MayaTlab
1000+ Head-Fier
So I've been using these for a few months already and so far it's a mixed bag for me. The TLDR is that :
- I don't like the default sound at all, for reasons that can be explained by on-head FR measurements.
- That being the case I find them strangely EQable, and right after creating an EQ profile, they are among my favourite HPs ever if not my favourite...
- but only for a short while as I have the impression that I'm having one of the worst "pads break in period" I've experienced so far, making this a constantly moving goalpost with, so far, no end in sight.
I've been practicing in situ, on head FR measurements for a few months and have found them useful to help in EQing headphones and to provide some plausible hypothesis in regards to the causes of my subjective impressions.
A first thing to note is that I've tried (and measured) several of them because of various QC issues. One sample for example arrived with a poorly assembled yoke, that would "grate" on the cup's plastic :
The one I decided to keep has a slightly asymmetrical headband arc shape, and one of the pads was poorly assembled :
I've received new pads from Austrian Audio via mail as a replacement.
So, my on-head measurements :
Using mic number 2 in that photo :
I'm getting somewhat decently low seatings to seatings variation during the same session (by "session" I mean that the mics weren't moved between the measurements), for example :
Five individual traces.
Not as good as some other headphones, but not too bad.
A pair of HPs with low on-head seatings to seatings variation (which is not the same thing as low variation when doing spatial averaging on a test rig) is a pair that provides repeatable measurements, and hence can be EQed quite precisely.
Comparing them on my head with mic number 1 above (top traces) and 2 (bottom traces) vs the HD560S, HD650 with Dekoni Elite Velour pads (not comparable to stock HD650), and H400SE, I'm getting this :
Averages of five traces during the same session (session means the mic wasn't moved between the HPs measurements), right channel only. X65 in turquoise, HD560S in blue, HD650 with Dekoni pads in orange, H400SE in dark fuchsia.
Please note :
- these measurements were done on my own head. Using the exact same methodology and instruments on your head may produce different results. These comparative results are not valid for you, they're just a way to illustrate my own experience.
- the probe mic, which is a small tube which orifice lands near my eardrum, is calibrated against a UMIK-1 in very near field conditions with a speaker. The results are uncompensated otherwise. It is a DIY attempt to get something similar to this : https://www.etymotic.com/product/er-7c/
- the absolute values are incorrect, so you can't say "5672Hz is 3.2dB higher than 3289Hz" - it's not that inaccurate for the probe, but shouldn't be trusted.
- the relative values between the four headphones have a degree of inexactitude that depends on seatings to seatings repeatability on my own head, sample variation, pads wear, pads warming up over the measurement session, the repeatability of the measurement method, the relevance of each measurement method for the application. You can't say "At 7629Hz headphones A are 7.45dB higher than headphones B", but rather "at around 7600Hz headphones A are around 7-8dB higher than headphones B", and it more specifically depends on each specific headphones for specific parts of the FR (I don't have the same degree of confidence in the 2-4kHz region for HPs A vs B as in the 4-6kHz region for HPs C vs D for example).
- The blocked ear canal entrance relative measurements aren't quite as trustworthy as the probe in the 2-4kHz region and most likely useless above 7kHz. Using the HD650 as a reference and plotting the difference with the other headphones I tend to get a pretty good agreement between mic 1 and 2 below 2kHz and between 4-7kHz (in the latter range mind the fact that it's the area where the X65 shows some seatings to seatings variation) :
The difference in the 2-4kHz region can be explained by the ear canal being blocked with mic 2. Above 7kHz they substantially disagree but some specific listening tests make me think that the probe is more trustworthy (I can't find a single transfer function for the blocked ear canal mics that consistently accurately locate the peaks in that range, while the probe as is consistently does so, and the probe is more consistent in terms of relative differences between headphones in that range from session to session, while the slightest modification of the insertion depth of the blocked canal mic changes the relative differences between HPs). Above 10kHz both methods are useless so far.
So the short version is that for me their on-head response differs quite a bit from Oratory or HeadphoneTestLab's measurements (particularly the elevated response at around 6kHz), but that similarly to them, while the peaks / dips tend to be quite significant in magnitude, they all have a symmetrical slopes and none of them are of the really high-Q and difficult to correct kind. Like HeadphoneTestLab's measurements I seem to not get too much in the way of sharp resonances. In fact I'm tempted to think that they may have the cleanest, least "intrinsically" resonant response I've measured so far.
I don't think that the discrepancy between my on-head and Oratory / Headphoneteslab's measurements (for example the significant 6kHz peak) is caused by sample variation. At least not to that degree. This is two different X65 samples during the same measurement session (the two bottom traces) :
Five individual traces for each, right channel.
Perhaps it's related to how they sit on my head vs. your typical test rig. I have a pretty extreme case of "wide temple + wide jaw + narrow neck" and the X65's design magnifies that issue :
The cups are quite significantly tilted upwards and rearwards, nearly hitting the limits of the yoke's range of motion in both cases. The large inner hole of the pads makes this worse as geometrically speaking it means that the lower rear quadrant sits further from my ear lobe in an even more depressed area.
This also means that the foam above the driver caresses my ears and may slightly deform my pinna.
Seal is, however, excellent.
Anyway, these are just idea thrown out there, I don't really know why.
So, theoretically only a few filters would make short work of these problems. And indeed it does to my utmost satisfaction - in fact post EQ they're among my favourite HPs so far - ... but only for a short while.
I was finding myself regularly in need to fine tune the EQ profile I made based on these measurements. Luck would have it I've been regularly measuring headphones with the exact same protocol around once a week in an effort to gather data en masse, and pulling all my X65's measurements done with the blocked ear canal mics highlighted a pretty significant issue : these may have one of the worst cases of "pads break-in" I've experienced yet.
This is the X65 compared to the HD650 with Dekoni Elite Velour pads (again, not comparable to stock) during the same period. The HD650's Dekoni pads were installed a couple weeks prior so not a direct 1:1 comparison of pads break in though :
What you're seeing here :
- each trace is an average of five individual seatings taken during the same measurement session.
- X65 at the top, HD650 with Dekoni pads at the bottom, both L and R channels. Please do note that these on-head measurements are inadequate to properly assess L/R channel matching.
- X65's L channel in fuchsia, R channel in turquoise.
- each measurement session was roughly a week apart. Both are my most used headphones at home so while I haven't logged the hours they both were pretty heavily used in that time.
The HD650 + Dekoni pads seems to have remained superbly stable in all that time (and it seems that the blocked ear canal measurements show decent repeatability for absolute values up to 7kHz or so).
The X65's variation in FR is directly correlated with measurement date. The 1600kHz dip has filled up over time (black arrow). The L channel seems to have departed from the earlier measurements further below 1kHz (red arrows). It's difficult to know why, but this is the pad that shows some deterioration of the fabric lining (cf. photo early in this post). While this seems correlated, I'm not certain that there is a causal relationship here.
So while the X65 is very EQable short term, that makes the X65 so far actually impossible to EQ long-term, as it's a constantly shifting goalpost. What worries me is that it has yet to reach stability.
I have yet to install the newer pads, but will do so I think quite soon, as well as installing new Dekoni pads on my HD650, and I'll repeat with more regularity these "pads break-in" measurements while trying to log in the hours. I may also get a completely new pair of Hi-X65 and compare all three.
That problem in the end is what may do it for me in terms of not keeping the X65 long term.
- I don't like the default sound at all, for reasons that can be explained by on-head FR measurements.
- That being the case I find them strangely EQable, and right after creating an EQ profile, they are among my favourite HPs ever if not my favourite...
- but only for a short while as I have the impression that I'm having one of the worst "pads break in period" I've experienced so far, making this a constantly moving goalpost with, so far, no end in sight.
I've been practicing in situ, on head FR measurements for a few months and have found them useful to help in EQing headphones and to provide some plausible hypothesis in regards to the causes of my subjective impressions.
A first thing to note is that I've tried (and measured) several of them because of various QC issues. One sample for example arrived with a poorly assembled yoke, that would "grate" on the cup's plastic :
The one I decided to keep has a slightly asymmetrical headband arc shape, and one of the pads was poorly assembled :
I've received new pads from Austrian Audio via mail as a replacement.
So, my on-head measurements :
Using mic number 2 in that photo :
I'm getting somewhat decently low seatings to seatings variation during the same session (by "session" I mean that the mics weren't moved between the measurements), for example :
Five individual traces.
Not as good as some other headphones, but not too bad.
A pair of HPs with low on-head seatings to seatings variation (which is not the same thing as low variation when doing spatial averaging on a test rig) is a pair that provides repeatable measurements, and hence can be EQed quite precisely.
Comparing them on my head with mic number 1 above (top traces) and 2 (bottom traces) vs the HD560S, HD650 with Dekoni Elite Velour pads (not comparable to stock HD650), and H400SE, I'm getting this :
Averages of five traces during the same session (session means the mic wasn't moved between the HPs measurements), right channel only. X65 in turquoise, HD560S in blue, HD650 with Dekoni pads in orange, H400SE in dark fuchsia.
Please note :
- these measurements were done on my own head. Using the exact same methodology and instruments on your head may produce different results. These comparative results are not valid for you, they're just a way to illustrate my own experience.
- the probe mic, which is a small tube which orifice lands near my eardrum, is calibrated against a UMIK-1 in very near field conditions with a speaker. The results are uncompensated otherwise. It is a DIY attempt to get something similar to this : https://www.etymotic.com/product/er-7c/
- the absolute values are incorrect, so you can't say "5672Hz is 3.2dB higher than 3289Hz" - it's not that inaccurate for the probe, but shouldn't be trusted.
- the relative values between the four headphones have a degree of inexactitude that depends on seatings to seatings repeatability on my own head, sample variation, pads wear, pads warming up over the measurement session, the repeatability of the measurement method, the relevance of each measurement method for the application. You can't say "At 7629Hz headphones A are 7.45dB higher than headphones B", but rather "at around 7600Hz headphones A are around 7-8dB higher than headphones B", and it more specifically depends on each specific headphones for specific parts of the FR (I don't have the same degree of confidence in the 2-4kHz region for HPs A vs B as in the 4-6kHz region for HPs C vs D for example).
- The blocked ear canal entrance relative measurements aren't quite as trustworthy as the probe in the 2-4kHz region and most likely useless above 7kHz. Using the HD650 as a reference and plotting the difference with the other headphones I tend to get a pretty good agreement between mic 1 and 2 below 2kHz and between 4-7kHz (in the latter range mind the fact that it's the area where the X65 shows some seatings to seatings variation) :
The difference in the 2-4kHz region can be explained by the ear canal being blocked with mic 2. Above 7kHz they substantially disagree but some specific listening tests make me think that the probe is more trustworthy (I can't find a single transfer function for the blocked ear canal mics that consistently accurately locate the peaks in that range, while the probe as is consistently does so, and the probe is more consistent in terms of relative differences between headphones in that range from session to session, while the slightest modification of the insertion depth of the blocked canal mic changes the relative differences between HPs). Above 10kHz both methods are useless so far.
So the short version is that for me their on-head response differs quite a bit from Oratory or HeadphoneTestLab's measurements (particularly the elevated response at around 6kHz), but that similarly to them, while the peaks / dips tend to be quite significant in magnitude, they all have a symmetrical slopes and none of them are of the really high-Q and difficult to correct kind. Like HeadphoneTestLab's measurements I seem to not get too much in the way of sharp resonances. In fact I'm tempted to think that they may have the cleanest, least "intrinsically" resonant response I've measured so far.
I don't think that the discrepancy between my on-head and Oratory / Headphoneteslab's measurements (for example the significant 6kHz peak) is caused by sample variation. At least not to that degree. This is two different X65 samples during the same measurement session (the two bottom traces) :
Five individual traces for each, right channel.
Perhaps it's related to how they sit on my head vs. your typical test rig. I have a pretty extreme case of "wide temple + wide jaw + narrow neck" and the X65's design magnifies that issue :
The cups are quite significantly tilted upwards and rearwards, nearly hitting the limits of the yoke's range of motion in both cases. The large inner hole of the pads makes this worse as geometrically speaking it means that the lower rear quadrant sits further from my ear lobe in an even more depressed area.
This also means that the foam above the driver caresses my ears and may slightly deform my pinna.
Seal is, however, excellent.
Anyway, these are just idea thrown out there, I don't really know why.
So, theoretically only a few filters would make short work of these problems. And indeed it does to my utmost satisfaction - in fact post EQ they're among my favourite HPs so far - ... but only for a short while.
I was finding myself regularly in need to fine tune the EQ profile I made based on these measurements. Luck would have it I've been regularly measuring headphones with the exact same protocol around once a week in an effort to gather data en masse, and pulling all my X65's measurements done with the blocked ear canal mics highlighted a pretty significant issue : these may have one of the worst cases of "pads break-in" I've experienced yet.
This is the X65 compared to the HD650 with Dekoni Elite Velour pads (again, not comparable to stock) during the same period. The HD650's Dekoni pads were installed a couple weeks prior so not a direct 1:1 comparison of pads break in though :
What you're seeing here :
- each trace is an average of five individual seatings taken during the same measurement session.
- X65 at the top, HD650 with Dekoni pads at the bottom, both L and R channels. Please do note that these on-head measurements are inadequate to properly assess L/R channel matching.
- X65's L channel in fuchsia, R channel in turquoise.
- each measurement session was roughly a week apart. Both are my most used headphones at home so while I haven't logged the hours they both were pretty heavily used in that time.
The HD650 + Dekoni pads seems to have remained superbly stable in all that time (and it seems that the blocked ear canal measurements show decent repeatability for absolute values up to 7kHz or so).
The X65's variation in FR is directly correlated with measurement date. The 1600kHz dip has filled up over time (black arrow). The L channel seems to have departed from the earlier measurements further below 1kHz (red arrows). It's difficult to know why, but this is the pad that shows some deterioration of the fabric lining (cf. photo early in this post). While this seems correlated, I'm not certain that there is a causal relationship here.
So while the X65 is very EQable short term, that makes the X65 so far actually impossible to EQ long-term, as it's a constantly shifting goalpost. What worries me is that it has yet to reach stability.
I have yet to install the newer pads, but will do so I think quite soon, as well as installing new Dekoni pads on my HD650, and I'll repeat with more regularity these "pads break-in" measurements while trying to log in the hours. I may also get a completely new pair of Hi-X65 and compare all three.
That problem in the end is what may do it for me in terms of not keeping the X65 long term.
. Yup. I need a few minutes to understand what just happened.....
