hd414 "classic" headphones

[Bob Carlyon <bob.carlyon@xxxxxxxxxxxxxxxxx>]

Just before christmas I circulated the results of some calibrations which
I'd performed on the "new" HD414 headphones recently re-issued by
Sennheiser, together with the calibrations performed on an "old" pair which
I had been using in the lab. I received some very interesting
correspondence which has prompted me to make some new measurements which I
present below, together with a description of some of the complicating
factors involved in making these measurements. For those of you with
insufficient time and/or inclination to read through the entire discussion,
the bottom line is that the distortion produced by the new 414s is not as
bad as I made out, but that their frequency response leaves a lot to be
desired.

Rob Hukin wrote to tell me that he'd read in "Studio Sound" (August 1995)
that the only similarity between the new and old 414s was in the mouldings,
and that the drivers on the new version were one of Sennheiser's latest
'off-the-shelf'
numbers. To quote the article, the new phones use  "selected versions of
Sennheiser's latest drive units employing lightweight Copper Clad Alloy
Wire driver coils, Neodymium magnets and OFC (oxygen-free copper) leads".
This could explain a lot.

Roel Smits pointed out that the level at which we calibrated the phones was
rather high, which I agree with. We used an input level of 0 dBV, which
corresponds to an output of about 106 dB SPL on the new phones, and about
98 dB SPL on the old ones. Further investigation revealed that the
distortion was highly level-dependent, which, given the different
sensitivity of the phones, complicates matters.  I should have made the
comparison at equal OUTPUT levels, and have now done so. The measurements
which follow were obtained at output levels of 90 and 100 dB SPL for one
"new" and one "old" earpiece. A further complicating matter concerns the
age of the cushions. As they wear out they produce less attenuation; I've
seen effects as large as 10 dB at 1 kHz, which makes it advisable to
re-calibrate your phones every year or so. The effect of cushion wear is
relevant to the measurement of distortion, whose level-dependence is,
presumably, determined by the level at the output of the driver (before
passing through the cushion). This means that if you use a worn-out
cushion, you need a lower "driver output" for a given SPL at the output of
the cushion, which will reduce the distortion compared to that measured at
the same SPL using a new cushion. The distortion measurements described
below used new cushions on both the "new" and "old" 414 tested. The numbers
in the table represent the attenuation of the 2nd and 3rd harmonics
relative to the primary. An asterisk indicates that the distortion was too
low to be measured on my HP3561A spectrum analyser.

                                "NEW"
FREQ     100 dB,H2       100 dB, H3      90 dB, H2       90 dB, H3
 125       43              51              53              61
 250       51              51              61              67
 500       59              62              *               68
1000       *               67              *               74
2000       *               *               *               *
4000       72              *               *               *
6000       *               *               *               *
8000       69              *               *               *

                                "OLD"
FREQ     100 dB,H2       100 dB, H3      90 dB, H2       90 dB, H3
 125       37              71             47               74
 250       58              73             68               *
 500       62              67             74               77
1000       65              75             76               *
2000       *               *               *               *
4000       73              *               *               *
6000       71              *               *               *
8000       72              77              *               *

So, the good news is that the distortion levels on the new 414s are quite
reasonable except at very high SPLs. Rather worse news is that the
frequency response of the new 414s shows some quite sharp peaks and dips,
which vary  across units  For example, the "new" 414 described above has a
response which rises gradually to a peak at 2775 Hz, drops by 4.7 dB as
frequency increases to  3175 Hz, rises by 7.8 dB to produce a sharp peak at
3700 Hz, dips again by 6.8 dB at 4375 Hz, remains fairly flat up to about
6800 Hz, before dropping steeply by 7.9 dB at 7300 Hz. The dip at 4300 Hz
is fairly typical across units.

If anyone has a use for 6 sets of headphones which look like the original
HD414s but aren't, I'm open to offers.

bob


------------------------------------------------------------------------------
Bob Carlyon
MRC Apllied Psychology Unit
15 Chaucer Rd.
CAMBRIDGE CB2 2EF
England

Phone: (44) 1223 355294 ext 720
FAX:   (44) 1223 359062

email: bob.carlyon@mrc-apu.cam.ac.uk
----------------------------------------------------------------------------
--