6 posts categorized "Music Recognition"
October 31, 2012
July 27, 2011
December 14, 2009
Demand Media (see this brilliant Wired piece), one of the biggest producers and distributors of online video (see the Mary Meeker / Morgan Stanley presentation here; page 42) produces 100s of 1000s of videos on topics that are solely determined by a proprietary algorithm that crunches data on popular search terms, keywords and their current rates on search engines, and information about how many web pages already cover the topic. If a topic is 'hot' and not yet covered, Demand Media commissions an army of freelance video makers, at $20 per video (!), to quickly produce short clips on the topic, e.g. on 'how to heel-flip on a skate board' etc.
Wired's Ryan Singel talks about AOL's similar new plan: "AOL’s new chief plans to combine algorithms, marketing partnerships and cheap freelance writers in order to turn the stale web property into a vibrant online content factory pumping out stories to fit the zeitgeist..." - all for the sole sake of taking advantage of the Google-page-ranking system i.e. to subsequently yield more advertising dollars.
With both examples, the idea is simple: to produce a huge a and hyper-distributed amount of fast, short - and above all - ultra-cheap content that is a perfect fit with the hottest and most expensive keywords on the web, today, so that the maximum advertising rates can be achieved at all times. In other words, this 'content' only exists as a way of garnering advertising revenues based on keyword popularity - hardly what I would consider 'adding value to the content ecosystem' ;)
In music, recommendations are already generated largely by software algorithms and data-crunching recommendation engines; some people even go as far as predicting whether a song will be a hit or not, using smart software engines (disclosure: I am on the advisory board of this company, uPlaya). Google's page-ranking system relies entirely on machine-intelligence, of course, and Twittercounter's top 1000 list is, of course, generated solely by data feeds - not by human editors (such as my own site, Futerati, which will, btw, be relaunched within the next 10 days).
Techcrunch's Arrington talks about the end of crafted content. Wired calls Demand Media a factory that stamps out money-making content. The Inquisitor talks about how this kind of approach is turning the web into an obese mess. The Washington Post sums it up, rather gloomy: "these models create a race to the bottom situation, where anyone who spends time and effort on their content is pushed out of business."
Here is what I believe:
- Content that is produced only because of keyword popularity and because eager and / or desperate producers (no blame there, btw, just stating a fact) are willing to work for exceedingly cheap rates may bring in the immediate bacon but in my opinion will not last in terms of continuous popularity and therefore in long-term revenues. And if they do, great for them - but it does not mean that all content production will move in this direction. Every single person that likes to eat fast-food still knows the difference between Wendy's and a nice meal: yes, it's more expensive and it takes longer but it's a much better experience, and it makes you feel better. Fast food chains simply co-exist with 'real' restaurants of all kinds, everywhere - and that's what we will have in the content industries, too. If you want to make a quick buck by starting a fast-food franchise, go ahead. I, personally, don't like to eat fast food, nor would I enjoy running a McDonald's franchise so I will go a different route.
- There will always be people who are willing to pay for better, deeper and more 'serious' content, and, in my opinion, increasingly so (mostly because of the trend towards mobile content consumption) - we just need to find new, web-native models of getting paid for content and translate the value of attention into tangible $. Yes, this is a real challenge, today, but new ways are emerging that will indeed provide plenty of resources for the continued creation of high-quality content. Let's have some imagination. Just because 100s of 1000s of aspiring teenagers want to see those free, ad-supported videos on pole-dancing does not mean we won't be selling video-on-demand, DVDs and books on more serious (or indeed, the same) subjects. Cheap, free and low-quality options have always co-existed with more expensive ones, and while the web has made this trend a lot more pronounced it will not spell the end of well-produced and high-quality content - the cheap stuff is simply first because so far we are lacking business models for the better stuff (for the most part).
October 31, 2005
Just ran across this new company that uses Wikipedia listings to drive music recommendations. Gotta check it out some more. "Q: How many artists and albums do you have in your database? bA: We had 75,786 artists and 394,399 albums last time we updated this page." I am always reluctant to download these kinds of clients but I may try this one ;). The company name is probably not a very good choice, though, imho... Brad Hill reviewed the service here:
"All in all, this is the best massaging of my music collection I’ve seen in quite some time. Now, if SonicBreakdown were to create a stand-alone music player with this explosion of information built into it—that would be something to really shout about..." he says
October 17, 2005
Good essay by Derek Slater
"....In this way, the tools may contribute to "semiotic democracy." Rather than just passively receiving music, consumers can actively engage it, altering and adding meanings and impressions of the musical works. When consumers stand beside (or replace) the traditional tastemakers, they also may diminish the control with which those traditional tastemakers had on how we engaged music. Consumers can shape the prism through which they view these works...."
April 13, 2005
claim to have a MATHEMATICAL solution to recognizing HITS, and to recommend similar music to consumers. Check it out. From the site: " April 7, 2005 (Barcelona, Spain) - On February 14th, 2005 Polyphonic HMI predicted that Aslyn's song, Be The Girl would break into the Top 30. Less than two months later it has. Polyphonic HMI predicted the success of Norah Jones' debut album, Come Away With Me, well before it won 8 Grammys in 2003. Soon afterwards the global microscope was directed at this firm's curious technology called Hit Song Science (HSS for short), focusing the music industry's as well as the media's attention on their ability to predict the likelihood of market success for songs well in advance of the music's release." Sounds too good to be true:) but let's take a look!