[csglinux]

Maybe FR should not be the biggest factor in evaluating a headphone...

But it is though I agree with everything else you've said. I even agree with your above statement - maybe it shouldn't be - but who has the time, the energy, or the sufficiently fine-grained parametric EQ to perfectly adjust each HP before evaluating? Perceived FR will be different for everybody, because we all have different ears, and will prefer different eartips, etc. But it's pretty important we find a HP with an FR that works for us. At least it is for me, because I want to be able to use a variety of sources, many of which don't allow any form of EQ.

The KSE1500 looks like an exception to this, because it has a really nice parametric EQ built in to its amp - but anytime you use this, you engage its internal DAC, which is something you might not want to do if, for example, you were running from the LO of a Chord DAC.

 

[headwhacker]

But it is though I agree with everything else you've said. I even agree with your above statement - maybe it shouldn't be - but who has the time, the energy, or the sufficiently fine-grained parametric EQ to perfectly adjust each HP before evaluating? Perceived FR will be different for everybody, because we all have different ears, and will prefer different eartips, etc. But it's pretty important we find a HP with an FR that works for us. At least it is for me, because I want to be able to use a variety of sources, many of which don't allow any form of EQ.

The KSE1500 looks like an exception to this, because it has a really nice parametric EQ built in to its amp - but anytime you use this, you engage its internal DAC, which is something you might not want to do if, for example, you were running from the LO of a Chord DAC.

I use LPG which has superb PEQ. Works very well with KSE1500 and the iSines. I actually find KSE1500 warm and a bit boomy for my taste. I played around with EQ and tried to make the KSE as neutral/flat as I can perceive it.

 

[castleofargh]

Which is the point where I was trying to go. Maybe FR should not be the biggest factor in evaluating a headphone as this can change from one guy to another and with different setup/environment (e.g. tip, temp, etc). Clarity, distortion level especially when driven at high volume, transient response, etc., I believe should take more weight as this can't be influence by FR. Like I mention a properly implemented EQ can almost fix the FR to one's desire/preference. In my experience, KSE1500 and the Audeze iSines should easily be able to match any headphone with a decent EQ.

I'm an EQ addict (I started measurements only to help myself with EQ routines and to verify the consequences of my applied EQ). but aside from objective limitations, many DAPs have crappy EQ capabilities, and even more people are just not using EQ because they have medieval understanding about digital audio, or simply because they don't know how to set up a nice EQ(IMO it takes a great deal of practice to do a good job, but of course if we never start practicing we go nowhere).

now a FR graphs is unreliable in many ways and people indeed have a different neutral for IEMs. this very topic is full of warnings of that nature, you'll find them every few pages. poor Crina can't do much more than that. well he could come to our home personally and headbutt us anytime we assume too much when reading his graphs, but then he wouldn't have any time left to measure new stuff. ^_^

 

[headwhacker]

I'm an EQ addict (I started measurements only to help myself with EQ routines and to verify the consequences of my applied EQ). but aside from objective limitations, many DAPs have crappy EQ capabilities, and even more people are just not using EQ because they have medieval understanding about digital audio, or simply because they don't know how to set up a nice EQ(IMO it takes a great deal of practice to do a good job, but of course if we never start practicing we go nowhere).

now a FR graphs is unreliable in many ways and people indeed have a different neutral for IEMs. this very topic is full of warnings of that nature, you'll find them every few pages. poor Crina can't do much more than that. well he could come to our home personally and headbutt us anytime we assume too much when reading his graphs, but then he wouldn't have any time left to measure new stuff. ^_^

To be clear i'm not trying to downplay @crinacle 's effort. I actually am thankful as I use his data as guide for this hobby. I'm fully aware that FR is not 100% accurate. But it is better than nothing. Like you said EQ is a valuable tool; but not all EQ are the same and sadly most DAPs don't have a usuable EQ and I think DAP-makers don't take EQ seriously.

 

[castleofargh]

EQ has been all the rage back in the days(analog but still), then it wasn't. and it's coming back as far as I can see, but maybe only because DAC chips now often come with an integrated EQ, so it would be silly for the DAP manufacturers not to offer all the bells and whistles when they have them anyway.

 

[bartzky]

Maybe FR should not be the biggest factor in evaluating a headphone as this can change from one guy to another and with different setup/environment (e.g. tip, temp, etc). Clarity, distortion level especially when driven at high volume, transient response, etc., I believe should take more weight as this can't be influence by FR.

Interesting point. Not to long ago I would probably have agreed with you.
Dozens of measurements, DSP experiments and several papers later, I'm sure that the importance of the FR is totally underestimated in our hobby.

Many things we like to attribute to technical abilities are greatly effected by FR. Take two IEMs that you think are on a different level of sonic performance, measure them and use DSP to make their FRs be the same, level match them and A/B them. I'm sure will be astonished - at least I was and still am when I do this.

Researchers nowadays use this method to investigate user preferences of headphones and IEMs. Previously their investigations were strongly affected by the bias of their subject group: you will know if somebody just put a LCD-2 or a Beats on your head judging from their weight and form factor (same is true for many IEMs). Hence there was the need for a better test methodology. Astonishingly they found correlations between user ratings of the real headphone/IEM and virtual headphone/IEM of 85-98%. That suggests that how much somebody likes any pair of headphones/IEMs highly depends on their FR.

To be an even bigger smartass: Transient response is strongly linked to FR - it's frequency and phase response in the time-domain. Given their transient response, e.g. their impulse response, you will know their FR and PR and vice versa.

but who has the time, the energy, or the sufficiently fine-grained parametric EQ to perfectly adjust each HP before evaluating?

It's not that much work when using REW and EQ Apo. I actually do this for many reviews.

 

[Cya|\|]

Interesting point. Not to long ago I would probably have agreed with you.
Dozens of measurements, DSP experiments and several papers later, I'm sure that the importance of the FR is totally underestimated in our hobby.

Many things we like to attribute to technical abilities are greatly effected by FR. Take two IEMs that you think are on a different level of sonic performance, measure them and use DSP to make their FRs be the same, level match them and A/B them. I'm sure will be astonished - at least I was and still am when I do this.

Researchers nowadays use this method to investigate user preferences of headphones and IEMs. Previously their investigations were strongly affected by the bias of their subject group: you will know if somebody just put a LCD-2 or a Beats on your head judging from their weight and form factor (same is true for many IEMs). Hence there was the need for a better test methodology. Astonishingly they found correlations between user ratings of the real headphone/IEM and virtual headphone/IEM of 85-98%. That suggests that how much somebody likes any pair of headphones/IEMs highly depends on their FR.

To be an even bigger smartass: Transient response is strongly linked to FR - it's frequency and phase response in the time-domain. Given their transient response, e.g. their impulse response, you will know their FR and PR and vice versa.


It's not that much work when using REW and EQ Apo. I actually do this for many reviews.

I agree with you. Frequency response is so underestimated. Other differences may matter more with speakers, where the frequency responses are most often very similar, and other factors may affect the sound. Concerning headphones and iems, where frequency responses vary greatly, I believe that the latter has more weight on how we perceive the sound. Of course I'm not saying it's the only factor, but probably the most important for headphones.

 

[Brooko]

Thank you Bartzky - its something I've been suspecting for a long time.

Point in case - I just completed a review for the Lyra II. Even some really experienced reviewers talked about the upgraded sound. When i got them here (along with a pair of the original Lyra), I listened with some quick switching, and was almost sure there were very slight differences. My next step was to measure volume to make sure they were matched, and surely enough, they were very slightly out. Volume matched, listened and no discernible differences (to my poor hearing anyway). Next step - left ear-piece from one (Lyra I), right ear-piece from the other (Lyra II), and volume match using the X7ii's channel adjustment. Now listen - perfect stereo image.

Finally - measure on my rig - practically a perfect frequency response match (once volume matched).

Conclusion - expectation bias is an absolute killer (new model- must be better), volume matching is essential, and I trust my measurement rig to be more accurate than the filter my brain puts over everything

 

[castleofargh]

Interesting point. Not to long ago I would probably have agreed with you.
Dozens of measurements, DSP experiments and several papers later, I'm sure that the importance of the FR is totally underestimated in our hobby.

Many things we like to attribute to technical abilities are greatly effected by FR. Take two IEMs that you think are on a different level of sonic performance, measure them and use DSP to make their FRs be the same, level match them and A/B them. I'm sure will be astonished - at least I was and still am when I do this.

Researchers nowadays use this method to investigate user preferences of headphones and IEMs. Previously their investigations were strongly affected by the bias of their subject group: you will know if somebody just put a LCD-2 or a Beats on your head judging from their weight and form factor (same is true for many IEMs). Hence there was the need for a better test methodology. Astonishingly they found correlations between user ratings of the real headphone/IEM and virtual headphone/IEM of 85-98%. That suggests that how much somebody likes any pair of headphones/IEMs highly depends on their FR.

To be an even bigger smartass: Transient response is strongly linked to FR - it's frequency and phase response in the time-domain. Given their transient response, e.g. their impulse response, you will know their FR and PR and vice versa.


It's not that much work when using REW and EQ Apo. I actually do this for many reviews.

I'm always impressed by how the EQ simulation in REW comes close to what I measure if I then apply that EQ in my measurement loop. for a free software that thing really rox.
and of course I agree that FR while not the only element, is one that impacts all our impressions. myto speed of the IEM, clarity, control, soundstage... and it's pretty obvious when we consider how we're meaty energy sensors. a FR change affects even the loudest signals at that frequency so of course the subjective impact will be strong too.
doesn't mean we have to spit on the rest and make the mistake to interpret everything out of FR graph, but I'm totally with you on how it's often a leading aspect of our impressions.

 

[crinacle]

Yeah, FR isn't the ONLY thing that matters for audio quality (I have to repeat this so many times, god people just LOVE to assume my mentality right off the bat), however it's the first thing that we notice as well as the most detectable characteristic.

Good FR does not always equal to good sound. I've seen good graphs and have gotten myself excited, but then am underwhelmed when I have a real listen. No one is saying that FR determines fidelity.

On the flipside, Crap FR = crap sound, 99% of the time. Oh, 10dB spikes everywhere? A hole in the middle of nowhere? A cliffdive after 5k? Huge, abrupt changes? Well, either the transducer is objectively mediocre or the fanboys are delusional. Not going to name names but twice now I've gotten my hands on hyped FOTMs and they were so bad that it left my head scratching on what the previous reviewers were listening to when they were gushing over them.

Okay this turned more rant-y than I'd like. Back to my dark corner of the room shoving things into tubes

 

[1TrickPony]

slightly of topic but figured it'd be a good opportunity to ask, would a listener/ expect a different fr curve (even slightest) if one were to measure iem "x" using say fiio x3ii vs fiio x5iii? Sure lots of factors are taken in account. I'm just curious as to the extent of any discrepancy if it'd first. Figure if throw it out there.

 

[Sinusoid]

Hi guys, I'm in the middle of something that I find disturbing... now it's two manufacturers that I have ordered a custom set of IEM's that both sound different from the Universal demo versions of theirs and well, not in a good way. Since I have some of the decent equipment available to me I have fooled around with measurements only to get to the point that there is probably no way to actually meaningfully measure the Universals as it seems there is no dedicated "sound" to it as the differences made by the tip of the Universals and the difference in drivers tubing length in the Customs mess everything up in the end?

This is the post I found here that gave me the inspiration to measure difference of the tips only to find in horror that the difference is huge and Custom sleeve or custom design in general messes up the sound.

https://www.head-fi.org/threads/shu...tatic-earphones.785377/page-207#post-13714398

FYI I use the SpinFit tips with the Universal demos. So, what do you measuring fanatics think about it? Do you agree with me and what do you think should I tell to the manufacturers for them to fix the sound so that it would actually resemble what people hear when they listen to the Universals and want "that sound"?

 

[Brooko]

slightly of topic but figured it'd be a good opportunity to ask, would a listener/ expect a different fr curve (even slightest) if one were to measure iem "x" using say fiio x3ii vs fiio x5iii? Sure lots of factors are taken in account. I'm just curious as to the extent of any discrepancy if it'd first. Figure if throw it out there.

If you were talking about the DAC section of the sources, and you didn't calibrate your system to flat before measuring, then yes you'd likely get roll-off from the DAC filters which would affect the IEMs you're measuring. Thats why most people measuring use a standard system which they've calibrated properly and specifically for the source they are using. The other parts to the equation are always using same settings on the measurement gear, and I always use the same tips and insertion depth.

 

[piotrus-g]

FYI I use the SpinFit tips with the Universal demos. So, what do you measuring fanatics think about it? Do you agree with me and what do you think should I tell to the manufacturers for them to fix the sound so that it would actually resemble what people hear when they listen to the Universals and want "that sound"?

Speaking from the "other side". I would say it'd be wise to ask manufacturer which universal sleeves will give you closest custom IEM sound. For us that's dual-flange T400 tips - which has been also backed by users' experience.
Each ear tip has different effect on sound due to size length or amount of isolation it provides. Custom has one correct setting for how it sits in the ear and how it produces sound. So if you are tip rolling on demo units you cannot really expect the custom to sound identical unless you order "custom universal" and use same tips you used on demo.

 

[csglinux]

I agree with @bartzky - FR might not be everything, but it's about 99% In particular, spatial imaging, soundstage width/depth, separation are all about timing, i.e., resolving micro-detail, which is all about resolving fast transients, which means high frequency extension. Assuming the drivers' masses are low enough to get there, EQ should make two IEMs sound virtually indistinguishable.
There's an important point which is almost always overlooked, and I need to give credit for this observation to my good friend @spook76 (who sadly just passed away: https://www.head-fi.org/threads/shu...tatic-earphones.785377/page-213#post-13800439) - most IEMs these days are using the exact same drivers (there are only two main BA manufacturers: Sonion and Knowles), so, assuming any multi-drivers have no phase issues, what would one expect to be different between two IEMs using identical drivers, other than FR?!

It's not that much work when using REW and EQ Apo. I actually do this for many reviews.

Agreed. I love REW. It's just that it's not practical to carry my PC around with me. My phone is awesome (you folks should check out the LG V30!), but has very limited EQ capabilities which only work on certain file formats. My DAPs (FiiO and Questyle) are also awesome, but also don't have very fine-grained EQ controls. This is why I have to have my preferred FR from my headphones without the need of EQ.