FIG. 1: The SVS system creates custom -surround simulations for each user and room, and it includes a head tracker that senses the wearer''s head position, so that the virtual speakers appear to remain -stationary as the head moves.
Among the potential applications is modeling any control room to which you have access. After spending a few minutes to capture your HRTFs in the selected room, using its speakers, you could then take the processor (or perhaps only a small SD card) to your bedroom or garage studio and hear your music as if you were in the original control room. How about monitoring in location trucks, where space is often way too tight for a surround speaker system? I'm excited by this technology, and I look forward to seeing it become commercially available.
Surround sound is becoming increasingly important for electronic musicians of all sorts. Movie and game soundtracks are routinely mixed in 5.1 surround, and studios often deliver multichannel music mixes for DVD-Audio and SACD. But installing a surround-sound speaker system is sometimes difficult, if not impossible, because of space and budget constraints. There's also the ire of family and neighbors when the volume gets too high. Finally, a surround system's sweet spot is very small, normally allowing only one person to occupy it.
To address those problems, several pioneers have tried to simulate surround-sound speaker systems using conventional headphones. Such simulations use mathematical algorithms called head-related transfer functions (HRTFs) to calculate the effect of sound waves diffracting around a human head from any given direction. The HRTFs are used to alter the amplitude, phase, and delay of an audio signal before it's sent to the headphones, and the listener perceives the sound coming from the desired “direction.”
Unfortunately, most of these simulation systems suffer from a couple of drawbacks. For one thing, the HRTFs are generic, based on the average of many individual measurements, which means they are accurate for no one in particular. Furthermore, as a listener turns his or her head, the simulated speaker system turns with it, which sounds unnatural. If you were listening to a conventional surround speaker system, turning your head would change the relative orientation of your ears with respect to the speakers, resulting in plainly audible changes to the sound.
A company by the name of Smyth Research (email@example.com) has come up with an ingenious solution to such concerns. Smyth Virtual Surround (SVS) is the brainchild of Stephen Smyth, who has been working on it for three years. Unlike other headphone-based surround-simulation systems, SVS starts by “capturing” the HRTF of an individual listening to a specific set of speakers in a specific room. The listener puts a small microphone in each ear, and a series of test signals are played from each speaker, which takes about two minutes for a 6.1-channel system. The resulting HRTFs are stored in a small file (roughly 1 MB) and used to simulate the effect of sounds coming from the modeled speakers.
As many as eight channels of audio are fed into SVS, which applies the selected HRTFs and sends them to headphones fitted with a small tracking device that includes motion sensors to gauge the wearer's head position at all times (see Fig. 1). That information is used by SVS to modify the HRTFs in real time so that the virtual speakers appear to remain in place as the listener's head moves.
For the pro-audio market, Smyth Research intends to introduce a standalone, dual-headphone processor with eight channels of balanced, analog I/O designed to be inserted into a mixer's bus. SD memory cards are used to load and transfer HRTFs from one unit to another. But the company's main goal is to license the technology to consumer-electronics manufacturers; in fact, they've successfully demonstrated SVS at various CE trade shows.