music when is album info not useful?
Maybe they don't want to duplicate the effort of MusicBrainz. MusicBrainz is about warehousing preexisting metadata (track listings, credits, etc.), whereas EchoNest specializes in using algorithms to extract novel features.
Yes, a good example of sticking with an idea and growing it slowly and steadily for years and years and years! Also, of pivoting and sticking to business-to-business solutions.
Earlier on, their projects involved creative remixing of music, but they extinguished them in favor of supporting other businesses [1]. We should all take a cue from them!
I've felt the same but the syllogism path is hard to explain. Basically, many of these services are founded the same way Amazon was founded - completely unsustainable and then hoping something magic happens to make it sustainable later. Amazon wouldn't have survived without the dotcom scene going crazy and they barely got through to the other side. The music services exist by seeking to convince end users that they are entitled to "mostly free" music, and they only reason they can supply that service is with debt funding. Their hopes is that the appetite they instill in the end user will give the business enough support to argue favorable licensing terms from labels and the government.
So when that happens, the party that hurts is the songwriter. You already have way too many people believing the canard that an indie songwriter should only expect to make money through touring and merchandise, and never from their recording efforts. That opinion is self-serving and pushed by those services "use your songs as a marketing expense!" when it is actually supposed to be the musician's livelihood.
So what happens when a musician gets less and less likelihood of getting any kind of financial remuneration from their actual songwriting? It's a disincentive to focusing on original craftsmanship, and an incentive to focus on churning out greater volumes of homogenous crap in an effort to capture a very tiny piece of a very large homogenous pie.
The end result from the perspective of the listener is that they never experience the counterfactual; they never know what they're missing. Sure, you might be enjoying that latest "indie" piece that is really just a max martin form with a decemberists diphthong singing tone combined with a couple of dubstep beats, but is that really original?
You've got some good points, but I actually believe it's a bit bleaker :)
I believe the role of the musician is too important to be simply cast as a typical capitalist job- e.g. do it 40 hours and get a salary. Although I'll accept this until the rest of sedentary culture reforms (not likely in this life).
The problem is our celebrity culture. We turn our efforts away from the 1000s of less famous musicians around us to the handful of celebrity musicians.
Classical music culture, with its reverence and deification of a handful of sponsored artists, provides the template for rock music. Hip hop culture successfully followed the template. Jazz mainly dodged it. Recorded media (sheet music and the record) catalyzed it.
Any technological aid to music culture can be judged by whether it leads to the (imo benevolent) fractioning of music - the re-directing of attention toward our immediate musicians - or whether it leverages and bolsters the celebrity culture. Any capitalistic enterprise is blind - or farcical - to these ends and will randomly support whatever ends lead to its own survival. However, this randomness must be biased toward existing resources, e.g. if not doing well, support the thread of celebrity culture to survive. This pattern continues for the individual musician, too, as you point out.
So, from this perspective, I appreciate services which connect individuals with musicians of their choosing, but it's not like connecting people to Primary Care Physicians, a geographically-constrained problem. Instead, it facilitates people following their own flawed decision-making processes, following the crowds. So it becomes an issue of individual freedom, which I adore, even though so many people will not use it optimally for themselves.
Ihat's why I listen to soundcloud exclusively. The artists there have very original material and you can see the music they are liking, playlists they are curating, and general community engagement.
Given that "the death of music" has been variously predicted to be caused by sheet music, player pianos, radio, home cassette tapes, MTV, MP3s, Napster, CD-Rs, Bittorrent etc I think Spotify may need to get in line...
It's too bad you were downvoted to oblivion, but a more reasoned response might have nurtured more conversation.
Anyway, having read lots of social theory on music, I'll bite. Go ahead and tell more. Specifically, death relative to what culture?
It's my opinion music as a practice before the sedentary age of agriculture will probably never be recovered. It is currently just as subject to capitalistic practice as health care and everything else. I don't see Big Data's leveraging of music for profit as being significantly worse or better than previous generations' exploitations of music.
In fact, I blame music for leading to its own demise. I believe music helped transmit the patterns and information which helped us become sedentary. I believe today's music rarely, if ever, soothes or nurtures us. There is another, truer form of music which attains that, but we have either lost it entirely or it is a well-kept secret.
I feel like I'm drowning in good music now thanks to the net. Soothed, nurtured, confused, educated, lifted...there's no limit to the number of verbs that the plethora of music I have access to now does to me.
Spotify and similar disservices mean death to a culture of music.
Care to expand on what you mean by that? I'll agree with you that the music industry is deeply messed up, but it has been for a very long time. If anything services like Spotify will allow artists to abandon parasitic record labels and release material by themselves.
People halo around the most visible artists. The various services promote the most popular artists, because that's what their audience wants. As the model shifts from a lean forward (I buy music that I like, so that I can listen to it) to a lean back (curated, data/discovery) model the top line hook for fairly un-engaged music consumers means that they discover less new music and hear more of what they expect to hear. Which is exactly what happened to radio, for exactly the same commercial reasons, over the past couple of decades.
Really? Because I have used Spotify for 2+ years and I discover a lot of new music through it. Not the kind of stuff that I find on the radio either. Being able to start a 'radio' with a specific song brings out some great stuff you wouldn't of found before
