algorithm

Tech blogger finds Net radio "inhumanity" makes him miss "Rodney on the 'ROQ"

Monday, July 1, 2013 - 1:40pm

"We've lost the genius of the great DJ, the human who exposes us to new music, who takes us on a trip every Sunday night," writes commentator Joshua Fruhlinger, feeling nostalgic for bygone on-air musical tour guides like KROQ's legendary Rodney Bingenheimer.

But it's not just the dearth of great DJs in broadcast radio. Fruhlinger says Internet radio that creates listening experiences algorithmically are "inhuman." Not in the literal sense (though that's true too), but in the sense they lack the capriciousness and serendipity that a human-guided listening experience entails.

He writes: "Imagine if we replaced, say, chefs with machines. 'We noticed you like cheeseburgers. Here, have a bacon cheeseburger.' That's a fair assumption, but when I eat a cheeseburger, I want fries. Sometimes I want a salad instead. People are weird like that, and no machine can take me from The Vaccines to Neil Diamond."

Fruhlinger is an Engadget editor whose blog "This Is the Modern World" explores "the culture of consumer technology." Read the piece here.

It takes more than a good algorithm to top Gizmodo's "Best Streaming Radio" ranking

Friday, May 17, 2013 - 12:20pm

Gizmodo contributor Mario Aguilar decided to find the best "automated DJ" on a streaming music or Internet radio service, pitting eight top services against each other for his "The Best Streaming Radio."  

His original intent was to find the service with the algorithm that created the best sounding user-generated station (he only considered services that offer "generative playlists" -- the ability for users to create "stations" on the fly by simply typing in a single artist, song title, or genre). Nearly immediately he realized picking a winner based solely on a good mix was futile, as they all, by and large, do a pretty good job at this task.

So he widened his considerations to other facets of the services -- "integration with social networks to the design and overall usability of each service's unique features" -- for the shoot-out.

I'll let you click through to see his ranking, but I'll include a few of his points here about specific services:

Turns out he's not a fan of Clear Channel's iHeartRadio service, which (he wrote), "does so little and doesn't do it especially well." He called it "radio in the most traditional sense," and didn't mean it as a compliment, since "regular radio stations are terrible, which is why we turned to the Internet in the first place."

Pandora, which is "showing its age," only fared a little better. He found the service "less evolved" with only "very basic" social integration -- but "if Pandora has a selling point it's simplicity."

And though Last.fm may still have something to offer, in 2013 it "feels ancient." In fact, its concentration on scrobbling (tracking what you listen to on other services) makes it "more of a recommendation engine than a polished way to listen."

Aguilar actually had some high praise for Slacker's human-curated stations, which he says offer the "kind of variety you can't get from a machine." But, alas, this shoot-out was for algorithm-driven "generative playlist" channels. And the newly-redesigned Slacker interface seemed "ambitious and very good- looking, but it's pretty confusing" and "needs streamlining." Perhaps even worse, he said Slacker "completely missed the potential of social" media integration.

So -- does any service offer anything he likes? See which topped Gizmodo's The Best Streaming Radio here.

Algorithmic intelligence still needs the "human touch," says NYT

Monday, March 11, 2013 - 12:10pm

As powerful and important as computer algorithms have become for any number of problems, The New York Times reports today that human judgement is still integral for nearly any service using them. Since "computers themselves are literal-minded, and context and nuance often elude them," it's still very necessary to have "people evaluate, edit or correct an algorithm’s work... assemble online databases of knowledge and check and verify them... (and) interpret and tweak information in ways that are understandable to both computers and other humans."

Even at Google, "where algorithms and engineers reign supreme," humans are contributing more to search results, writes The Times (Just one example: Type a celebrity's name in the Google search bar, and you'll probably see a summary about that person on the right-hand side of the results page... those are drawn from human-edited databases.).

Read more in The New York Times here.

Firms like The Echo Nest use algorithms to assemble databases of "music intelligence." Leading webcaster Pandora uses its own combination of music analyzed by humans but assembled in playlists by algorithm (though likely with significant influence of the "thumbs up/thumbs down" listener ratings it collects). As a marketing strategy, several other webcasters recently have positioned themselves as "curated by music experts, not algorithms" (more in RAIN here).

More RAIN news "quick hits": Talk radio on the Internet; music's positive effect at work; the post-IPO Pandora; M. Ward's Net radio app

Friday, August 17, 2012 - 12:30pm

-- Paragon Media Strategies' "The Blog" examines Internet radio talk programming. "For the aspiring talk show host trying to maximize the number of listeners, Internet Talk sites may offer more potential than the small audiences that the overwhelming majority of AM stations can muster," Paragon's Larry Johnson writes. "The best strategy may be to have an ‘aggregator’ or site pick up the show. There seems to be plenty of Internet Talk sites including Stitcher, Voice America, Live 365, WSJ Radio, and TuneIn." Read more here.

-- The New York Times reports on various studies of the positive effects of music in the workplace. "In one study involving information technology specialists, (University of Miami assistant professor Teresa Lesiuk) found that those who listened to music completed their tasks more quickly and came up with better ideas than those who didn’t, because the music improved their mood." Let's hear it for more Internet radio at work! Read more here.

-- A former Pandora employee discusses how much different work (and, supposedly the way the Pandora music algorithm works) changed post-IPO. "It used to be that you’d put in Modest Mouse and then hear all these crazy college indie bands. That was how it was created. It was great. But people in the Midwest hated it. Now, you put in Modest Mouse and you hear Maroon 5. It’s much more like radio. Some people got angry, but the majority like the changes," the anonymous former employee told BuzzFeed, here.

-- The musician M. Ward has created his own iTunes Internet radio app to allow anyone to locate almost 1,000 of the country's best, independent radio stations -- handpicked by Ward -- and easily stream them online. Read more in The Huffington Post here.

Taking a side in "man vs. machine": Slacker employs radio programmers to craft stations

Tuesday, August 7, 2012 - 12:05pm

In the company blog, Jacobs Media president Fred Jacobs suggests "the human element" -- something that's been a part of broadcast radio since the beginning -- might be key for pureplay webcasters like Slacker to compete against the 600lb. gorilla that is Pandora.

As Jacobs mentions, Slacker’s senior radio program manager Mat Bates, a veteran of broadcast radio, spoke at our RAIN Summit Midwest event at The Conclave last month. Bates spoke of lessons learned in broadcast, and how they could benefit a pureplay webcaster like Slacker: namely, music presentations crafted by knowledgeable and passionate human beings, and not computer algorithms.

The first wave or two of online music services seemed to us to be a reaction to everything bad that broadcast radio had become: lowest-common-denominator playlists with no surprises, an overload of commercials, and air talent relegated to reading promos. The (largely) non-radio people (quite often from the tech world) were the entrepreneurs of the first generation of online music and radio services, and they developed products that that renounced the "evils" of commercial broadcasting. Some would argue that in doing so, their services were prone to veer in the opposite direction: they often had unfocused playlists, no clear plan for monetization, and lacked any sense of "humanness."

And it really brought to the fore the question: Does the human insight bring something to music programming that we can't (yet) replicate with algorithms and machines? And, what we think is more interesting: does the consumer truly benefit? Is the listening experience so improved as generate increased (and monetizable) listening so as to justify the costs of employing human music experts? And, is this a worthwhile branding advantage (in other words, does the listener realize and care whether her music is "curated" by a passionate musicologist, or cranked out by an algorithm and a database?)? 

Naturally, radio programmers will cling to the notion that what they do with a 300-song playlist simply can't be replicated electronically; likewise, technophiles will smugly chuckle at them. Slacker's strategy seems to acknowledge the value of broadcasting's human element. The service employs experienced broadcast radio programmers (some still working in radio), and they even insert occasional brief "jock" breaks between songs on some channels. What we find interesting is that Slacker doesn't explicitly promote this to the consumer. We can't find anything on the site, nothing in the programming itself, that makes it plain to the listener that "Hey, you're hearing this song or artist because we like it, and we think you'll like it too."

Has Slacker concluded that human curated music programming is so superior to algorithms as to be self-evident -- simply, the proof is in the listening? And that promoting the fact that "hey, we have humans crafting your listening" is just more promotional noise? 

Read Fred Jacobs blog post here.

ReadWriteWeb: As automated as music recommendations get, there'll always be a need for "human intervention"

Friday, June 29, 2012 - 11:45am

Computer musicComputer algorithms and automated services are currently "the dominant approach" for web radio music recommendation and discovery services, writes ReadWriteWeb. To some extent, Pandora, Last.fm and The Echo Nest (along with the services it powers, like iHeartRadio) all take this approach.

"But they're by no means perfect," argues ReadWriteWeb. "We still need human brains to do some of the listening and interpretation that goes into music recommendations. That's why Pandora's algorithm works so well and why Songza has such promise. Both services, to varying degrees, rely on human intervention to help curate and cross-reference songs."

The article points to Spotify's app selections as a good example: it offers automated services from Last.fm and MoodAgent, as well as curated apps from Pitchfork, Rolling Stone and others.

"Algorithms and APIs can do amazing things. But at the end of the day, determining what music - as well as art, movies, books and other media - people will like still requires human beings with real ears connected to real brains. The future of music discovery will rely on both."

You can find ReadWriteWeb's article here.

Syndicate content