HeadRoom Supreme headphone amplifier Measurements part 2

So, how to examine the HeadRoom's response shaping and crossfeed? In its bypass mode, the amplifier was flat in response up to the mid-treble, with then a gentle ultrasonic rolloff. This can be seen in both the rounding of a 10kHz squarewave's leading edges (fig.5) and in the amplitude response (fig.6). The -3dB point lies at 27kHz, meaning that audio-band response is down 2.2dB at 20kHz. This will be audible as a slight softening of the overall sound—not a bad idea with headphones! The effect of the filter is also shown in fig.6; it can be seen to shelve up the mid-treble by a mild 3dB.

Fig.5 HeadRoom Supreme, 10kHz squarewave at 1W into 8 ohms.

Fig.6 HeadRoom Supreme, frequency response at 1V into 100k ohms (bottom trace at 1kHz) and with Filter engaged (top) (right channel dashed, 2dB/vertical div.).

Driving just one channel with Process engaged but looking at the output from both channels shows you the nature of the crossfeed signal. As can be seen from fig.7, the result is not too similar to the Bauer circuit responses (fig.1 in the main text). The driven-channel response is basically the same as in Bypass (fig.6), while the crossfeed signal is nearly 12dB down in level and starts to roll off above 1kHz. (This measurement doesn't reveal the time delay, of course.)

Fig.7 HeadRoom Supreme, frequency response of both channels at 1V into 100k ohms with left channel only driven (right channel dashed, 2dB/vertical div.).

What happens when the HeadRoom is handling signals in both channels? The Audio Precision System One allows you to drive a stereo component with both in-phase and out-of-phase dual-mono sweep signals. The results are shown in fig.8. When a signal is the same in both HeadRoom channels, it is shelved up 2dB below 1kHz, with rolled-off highs and a broad 2dB notch centered on 2kHz. This is for the amplifier customized for use with Sennheiser '580s; the bass shelf is 1dB higher for the 20k/6800pF version that works best with Beyerdynamics; and 1dB higher still for the 10k/8200pF version for use with Etymotics and as a line-level processor for Stax Lambda Pros. For a signal that is identical but out-of-phase in the two channels, this is equalized in the complementary manner.

Fig.8 HeadRoom Supreme, frequency response at 1V into 100k ohms with the two channels in-phase (top trace at 100Hz) and out-of-phase (bottom at 100Hz) (right channel dashed, 2dB/vertical div.).

Another way of looking at the HeadRoom process's effect is to drive the amplifier first with dual-mono pink noise, then with completely uncorrelated pink noise (ie, the two channels are completely different, not just out-of-phase). Track 15 on Stereophile's Test CD 2 was ideal for this test. When the pink noise was the same in both channels (fig.9), the same sum-channel trace was obtained as in fig.8. (Actually, this test was performed on an earlier version of the HeadRoom; the 2kHz valley is correspondingly about twice as deep.) But when there is no relationship between the signals in the two channels—ie, an instrument or voice is panned hard left or hard right—fig.9 confirms the fig.7 result that each channel is reproduced with a flat response. (The bumps and dips in this curve should be ignored; noise signals only measure truly flat when you can average the power at each frequency over a long period of time.)—John Atkinson

Fig.9 HeadRoom Supreme, 1/3-octave frequency response at 1V into 100k ohms using dual-mono pink noise (bottom trace at 1kHz) and completely uncorrelated pink noise (top at 1kHz) from Stereophile Test CD 2 (right channel dashed, 2dB/div.).

611 North Wallace Avenue
Bozeman, MT 59771
(800) 828-8184