Update, April 8, 2013: Can you hear it? Take a watermark listening test.
A while ago I posted about my confusion regarding Weird Spotify Compression Artifacts. It turns out the artifacts are not due to compression, but a result of audio watermarks that Universal Music Group embeds in digitally distributed tracks. The artifacts appear on UMG tracks at Rdio, Spotify, iTunes, Amazon, and others. These watermarks can also be heard in Universal tracks broadcast over FM radio. Universal Music tracks make up about 25% of most online catalogs and its labels include Interscope, The Island Def Jam, Universal Republic, Verve, GRP, Impulse!, Decca, Deutsche Grammophon, etc.
Just to show that it's not limited to Spotify, here is an example of the watermark artifacts through another distributor.
What the watermark sounds like
Spectrogram of the difference between a watermarked and unwatermarked UMG track. The energy is concentrated in two bands between about 1 khz and 3.5 khz - where the human ear is most sensitive.
UMG uses a spread spectrum watermark, a technique explained in detail in this Microsoft research paper. The watermark scheme modulates the total energy in two different bands, 1khz to 2.3 khz and 2.3 to 3.6 khz. The energy is concentrated in the most perceptually sensitive frequencies because that makes it more difficult to attack or remove without significant audible distortion.
The energy is increased or reduced in 0.04 second blocks. The result can be characterized as a fluttering, tremolo sound. Listen closely to the original vs. watermarked audio samples and try to focus on the 1 khz to 3.6 khz noise range. It helps to wear headphones in a quiet environment.
Audio samples
Here is a short sample (excerpt: Three Doors Down - When You're Young). These are lossless original and watermarked files; what you hear is not a result of compression.
Original:
Watermarked:
If the difference between the two isn't clear, here it is by itself:
Difference:
The character of the watermark may seem subtle during this short sample, but through the duration of an entire song it becomes more familiar and more annoying. Check out my original post on the subject for more examples.
Technical details
The watermark does not start until 1 second into the audio. After this the signal is divided into 0.08 second blocks. Each block is divided in two: some amount of energy is added to the first half and the same amount is subtracted from the second half. This coding scheme allows blind detection (without access to the original file). The actual information in the watermark is not easily recovered because it is modulated by a pseudo random sequence, which is generated by a secret key.
I have done a little searching and I think UMG's watermarking technology is provided by MarkAny, a Korean company that has developed their own watermarks out of university research, and purchased some watermarking patents from Digimark.
Removing the watermark
Since the watermark creates audible distortion, it's worthwhile to try to reduce it. I wrote a script that analyzes the block energy and applies some smoothing. This is the result so far.
Watermarked:
Restored:
More discussion
Hydrogenaudio forums on watermarking
UMG Watermarks audiophile files, pisses off paying customers
Why do labels watermark tracks? Watermarking simplifies copyright enforcement by letting a company track music on peer-to-peer networks. "It gives them the ability to put pressure on policy makers and ISPs to do filtering," says Fred Von Lohmann, an Electronic Frontier Foundation attorney. That may be about the best explanation you will find. See DRM Is Dead, But Watermarks Rise From Its Ashes
I don't have anything against watermarking, but I have a problem as a consumer when it is poorly implemented and destroys music I've downloaded legally.
Why isn't UMG's watermark talked about more? Maybe people think the audio quality problems are due to some kind of lossy compression, as I did, and ignore it completely, or blame the streaming service/distributor. The problem here is that the UMG watermark degrades the audio to about the equivalent of a 96 kbit MP3. My guess is that if consumers were informed about what is going on, they would care. Especially those who pay full retail price for digital downloads advertised as lossless audio.
Alex Jimenez
/ February 4, 2012Thank you for figuring this out. I use Spotify and like many had thought something was wrong with the streaming codec. It's also being used on iTunes, judging by a recent LA Philharmonic download of Wagner excerpts--the irritating flutter is there too.
Taylor Perkins
/ June 12, 2012This is great analysis.
bozmillar
/ September 10, 2012I've been wondering what this flutter was for a long time. Thanks for putting this up. I always just figured it was a compression issue, but when I upgraded spotify to the "high quality streaming" and it was still there, I got suspicious and found this.
George
/ September 28, 2012Hi;
How did you obtain a difference wav ? What tool did you use ? I have a CD and an official download and wanted to check for myself.
Thanks;
George
Kasper Holbek Jensen
/ October 4, 2012Very nice brief analasys for non-soundtechies like myself, I was wondering if the smoothing script is of up grabs. Or would that violate some term/law?
Matt
/ October 4, 2012I didn't post the script since distributing it would probably violate the DMCA, but it is pretty easy to duplicate if you know how to work with audio in the frequency domain...
Matt
/ October 4, 2012I generated the difference audio by aligning the signal of the original and watermarked audio and subtracting them. You can do this in any wave audio editor like Audacity or Adobe Audition, but the sample alignment is the most important part.
Kasper Holbek Jensen
/ October 5, 2012I wouldn't know where to start. It is a strange thing that the real quality audio is now reserved for the pirates. This industry really knows how to hit a target.
B. B. Pedersen
/ May 13, 2013Thanks, shared.
George W
/ May 13, 2013I think they thought they were really clever this time. Since virtually all music consumption today is via online listening, (read Youtube &Co) so people neither notice nor care.
Most of the few who actually pay for music are playing it through the laptop USB speakers, (read Spotify &Co) which often aren't exactly HIFI. And most grown-ups don't hear the difference anyway. (I tested the given samples with Sony headphones. And my hearing is flat to a 12kHz personal cutoff, which is average for my age.)
So, the only ones to really suffer must have good hearing and a proper stereo set. But many of them do actually buy CDs, so that's even better for the industry. :-/
Now, I do agree about watermarking in general. I think it is OK to watermark your intellectual (or alas, your artists' intellectual) property, but it shouldn't degrade paid-for music.
At any rate, this is a step towards sanity compared to copy protection. I think we are on our way to a better future because every year an old boss retires, and instead comes a younger one, who's presumably seen computers as a child.
Matt
/ May 14, 2013George - I want to offer a small counterpoint to a few of your statements. It might be that listeners with poor hearing or cheap stereos actually hear the watermark more prominently. Consider that a cheap stereo system will have a low and high frequency roll off, such that only the midrange is left over; the watermark is concentrated in the middle, unaffected. However, played over a hi-fi system, there's more energy in the highs and lows, and that's just more energy that might mask the watermark in the middle. Likewise, for someone with some hearing loss, unless their hearing loss lands right on 1 to 3.6 kHz, they're simply going to be deaf to acoustic energy that would otherwise serve to mask the watermark.
Pål Bråtelund
/ May 15, 2013Thanks for this analysis, Matt. All this is in a lossless domain - got any thoughts on how the most popular codecs (AAC, Vorbis and MP3) reacts to spread spectrum watermarks?
Matt
/ May 15, 2013The watermark is designed to withstand lossy compression. You'll still be able to hear it in a 64 kbps MP3, for example.