I've been planning to do more write-ups from past hackdays, such as building a functional Shazam experience in client side JavaScript and other visualisations. For example one time, I repurposed a WebGL globe to plot music recognitions over it. I don't have access to the Shazam tag stream anymore, but you can get a feel for how this looked as I've applied near real-time Wikipedia edits on the globe instead: https://umaar.com/globe/
For the data maps which the Medium article displays, I used: https://github.com/ericfischer/datamaps initially. If there's interest, I can try to get my scripts onto GitHub. They:
a) Include Node.js scripts for converting raw big data (the raw Shazam stream) into something usable for visualisations.
b) Contain Node.js scripts & utilities for automating the creation of maps focussed on particular geographic locations.
c) Contain client side JavaScript for displaying all the individual map tiles onto an interactive map & Google Maps API code to query for places of interest (like a club, shopping center) which may explain heavy usage of Shazam in a particular location.
Thanks for posting this! I've always wanted to build an audible ad-blocker that can mute/replace the audio and analyzing a background/silent feed to restore volume. Unfortunately all product paths I can forsee will surely lead to targeted replacement ads, played at higher volume
Shazam is the one app that has been around since about the birth of the iPhone that honestly still blows my mind in it's capabilities. The fact it can so quickly figure out any song in so many environments is incredible. Call me simple, but I use it almost daily, and it is accurate close to 98% of the time. I absolutely love Shazam's iOS app, and they have done a great job keeping up with UI/UX treads for it over the many years it has been around. Easily one of my favorite and most used apps.
With all that said, PLEASEEEEEE don't ruin it, Apple. Specifically meaning I love it's integration with Spotify. I will be bummed if that goes away for an exclusive Apple Music integration.
> With all that said, PLEASEEEEEE don't ruin it, Apple
I don't think the overall UX, song recognition, tag tracking will be ruined
> Specifically meaning I love it's integration with Spotify. I will be bummed if that goes away for an exclusive Apple Music integration.
Well that was pretty much what Apple paid for. I can already see that all my shazam tags automatically show up in Apple Music for me. Clearly this is just becoming another growth channel for Apple music which means their Spotify integration will likely be diminished or end.
Yeah, Shazam was the last Zero to One moment I had with Tech. A true, visceral in-the-moment 'wow' that no social media giant or app has come close to emulating.
The user value to time investment ratio is off the charts for Shazam. I love music and what they have done for it, especially in the context of other music apps like Spotify et al.
Second the Don't Ruin It Apple plea, but congrats to the creators on a well-deserved exit.
Google Now natively supports song detection in surroundings. It is also pretty good at it, and Google will do it passively without ever needing to open it as an app.
I used Shazam a long time ago, so dunno how it had changed, but it certainly isn't a novel idea.
Just to jog people's memory, patent threats were sent to others doing similar things [0]. Not sure if that represents Shazam or not, but if so, they'll fit in perfectly with Apple.
I agree with the lauding of Shazam, but I am interested in your statement "I use it almost daily" -- what for? I've always viewed Shazam as a novelty and have installed and uninstalled it a few random times, but as Jobs would say it seems like a feature not a product.
As a DJ constantly looking for tracks to add to the collection. I listen to streaming music pretty constantly. For individual tracks with all of the metadata, the artist/track name info is provided. For listening to other DJ mixes, you just get the DJ name and mix name. Shazam tells me the specific track.
I also use My Shazams as a potential mix list. If I have an item in My Shazams, it caught my attention for some reason. Just putting those tracks together into a set has worked out pretty well. It gets tricky if I'm using my phone as the streaming source. At that point, I usually just take a screen grab. If it's an individual track, I have artist/track name, and if it's a mix, I can use the playhead position to narrow down which track it was later.
Mixcloud actually had that. Then they've switched over to only displaying the current track. Then they've made some shitty version of music recognition similar to Shazam, which simply does not work for this use case (every single time it detects something, which happens almost never, it disregards that it's an edit/bootleg/remix of a song and displays the original instead).
In fact, as an uploader, you can still see it (I don't know why), but not as a listener. As an example, here's a screenshot of me viewing my own upload: https://imgur.com/a/REpxOgx
There's also a browser extension that modifies Mixcloud pages to display the tracklists again, but they're useless since no uploader has an incentive to go through the painful process of entering timestamps, and that process is now crowdsourced anyway thanks to 1001tracklists.
>As a DJ constantly looking for tracks to add to the collection. I listen to streaming music pretty constantly. For individual tracks with all of the metadata, the artist/track name info is provided. For listening to other DJ mixes, you just get the DJ name and mix name. Shazam tells me the specific track.
So a pretty narrow demographic for the feature though...
>So a pretty narrow demographic for the feature though...
If you're only going to take one person's usage then, yeah, I guess it's pretty narrow. How about another example for from the same person. As a video editor I'm constantly having to find "hip/trendy/cool" music for video edits. These clients typically like music that I do not normally listen to, but I know it's something they would like when I hear it. I'll let it live in My Shazams until one day I need to find something. I've impressed more than one client with the ability to pull out a "fresh" song for them. Clients appreciate that sort of thing.
>How about another example for from the same person.
How about it? We've just added video editor looking for cool music to DJ. We could add tens such cases more, but all those combined are still narrow demographics.
Diabetes apps operate on one of the most lucrative markets (health), and are essential and even life-saving for their demographic. Not a mere nicety for some creative professions...
Also, diabetes narrow? 30 million have it in the US, I'd say those are way more than DJs that need to recognize tunes, and have much more motive to use an app that helps them keep track of something that can have huge impact on their health...
I'm not a DJ, but i listen to electronic music mainly and it usually is on soundcloud in the form of sets. If 1001tracklist doesn't come through, Shazam is my go to app.
This "feature not a product" belief only makes sense if you pretend that the product doesn't provide the answer sought by its primary user and as byproduct produces a truly real-time stream/map of what music-enjoying people are interested in finding out more about.
This data would be incredibly valuable to companies trying to market and sell music. Like labels, and now, Apple. Not to mention keeping that data now out of the hands of others.
I mean a feature in a sense that it seems (although I assume I wrong and thus am asking) like something that one uses extremely rarely and offers limited utility. I have used Shazam a few times, and each time quickly install, use it, and remove it. I would have used it more if I could just ask Siri or Ok Google what a song was.
There is another reply to me that snarkily opines that the average HNer doesn't engage in social events, which is a bit humorous really because a more correct statement is that most people are socializing at social events, not doing inference on background music. I can say that I have never, ever seen another person pull out shazam in public.
You can ask Siri what the song is. One UI flaw is that music app is open or playing on your phone, she will name the active track. Otherwise, she will say ‘Listening...’ and then identify the song and artist, I think with iTunes links (or course!).
I don't use Shazam since I have a Pixel 2 that does the same thing in the background, but I use track recognition every day to automatically populate a Spotify playlist with all the music it recognizes throughout the day. It's really nice to have a playlist to scroll back through when I find myself asking "what was that cool song I heard yesterday?”
I can't speak for who you're replying to, but I find myself using it at restaurants, bars, and public locations where music is playing, as well as friends' cars, and other social gatherings. I know these sort of situations are rare for the average HN user.
I don't go out much, but when I do, most music I've experienced seems to be tolerable at best and ear-assault at worst. I can only think of maybe 2 or 3 times in the past decade where I've heard music in a public location and was interested enough to want to know what song it was.
Have you ever heard about the concept of the "joke?" Novelty thing, I tell you. Admittedly, it might be a bit too recent of a concept for the average person to be aware of, especially with how radical it is.
Here's a recent book of them, to help you get "in" on the "hip slang" of tomorrow:
An Abderite sees a eunuch talking with a woman and asks him if she's his wife. The guy responds that a eunuch is unable to have a wife. "Ah, so she's your daughter?"
A correct statement based on three wrong assumptions.
1) That I've lost interest in music myself (I didn't).
2) That whether some other people might not lose music was disputed. Perhaps you imagined me as some strange creature in a cave that has never heard or known of people that maintain their interest in music to their late age.
3) That the fact that some other people don't lose interest in music is relevant. I was making a general statement ("it's also rare"). Most people do lose interest in music compared to their younger years. That's not some fringe claim, it's borne out by buying patterns and studies. People tend to stick to the music of their youth time and again. Even the new artists they discover tend to be of the same genres and styles in vogue when they were younger.
Even if so, why would that mean older people have no use for Shazam? Have you never heard an old song you half-recognized and thought “huh, what band is that again?”
>why would that mean older people have no use for Shazam?
It doesn't mean "older people have no use for Shazam".
It means "as it has been observed statistically, _most_ older people don't seem to have a use for Shazam and any random older person is more likely to fall into that category than the rest of the older people".
Carl Craig is old too.
>Have you never heard an old song you half-recognized and thought “huh, what band is that again?”
Yes, but this is not such a large part of the older demographic's lifestyle for the app to matter much to them.
Old people also want to send their dickpicks too, but they don't flock to Snapchat.
I would dispute 3, listening to the same genre doesn't imply you lost interest in music, because being interested doesn't comprise only of listening to the different genres. According to your definition if you listen to the same genres 7/24 your whole life, it means you lost interest in music after 30?
Listening to the same genre means that you're not as willing to search anymore. So even if you're still interested in _your_ kinds of music, you're not interest in exploring the musical landscape as much as you did -- your preferences have solidified.
But, honestly, I don't know how all this is controversial, unless someone is 20 and all their friends are too.
I'd expect anybody over 30 or so to verify that even if music is still a big deal for them, it's not as much for most of their friends as it was in their teen years and early 20s.
Unless you're 30 and all your friends are eg. DJs and techno-heads or heavy-metal fans still rocking those leather pants, the usual course is that they are more interesting in their careers, families, new babies, and binge watching TV and surfing, than following music, even if that was a big part of their identity in their teens. Is that inaccurate?
Soundhound is comparable in terms of detection quality and probably will keep their Spotify integration running in case Shazam locks people into Apple's ecosystem.
While I find Shazam to be amazing, I find SoundHound's capabilities even better in terms of identifying a song based on humming the tune or hearing a live version.
I was at a live jazz concert last night, and the trio started playing a riff on tune I recognized but couldn't exactly place. SoundHound nailed it (Sonny Rollins "Four"), while Shazam had no idea.
That's because that's not how Shazam's recognition works (unless they have a secondary mechanism). Shazam is very sensitive to time and frequency features of specific recordings, which is also what makes it so robust in noisy environments. Having both apps is handy for their relative strengths.
Birth of the iPhone?! I remember using Shazam on my dumbphone driving home from university. That must have been 2002/3. Blew my mind. Pretty incredible really. Good for them!
I just hope they don't shit on the android support like Snapchat. It's the first ever paid app i got and i'd like to think that's saying something for an android user. Probably far fetch but it would be nice the android version became open source if that should ever get dropped by apple (so community can maintain it)
I can't really be bothered to install an app for this purpose but I'm constantly disappointed that Google's assistant is so bad at picking up music. I would say it works less than half the time. It only works in public areas if it's pretty quiet.
My experience with shazam is that it finds about a quarter of the tracks I'm looking for. It's really really incredibly good at finding anything that airs on the radio or exists in spotify. But outside of that the results drop off quickly.
Yes, that's my experience as well. I listen to and discover a lot of obscure, often rare music and Shazam isn't very helpful in these niches. Although that might be expected.
On a sidenote: It's surprising that Google isn't more useful in those searches either. Often I have to search for rare releases by relying on lyrics and Google's results are plain terrible. I understand that the lyrics are probably nowhere listed and idnexed, but considering that even the rarest releases are uploaded by someone on Youtube, some kind of auto transcription feature would be very helpful.
Sure, it's a niche interest, but the Discogs user base is still huge.
Is Shazam really ML/AI/DL? I thought it was more of a signal processing and search engine system. You transform snippets of audio signals into text, pump it into a distributed database, and then perform a search against that.
Maybe there is some AI involved in noise reduction, but that seems like a rather small component.
"Music information retrieval" is what researchers call the overarching field. You might say the Shazam algorithm is akin to those fields in the sense that it relies on feature extraction. It's possible they may train a classifier to do the matching, but yeah, probably not purely ML/AI/DL.
They do something that gets results. It doesn't matter if it's ML, AI, DL, linear regression, or just fingerprint+cosine similarity. There's some talent for getting some complex shit done.
Second this. And not just for english music. I have lost track how many times I have pulled up shazam to identify a bollywood song or an Indian regional song which I recognize vaguely from my childhood. It has always recognized it.
does Shazam also sync up lyrics? I havent used Shazam in years since I discovered Soundhound. Soundhound has a better ux imo and it also brings up the lyrics for popular songs and they are synced to the music as in they scroll by at the right place like karaoke
Tl;dr You're kinda boned. Apple Music integration will get better, but Spotify should be looking for an exit - before they get kicked out the door.
Not saying I agree with this, but given how most acquisitions go, other services don't survive long. Apple Music has an Android app, but that's more of a halo effect kind of thing to get people in the ecosystem and (eventually) switch to iOS.
I frequent mostly heavy metal pubs and for the last 5 years neither shazam nor sound hound has detected a single song for anyone near me. Complete crap. We've been using the ancient technique of asking the DJ on the way to the bar.
Some genres are definitely trickier to identify than others. EDM and heavy metal being two of the most difficult for example. I find the short segments are a little less unique. Try letting go the app get the full 10 seconds - the longer the better. Source: I used to work on algos at SoundHound. And I like these genres too :(
Shazam is the closest thing we have to witchcraft. I'm sure plenty of HNers will chime in with "well acktually..." type comments about how they could write it in their basement in a weekend, but to do what it does, as quickly as it does, on the ridiculously enormous library of music that it does is really something special.
I agree with you, and I work in music audio signal processing.
Shazam was founded in 1999-2000 ish and was live by 2002, longer ago than many people realise. The fingerprinting method they used was a novel one, and they successfully deployed it in a couple of years, with a database of two million tracks, in a system that could be used by dialling a number from any phone. It's an extraordinary and rare example of taking a research method and getting it to work at scale in the face of real constraints, in the process producing something that most people would never have imagined could be done.
I was surprised that they didn't use a Jaccard distance when I looked into the paper, but I never really tested out why their more complicated method was required. The core insight (that note timing / beat is a fingerprint) is money, but the rest of the math seemed unnecessarily complicated.
Would really be interested if someone has tried other approaches and can comment.
Me too. Are there any similar startup with functional prototypes or products? Something like Shazam, or Theranos if it were actually legitimate and had a working product.
Philips had similar technology that predates Shazam. That said, Shazam's algorithm scales better as they use "combinatorial hashing". This coupled with the fact they can up the hashes in a key-value table makes them better.
This is a really nice paper as well, but the Shazam method is more exciting because of its focus on the search part of the problem. Audio fingerprinting wasn't an entirely new field of course - the Philips paper also cites several prior publications.
The big limitation of all these methods is that they are robust to EQ and other degradations but not at all to variations in timing between frames. So they work well for identifying different instances of the same recording, but not at all for matching different performances, or for "query-by-humming".
After I learned how it roughly works, it seems no longer like witchcraft, but it is still impressive.
The remaining thing I don't understand now is how the heck they managed to negotiate the license for all the songs they have in the database. A friend of mine had a major website for lyrics, and he got some legal trouble from the record companies. When he tried to buy a license to display the lyrics, after some back and forth, they basically said they're not interested in licensing. This was in the early 2000s before streaming took off. Since that time I have the impression that often legal and social problems are much harder to solve than technological ones.
They are likely able to show that the DB is comprised of simply the fingerprints, and not the full songs - thus they wouldnt be able to "use" the DB to profit from play-back of the songs it can identify.
my recent experience with shazam is that it's much less accurate than Google when picking out random background music from a noisy area.
that said, i still prefer shazam because it's an actual app, whereas google's implementation is a button that sometimes appears in the google app, and sometimes doesn't, seemingly at random.
I don't mean to take away from Shazam (because I too was amazed when I first used it) but the Pixel phones have Shazam-like ability built in which uses minimal battery and without internet... and it's constantly running any time music plays.
Not really "magic" how it works but still pretty amazing that I can pull the phone out of my pocket any time I hear a song I like and immediately see the band/title.
Coming soon, Pixel will be able to recognize over 1 million individual voices in your area! Using our enormous database of phone calls to your area code and individualized location tracking, you will be able to find out everything about a person simply by holding your phone near their voice!
That's really sweet—I wonder how much storage that requires? I would have expected it to require a lot, but if it's on a phone, then maybe not that much?
Google used to have song identification via voice search and a widget on all Android phones since Lollipop. Then they killed it for no reason other than to screw over existing users. Widget is still there in Nougat but doesn't do anything.
what, the functionality is still there in Nougat, I access it mostly through the assistant which is good enough to detect a song is playing and suggest What is this song? and has great accuracy
I haven’t had it misidentify anything — it just won’t identify a song. Granted I do this when in a public place and usually it is something like a piano rendition of another song. I really can’t blame it for not being able to hear the music as it is hard for me to.
Agreed 100%. It's the one app that truly does feel magical to me. And it still impresses people I show who don't use it regularly when we are out somewhere and they ask the age old "what song is this?" Pull out my phone, click one button, and in less than 5 seconds it's there. Blows my mind every-time.
even more mystic in its infancy when it was an abbreviated phone number you would call, hold up the phone mic near the source, get hung up on and then recieve a txt with your song title and artist.. in 2002
I have worked on integrating a shazam like library inside an app.
We have looks at more than 12 solutions before finding a company with a song recognition library good enough to be called shazam like.
It is by no mean a small feat you can reproduce during a weekend.
I mean, it's never worked for me even once that I can recall. I've mostly tried to identify music playing in commercial spaces (retail, cafes, etc) via Siri and I don't think it's ever given me an answer.
Exactly! Maybe if you listen to the billboard hits you'll always find a match, but IME it rarely matches the songs I actually don't know and can't find details on.
Ooook I don't want to be that guy, but if you take a signal processing course in a university, you'd see that algorithms that Shazam probably uses aren't blackmagic. In UC Berkeley one of the projects we did in our Linear Algebra class was doing something like Shazam; the input was a 5 seconds of a random song tuned down so much that human hear can barely hear and white noise added. Then the problem is matching this record with the song and specify start and end time (e.g. 2:30 to 2:35). Shazam is probably much more complex than this (I don't use it) but if college freshmen can do this...
I worked in email marketing where at its core, we were just doing string replacement from database lookups. Now, we did those string replacements over 300 million times a day. As my boss said "Everything at scale is hard".
I only have a cursory knowledge of DSP, but it sounds like this was implemented by using the 5 second clip to build a matched filter kernel. This would require convolving the samples of the 5 second clip with the entire song to find out where it matches, which should work super well, but I think it would be really tough to scale this up to millions of songs. Is this how it was implemented?
Please don't be a jerk on HN, even if another comment seems a bit pedantic.
Separately: there was nothing wrong with gnulinux's comment. It was obviously posted for the pleasure of sharing information, and didn't put down either the GP or Shazam.
Why can't I delete comments in HN? If from the responses to my comment I learned that I add nothing to do conversation, I want to be able to delete my comments. Am I missing something?
For what it's worth, I think it's neat that you shared that the knowledge necessary to build a Shazam-like thing is something an undergraduate might do as a homework assignment. I had thought it would be a lot more complicated. :)
If you're interested, you can search Berkeley EE16A and if you're lucky you'll find the homework assignment (Berkeley's class materials are always open to public). It was like a Jupyter notebook and the "song" was I think random 5 seconds of Universal Declarations of Human Rights, or something like that. I'm writing this in case it helps your search, I'm at work so I don't have time to search it myself. I took this class a while ago and am alumni so my memory is fuzzy about the details of this assignment.
It's mostly pretty standard linear algebra (from matrix operations to vector spaces to eigenvalue/eigenvectors) and some circuit theory (KVL, KCL, op-amps, circuit design etc) and some signal processing (cross-correlation, OMP, Sparse OMP, least squares etc). All EE/CS students take this class in Berkeley, it is strongly recommended to be taken as freshman (unless you take Math 54 which is an equivalent course with no circuit theory). It is followed by EE16B which includes further topics in linear algebra (SVD!), circuits (transistors!) and signal processing (Fourier Series!), and CS70 which includes discrete mathematics and probability theory. This series is usually taken along with programming based classes CS 61A and CS 61B; and after finishing both of these series students start taking "upperdivision" classes.
I think you have a few minutes to delete comments, but as soon as someone replies to you that link goes away.
I think it's a good thing, because often there are good comments in response to downvoted comments. Letting people delete comments would make the discussion hard to follow.
You don't have to, the votes will gradually adjust comment visibility based on its perceived (de)merit. Besides, an unpopular comment doesn't mean it's incorrect, e.g. the content might be useful but the tone irritating.
It creates a time series of frequencies then checks a db for that same series. A problem that is easily parallelizable. Probably does some filtering first. It's a very useful and clever application, but not that technically challenging. The most amazing feat is they got access to the catalogs of the record labels.
I wish someone would do the same thing for commercials on TV. Auto detect the commercial and advance the DVR by its exact duration
That's why TIVO (or whichever DVR company) had the skip 30 seconds button. If it was a 60, you hit it twice. However, they also sell 15 second blocks. Now, it's just fast forward at ludicrous speeds.
Yes shazam's algorithm is aktually pretty simple math, but there's nothing wrong with creating a busines sbased on simple tech!
Like most companies in tech, it's the business part that is hard, not the technical part.
When Twitter came out I was like "well it's a glorified shoutbox, I can hack that in an afternoon". Actually it still is, any developer can create a Twitter clone in hours. However Twitter managed to have the network effect, and (eventually) solve the huge scaling challenge that a system that transfers billions (?) of messages worldwide every day.
Twitter is hard to scale, you have to make trade-offs all over the place.
AFAIK that's not the case of Shazam where a read-only database would be enough.
Sure eventually you have to manage multiple versions of your app and be able to gradually update your DB, but that's still simple stuff compared to the problem of how to make money.
>> Shazam is the closest thing we have to witchcraft.
I've never used Shazam, but I used a similar feature on my LG Chocolate back in the day. How is Shazam any different than what my LG Chocolate did (aside from the Spotify integration)?
Apparently Verizon's LG Chocolate came with a service called VCast Song ID, that did the same thing as Shazam: "On certain phones, users can record a 10-second clip of a song, have it automatically identified, and optionally buy a copy of it for playback on a phone or computer." (https://www.digitaltrends.com/mobile/name-that-tune-with-ver...)
That said, Shazam is actually older, having started as a shortcode service in the UK way back in 2002.
I am equal parts amazed and horrified when I hear that some app started out as an MMS hack on feature phones. Dictionary lookups, directions, and multimedia, all on "candybars". It's one area the US never really got behind and jumped straight to smartphones and discrete apps.
Well acktually, in my admittedly limited experience, Shazam is not capable of recognizing any music at all without the digitized metadata encoded with sound files. I've tried multiple times to sing very popular songs to it. It never recognized them. You could attribute it to my poor musical skills (and I'm fine with that), but 100% of humans would recognize the songs in < 10s timeframe.
As far as I know, the way Shazam uses fingerprints it is not supposed to recognize any song this way. Or even when a cover or a live version of a song is played.
I remember Adam Neely (A musician/jazz composer/youtuber) mentioning in a Q&A, that for songs they played live for a television broadcast, backing tracks were added underneath and it was mixed as close as possible to the originals, among other reasons for the purpose of them being better "shazamable".
It doesn't recognize you singing because it doesn't have the same spectral content as the full song. Shaman doesn't really identify tunes, it identifies recordings. (And it is amazing to me that it works at all)
To the best of my knowledge, Shazam works on frequency spectrum analysis and how it changes over time. This includes all aspects of the music and not just melody.
I once read in an article (that I unfortunately can't find right now) that the the worth of Shazam lies in the fact that it can recognise, which songs are trending. The value chain is approximately like this:
1)New song is released
2)It starts being "shazamed" by various people - it's becoming trendy
3)Large stores that differntiate themselves by being trendy want to play trendy songs first
4)They pay Shazam to tell them which songs are trendy
5)The stores play them
6)Customers "shazam" those songs in the store
7)Customers view the store as trendy
8)Now four instances have profited: Shazam, the store, the musician and the customers (for having found a trendy song very early on)
The UK CEO presented some of their different data capabilities at the agency where I used to work: they were encouraging advertisers to use this data to soundtrack ads.
The different 'shazam'd songs were also connected to locations and user demographics so if you wanted a track that was going to be big in a specific community you could get a good guess.
I’m not sure what the middleman is, or which way the money flows, but there are also services that provide new music to retailers. The retailers get access to the current trendy ambiance, and the artists get exposure in an environment where people are consciously accessorizing. Half the time I use shazam it’s in a store (the other half being lyfts).
A great app that PRE-DATES smart phones. I remember using Shazam when it was a dial in service (dial 2580 in the UK), hold your old brick mobile phone up and you would get a txt message a minute later with the song. Seemed like magic then and still does now.
I was unaware that this was a thing. That's pretty cool, and it's always nice to see a company like that able to adapt and thrive with the advent of smartphones
I remember coca-cola did a large marketing campaign in some parts of Europe with them somewhere before the first iPhone (or at least before smartphones were common).
>You can bet they will kill integration with Spotify though.
Will they? Don't forget, the information flow is 2-way here, users get something useful out of it but Shazam (and now Apple) also learns what stuff is trending. If anything I actually lean towards them keeping or even enhancing Spotify and other service integration, at least as long as those services are viable: it effectively creates a direct channel into competing services that tells Apple exactly what stuff people want out of them, which they could then use directly in making Apple Music more competitive in turn.
Even if we take as granted that boosting Apple Music is a major concern to Apple, doesn't that at least maybe seem like an equally or more valuable way to do it? It doesn't make any users angry either or act as bad PR, they aren't taking anything away, and they get a massive information channel into every competitor and even real-world venues.
>And maybe no more Android support.
By the same token, if Apple wants to boost Apple Music in general then having information on music popularity and trends coming in from the population of Android users too also seems helpful.
Being able to see trending songs regardless of the platform is definitely a benefit for Apple, but I would also say that another possibility is that Spotify itself cuts off the integration now that Shazam is owned by their competitor.
This is immediately where my mind went, the spotify integration is without a doubt the best thing (IMO) about Shazam. Find a song, in two screen taps I can add it to a playlist and come back to it when I'm queuing up tunes for the drive home from work.
I will be very disappointed if this comes to light, and the only integration we're left with is Apple Music.
Probably not likely, but probably not improbable either.
Google Play Music has a Shazam feature that's just as good (though I haven't rigorously tested it, it's never failed me), so lack of Android support is no real loss.
Hopefully they'll rewrite Shazam's Apple Watch app using the private first-party APIs so that it actually works rather than randomly crashing 50% of the time.
The state of third-party apps on the Watch is awful. It usually takes 10 seconds for an app to start on a Series 2, and sometimes they just get stuck on that loading screen. If you manage to get an app running, the UI library has ridiculous bugs like accidentally enabled screen rotation where turning your wrist flips the content by 90°. (Unlike a phone, there's no reason at all for a watch to react to orientation changes.)
I could sort of understand these deficiencies for a hurried 1.0 product, but it's been three years and it's still this bad. There's no API for watch faces either. I get the impression that Apple truly doesn't want third parties on the Watch at all.
If they "truly" didn't want third parties on the Watch "at all" they wouldn't have released the watchOS SDK or made it possible to write, build and publish watch apps on the App Store.
This wouldn't be the first time that Apple releases an SDK, decides internally it's actually not something they want to support, and then just leaves it languishing in a semi-broken state because it would be too much trouble to cancel it outright.
This was essentially announced at the end of last year, and is now being closed.
You can now search for songs on Apple Music by typing in the lyrics to the song.
The part I love about Soundhound is that I can basically hum something to it and expect a good result within a few attempts. Shazam can't do anything like that.
Yeah I love the soundhound app as well. My wife is a professional singer, so often times she can sing or whistle a song and soundhound will pick it right up.
Agreed. Shazam has failed for me too often where Soundhound has returned the result, that the only reason I use Shazam anymore is out of curiosity after Soundhound, to see if it can also find the track.
It’s weird to think that I would find it strange if Apple acquired a company like Shazam and left the ads on, isnt it? In a day and age when all of their competitors are turning to ads to increase revenues, Apple is making great customer experiences ad-free.
I think the idea that making Shazam ad-free “makes sense” is actually fascinating because if it were google or amazon buying Shazam, I would assume it would remain littered with ads.
Yeah it's on-brand with Apple's approach for privacy and the like. Plus, the end-game to me seems like tighter integration with Siri and the AI ecosystem they're building.
It's hardly "littered" though. I use Shazam a lot and always found the ads very unintrusive. It is good news that they're going completely ad-free however.
Apple doesn't really need to mess with the tech behind Shazam too much, I've personally found it to be lightyears ahead of the competition. I switched to Soundhound for a few years back in 2010 when both were relatively new, but the difference is unbelievable now. Shazam can get popular songs in less than a second, it's like some sort of musical L1 cache.
The journal article by the Shazam creators titled, "An Industrial-Strength Audio Search Algorithm"[0], was one of the first CS papers I ever read. It was very easy to follow and I remember even making a personal utility that implemented their algorithm.
I would love to see something like the Pixel's Now Playing feature come to Apple devices. If you haven't seen this, it periodically, passively listens for background music and displays the currently-playing song on the lock screen, using a local database of IIRC 20,000 songs. There's a paper about it (https://static.googleusercontent.com/media/research.google.c...) – but I'd have to guess it's patented and we won't see it soon.
The always-on nature of this is the really killer aspect of it, especially when combined with the always-on OLED screen. If something is playing nearby I don't even have to touch my phone to know what it is - just a glance will tell me what's playing.
Tangent: I recently discovered SoundHound which supports singing / humming / whistling as input for song ID. IME it's not as good as Shazam for capturing "real" (pre-recorded / radio) tunes though.
Shazam is one of those apps that has truly changed my life. I still remember as a teenager times where I heard a song on the radio or a mix show and missed the name of it. I have purchased entire albums thinking they had a song I liked and have been WRONG. Now I can let my phone figure it out over a noisy background like a bar or restaurant and it still knows what song I am tagging. This app is one that truly feels like magic. Even almost a facade later.
Yes, Shazam has done a lot for genre, and diminishing the care for genre at all. Now for people that have spent way too much of their life defending or living a genre, what I mean is that it allows artists to be more creative and less pigeonholed, and it allows people to explore more of what they want.
People want certain sounds, and want more of that sound.
Shazam has been a key component of the discovery process.
Sorry to hijack this post, but if anyone is interested in content recognition, we built one of the largest reverse search engines for video/audio [0] and are always looking for skilled engineers to do more.
I think our scale makes things very interesting. We currently index over 20 hours of video every second and to date we've indexed over 11.5B videos.
Feel free to reach out to me directly at r@pex.com.
Apple Music's recommendation and discovery is pretty abysmal. I guess this will fix that issue. I'm amazed at how bad it given they know the standard that has been set by their competitor Spotify.
When Apple acquired Beast we heard about how Jimmy Iovine and Dre were going to bring their understanding and expertise to Apple's music service. This now seems to have been a lot of bluster though. Apple had such a head start with iTunes, missed the boat with streaming, spent a fortune on Beats and finally seemed to be getting serious about their streaming music offering. And the product has been somewhere between underwhelming and "just OK."
It's interesting to see that Apple's solution is to just buy their way out of their discovery problem. For a company with unlimited resources and great design skills it's strange that Apple is not really good at building services.
I don’t think Apple Music for Android was a play for Android users. It’s more of away not to be at a disadvantage when Apple is trying to sell a Family Plsn for Apple Music to a family of mixed Android/iOS users.
Back in the day, people didn’t realize the competitive advantage it gave MS to have a version for Macs. They could sell into large enterprises that may only have 5-10% Macs with a cross platform offering. Their competitors couldn’t.
You're absolutely right! And I hadn't thought about MS Office for Mac in that way before; thanks!
My comment was more geared towards OP's "well, now they're going to cancel the Android version" when Apple literally just released a new version of Apple Music for Android with platform-specific features: https://www.macrumors.com/2018/09/19/apple-music-android-aut...
I believe Shazam is very much an acquisition for its patent and algorithm sauce. The technology behind it is relatively simple based on audio fingerprinting
It was either Google Music, Spotify or Apple Music that would have acquired it.
I’m a 100% sure this will end up as a deep integration with Siri and the existing Apple Music app.
Apple has always done well with its acquisitions. Buy small companies that do one small thing and do it extremely well, and fuse that with the Apple ecosystem, quality and rigor.
Congrats to the Shazam team for going with Apple. Google would have most definitely messed it up like they did with Songza. I will never forgive them for that.
For those wondering why a seemingly normal acquisition took a year to close, it's because the EU was investigating it as a potential anti-competitive move.
I'm hoping that this extra access to Apple's extensive music collection will make this usable for music other than what you would hear on MTV or the classic rock station. The potential of the application is great, but as somebody with limited interest in pop, I find myself lacking for a way to identify new music discoveries.
Am I the only one that feels Shazam is no longer relevant? I don't think I know anyone has iShazam or SoundHound installed anymore. I feel all our music is on our phones now, so we already know what it is.
Really? I still find it highly relevant. As someone who goes to a lot of gigs, listens to a lot of DJs & always has some sort of set playing in the background while I work—I find it invaluable for tracks I can't identify and want to quickly note/share with friends.
I first heard about Shazam when I worked in a project to integrate Shazam API to Hello Moto J2ME app back in 2007~2008.
And it was amazing to see how they were able to move to the smartphone world so gracefully!
This functionality has been around for a bit actually, you may have missed it. You have to connect your Spotify in the Shazam app. And every time you Shazam something it auto adds it to a "My Shazam Tracks" playlist in Spotify. Been using that feature from them for about a year now! :)
You can also just ask her what song is playing. You just have to make sure she isn’t confused by any music app you might have open, because he will name the active track in that first.
I agree a drawback is there is no history. I take a screenshot of the results.
Google has this built right into the assistant now and detects music is playing well enough to suggest "What is this song?" as a query - and it has really really good accuracy in my experience, and I am impressed mostly by it being able to pick out the song in a somewhat noisy room (a bar for example)
Who needs shazam? Apple's just looking for more AI chops. G's already all over it.
I'm guessing this thing is going to run all the time and send Apple a continuous stream of what music is actually playing around people, so they can use the data to better market music.
Having song identification built into your iDevice is a nice value add for the consumer though
To anyone curious how Shazam works, it looks for similarity in the spectrogram of the music. This is not too different from how we identify certain chemical compounds with mass spectrometry.
I remember feeling that way the first time I used Uber (in SF where getting a taxi was difficult and unreliable).
A few taps on my phone and a car shows up to my exact location in 2 minutes?! And that's it? No calling dispatch and hoping they'd show up eventually (sometimes they didn't). No fumbling with cash or tip or anything.
A shame that uber and the whole gig economy is so shit, circumventing basic employment rights and both minimum and living wages. I mean I like the basic idea and I agree that the taxi business needed the disruption, but it's opened the gates to people working well below a living wage because there's nothing else. That is, there's plenty of people that are willing to work below minimum wage because it's the only thing they have left. And companies - including Uber - are more than happy to offer that kind of work.
Taxi drivers were already considered contractors and working for shitty wages. Uber just followed the existing model. In fact, it was worse, as drivers had to pay to work, by renting the cars with medallions, rather than being able to use their own.
The change for taxi drivers has only started to come after Uber was already making it big, e.g.: "For what the union involved says is the first time ever, the National Labor Relations Board (NLRB) declared taxi drivers are legally “employees”"
My uncle was a taxi driver around the same time (maybe a little earlier?) and was paid a livable wage plus benefits.
>"There are so many drivers available that it's an owner's market," said Edward G. Rogoff, a professor of management at Baruch College who studies the taxi industry. "Drivers are competing for these low-prestige, low-pay jobs, which pushes their incomes down further."
You also missed the main part of the article. Simple stated there were too many taxi drivers. Now I don't know about you but if I were to have that many people willing to work for me its either the pay is good or working conditions.
This sucks because like the article mentions, it becomes an owner's market. On top of this it's also NYC. No taxi driver is gonna be able to afford many of the places they get fares from.
>"In '87, '88, you could make $100 a day, maybe more," said Sam Khan, a 37-year-old driver from Pakistan. But now, he said with a grimace, "maybe $60, $65."
Not all countries have the medallion system (or that system the way it works in parts of the USA), and in some places drivers felt they were paid pretty adequately the way things were, and so they reacted very negatively to Uber drivers showing up and undercutting them.
In places where medallion or similar systems were in place, most of the protesters tended to be medallion owners. Who may or may not themselves be drivers, but even if they were were probably still deriving significant income from non-owner drivers renting their cars and medallions.
I don't understand this logic generally and in this case it makes even less sense. 10 years ago, Uber literally didn't exist. How, in any way, is Uber forcefully underpaying it's employees/contractors? People flocked to drive for Uber; where is the oppression here?
This is the same "Walmart doesn't pay its workers and destroys local businesses argument". People work there, and people shop there. Walmart didn't make anyone do anything, and neither does Uber.
In a free market, people do what is valuable to them. How can you say those people aren't? If it's not a free market, then how is it the corporations fault? Don't hate the player, hate the government.
In a completely free market, corporations are free to distort economic and labour policy so that poor unemployed people are desperate to get any job they can, regardless of how poorly it pays.
To add to this: if workers are underpaid relative to the cost of living in their area, they'll likely end up on some government assistance program.
Effectively, Walmart/Uber/etc get to double-dip -- they pay their employees less than a living wage, and based on the US tax structure, don't have to pay for the government assistance programs that their employees end up needing to make ends meet.
> To add to this: if workers are underpaid relative to the cost of living in their area, they'll likely end up on some government assistance program.
But that was already the situation before Uber came to the market? For all we know, driving Uber might have moved some of the people less dependent on the government assistance programs.
The Wikipedia article on minimum wage laws [0] has a pretty good overview of some of the most common arguments for and against. Really, though, the whole article is pretty interesting for a Wikipedia piece and illustrates the level of disagreement between economists over the effects these laws actually have.
Some jobs are viable at $8 /hr but not $15 /hr so the business have to eliminate positions to stay solvent. That means someone who could have taken the lower paying job now has no job at all.
Also, inflation. When everyone is making at least $15 /hr then that $15 means less than it used to. All costs of goods go up to match the new supply of "wealth". Burger flippers make $15 /hr but burgers now cost $10.
In more meta terms: manipulating the market is never easy if it's possible to do successfully at all.
My opinion is that having a minimum wage influences people to not make their own target, since an authority figure essentially deemed one hour of work is worth X dollars. Removing the minimum wage would make workers think for themselves how much they should earn and I believe that inclusion in their earning process would lead them to plan better, work better, and therefore earn more.
Essentially, I believe that when the government thinks for the people like declaring a minimum wage the people tend to avoid thinking about it themselves.
Free markets rely on educated consumers operating in a transparent market.
Ubers model works because they externalize a lot of operating costs to ignorant (under-informed) "contractors", and obscuring the economic realities from customers.
That's assuming there's an infinite supply of jobs.
When you're in financial difficulty - say your 0 hour contract doesn't bring enough money to pay the bills - you'll take whatever terms and conditions Uber is giving to you. Because you need to fulfill your basic needs.
I don't think that many Uber drivers are doing it by choice, they just don't talk about it to avoid bumming clients and getting a poor rating.
I'd like the slickness of the Uber app with a bit more ethics and I would be a good client.
> That's assuming there's an infinite supply of jobs.
Not only did I not assume that, the exact opposite is true.
Uber _just_ created all of their jobs; they didn't used to exist. Everyone started working there because they wanted to. It's valuable to them to spend their time driving. It's not like Uber forced closed other companies at a 1:1 ratio and force hired their workers for less.
I mean, if you're a libertarian that makes sense. If you're not then wonder about why things like minimum wage exists, surely if people are willing to work for less than minimum wage why stop them?
Isn't this a case against minimum wage laws? Obviously people's paychecks need to be paid from somewhere. Not all business models are going to work when you have to pay $15 /hr. So in that case, no job gets created and the person who could have done the work gets no money.
By removing the artificial market controls, people can voluntarily choose to get paid less, and make something (which is better than nothing).
I think people that complain about Uber not providing "basic employment rights and both minimum and living wages" don't really know what it's like to be poor. Talk to an Uber driver sometime, they're grateful for the opportunity to make some cash on their own terms.
> it's opened the gates to people working well below a living wage because there's nothing else
Haven't you just walked into an argument for abolishing the minimum wage here?
(I'm not arguing for or against minimum wage. I'm just observing that the tone of your comment seems to suggest you don't like the idea, but this sentence seemed to be a good argument for it.)
Could someone conceivably build a reference implementation FOSS client and server so that anybody could start their own Uber / Lyft. It would eliminate the friction to get going.
It's fascinating how you can fully spell out that Uber is giving people in a shitty situation a better life, and still have no doubt they're doing something bad!
Uber charges you for "wait time" after you request a car...which is insane. Not that Lyft is much better, but I've never seen that kind of nonsense with them.
It feels like they're double dipping. Unpopular as it might be, I'd guess people would be more than willing to pay extra so that they could shitcan this "feature".
I don't think it's arbitrary. You're paying for the (vehicle, driver) time and distance and the closer your are to vehicle the shorter the trip. If the vehicle is far enough away to make it vary greatly just cancel. The seemingly arbitrary charges and price volatility are not supposed to start until after you're in the vehicle. This may just be a bug.
In my case I stood there in the rain for 15 minutes waiting for a car that never came and eventually canceled, so I walked around till I got to a taxi stand. A little less than magical.
I've been planning to do more write-ups from past hackdays, such as building a functional Shazam experience in client side JavaScript and other visualisations. For example one time, I repurposed a WebGL globe to plot music recognitions over it. I don't have access to the Shazam tag stream anymore, but you can get a feel for how this looked as I've applied near real-time Wikipedia edits on the globe instead: https://umaar.com/globe/
a) Include Node.js scripts for converting raw big data (the raw Shazam stream) into something usable for visualisations.
b) Contain Node.js scripts & utilities for automating the creation of maps focussed on particular geographic locations.
c) Contain client side JavaScript for displaying all the individual map tiles onto an interactive map & Google Maps API code to query for places of interest (like a club, shopping center) which may explain heavy usage of Shazam in a particular location.
For the globe, which is unrelated to the datamaps, here's the original WebGL globe which someone did in 2011: https://github.com/cedricpinson/globetweeter
The Wikipedia edits version of the globe I did: https://github.com/umaar/wiki-globe - happy to do a write up of this if people think it'll be useful.
This thread is a good one, though, so we won't bury it as a dupe.
With all that said, PLEASEEEEEE don't ruin it, Apple. Specifically meaning I love it's integration with Spotify. I will be bummed if that goes away for an exclusive Apple Music integration.
I don't think the overall UX, song recognition, tag tracking will be ruined
> Specifically meaning I love it's integration with Spotify. I will be bummed if that goes away for an exclusive Apple Music integration.
Well that was pretty much what Apple paid for. I can already see that all my shazam tags automatically show up in Apple Music for me. Clearly this is just becoming another growth channel for Apple music which means their Spotify integration will likely be diminished or end.
That was happening before the acquisition as well
The user value to time investment ratio is off the charts for Shazam. I love music and what they have done for it, especially in the context of other music apps like Spotify et al.
Second the Don't Ruin It Apple plea, but congrats to the creators on a well-deserved exit.
I used Shazam a long time ago, so dunno how it had changed, but it certainly isn't a novel idea.
Just to jog people's memory, patent threats were sent to others doing similar things [0]. Not sure if that represents Shazam or not, but if so, they'll fit in perfectly with Apple.
0 - http://royvanrijn.com/blog/2010/07/patent-infringement/
I've never heard of Apple threatening individuals in the Open Source community -- am I missing something?
Not particularly. With conditionals added, we can differentiate any behavior and pretend similarities don't exist.
I also use My Shazams as a potential mix list. If I have an item in My Shazams, it caught my attention for some reason. Just putting those tracks together into a set has worked out pretty well. It gets tricky if I'm using my phone as the streaming source. At that point, I usually just take a screen grab. If it's an individual track, I have artist/track name, and if it's a mix, I can use the playhead position to narrow down which track it was later.
If only Soundcloud/Mixcloud integrated it so you saw the current track while the mix played - likely avoided for copyright deniability?
In fact, as an uploader, you can still see it (I don't know why), but not as a listener. As an example, here's a screenshot of me viewing my own upload: https://imgur.com/a/REpxOgx
Editing tools still allow you to post accurate timestamps: https://i.imgur.com/gZabJUx.png
There's also a browser extension that modifies Mixcloud pages to display the tracklists again, but they're useless since no uploader has an incentive to go through the painful process of entering timestamps, and that process is now crowdsourced anyway thanks to 1001tracklists.
So a pretty narrow demographic for the feature though...
If you're only going to take one person's usage then, yeah, I guess it's pretty narrow. How about another example for from the same person. As a video editor I'm constantly having to find "hip/trendy/cool" music for video edits. These clients typically like music that I do not normally listen to, but I know it's something they would like when I hear it. I'll let it live in My Shazams until one day I need to find something. I've impressed more than one client with the ability to pull out a "fresh" song for them. Clients appreciate that sort of thing.
How about it? We've just added video editor looking for cool music to DJ. We could add tens such cases more, but all those combined are still narrow demographics.
Also, diabetes narrow? 30 million have it in the US, I'd say those are way more than DJs that need to recognize tunes, and have much more motive to use an app that helps them keep track of something that can have huge impact on their health...
I'm not the original commenter, but I'm going to go out on a limb and suggest—they use it to find out what a song is that they hear & like?
This data would be incredibly valuable to companies trying to market and sell music. Like labels, and now, Apple. Not to mention keeping that data now out of the hands of others.
There is another reply to me that snarkily opines that the average HNer doesn't engage in social events, which is a bit humorous really because a more correct statement is that most people are socializing at social events, not doing inference on background music. I can say that I have never, ever seen another person pull out shazam in public.
It also makes sense if the scope of the product is not enough for it to make anyone pay for it (except as part of a larger product).
That's not the case for an app you might take out a couple of times a day (at best) to check some tune for 10 seconds.
Behold, the average HN user I was referring to
A generalization if I ever saw one.
And the original Senator Claghorn that Foghorn Leghorn was channeling: https://www.youtube.com/watch?v=SBE92naflJ0
Here's a recent book of them, to help you get "in" on the "hip slang" of tomorrow:
https://en.wikipedia.org/wiki/Philogelos
/s
/s
1) That I've lost interest in music myself (I didn't).
2) That whether some other people might not lose music was disputed. Perhaps you imagined me as some strange creature in a cave that has never heard or known of people that maintain their interest in music to their late age.
3) That the fact that some other people don't lose interest in music is relevant. I was making a general statement ("it's also rare"). Most people do lose interest in music compared to their younger years. That's not some fringe claim, it's borne out by buying patterns and studies. People tend to stick to the music of their youth time and again. Even the new artists they discover tend to be of the same genres and styles in vogue when they were younger.
It doesn't mean "older people have no use for Shazam".
It means "as it has been observed statistically, _most_ older people don't seem to have a use for Shazam and any random older person is more likely to fall into that category than the rest of the older people".
Carl Craig is old too.
>Have you never heard an old song you half-recognized and thought “huh, what band is that again?”
Yes, but this is not such a large part of the older demographic's lifestyle for the app to matter much to them.
Old people also want to send their dickpicks too, but they don't flock to Snapchat.
But, honestly, I don't know how all this is controversial, unless someone is 20 and all their friends are too.
I'd expect anybody over 30 or so to verify that even if music is still a big deal for them, it's not as much for most of their friends as it was in their teen years and early 20s.
Unless you're 30 and all your friends are eg. DJs and techno-heads or heavy-metal fans still rocking those leather pants, the usual course is that they are more interesting in their careers, families, new babies, and binge watching TV and surfing, than following music, even if that was a big part of their identity in their teens. Is that inaccurate?
I was at a live jazz concert last night, and the trio started playing a riff on tune I recognized but couldn't exactly place. SoundHound nailed it (Sonny Rollins "Four"), while Shazam had no idea.
On a sidenote: It's surprising that Google isn't more useful in those searches either. Often I have to search for rare releases by relying on lyrics and Google's results are plain terrible. I understand that the lyrics are probably nowhere listed and idnexed, but considering that even the rarest releases are uploaded by someone on Youtube, some kind of auto transcription feature would be very helpful.
Sure, it's a niche interest, but the Discogs user base is still huge.
This looks to me like an acquihire for the ML/AI/DL talent. So basically, yes they will be diverted to other projects.
------
[1]: https://www.ee.columbia.edu/~dpwe/papers/Wang03-shazam.pdf
Not saying I agree with this, but given how most acquisitions go, other services don't survive long. Apple Music has an Android app, but that's more of a halo effect kind of thing to get people in the ecosystem and (eventually) switch to iOS.
Shazam was founded in 1999-2000 ish and was live by 2002, longer ago than many people realise. The fingerprinting method they used was a novel one, and they successfully deployed it in a couple of years, with a database of two million tracks, in a system that could be used by dialling a number from any phone. It's an extraordinary and rare example of taking a research method and getting it to work at scale in the face of real constraints, in the process producing something that most people would never have imagined could be done.
Would really be interested if someone has tried other approaches and can comment.
This is a really nice paper as well, but the Shazam method is more exciting because of its focus on the search part of the problem. Audio fingerprinting wasn't an entirely new field of course - the Philips paper also cites several prior publications.
The big limitation of all these methods is that they are robust to EQ and other degradations but not at all to variations in timing between frames. So they work well for identifying different instances of the same recording, but not at all for matching different performances, or for "query-by-humming".
http://coding-geek.com/how-shazam-works/
https://github.com/jvbalen/soundofshazam/blob/master/sound_o...
And an example output (try Shazam-ing it!):
https://soundcloud.com/sample_noise/shazuffle-ii-shazam-me
He also got some legal threats just for publishing code he created himself.
[0] http://royvanrijn.com/blog/2010/06/creating-shazam-in-java/
The remaining thing I don't understand now is how the heck they managed to negotiate the license for all the songs they have in the database. A friend of mine had a major website for lyrics, and he got some legal trouble from the record companies. When he tried to buy a license to display the lyrics, after some back and forth, they basically said they're not interested in licensing. This was in the early 2000s before streaming took off. Since that time I have the impression that often legal and social problems are much harder to solve than technological ones.
I actually bought Shazam on android after repeated attempts at getting google to find music in the same way - they're laughably bad.
that said, i still prefer shazam because it's an actual app, whereas google's implementation is a button that sometimes appears in the google app, and sometimes doesn't, seemingly at random.
Not really "magic" how it works but still pretty amazing that I can pull the phone out of my pocket any time I hear a song I like and immediately see the band/title.
For anyone else finding this as implausible as I was, apparently the Pixel has a local database of 17,300 tracks. Still pretty impressive though.
https://uk.pcmag.com/news/91606/googles-pixel-2-phones-recog...
Is it on by default? Cos that sounds really creepy.
[1] https://motherboard.vice.com/en_us/article/8q8ee3/shazam-kee...
That's really sweet—I wonder how much storage that requires? I would have expected it to require a lot, but if it's on a phone, then maybe not that much?
https://ai.googleblog.com/2018/09/googles-next-generation-mu...
It's also pretty garbage for classical music.
(o) https://en.wikipedia.org/wiki/Shazam_%28application%29#Early...
I have worked on integrating a shazam like library inside an app. We have looks at more than 12 solutions before finding a company with a song recognition library good enough to be called shazam like. It is by no mean a small feat you can reproduce during a weekend.
The interesting thing is: most of these songs are on YouTube, so in theory Google could build a better Shazam.
Separately: there was nothing wrong with gnulinux's comment. It was obviously posted for the pleasure of sharing information, and didn't put down either the GP or Shazam.
I think it's a good thing, because often there are good comments in response to downvoted comments. Letting people delete comments would make the discussion hard to follow.
It was informative to myself at least.
I wish someone would do the same thing for commercials on TV. Auto detect the commercial and advance the DVR by its exact duration
They've had it for years now. I was skipping commercials on my HTPC back in 2007-ish. I think the older Tivo's also had this ability.
AFAIK that's not the case of Shazam where a read-only database would be enough. Sure eventually you have to manage multiple versions of your app and be able to gradually update your DB, but that's still simple stuff compared to the problem of how to make money.
I've never used Shazam, but I used a similar feature on my LG Chocolate back in the day. How is Shazam any different than what my LG Chocolate did (aside from the Spotify integration)?
Expecting people to know how is it different from an LG Chocolate is just absurd...
That said, Shazam is actually older, having started as a shortcode service in the UK way back in 2002.
How exactly did this come to be?
That's the anthology I want to read.
I remember Adam Neely (A musician/jazz composer/youtuber) mentioning in a Q&A, that for songs they played live for a television broadcast, backing tracks were added underneath and it was mixed as close as possible to the originals, among other reasons for the purpose of them being better "shazamable".
https://www.youtube.com/watch?v=zH_5j3Wm9qM&t=11m38s
What are you referring to here?
It doesn't recognize you singing because it doesn't have the same spectral content as the full song. Shaman doesn't really identify tunes, it identifies recordings. (And it is amazing to me that it works at all)
1)New song is released
2)It starts being "shazamed" by various people - it's becoming trendy
3)Large stores that differntiate themselves by being trendy want to play trendy songs first
4)They pay Shazam to tell them which songs are trendy
5)The stores play them
6)Customers "shazam" those songs in the store
7)Customers view the store as trendy
8)Now four instances have profited: Shazam, the store, the musician and the customers (for having found a trendy song very early on)
Predicting a Billboard Music Hit with Shazam Data - Cait O'Riordan (https://www.youtube.com/watch?v=mcTPvxo8SXY)
The different 'shazam'd songs were also connected to locations and user demographics so if you wanted a track that was going to be big in a specific community you could get a good guess.
Probably, one of the best results from this acquisition.
Will they? Don't forget, the information flow is 2-way here, users get something useful out of it but Shazam (and now Apple) also learns what stuff is trending. If anything I actually lean towards them keeping or even enhancing Spotify and other service integration, at least as long as those services are viable: it effectively creates a direct channel into competing services that tells Apple exactly what stuff people want out of them, which they could then use directly in making Apple Music more competitive in turn.
Even if we take as granted that boosting Apple Music is a major concern to Apple, doesn't that at least maybe seem like an equally or more valuable way to do it? It doesn't make any users angry either or act as bad PR, they aren't taking anything away, and they get a massive information channel into every competitor and even real-world venues.
>And maybe no more Android support.
By the same token, if Apple wants to boost Apple Music in general then having information on music popularity and trends coming in from the population of Android users too also seems helpful.
I will be very disappointed if this comes to light, and the only integration we're left with is Apple Music.
Probably not likely, but probably not improbable either.
It's almost guaranteed.
>Exclusive?
>Start charging money?
>Patents?
Not to mean that people shouldn't make money, but acquisition means that there will be change.
The state of third-party apps on the Watch is awful. It usually takes 10 seconds for an app to start on a Series 2, and sometimes they just get stuck on that loading screen. If you manage to get an app running, the UI library has ridiculous bugs like accidentally enabled screen rotation where turning your wrist flips the content by 90°. (Unlike a phone, there's no reason at all for a watch to react to orientation changes.)
I could sort of understand these deficiencies for a hurried 1.0 product, but it's been three years and it's still this bad. There's no API for watch faces either. I get the impression that Apple truly doesn't want third parties on the Watch at all.
Congrats to Shazam though!
I think the idea that making Shazam ad-free “makes sense” is actually fascinating because if it were google or amazon buying Shazam, I would assume it would remain littered with ads.
[0] https://www.ee.columbia.edu/~dpwe/papers/Wang03-shazam.pdf
People want certain sounds, and want more of that sound.
Shazam has been a key component of the discovery process.
I think our scale makes things very interesting. We currently index over 20 hours of video every second and to date we've indexed over 11.5B videos.
Feel free to reach out to me directly at r@pex.com.
[0] https://softwareengineeringdaily.com/2018/06/22/video-search...
When Apple acquired Beast we heard about how Jimmy Iovine and Dre were going to bring their understanding and expertise to Apple's music service. This now seems to have been a lot of bluster though. Apple had such a head start with iTunes, missed the boat with streaming, spent a fortune on Beats and finally seemed to be getting serious about their streaming music offering. And the product has been somewhere between underwhelming and "just OK."
It's interesting to see that Apple's solution is to just buy their way out of their discovery problem. For a company with unlimited resources and great design skills it's strange that Apple is not really good at building services.
Back in the day, people didn’t realize the competitive advantage it gave MS to have a version for Macs. They could sell into large enterprises that may only have 5-10% Macs with a cross platform offering. Their competitors couldn’t.
My comment was more geared towards OP's "well, now they're going to cancel the Android version" when Apple literally just released a new version of Apple Music for Android with platform-specific features: https://www.macrumors.com/2018/09/19/apple-music-android-aut...
http://royvanrijn.com/blog/2010/06/creating-shazam-in-java/
It was either Google Music, Spotify or Apple Music that would have acquired it.
I’m a 100% sure this will end up as a deep integration with Siri and the existing Apple Music app.
Apple has always done well with its acquisitions. Buy small companies that do one small thing and do it extremely well, and fuse that with the Apple ecosystem, quality and rigor.
Congrats to the Shazam team for going with Apple. Google would have most definitely messed it up like they did with Songza. I will never forgive them for that.
https://www.ee.columbia.edu/~dpwe/papers/Wang03-shazam.pdf
I wonder if Apple will allow them to keep this feature.
It has to have been published "properly" though, which in the day and age of self-publishing directly to the net doesn't work so well.
http://willdrevo.com/fingerprinting-and-audio-recognition-wi...
The platforms will swallow our options until we have no recourse but to jump platforms.
Unless they make that too difficult, too.
I wonder if that will be more or less possible now. Maybe with a Shortcut?
https://www.fox.com/beat-shazam/
It's literally a 30-minute commercial for the app (including a play-along game you do with the app).
I agree a drawback is there is no history. I take a screenshot of the results.
What happened 2 years ago that you think affected its recognition quality?
One specific thing I recall is that more false negatives started popping up - which is bad over UX (vs. no matches at all).
The change was (perceivedly) strong enough for me to stopping using Shazam.
Who needs shazam? Apple's just looking for more AI chops. G's already all over it.
Having song identification built into your iDevice is a nice value add for the consumer though
A few taps on my phone and a car shows up to my exact location in 2 minutes?! And that's it? No calling dispatch and hoping they'd show up eventually (sometimes they didn't). No fumbling with cash or tip or anything.
It was like magic.
https://www.nytimes.com/1995/04/09/nyregion/driving-a-taxi-d...
The change for taxi drivers has only started to come after Uber was already making it big, e.g.: "For what the union involved says is the first time ever, the National Labor Relations Board (NLRB) declared taxi drivers are legally “employees”"
http://www.peoplesworld.org/article/nlrb-declares-taxi-drive... (2015)
My uncle was a taxi driver around the same time (maybe a little earlier?) and was paid a livable wage plus benefits.
>"There are so many drivers available that it's an owner's market," said Edward G. Rogoff, a professor of management at Baruch College who studies the taxi industry. "Drivers are competing for these low-prestige, low-pay jobs, which pushes their incomes down further."
You also missed the main part of the article. Simple stated there were too many taxi drivers. Now I don't know about you but if I were to have that many people willing to work for me its either the pay is good or working conditions.
This sucks because like the article mentions, it becomes an owner's market. On top of this it's also NYC. No taxi driver is gonna be able to afford many of the places they get fares from.
>"In '87, '88, you could make $100 a day, maybe more," said Sam Khan, a 37-year-old driver from Pakistan. But now, he said with a grimace, "maybe $60, $65."
They're more like landlords than employees.
This is the same "Walmart doesn't pay its workers and destroys local businesses argument". People work there, and people shop there. Walmart didn't make anyone do anything, and neither does Uber.
In a free market, people do what is valuable to them. How can you say those people aren't? If it's not a free market, then how is it the corporations fault? Don't hate the player, hate the government.
This is why minimum wage and other laws exist.
See also: race to the bottom.
Effectively, Walmart/Uber/etc get to double-dip -- they pay their employees less than a living wage, and based on the US tax structure, don't have to pay for the government assistance programs that their employees end up needing to make ends meet.
But that was already the situation before Uber came to the market? For all we know, driving Uber might have moved some of the people less dependent on the government assistance programs.
This is not a settled question in economics. The idea that minimum wage laws help the poor is at the very least debatable.
0: https://en.wikipedia.org/wiki/Minimum_wage#Debate_over_conse...
Also, inflation. When everyone is making at least $15 /hr then that $15 means less than it used to. All costs of goods go up to match the new supply of "wealth". Burger flippers make $15 /hr but burgers now cost $10.
In more meta terms: manipulating the market is never easy if it's possible to do successfully at all.
https://www.youtube.com/watch?v=x-BGi4NIFww
Essentially, I believe that when the government thinks for the people like declaring a minimum wage the people tend to avoid thinking about it themselves.
Ubers model works because they externalize a lot of operating costs to ignorant (under-informed) "contractors", and obscuring the economic realities from customers.
So I think it's fair to hate the player.
When you're in financial difficulty - say your 0 hour contract doesn't bring enough money to pay the bills - you'll take whatever terms and conditions Uber is giving to you. Because you need to fulfill your basic needs.
I don't think that many Uber drivers are doing it by choice, they just don't talk about it to avoid bumming clients and getting a poor rating.
I'd like the slickness of the Uber app with a bit more ethics and I would be a good client.
Not only did I not assume that, the exact opposite is true.
Uber _just_ created all of their jobs; they didn't used to exist. Everyone started working there because they wanted to. It's valuable to them to spend their time driving. It's not like Uber forced closed other companies at a 1:1 ratio and force hired their workers for less.
These are all voluntary transactions.
Depends if it's a sellers' or buyers' market.
By removing the artificial market controls, people can voluntarily choose to get paid less, and make something (which is better than nothing).
I think people that complain about Uber not providing "basic employment rights and both minimum and living wages" don't really know what it's like to be poor. Talk to an Uber driver sometime, they're grateful for the opportunity to make some cash on their own terms.
Haven't you just walked into an argument for abolishing the minimum wage here?
(I'm not arguing for or against minimum wage. I'm just observing that the tone of your comment seems to suggest you don't like the idea, but this sentence seemed to be a good argument for it.)
It's not their second Android app either.