0
$\begingroup$

and how does it work?

http://wmot.org/post/music-money-and-metadata-promise-blockchain#stream/0

claims that this is an answer to the metadata woes of the music/recorded media industry.

I don't see how it can possibly prevent rewriting the file.

$\endgroup$

closed as off-topic by e-sushi Mar 29 '17 at 18:06

This question appears to be off-topic. The users who voted to close gave this specific reason:

If this question can be reworded to fit the rules in the help center, please edit the question.

2
$\begingroup$

What is the fuss about 'dot blockchain' and music metadata?

Publicity man creates his own fuss.

claims that this is an answer to the metadata woes of the music/recorded media industry.

There are no such woes. Sure there is piracy but that's not because songs can't be identified using a centralized interface (they can, and more efficiently than watermarking).

I don't see how it can possibly prevent rewriting the file.

It's a watermarking scheme. You can't just get rid of these watermarks by converting to PCM and re-encoding because the watermark is present in the audio data itself. Consider, for example, shifting the center point (DC offset) of the PCM samples slightly - to either "up" or "down" states each lasting a few tenths of seconds, and using this to embed a repeating 128-bit identifier. No quality loss because DC offset is practically ignored by the end output analog playback system and you only shift it slowly.

Then consider fancier tricks like partially undoing the DC shift by selectively amplifying some of the lobes on the other half of the sample space and you get something that requires very advanced processing to remove without introducing artifacts.

The even meaner thing to do if you run an online store is to embed different identifiers for each user when they download the song ... then when some of these downloaded songs, each unique, appears on a file sharing site you can do "traitor tracing" and send in the FBI.

In terms of the other stuff they talk about, enforcing the rights, well, that's just not doable with today's tech because it will come out in analog at some level. Even assuming a secure encrypted link between a computing device and a set of speakers, at the most basic level you can just capture the analog voltage going to voice coil and process that somehow back into an audio file to drop DRM restrictions.

Note however that a good watermarking scheme will survive a few iterations of this. Applying DRM restrictions automatically based on watermarking sort of works in theory but requires a doomsday trusted computing scenario where users have absolutely no ability to run arbitrary programs on their computing devices (i.e., programs that don't necessarily obey DRM restrictions).

Watermarking is quite an old technology. It appears what they consider to be the main benefit of putting the metadata on the public blockchain is that it cannot be revoked (as in, the publisher who applied the watermark going out of business and lookups to their website/API failing to provide data for audio files). This is a mildly interesting idea for Indie media, etc, but of course big media does not assume they are going to go out of business any time soon so there is no impact on their business model from this technology.

$\endgroup$
  • $\begingroup$ You've posted your own solution to nullifying this technique which I'll take forward elsewhere. You release a filter app that either 1) emulates a DC blocking capacitor and subsequently recodes the track or 2) adds a random audio watermark of it's own which obfuscates the original one making the metadata useless. Isn't that already done on youtube videos where all that weird flashy stuff is added over the movies? Fortunately it's doomed. $\endgroup$ – Paul Uszak Mar 29 '17 at 10:30
  • $\begingroup$ @PaulUszak 1) only defeats the first naive scheme and not the more obfuscated one. 2) is more general but probably counts as a DMCA-prohibited protection-cracking tool in the U.S. $\endgroup$ – DepressedDaniel Mar 29 '17 at 20:44
  • $\begingroup$ The ultimate watermarking scheme would also be able to overcome 2) by not allowing exact recovery of all watermark components - for example by injecting a large number of signals with a very low signal to noise ratio and heavily relying on an error correcting scheme to recover the watermarked ID from recovered code signals that can include false positives. $\endgroup$ – DepressedDaniel Mar 29 '17 at 20:50
  • $\begingroup$ @PaulUszak I'm not a big fan of ultra-left views. Musicians are just trying to make money from their craft - you might as well say that a minimum wage worker at McDonald's is motivated by "greed". Sure digital media can be copied at zero cost. But creating the original still represents someone's work and toil. We could apply the principle of "everything at cost" to the entire economy and conclude that a blacksmith should not be allowed to sell pickaxes at more than the cost of the materials and coal for his furnace. But in such a system nobody makes any money so how can anyone buy anything? $\endgroup$ – DepressedDaniel Mar 30 '17 at 1:07

Not the answer you're looking for? Browse other questions tagged or ask your own question.