Billboard magazine has a fascinating piece on the rise of the use of deepfake vocals to create what the piece terms “faux songs”: ‘It’s Fan Fiction For Music’: Why Deepfake Vocals of Music Legends Are on the Rise.

The piece showcases the recent rise of deepfake vocals (vocals produced by computers that closely resemble the voices of well known artists) in new songs that are generally produced without involvement of the artists who’s voices are reproduced¹. It is a super interesting introduction into a world of music production full of delightfull wired-ness. I particularly like this track featuring a recreation of Travis Scott’s voice, which combines glitchy aesthetics with what the agency producing the song “considered to be cohesive lyrics”:

Sadly the main thesis of the piece (“Faux songs created from the original voices of star artists are becoming more popular (and more convincing), leading to murky questions of morality and legality”) does not do this phenomenon much justice. The idea that this type artistic production enabled by digital manipulation techniques can somehow be understood though concepts like “faux” vs “orignial” strikes me as outright silly. Quite obviously the “faux” works are just as orignal as any of the “originals” that the voices are derived from.

Even more problematic are continuing attempts to fit this type of creativity into existing concepts like authorship and copyright. The realities of creative production today are so far removed from the realities in the late 19th century (when modern ideas about copyright and authorship were codified) that it really does not make much sense to try to analyse them throught the lens of these concepts.

We urgently need new concepts that do a better job at recognising artistic innovation while at the same time ensuring that the value created in the process is fairly distributed.

Turns out that theverge does music reporting of sorts. their recent history of dubstep music (‘beyond lies the wub‘) contains two short passages that highlight the impact of digital technology on art (in this case dubstep music).

The first passage highlights how digital piracy killed the music industry might have actually driven the quality of music production in the past dececade or so:

“They cost nothing if you know how to get them for free, which most people did,” says Martin Clark. “The VSTs [Steinberg’s audio plug-in architecture] were available on peer-to-peer sites. So suddenly it’s democratized, right? You have zero cost to acquire a studio. You have this like, infinite [potential]. Anyone can be a producer if they can get a hold of these. From that pool you have a much larger pool to select who makes interesting music, as opposed to just who can make music. It’s no longer a question of whether you can make music, because the software is distributed, it’s accessible.”

On the other hand it seems that the same technology has also contributed to the decline of the very pirate radio stations which were once considered the backbone of London’s urban music culture. turns out that the same developments that allow pretty much everyone to become a producer, make running a pirate radio station a pretty silly exercise:

“It’s totally changed,” according to Boomnoise. “I mean, if you sit down at a computer with an internet connection, you can find pretty much anything you want to listen to, whether or not it’s in the radio format. And what happened was, I think with the internet was a shift from the local to the global, essentially, a shift away from having a very localized audience.”

Read the full article here

Had a bit of a deja-vu this morning when browsing through the economics section of friday’s NRC Handelsblad: Page two of that section contains a full page article (‘free jukeboxes against piracy’) on music streaming services such as spotify.com (click to enlarge, no online version available):

As you can see from the photo above, the article does come with a nifty info-graphic that illustrates how little artists earn from their music being available on services such as spotify (the number 4.882.758 in the green circle indicates how many times a single song needs to be played by users of spotify in order generate an income that is equal to the minimum wage in the Netherlands).

Now there is nothing wrong with this info-graphic as such, but there are two rather dubious aspects: the article is not providing any information with regards to the source of the data used and i had the strong impression that i had seen this info-graphic before (the deja-vu mentioned above).

A quick google query reveals that this was indeed the case: in april 2010 informationisbeautiful.net published a strikingly similar info-graphic (‘How much do music artists learn online?’) that obviously served as the basis for the illustration in the NRC:

The overall numbers are different, but that is simply the result of the fact that the NRC article is using dutch minimum wage (€1.416) as a reference point while the original used the US minimum wage ($1.160). this results in different sales numbers required to generate minimum-wage level income but for the rest the data used to illustrate the NRC article is identical (even worse they simply took the original artists revenue numbers that were expressed in dollars and simply re-stated them in euros).

It is bad enough that the NRC simply lifts these numbers from a website (which itself did took most of the figures from another blog, but makes that very clear by giving credit both to the originating blog and pointing out additional data sources) without giving credit. This is not only plain old-plagiarism (especially dumb if the subject of the article is ‘piracy’) but also grossly misleading: If the NRC article would have given credit to the source readers would have had the opportunity to take a look at the source themselves and would have learned that the figures presented by the NRC are missing an important caveat that is present in the original blogpost on informationisbeautiful.net:

Note: these figures do not include publishing royalties (paid to composers of songs). The full spreadsheet of data does though. You can see all the numbers and sources here: http://bit.ly/DigitalRoyalty

If you click through to the google docs spreadsheet with the full numbers you will see that if you count publishing royalties the amount of plays of a song on spotify required to make minimum wage is reduced by almost 75%. This still means that you need north of 1 million plays (which still is outrageous) but apparently 4.8 million required plays look a bit better when you want to illustrate an article.

This once again shows that the NRC really needs to realize that the days where they were the ‘quality newspaper’ more or less by default are over. Pointing out your sources is one of the key ingredients of credible journalism and unfortunately for the NRC it seems that new media outlets such as informationisbeautiful.net are lightyears ahead of the NRC here…

Just came across this brilliant foto while reorganizing my photo folder (i am locked away in a museum basement without any form of internet access: no wireless & no mobile phone reception):

I took this at a stall selling bootleg CDs at the Kwakoe Festival in Amsterdam southeast in the summer of 2006. To me it perfectly captures the essence of the revolution caused by napster when it launched 10 years ago (give or take a couple of weeks): If your work is not available on napster is is probably not good either. accept to be copied or become irrelevant.

Given that this has been going on it is stonishing how few people seem to have understood that this basic shift in how we relate to cultural goods has already happend….

I have always had an unspecified strange feeling about Tribe of Noise. Tribe of Noise is an Amsterdam-based online music platform that allows musicians to upload and share their work as long as they agree to make it available under a Creative Commons Attribution ShareAlike license (CC-BY-SA). Simply put this license allows everybody to redistribute the songs on the platform, make remixes of them and redistribute these remixes under the same licensing terms. In all cases credit needs to be given to the original artist(s). It explicitly allows for commercial (re)use of the licensed works and it is one of the least restrictive Creative Commons licenses (and the license that has recently been chosen by a huge majority of wikipedia editors to apply to all text on wikipedia). I like this license.

I am writing this (rather long) text because I have come to the conclusion that the way Tribe of Noise uses this license is confusing to people contributing to the platform and can in the end be harmfull for the reputation of the Creative Commons Attribution ShareAlike license and the Creative Commons licensing model as a whole.

Even though i have endorsed Tribe of Noise back when it was launched (something i should have never done and which i am obviously retracting by writing this) the exclusive choice for the CC-BY-SA license made by Tribe of Noise never felt in line with the way the Tribe of Noise (TON) chooses to present itself: It is an aggressive start-up and the founder (and selfdeclared ‘Chief of Noise’) Hessel van Oorschot aggressively markets it as such. They have been quite successful in getting media attention and have known to attract a substantial number of artists to their platform (at the time of writing there are 5878 members). The platform is promoted to artists as a way to get in contact with commercial users of music.

I have met with Hessel on a number of times in the past and among others i have invited him to the ‘Filesharing: Up or Down?‘ discussion that i organized at de Balie in Amsterdam in the wake of the Pirate Bay trail and the publication of the Ups and Downs study on the economic impacts of file sharing. That evening Hessel explained that one of the components of the Tribe of Noise business model is to license (for a fee one assumes) the music repertoire posted to Tribe of Noise to video hosting services so that they can use it in services like youtube’s audio-swap or offer the music to video makers that are looking for copyright un-encumbered music tracks to use in their videos.

However, such a business model is rather difficult to carry out based on the rights granted by the CC-BY-SA license. One of the key features of this license is that it requires derivative works of the original works to be licensed under the CC-BY-SA license as well (the share alike mechanism). This means, that every time a work licensed under a BY-SA license is integrated into another work (or the other way around) the resulting work needs to be distributed under a the CC-BY-SA license as well. If a video maker uses a short snippet of CC-BY-SA licensed music in a (long) video that is otherwise completely made by herself she needs to release the entire video under the CC-BY-SA license or she is in breach of the license (and thus infringing on the copyright of the musician in question)¹. This means that CC-BY-SA licensed music is pretty much useless for purposes like sound-swap unless the provider of the service intends to force the video makers to use a CC-BY-SA license themselves.

Given this Tribe of Noise probably has a hard time selling the repertoire uploaded to the platform to video hosting services (at least as as long as you assume that they would base these transactions on the CC-BY-SA licenses granted by their uploaders².

Back in April i did not really notice this contradiction. Tribe of Noise came back to my attention two weeks ago when i read about a Creative Amsterdam Award they had won at the ‘Creative Company Conference‘ in Amsterdam:

[…] The runner up, Tribe of Noise, was also given an honorary mention for their brilliant concept of music library in which creative commons licence (sic!) could be used for commercial purposes.

The fluffy language in the conference summary triggered my interest. What exactly is so brilliant about Tribe of Noise’s concept? Given the characteristics of the CC-BY-SA license explained above i failed to see how there could be a ‘brilliant concept of music library in which Creative Commons license could be used for commercial licenses’. Sure there are three CC licenses that allow for commercial use of the licensed works but it is hardly brilliant to allow people to upload content to your platform under one of them.

Given the limitations of the Attribution Share Alike license outlined above, one is inclined to assume that there are other parts to the business model that allow it to function and looking at the terms of use of Tribe of Noise it quickly becomes apparent that the business model of Tribe of Noise is not based on the rights granted by the uploaders via the Creative Commons licenses. Instead it relies on a much broader (and much less advertised) non-exclusive license granted to Tribe of Noise. Section 9 of the ToN terms of use, that you have to accept when you open a tribe of noise account contains these two sub-clauses:

1. Licenses Granted by the User

9.1 If you upload any content to Tribe of Noise or post any content on the Website, you grant:

1. a worldwide, nonexclusive, royalty-free, transferable license (with the right to sublicense) to Tribe of Noise for the use, reproduction, distribution, demonstration, making available to the public and performance of, and creation of derivative works from, that content in relation to the provision of the Services, and otherwise in relation to providing the Website and in relation to Tribe of Noise’s business operations, including the promotion and further distribution of all or part of the Website (and works derived from the Website or part thereof), in whatever media-format and through whichever media channel, now known or hereinafter invented;
2. andto every user of your work on the Website the following Creative Commons license: CC 3.0 By – Share Alike.

What is of interest here is the first of these two clauses. It essentially grants Tribe of Noise the (non-exclusive) right to do whatever they want with the uploaded music. For example they can sell (non-exclusive) licenses to third parties without having to pass on parts of the revenues generated to the musicians that have uploaded the music. Also Tribe of Noise can allow third parties to do whatever they please with the music that has been uploaded to the platform (without having to require them to give attribution or redistribute derivative works under a CC license). In short, by uploading a work to Tribe of Noise the artist grants Tribe of Noise a very broad license that allows them to commercially exploit the work while not getting any right of compensation in return³. In the same section of the terms of use the uploaders also grant all users of the Tribe of Noise platform the right to use the uploaded works under the Creative Commons Attribution ShareAlike license.

This dual license grant is nothing that is specific to Tribe of Noise. Almost all web platforms ask more rights in the content uploaded by their users than what uploaders are willing to grant to the general public (see for example the terms of use of youtube which are very similar to those of ToN). Some services (jamendo, blip.tv) specifically ask for the right to grant commercial licenses to third parties or run ads in connection with the content but in return they promise to share the revenues generated through such transactions with the uploaders, which Tribe of Noise does not do.

Having users agree with terms of service that include such an unbalanced license grant can hardly be called a ‘brilliant business concept’ and definitely has noting to do with ‘using a Creative Commons use for commercial use’: Looking at the Terms of Use of Tribe of Noise one has to conclude that using a Creative Commons license has nothing to do with subsequent commercial exploitation of the uploaded works by Tribe of Noise as the commercial exploitation is enabled by the parallel license grant to Tribe of Noise.

Even worse, i get the impression that Tribe of Noise uses the Creative Commons licenses in order to hide the fact that they are indeed trying to obtain a much wider license grant from the members of the platform. Apart from the above quoted section of the Terms of Use there is no mention of the additional license grant to Tribe of Noise on their website. Certainly not in the FAQ or the more info movie aimed at musicians (the two places where one would expect to find information about what rights are granted by simply uploading a work). Instead, in the video with more information for artists, Hessel van Oorschot states that they have solved a ‘legal challenge of sharing music with companies’ by using the CC-BY-SA license⁴:

[…] sharing music with other musicians and companies around the globe and getting more attention that is a legal challenge. But we came up with a solution so let me know how it is done on tribe of noise: Sharing music online even for commercial purposes is legal!! with help from legal advisors laywers (sic!) and creative commons!! Creative Commons Attribution Share Alike 3.0 Unported⁵.

Of course Tribe of Noise is free to ask their users for whatever license grants they want (and one could imagine that some artists do not object to give away the right to commercial use in exchange for exposure of their work on Tribe of Noise). However one would assume that a site that states ‘Tribe of Noise means music and respect!‘ openly informs its uploaders what rights they are granting to the platform in exchange for being allowed to upload a work to the platform. Hiding such information in the legalese of the Terms of Service does not really show respect for the musicians using the platform.

First of all this is objectionable because it relies on the lame old trick of hiding stuff in the Terms of Use that one has to click though during a registration process and then using a CC licenses in order to imply that the site does respect everybody’s rights. However, the conduct of Tribe of Noise is also objectionable on a more profound level as it shows that the team behind Tribe of Noise apparently thinks that it is ok for them to make commercial deals with works authored and performed by other people without reimbursing them for such uses. While writing this i have asked Hessel van Oorschot if Tribe of Noise has a revenue sharing model in place and he has responded that they will certainly start working on a honest sharing mechanism. If such a revenue sharing mechanism gets introduced to the platform in the future that is certainly a step in the right direction but it does not aliveate my other point that Tribe of Noise is far from transparent when it comes to dealing with the copyrights of the members of the platform.

The English translation of the Ups and Downs report on the Economic and cultural effects of file sharing on music, film and games is out [thanks Natali!]. You can download the 130 page report written by a research consortium formed by TNO, SEO & IvIRon the IVIR website [pdf].

I have mentioned the report (commissioned by the ministry of Economic Affairs, the Justice Department and the ministry of Education Culture and Science here before [first thoughts, popular science edition] and as far as i am concerned it is well worth the read… [thanks Natali!]

I just finished reading ‘Ups and Downs – the economic and cultural impact of file sharing for music, film and games’ (see my earlier post for context). All in all the full version does not contain a lot of surprises when compared to the executive summary (which my first post was based on): It is a well written report that, although it makes a lot of sense to someone familiar with the subject, does not really come up with much new insights either. The strength of the report is that it places file-sharing within the wider social and economic context (as opposed to placing it solely within the economic logic of the entertainment industry). While they sometimes appear naive (it does not seem to occur to them that buying CDs or renting DVDs from the video-rental-shop is rapidly becoming obsolete from a technicals point of view) the researchers do seem to have a fairly good understanding of what is going on.

The core of their argument (to be found in sections 5 & 6) is that there is no direct causal relationship between file-sharing and the decline in revenues in the music industry. On top of this the researchers argue that even tough it is likely that there is a substantial decline in revenues for the recording industry as a result of file sharing, this is offset by an even more substantial increase in welfare for the general public (or at least that proportion of the general public that downloads musical works). This finding is based on an economic model that is summarized in figure 6.1:

Figure 6.1 from ‘Ups and downs’ – blue boxes and grey arrows and labels mine (personally i am a bit surprised by the relative amounts of lazy and smart peple implied by this figure. life experience tells me to expect the opposite distribution).

1. The orange block represents the revenue generated by selling recoded music in the absence of file sharing, which equals the maximum possible revenue for the recording industry. In this situation the rich people(a.k.a stupid people) profit (save money) because they would have been willing to pay more than the market price. All the people to the right of the orange colored block simply could not afford to buy recorded music.
2. With the possibility of file sharing available to consumers we see a shift: a certain amount of people who used to buy recorded music now download it for free (‘cheap people‘). In addition the smart people (a.k.a poor people) now have the same access to recorded music as all the others and finally there also is a group of lazy people who simply cannot be bothered to download because they perceive the process as too burdensome.

When comparing the changes between (1.) and (2.) in economic terms the researchers conclude that while there is a negative impact on the recording industry (caused by the cheap people) the fact that the smart people now also have access to recorded music represents a much bigger increase in economic welfare (and does not hurt the recording industry as it is ‘demand without purchasing power’ that is being met)¹. As mentioned in my earlier post the researchers value the damage to the recording industry at a maximum €100 million p.a while they value the socio-economic gain caused by the increased access to recorded music at at least €200 million p.a.

Personally i am not sure if this will be of any consolation to the recording industry, but as far as i can see it is a fairly adequate description of the current transformation process: A business model anchored in an outdated means of distribution is (partially) being replaced by a social practices that are (a) more in line with the technological state of the art and (b) provide greater socio-economic benefits to society at large.

For the rest the report does not contain much news: Chapter 3 (‘the legal framework’) gives a solid and up to date (it even includes last years legislative battle around the EU’s telecom review) overview of the legal implications of file sharing (in the Netherlands) and Chapter 5 gives an overview of recent studies on the economic impact of file sharing². Apart from the economic model described above chapter 6 also lists a number of ‘dynamic and indirect’ effects of file sharing that are fairly obvious but nevertheless worthwhile to repeat: The researchers argue (p.123) that while it is likely that file sharing hurts big successful artists (as cheap people will buy less CDs from them) it has a positive impact on smaller artists (as it allows more people to sample their works, which will turn some of these people into buyers of their CDs or make them attend concerts). More interestingly the researchers also argue (p.125) that acceptance by consumers of the substantial increases in ticket prices for live-concerts has to be seen in the context of file-sharing: The increased willingness to pay high prices for concert tickets may be due to the fact that consumers are aware that they are spending less on recorded music (or the other way around: as they have to pay more for concert-tickets consumers are less willing top pay for recorded music and resort to file sharing).

When it comes to their conclusions the researchers note that file-sharing is here to stay and that we (the recording industry) are beyond the point of no return: It is impossible to build a successful business that is solely based on trading recorded music. According to the researchers is is also highly unlikely that there will be a point in the future where all music will be obtained from authorized sources (p.136). Given this they argue (inter alia, their official recommendation comes down to a pathetic paragraph where they make a plea against criminalization of end users and for more awareness building among file sharers) for a model where internet service provides offer internet subscriptions that include a fee for the access to copyright protected content (a.k.a the content flatrate).

Early last year the Dutch government (the ministry for Economic Affairs, the Justice Department and the ministry for Education Culture and Science to be precise) commissioned a research report on the socio economical aspects of (peer 2 peer) file sharing. Last week the research consortium formed by TNO, SEO & IvIR published the final version of the report titled ‘Ups and Downs – the economic and cultural impact of file sharing for music, film and games’. This 141 page report looks into the economic and cultural consequences of file-sharing for the music, movie and games industries. The central conclusion of the report is that:

The research shows that the economic impact of file sharing on the Dutch economy is strongly positive when viewed from both short term and long term perspectives. As a result of file sharing consumers get access to a wide range of cultural products. This has a positive impact on the economy […] According to estimates the positive economic effects for consumers amount to 200 million euro per year. On the other side the maximum decrease in revenues for producers and publishers of sound recordings is 100 million euro per year. [page 3 of the report, translation mine, an official english translation hereof the entire report is forthcoming]

It is refreshing to see a government sponsored report that recognizes that while one part of the entertainment industry (music) suffers some losses, these don’t necessarily outweigh other – positive – effects of file-sharing: According to the researchers, file-sharing gives access to a wide range of cultural goods and is often used to sample works that are bought later. Most file-sharers would have never bought all the content they downloaded, and having access to such a large media library has positive effects on the social well and economic position of downloaders and the society as a whole.

One of the most interesting observations in the report is that while revenues related to the sale of music are steadily declining, the overall amount of money spend by consumers on media for entertainment (ie music, videos and games) is relatively steady. It appears that money that is not spend on music is instead spend on video games. This can be seen as an indication that the real cause of the decline of revenues in the music industry is not primarily caused by file-sharing consumers, but by intra-industry competition: people simply spend their entertainment euros differently.

Instead of music CDs consumers buy Guitar Hero or Rock Band (plus extension) packs these days. This is one more reason why the recording industries’ push for stricter IP enforcement will probably not do them much good: It gives consumers little reason to not spend money on games and go back to spending it on music CDs. From the consuer perspective a €50 game is much better value for money than a CD that contains one or two really good tracks.

After the presentation of the report on Saturday in Groningen a Buma/Stemra representative called the report ‘scary’ [‘greizelig’ in Dutch]. One can only hope that his fear will transform itself into the insight that the industry will need to change if it wants to ensure it’s survival. But if the past is any indication the most likely reaction to this fear will be a counter study that comes to the conclusion that downloading is extremely bad for the economy and that we need much stricter IP-enformcement. In the meanwhile one third of the Dutch citizens will continue to download and go to concerts and buy CDs and buy DVD and go to the movies and fail at becoming Guitar Heros…

The fabulous electronic performance arts venue/center/place/non-place [it is hard to describe what this place really is] steim in amsterdam is in danger of loosing it’s funding. over the past years steim has received funding from both the dutch ministry of culture education and science and the city of amsterdam. for some reason (probably because they never go there or because their idea of culture is quite limited) the advisory bodies for both the ministry and for the city of amsterdam have decided that steim should not be supported under the upcoming (2009-2012) 4 year plans for culture (yes they do still have soviet style 4 year plans for culture here in the Netherlands).

Needless to say this would be quite a bad thing to happen. steim is one of the very few places in amsterdam that are unique and even if it caters to a ‘niche audience’, it manages to bring in a remarkably diverse set of artists from all over the world that have made it one of the best places in town to hang out and broaden your horizon. for steim the loss of structural support would probably be quite devastating and the steim crew is calling for support:

Things are not well at STEIM. We are in the danger of losing our structural funding from the government, based on a review from the advisor board which called us ‘closed and only appealing to a niche audience’. The outlook isn’t exactly bleak, but at the moment our future is unclear.

As we see you as an important friend and colleague of STEIM, we would like to ask you to help us present our case that we are connected to a diverse network of professionals and that our work has significant influence on both a Dutch and an international community.

Will.i.am from the Black Eyed Peas brought together a bunch of american musicians and actors to record a video clip (song) based on the ‘yes we can’ thank-you speech held by Obama on January 8th in Nashua, New Hampshire. The whole clip (‘yes we can song‘) is extremely well done (you might want to call it ‘slick’) and certainly makes me want him to win as many delegates as possible tonight. Especially since, towards the end, there is some proud display of a pirate flag tattoo by one of the female vocalists (excuse my ignorance, but i have absolutely no clue who most of these people are, but then i have mostly listening to Bach in the last couple of days):

Bonus points for the reader who tells me the name of the depicted person.