has anyone else experienced this or could it be just my pair?
I also experience that problem, even for pairs which pads weren’t crumpled by the poorly designed packaging (now solved on more recently shipped samples).
I think that this phenomenon can be illustrated with in-ear mics. Here’s how three headphones compare on my own head (HD650
with Dekoni Elite Velour pads - not comparable to stock in absolute terms, Airpods Max - as another example of a really poor design concept that’s partly saved by decent execution and its feedback mechanism, K371 - with non-crumpled pads), measured with blocked ear canal entrance mics, in terms of :
- Seatings to seatings variation (while keeping my head straight up and looking forwards, I completely remove and re-position the headphones, as I would normally wear them between each measurements - of note here is that this is a conceptually different exercise from the sort of spatial averaging that some reviewers perform when measuring headphones on ear simulators, as I make no attempt to re-seat them in unrealistic - to me - positions, so don’t compare them).
- Positional variation while leaning my head (I do not remove the headphones between measurements, but I lean my head front / back / left / right).
Please note that :
- The absolute values are incorrect and that the same methodology may yield different results with your own samples (or even the same) on your own head.
- Also, the relative values between the headphones may be increasingly incorrect as the frequency rises (particularly around 2-3kHz and above 7kHz, because of the mics used), please only compare the headphones with themselves.
- I’ll also show both R and L channels for the K371 as they behave differently, but this on-head methodology is inappropriate to compare R and L channels matching above 1kHz, and below that is only appropriate to assess the delivered channel matching on a real human with an asymmetrical head, not how they’d behave on a symmetrical testing apparatus.
So, first, here’s the seatings to seatings variation.
10 individual traces each, 1/24 smoothing. Traces were
not normalised.
I found out that the K371 was more consistent if I let the pads warm up a few minutes before measuring them, that’s what I did here.
For the Airpods Max pink noise was played in between the sweeps.
For the HD650 + Dekoni pads, right channel :
Airpods Max, right channel :
K371, right channel :
K371, left channel :
You’ll notice that the left channel has a lower bass output than the right, that’s at least partly caused by seal I believe. I’ve added a single red trace to show how they behave under a little bit of pad compression.
Second, the positional variation while leaning my head front / back / left and right.
The blue trace is how they naturally sit on my head while looking forward, the red traces 2 or 3 measurements for each variation in head position.
For the HD650 + Dekoni pads, it looks like this :
A better way to represent it, in my opinion, is to flatten to zero the default position and only show the difference between it and the variations in head position.
Here’s what it looks like for the HD650 + Dekoni pads, right channel :
Of note here is that some of the variation in one specific orientation (when I lean my head towards the right) comes from my right shoulder interacting with the open rear, not the headphones themselves.
Airpods Max, right channel :
Of note here is that the spike you see at 8kHz is not as scary as it looks, it fills a fairly deep, high-Q null.
K371, right channel :
K371, left channel :
So, to sum it up :
- The HD650 with Dekoni Elite Velour pads (but that also applies to the default pads as well) is a very unproblematic design for me, both in terms of seatings to seatings variation and in terms of how they vary when moving my head. Both channels behave similarly.
- The Airpods Max’s seatings to seatings variation is low. The variation with head position is higher than the HD650, with one position in particular being problematic (when leaning my head backwards, they seem to shift more so than in the other positions, possibly because of their weight and my head shape). In the other positions it isn’t shifting that much. The feedback mechanism makes variation a total non-problem below 800Hz, but it's a double edged sword as it means that the response floating above that frequency "tilts" the response more so relative to the feedback range. Both channels behave similarly.
- The K371’s seatings to seatings variation is higher on the left channel. The variation with head position, however, is much more problematic, particularly since both channels behave in quite a dramatically different fashion. On the right channel a particular head position (leaning left) breaks seal quite badly. On the left channel SPL varies across a fairly large range in a way that isn’t consistent with the same position on the right channel. The resulting sensation when listening to music is quite weird as the sound audibly changes when tilting my head forward to look at my computer on my desk, for example.
Addendum :
I've made a little GIF illustrating the variation, not normalised, for both the R (blue) and L (green) channels with the K371, for each head position, below 1kHz :