Of course. How else are Shazam going to make money? They need to listen to you, process the audio using a natural language processing procedure, and then target advertising at you.
Can confirm that they're pursuing this business model: particularly to see what TV ads you're being exposed to. Coverage from existing sources is panel based with low sample sizes that do not extend well into more targeted research.
Yes, but there is an on/off switch which presumably stops this activity. The energy impact in mavericks while "listening" is hovering around 2 on my machine. For comparison, Safari is at 8.
My first thought as well. This could be used as constant surveillance tool if Shazam chose/was forced to shared the uploaded data with GCHQ/NSA </paranoid-theory>
They can listen in the background and Shazam will do this if you enable Automatic mode, though the iOS status bar turns red while the microphone is being accessed by a backgrounded app.
You can do this (and a ton of different cool stuff) with RogueAmoeba's Audio Hijack Pro app, which lets you route audio from one app to another, optionally applying some filters to it. I sometimes use it to fix audio out-of-sync problems in certain videos.
Edit: As commenter above noted, the audio routing functionality is also available with the free Soundflower app.
SoundFlower does this, it's open source. It contains a decent chunk of audio routing features that are inexplicably missing from the core OS X.
For instance, you can route application audio out to a Soundflower channel and use that as input for another app.
Edit: just checked, Soundflower rerouting works perfectly with the Shazam Mac app, tried it out on a couple of YouTube songs, and it got it right on 47 out of the current top 50. It's a shame that Shazam is still only useful for mainstream stuff, when I tried it on some random metal, Goa, soundtracks, and industrial... 0 out of 15.
It's the countdown until the mic switches off again. You may not want to keep the line open all the time, it may prevent your computer from going into power saving mode (not sure), and Shazam probably doesn't want everyone to hit their servers constantly either.
But if you switch it off and on again the timer starts anew, so it's not a per-user limit.
The actual identification of individual songs is done using the Echonest API (they have a huge database of song hashes).
Before this can be done however, the placement of each song within the audio track as a whole needs to be determined, which is done with a combination of ffmpeg and C. Currently this process is quite slow and inaccurate - I'm trying to learn some ML theory at the moment with the hope of improving it!
Great work and a really interesting read! Really nice performance too - for reference Echonest advises samples be 30s or longer for optimum recognition.
With dark mode in Yosemite I've got a feeling devs will need to either do a dark and light version of menu bar icons or go all color. Currently my menu bar is useless in dark mode as I can't see any icons.
Personally I would never allow any app to use my microphone all the time, effectively being able to monitor a certain area around the device it is running on all the time. This is gross.
This is all the more true for an app from a company that did transfer sensitive user data to third parties in the past.[0]
In Germany, under certain circumstances, it is also illegal to make sound recordings of persons without their knowledge.[1] This might be the case in other jurisdictions too.
I wonder what the most common use case is for this? It's rare that I'm in front of my computer, there's music, and I have no way of determining what it is.
My first job was at a record store. The owners were an old man and his wife.
The old man was essentially Shazam personified. It was amazing. He could name any song, composer and artist after listening to a few beats. Except he would do it with old jazz artists and big band music.
Coincidentally the old man had huge ears. Enormous.
And when you'd ask him how he could remember all those songs, the wife would chime in, "It's cause of his ears. The songs all stay in his ears. Bouncing around."
The record store is long gone and the space is now occupied by a Cold Stone Creamery.
42 comments
[ 4.6 ms ] story [ 96.8 ms ] threadThat and, constant audio surveillance (even with "an option to turn it off")? No thanks...
Edit: As commenter above noted, the audio routing functionality is also available with the free Soundflower app.
For instance, you can route application audio out to a Soundflower channel and use that as input for another app.
Edit: just checked, Soundflower rerouting works perfectly with the Shazam Mac app, tried it out on a couple of YouTube songs, and it got it right on 47 out of the current top 50. It's a shame that Shazam is still only useful for mainstream stuff, when I tried it on some random metal, Goa, soundtracks, and industrial... 0 out of 15.
I must have skipped over the article as I assumed it could pick up on internal audio. Seems like an oversight...
But if you switch it off and on again the timer starts anew, so it's not a per-user limit.
Time to evolve!
Before this can be done however, the placement of each song within the audio track as a whole needs to be determined, which is done with a combination of ffmpeg and C. Currently this process is quite slow and inaccurate - I'm trying to learn some ML theory at the moment with the hope of improving it!
https://github.com/worldveil/dejavu
This is all the more true for an app from a company that did transfer sensitive user data to third parties in the past.[0]
In Germany, under certain circumstances, it is also illegal to make sound recordings of persons without their knowledge.[1] This might be the case in other jurisdictions too.
[0] http://www.google.com/translate?hl=en&ie=UTF8&sl=auto&tl=en&...
[1| German Civil Law Code § 201 – Breach of confidentiality of the word: http://www.google.com/translate?hl=en&ie=UTF8&sl=auto&tl=en&...
The mobile version makes -more- sense to me.
Students?
The old man was essentially Shazam personified. It was amazing. He could name any song, composer and artist after listening to a few beats. Except he would do it with old jazz artists and big band music.
Coincidentally the old man had huge ears. Enormous.
And when you'd ask him how he could remember all those songs, the wife would chime in, "It's cause of his ears. The songs all stay in his ears. Bouncing around."
The record store is long gone and the space is now occupied by a Cold Stone Creamery.
I've made my own audio fingerprinting software in Python under the MIT license, check it out!
https://github.com/worldveil/dejavu