Exclusive Interview: AMG President Karl Ryser

By Dave Califano, Associate Editor

All Media Group (AMG) President Karl Ryser talks about the company's technologies for assisting consumers with searching and categorizing their voluminous digital media archives.

Consumer Electronics Daily News: Give our readers a little bit of background about All Media Guide.

KR: At AMG, our goal is to assist people in the enjoyment of the arts and entertainment, specifically with music, movies and games. With the vast amount of content available in the digital world, we provide the information and editorial that helps people discover and navigate the universe of entertainment content. This information is available through our Web sites [AllMusic.com, AllMovie.com, AllGame.com], through business-to-business licensing to our partners and through our technologies [Lasso and Tapestry].

CEDN: AMG is described as a critical partner for any music-based service or device. Why?

KR: For a consumer or user, arriving at a store, subscribing to a service, or filling up a device is just a first step. What to do with a seemingly endless amount of choices is the difficult part. AMG assist(s) the user/consumer in answering questions like who played on this album, what tracks are on this album, what's the best album or songs by this artist, what are the best new releases in my favorite genre, and what do the experts have to say about this release. Our Lasso technology recognizes the physical media and provides clean and standardized tags for digital media—the essential first step. The Tapestry technology gives the user the ability to navigate through their content by choosing the best music for a mood or situation, and also makes recommendations for further discovery.

CEDN: AMG’s recently-announced a partnership with IBM and Avoca. Tell us about it.

KR: As music collections get larger and larger, navigating through these big collections can get to be a real chore. One of the most immediate ways to be able to access the exact artist, album or song that you're looking for is to ask for it by name. We've gotten requests, primarily from the automotive industry, to be able to allow their users to select music using voice control and search technology, so we partnered with Avoca Semiconductor who has done significant work in the area of creating user interfaces for music collections. Their technology is based on IBM's ViaVoice, and we are enabling music fans to be able to instantly access their favorite song or artist, not just by the exact name, but by tying these commands to things like nicknames and aliases, so the user can just say Play The Boss and they will get Bruce Springsteen, or Play CSNY and get Crosby, Stills, Nash and Young.

CEDN: Can this voice recognition functionality be built into a device such as an MP3 player or stereo system, or is designed primarily for search at the digital media point of purchase?

KR: This technology can be used anywhere people are trying to find music. One of the interesting aspects of using voice control in the car is the issue of safety—a user who is trying to flip through a thousand or more songs on an MP3 player or internal hard drive is likely to take their eyes off of the road for extended periods of time. With safety being a big issue in the automotive telematics industry, we feel as though this will become a must-have for upcoming model years.

CEDN: What’s on the horizon for AMG?

KR: An interesting product that we are very excited about is Tapestry, which is our music discovery and recommendation technology. Basically AMG recognized that the way people are purchasing, storing and listening to music has been changing. Albums as we know them are not necessarily the preferred method of obtaining music anymore and many people are choosing to listen to music on a song-by-song basis.

We recognized that to accurately be able to sort music and make accurate recommendations, there needs to be a good description of the music on the song level.  So our editors sat down with headphones and started profiling the music based on a number of different factors—the Style of the music, the Mood of the music, the Themes of the music, which events, environments or activities the music is appropriate for, the instruments being played, the tempo of the song, etc. Tapestry is then able to analyze the information associated with a song and bring back accurate recommendations for more songs that could be purchased, or build a playlist from a user's collection based on Latin+Pool Party or Jazz+Romantic Evening or even create a streaming radio station based on the user's specific mood.

We are in talks now with several companies that are extremely enthusiastic about giving their users this kind of experience. An online demo can be seen here.