kyma•tweaky: Share StereoCrossFeedDiscussion

This is an improvement for headphone channel crossfeed. Included are two versions of this crossfeed for stereo headphone listening. The first is based on MIT measured HRTF (head related transfer functions) feeding each side with a mix of sound from azimuth 30 deg and -30 deg from the frontal direction. The HRTF's are incorporated into a pair of FIR filter Sounds feeding each side. The problem with this approach is that it kills the bass response.

However, after comparing the dB amplitude plots of each of these filters, it was found that to a very good approximation, the underlying difference between the direct and cross fed sound is a 1/F slope (ignoring the various bumps and notches). This makes physical sense. It means the energy of the sound crossing to the opposite ear varies as 1/F, so the amplitudes vary as 1/Sqrt(F). Hence we need a -3 dB/octave rolloff.

So using my "DirectIIR" microsound, and some Mathematica analysis, it was found that a truncated pinking filter, accurate to 0.5 dB, provides an appropriate -3 dB/octave slope beyond 300 Hz. At the low end, the filter is flat to 200 Hz. The DC response is at -2 dB. After delaying the cross channel by 0.25 ms, corresponding to the delay appropriate for 30 deg speaker placement, the delayed sound is sent through this truncated 1/F filter and mixed with the opposite direct channel on each side.

The result is a very nice, full bass-response, stereo cross feed that makes listening to stereo tracks with wide separation very easy. No differential ear pounding. The 1/F crossfeed has a toggle to allow you to A/B the sound coming through. Just pump your favorite wide separation tracks into channels 1&2 and have a listen.

-- DavidMcClain - 03 Dec 2003

k y m a • t w e a k y — the kyma collective || view the current website || February 2015 Archive