Sunday, October 06, 2002

I would appreciate any feedback on this post from readers in the know on copyright law, piracy legislation etc. The following arose from a conversation with The Good Lady Wife. It’s a frequent case that her sharp questioning and insistence on polite discourse means I must formulate reasonable arguments, logically buttressed. For this, and so many things, I thank her.


It is a common that people can be identified by characteristics not detectable by human senses. DNA, fingerprints and voice prints are accepted to one degree or another in many jurisdictions. It is not necessary to prove that this is a gross or obvious characteristic. Indeed, the very obscurity of the ID contributes to the reliability of the marker.

However, this might be a two-edged sword. Soon, the science of animation and reproduction will reach a point that it will be possible to create a copy of an image or a voice that is indistinguishable from the original, at least to the limit of human senses. I would like to focus today on voice and music simulation, as I believe this will be accomplished first, simply because the programming and processing load is smaller.

Many of us are familiar with voiceprints; each human voice can be differentiated to, I think, the accuracy afforded by fingerprinting. No matter what non-technological steps the speaker takes, analysis of the recorded voice will show the sub-vocals and harmonics that identify the unique person. Similarly, any impersonator can be undone regardless of the talent employed.

Let’s propose for a moment that some Bright Boy manages to synthesize the singing voice of Robbie Williams. At the same time, he installs a “watermark” into the recording so that even basic computer analysis will show that this is a fake. The music is not marketed under the name “Robbie Williams”, but as “XYZ”, singing an original composition of Bright Boy. To the unaided ear, the music is indistinguishable from Robbie Williams voice; but to a legal standard often accepted at the higher criminal standard of proof, it is easily shown to be not Williams.

The question on the table is: Does this constitute piracy? And if so, how?

Perhaps as interesting, is the question of whether it will impact on music sales. The GLW is certain, and I tend to agree with her, that it will not. She feels that much of music sales, particularly at the extreme end of the popularity curve, is all about gaining a piece of a person’s celebrity. To buy a genuine CD is to buy a slice of the artist, however far removed.

I would add a caveat applying to internet downloading. It would be reasonable to assume that an MP3 file virtually identical to Robbie Williams would be a popular item on the peer-to-peer networks. My experience is that file tags are at best haphazard, and at worst useless. So I would expect to see those files quickly re-tagged from “XYZ” to “Robbie Williams” and sent on their way. Inaccurate tagging is not a crime yet, and any competent lawyer could bounce a client out of a charge of piracy if a moment’s work was all that was required to identify the file as non-Williams.

This would make some inroads into CD sales, and legal internet downloads (assuming such a system becomes available), but how much? If the availability of genuine tunes via P2P systems is choked off by the record companies, the inexorable laws of supply and demand will see traffic in the remaining legal files skyrocket. Further, if and when this kind of technology becomes available to the mass market, it is likely we would see a seismic shift in the recording industry as anonymous people manipulate the voices of the rich and famous into anything they like.

The bizarre US Supreme Court ruling is the first major case decided on virtual images. If the Supremes feelthemselves unable to protect the public from people who are fundamentally child pornographers, can multi-billion dollar entertainment concerns expect to be better treated?

Comments: Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?