Monday, December 3, 2012

You Can't Just Fingerprint Money: Why A Carefully Considered Public Performance Tracking Strategy Is In Our Best Interest

There's nothing like a well-implemented technological innovation. Whether solving a problem or enhancing a process, when the creation is thought out in full and applied to the right situation the results can be astounding. Efficiency. Opportunity. Productivity. Equity. Transparency. Advancement. The benefits go on and on. And that, of course, is what drives society's continued investment in new technology.

But while the fruits of technology-done-right can be far reaching and plentiful, the shortcomings of half-baked creations can be equally detrimental (I'm looking at you, DRM from the immediate post-Napster era). When new technologies create new problems the benefits carry less weight, and at a certain point if it doesn't work well, it doesn't work at all.

This is the exact dilemma that sits at the heart of ASCAP, BMI and SESAC's strategy for using fingerprinting technology to track TV performances. The systems they currently employ to license the rights, identify usages, and distribute monies associated with the public performances of the works they represent are well established and proven. To put things in context, the three PROs together process hundreds of BILLIONS of performances and distribute close to $2 BILLION to songwriters, composers and publishers every year. The system clearly works. But that said, the PROs also acknowledge that there is still room for improvement and are carefully considering new technologies to aid in more accurate and transparent accounting.

Last month TuneSat, one of a handful of audio fingerprinting and usage detection companies servicing the television market, issued a press release that took aim at ASCAP and BMI's methods. Both ASCAP and BMI have so far refused to accept performance tracking data as collected by TuneSat, and TuneSat is spinning this to paint the PROs as entrenched, archaic institutions.

TuneSat states that "up to 80 percent of music on TV goes unreported or misreported" and that manual reporting is prone to error and manipulation. They believe that tracking through digital fingerprinting is the way of the future. Manual reporting, ultimately left in the hands of licensees, is prone to error. After all, we are only human. Likewise, it is prone to manipulation, as reported by the recent allegations against music supervisors and producers at CBS. But it's important to note that these are far and away the exceptions and not the rule. Before jumping to conclusions about ASCAP and BMI's refusal to accept TuneSat's data, it's important to look at both sides of the issue. Audio fingerprinting will be an important part of the PROs' strategies moving forward (a prospect we're very supportive of too), but only when the benefits are clear and inarguable. Unfortunately, the technology has yet to prove itself in this regard.

There are several considerations that must not be overlooked when considering how PROs employ audio fingerprinting technology. Many composers today rely on industry standard instrument and sound samples to create their compositions. With that in mind, it's not uncommon for a digital tracking service to mis-identify a composition based on a commonly used sample. In order to make the proper royalty distribution, these "false positive" IDs must be rectified, which can be a very costly and time-consuming process. In some cases, this reconciliation costs more than the value of the performance itself, which ultimately harms content owners instead of helping them.

But there's a larger issue here. A non-starter, if you will. As it stands now, digital tracking technology cannot identify the TYPE of performance (i.e. featured, background, underscore, commercial, promo, theme etc.). Since this usage type information is paramount in determining the value of the usage, additional work would again be needed to properly reconcile the IDs - making these systems unusable.

Like so many things, there are always several sides to an issue.  Ultimately, the PROs exist to serve their members, and every decision they make must be in their members' best interests. There are many potential benefits to a well implemented digital tracking service. The ability to analyze dirty audio, to instantly detect and report usages, and to quickly capture and post audio samples of every usage are all features with great potential. And make no mistake - we all want it to work. But the decisions that the PROs make regarding digital tracking services for television are not merely matters of adopting new technology. They have proven their willingness to do so in several other areas, including radio, promos and commercials. The decisions are matters of how well the technology works in serving their needs and their membership.


  1. Excellent analysis. But there are even more issues that could have been covered. With 80 years of music in motion pictures and 60 years of TV music, who is going to fingerprint all the back catalogs that still get performances? And anything performed live on TV has no reference copies on file for digital comparison. As a prospective technology, fingerprinting will be a great help to the PROs to be more (but still not 100%) accurate. But for music already contained in programs, it is unlikely that fingerprinting will ever take place, leaving the current cue sheet system as the only method for tracking and reporting those performances.

    1. Great points Steve. Thanks for adding these in. It seems clear that, at least as things stand now, it is most prudent to look at fingerprinting as a potential supplemental technology as opposed to a direct replacement technology. Of course, we'll be watching this closely.

  2. Hi Joel,

    I'm curious to learn if ISRC codes can help the tracking of music performances on TV shows (background cues, etc).

    Is this technology being used anywhere in the world for this type of thing?

    1. Hi Alan. Good question. ISRC's are not currently required when registering songs or compositions in the U.S. If used more stringently, they may be helpful to ensure that if the recording is identified correctly, the composers, songwriters and publishers will be too. But they still don't necessarily ensure that the usage will be identified correctly in the first place, whether by a fingerprint tracking service or a human being creating a cue sheet. The best way to address this is to make sure that the information provided to licensees is as clear and complete as possible. Similarly, it is important to keep in regular contact with licensees to confirm the music they are using and that they are correctly marking your music on their cue sheets. It's always in your best interest to ask for a copy of the cue sheet yourself so you can make sure it's accurate and that it has been submitted to the appropriate PROs.

  3. Two quick points:

    Audio recognition isn’t foolproof. In my experience working at the ill-fated Landmark (BMI), the false positives created by sampling, re-titling and other sound anomalies never amounted to more than 1% of the overall reported broadcasts in a given period (one month, 24/7 monitoring, 400+ FM stations). I agree that the manual reconciliation of these inaccuracies can be costly, but c’mon… we have to start somewhere.

    Also, the TYPE of performance (Feat, Visual Vocal, Background, Advert, etc) can be gathered by marrying the context of the specific performance timestamp with the known broadcast schedule of the source audio. This radio and TV data can be licensed or compiled. Again costly, but once you’ve ramped up and streamlined the process, the benefits will eventually outweigh the costs.

    Maybe after ASCAP and BMI merge we’ll see some progress made in the “honesting-up” of performance identification. Until then, it’ll be smoke and mirrors and silence.