FIG. 1: A parametric EQ can go only so far in correcting room anomalies (a). MultEQ creates a far more precise filter that is the inverse of the room''s measured response, resulting in a much flatter final response (b).
Photo: Courtesy Audyssey Laboratories
As all electronic musicians know, reproducing sound from speakers in most rooms can be fraught with problems. In particular, the sound reflects from the walls, ceiling, and floor, interacting with the sound coming directly from the speakers as well as other reflections. These interactions can seriously alter the spectral profile of any sound emanating from the speakers, causing what you hear to be distorted.
In addition, the distortion is different depending on where in the room you are located. For example, standing waves occur at frequencies whose wavelengths are simple multiples of a room's dimensions, causing nulls at some locations and peaks at others. These so-called room modes are the bane of all recording engineers and the main reason that recording studios are so carefully designed.
For many years, acousticians have used graphic and parametric equalizers to correct such problems, which can persist even after the room has been acoustically treated. But these remedies go only so far. In most cases, conventional equalizers don't have enough bands, and the width of each band in a graphic EQ is not controllable. As a result, the correction is less than perfect (see Fig. 1a).
A company named Audyssey Laboratories (audyssey.com) has come up with a different approach to room correction that goes much further than conventional solutions. The system, dubbed MultEQ, began as an academic research project at the University of Southern California under the direction of professor Chris Kyriakakis and one of his PhD students, Sunil Bharitkar. In 2002 they went on to found Audyssey with USC Audio Lab researcher Phil Hilmes and USC professor Tomlinson Holman (of THX fame).
The basic idea is to play test tones through the sound system in a room and measure the room's acoustic response at multiple listening positions. MultEQ then creates a precisely tuned inverse filter that counteracts most room-based distortions (see Fig. 1b). The technology, rather than using a few EQ bands, uses finite impulse response (FIR) filters with hundreds of control points, allowing a much more accurate and comprehensive correction in the time and frequency domains. A majority of the system's correction power is dedicated to the low end of the frequency spectrum, because that's where the worst acoustic anomalies arise.
Instead of simple averaging, the MultEQ algorithm accounts for the differences at each listening position using a technique called fuzzy clustering, which assigns locations with similar responses to a group or cluster. The process also calculates weighting factors that reflect the relative importance of each location to the other clusters. This allows the system to devise one filter that corrects the anomalies at various locations over a large listening area. The number of measurable locations is limited only by the amount of memory and DSP horsepower available.
MultEQ has been implemented in many modern A/V receivers to automatically calibrate the sound for up to eight locations in a consumer's listening room, and it has enjoyed accolades from many reviewers. Now Audyssey hopes to make similar inroads with recording studios. Two products are currently available: a hardware processor from Audyssey called the Sound Equalizer (which requires a Windows PC to provide the user interface) and a software plug-in from IK Multimedia called Advanced Room Correction (ARC).
I've heard the result of MultEQ in consumer systems, and it's very impressive. The sound is clean and clear, and the difference between turning it off and on is like night and day. This system could be a real boon for recording musicians who want to hear their music exactly as they intend it, without the unwanted contributions from their rooms.