[mark_h]

It's answering the question you asked about frequency response and eardrum pressure and the effects of 'physiological noise at low frequencies' in IEMs

 

[Spyro]

Zalithian...terrific post and great analogies.

 

[tinyman392]

Quote:

"Charts are for nerds"
 
No, they just don't lie and aren't affected by the subjectiveness of the listener. Surely that's kind of like saying "get off the whole science thing, the earth is flat". Empirical data provided by charts is kind of what I'm after as I'm slightly sick of everyone just talking about how treble lacks sparkle, mids aren't forward enough, bass lacks punch, I mean, how many words can you come up with to describe such things? Surely, a flat frequency response, is what's desired. A deviation from this is colouring the sound against how the producer wanted it to be. Please, tell me what I'm missing here.. what is the "tactile effect of bass"? Airflow/vibration?

 
Charts don't lie, people do.  The charts are, as you say, empirical data that is provided (assuming their accuracy).  However, that's where it ends.  The interpretation of such a chart, to date, is 100% subjective.  Why?  Because the terms used describe the headphones through the chart is based on subjective definitions.  Sparkle, sweetness, impact, texture all hold subjective definitions, not objective.  They are based on ear perceptions, not chart.  There is a correlation between many things in audio, but none of them work 100%; they stand as correlations.  Every single subjective definition needs to be reworked to fit both the subjective ear as well as the objective.  Saying that boomy bass is a bump around the lower bass is just half the story.  We have no relativity to any other frequency range.  So when a person states that they have objective evidence that headphone X is punchy, it means it fits the correlation; it's not objective, it's subjective.  Kind of like how you can skew numbers to match the information you want to portray.  "Objective" measurements are as bad as subjective.  
 
I do also want to add in that the perceptual graphs, the compensated ones, may or may not work.  They run on one assumption, that the human perceptive hearing of sound pressure changes (dB) is linear.  This is a huge assumption which takes a lot of these compensated graphs far back.  This is one of the things that can make graphs look 100% off when it comes to actual perception vs compensated to perception.  So, when looking at graphs, it's important to take raw data into consideration, not the compensated (often time colored).  
 
So it's true, the graph doesn't lie, but people can.  An objective graph can also impact a persons interpretation of a headphone.  
 
To answer your question in the OP though, there are a lot of factors that determine the bass of an IEM.  Seal is the largest one that is not only determined by insertion depth, but tip selection as well.  Much research has come along with the missing 6 dB as well that states that the human ear naturally creates some sort of amplification in the sub-bass areas by about 6 dB.  When the IEM is inserted, it bypasses these areas.  This could very well be the reason why people har it anemic.  
 
To answer your second question in the OP, this is quite personal.  What we really deal with are 3 bars to match by (maybe more).  The first one, accuracy.  This deals with how tonally accurate (tonally neutral) an IEM will be.  In the case of W4 vs Etymotic here, the Etymotic will win hands down.  This neutrality is really second to none in the IEM business.  
 
The second bar would be musicality, the ability to reproduce musical aspects.  The idea of this runs further from neutrality as a non-neutral headphone can indeed be "musically" accurate (IE producing all the aspects of music, a song).  Basically what happens is that the IEMs will be tested against the subjective descriptors listed in the glossary of terms, if it can reproduce all of them without overdoing it (turning it into a negative aspect), then it becomes "musically" accurate.  Many neutral IEMs are not "accurate" in this sense, but can be.  In the case of the Etymotic (HF2, I don't have an ER4) vs W4, the W4 takes the win slightly in terms of musical reproduction.  
 
The last bar is personal preference.  Many here have talked about this one, personal preference will play a huge role in how one perceives and IEM, and how one can become better than the other.  Normally, a signature type plays the huge role.   
 
There may be other ways to compare headphones by, the first two would be the main ones I see on Head-Fi, I personally use the second one.  The third one ends up being used a lot as well with hype and FOTM threads to give an IEM an edge against a moving target.  I really haven't seen too much more, but they may exist.  At the end of the day, all of it does fall under number 3 though.  The idea that an IEM is better for reproducing neutrality is personal, reproducing musical elements is also personal as well.  Personal preference is personal.  It's important to know now just why a person ranks an IEM against another, but how he does it (how does he determine "better").  

 

[bobeau]

Quote:

 I don't believe producers produce music expecting everyone to listen to it on a perfectly neutral and balanced system. Actually, if I were to bet, I'd bet they expect most people to be listening to it on some of the crappiest sound systems possible. 

 
Wow, excellent post.  I swear I've seen guys in the sound science forum claiming to be audio engineers say this isn't the case - yet I've read articles from (I guess pop or rock) producers saying indeed when doing final mix-downs, they do testing on a variety of systems and generally target the final mix to the type of equipment their main demographic will listen to.  That final mix on their near field monitors can sound absolutely horrid.

 

[planx]

Simple, try the ER-4 yourself and find out. I certainly think that the ER-4 is the better sounding IEM to me while others will disagree and say that the W4 is the better sounding IEM to them. Same story with the PFE232, I find the PFE232 to sound better whereas others will find that the W4 sounds better to them... In simple, opinions from others do not matter, trust your own ears and simply enjoy the music. Don't take Head-Fi and Head-Fiers TOO seriously 

 
Quote:

http://saha.searle.northwestern.edu/bibcat/pdf/1012.pdf

 
Excellent read.

 

[Inks]

It IS about tastes.  No matter how well it is tuned many think (me included) a single BA has severe limitations in realistically and adequately covering the entire frequency response curve while creating great realism and dynamics.  It's much more than marketing. 
 
I've never heard any "live" music sound anything like what I hear with an Etymotic earphone but I might just have bad ears.

  The highest bandwidth single BAs have just as much treble extension as many multi-drivers and bass extension is only missing the lowest depths, which even some multi-BAs don't cover that well either. I don't see anything "severe". Take for example a UM3X, 3 drivers yet the ER4 series covers just as much and more adequately. Etymotic are very realistic, the only thing missing is bass rumble and 3k of the extremes trebles. The W4 doesn't sound as realistic because it has too much midbass that thickens notes too much and veils the lower midrange a bit, though it does have better high treble than the ER4 series due to better extension [w/impedance and wide bore tips]
 
 

That's not very helpful. I'm asking for a technical explanation as to why other earphones are considered better, because surely you could just EQ one with a totally flat response to your taste? What other technical factors make other IEMs superior? I'm sure there are some, I just don't know them.

Audible distortion, bandwidth,  transient speed and polarity are other aspects to look into. Transient speed and bandwidth are specially important. You could EQ a MH1C and W4/UE900 flat, yet the MH1C will sound slower, muddier in it's bass  because of it's uneven speed. 
 
Originally Posted by Zalithian /img/forum/go_quote.gif

"Surely, a flat frequency response, is what's desired." - This is a false claim/assumption. Desired from who? Consumers? Surely natural food is desired from consumers, with no taste beyond what is naturally occurring, as mother nature intended. Why would anyone ever want something to have more taste than it has naturally? Why do some people prefer bananas to strawberries? Science should be able to tell us. Sir, can you please explain to me scientifically why people prefer bananas to stawberries? Why do some people like MSG on their food? Come on, man.
 
Charts, like statistics, do not tell the whole story. Have you ever watched sports? Have you ever seen an athlete with a good stats and an athlete with bad stats? Some things cannot be measured. One may have good stats but suck in crunch time. One may have bad stats but come through in the clutch. Where do you see leadership or the affect on teammates in statistics? Does it measure how he might coordinate his team or lead them to make changes during a game? How do you measure the width of headstage in frequency response? How do you measure how wide it sounds with frequency response? Or anything? 
 
"A deviation from this is clouring the sound against how the producer wanted it to be." Again, false. Do you have something to back this up? I don't believe producers produce music expecting everyone to listen to it on a perfectly neutral and balanced system. Actually, if I were to bet, I'd bet they expect most people to be listening to it on some of the crappiest sound systems possible. Walk around and tell me how many people you see listening to music on ibuds or some crappy $10 JVC gummies.
 
People in general, like excitement. It just might explain why movies are a more popular form of entertainment than documentaries. What is more realistic is not necessarily what people prefer.

Things aren't perfect, but a flat response [diffuse field] is the more pleasant to listeners, there have been plenty of studies for this. Does that mean you want to reach a specific curve? no, because of many factors, sometimes an IEM with less or more treble is preferable [things such as mastering of the music], the thing is, it has to at least be a linear curve to avoid unpleasant resonants. The "graphs don't tell everything" argument is getting old, of course they don't, but understand that they portray specific technical strengths or weaknesses that will be potentially perceived. If there's a severe dip or peak, lack of extension, etc. it's an objective data of the IEM's performance. 

 

[Inks]

Quote:

"Surely, a flat frequency response, is what's desired." - This is a false claim/assumption. Desired from who? Consumers? Surely natural food is desired from consumers, with no taste beyond what is naturally occurring, as mother nature intended. Why would anyone ever want something to have more taste than it has naturally? Why do some people prefer bananas to strawberries? Science should be able to tell us. Sir, can you please explain to me scientifically why people prefer bananas to stawberries? Why do some people like MSG on their food? Come on, man.
 
Charts, like statistics, do not tell the whole story. Have you ever watched sports? Have you ever seen an athlete with a good stats and an athlete with bad stats? Some things cannot be measured. One may have good stats but suck in crunch time. One may have bad stats but come through in the clutch. Where do you see leadership or the affect on teammates in statistics? Does it measure how he might coordinate his team or lead them to make changes during a game? How do you measure the width of headstage in frequency response? How do you measure how wide it sounds with frequency response? Or anything? 
 
"A deviation from this is clouring the sound against how the producer wanted it to be." Again, false. Do you have something to back this up? I don't believe producers produce music expecting everyone to listen to it on a perfectly neutral and balanced system. Actually, if I were to bet, I'd bet they expect most people to be listening to it on some of the crappiest sound systems possible. Walk around and tell me how many people you see listening to music on ibuds or some crappy $10 JVC gummies.
 
People in general, like excitement. It just might explain why movies are a more popular form of entertainment than documentaries. What is more realistic is not necessarily what people prefer.

Things aren't perfect, but a flat response [diffuse field] is the more pleasant to listeners, there have been plenty of studies for this. Does that mean you want to reach a specific curve? no, because of many factors, sometimes an IEM with less or more treble is preferable [things such as mastering of the music], the thing is, it has to at least be a linear curve to avoid unpleasant resonants. The "graphs don't tell everything" argument is getting old, of course they don't, but understand that they portray specific technical strengths or weaknesses that will be potentially perceived. If there's a severe dip or peak, lack of extension, etc. it's an objective data of the IEM's performance. 

 

[Inks]

Quote:

Charts don't lie, people do.  The charts are, as you say, empirical data that is provided (assuming their accuracy).  However, that's where it ends.  The interpretation of such a chart, to date, is 100% subjective.  Why?  Because the terms used describe the headphones through the chart is based on subjective definitions.  Sparkle, sweetness, impact, texture all hold subjective definitions, not objective.  They are based on ear perceptions, not chart.  There is a correlation between many things in audio, but none of them work 100%; they stand as correlations.  Every single subjective definition needs to be reworked to fit both the subjective ear as well as the objective.  Saying that boomy bass is a bump around the lower bass is just half the story.  We have no relativity to any other frequency range.  So when a person states that they have objective evidence that headphone X is punchy, it means it fits the correlation; it's not objective, it's subjective.  Kind of like how you can skew numbers to match the information you want to portray.  "Objective" measurements are as bad as subjective.  
 
That first part makes no sense to me. Take for example, the 10k region, these are "sparkle" regions because they are fundamental regions which will cymbal crashes lie in. There is no subjective in what the 10k region does (boost it and cymbal crashes do indeed sparke more), if there's somehow a disagreement on what "sparkle" is, it's the terms themselves that need to be fixed, the effect of the 10k region remains objective. Sparkle is fine but terms like "sweetness" are useless..imo and just cause confusion. If used/read correctly, FR graphs are an amazing tool, just note of transients/decay extension and you've got a pretty concrete idea of what to expect. 
 
 
I do also want to add in that the perceptual graphs, the compensated ones, may or may not work.  They run on one assumption, that the human perceptive hearing of sound pressure changes (dB) is linear.  This is a huge assumption which takes a lot of these compensated graphs far back.  This is one of the things that can make graphs look 100% off when it comes to actual perception vs compensated to perception.  So, when looking at graphs, it's important to take raw data into consideration, not the compensated (often time colored).
 
Compensations are based on resonants of the ear canal and concha, make a an IEM flat purely on it's raw data and it will sound veiled because of it [ML Mikros/Etymotic ER2] and it be a horrible idea to use a db changes of volume to tune and IEM, read this from Toole. 
[size=x-small]"The curves tell us that different frequencies at the same sound level may be perceived as having different loudness. This is not a message that anything needs correcting. We live with these characteristics from birth, and they are a part of everything we hear, whether it is live or reproduced. That is why audio equipment must exhibit flat-frequency responses—uniform output at all audible frequencies—so the sounds we perceive have the correct relative loudness at all frequencies, assuming they are reproduced at realistic sound levels."[/size]
 
So it's true, the graph doesn't lie, but people can.  An objective graph can also impact a persons interpretation of a headphone.  
 
To answer your question in the OP though, there are a lot of factors that determine the bass of an IEM.  Seal is the largest one that is not only determined by insertion depth, but tip selection as well.  Much research has come along with the missing 6 dB as well that states that the human ear naturally creates some sort of amplification in the sub-bass areas by about 6 dB.  When the IEM is inserted, it bypasses these areas.  This could very well be the reason why people har it anemic.  
 
Not exclusive to IEMs, headphones in general. No matter what, they can't be perceived as speakers, best a manufactuer could to is maximize extension just as Sony did with EX1000 and Audeze in the LCD2. 
 
To answer your second question in the OP, this is quite personal.  What we really deal with are 3 bars to match by (maybe more).  The first one, accuracy.  This deals with how tonally accurate (tonally neutral) an IEM will be.  In the case of W4 vs Etymotic here, the Etymotic will win hands down.  This neutrality is really second to none in the IEM business.  
 
In terms of IEMs, there are some that are just as neutral as the ER4 series, modded UE9000-to ER4S, RE272 to ER4P. 
 
The second bar would be musicality, the ability to reproduce musical aspects.  The idea of this runs further from neutrality as a non-neutral headphone can indeed be "musically" accurate (IE producing all the aspects of music, a song).  Basically what happens is that the IEMs will be tested against the subjective descriptors listed in the glossary of terms, if it can reproduce all of them without overdoing it (turning it into a negative aspect), then it becomes "musically" accurate.  Many neutral IEMs are not "accurate" in this sense, but can be.  In the case of the Etymotic (HF2, I don't have an ER4) vs W4, the W4 takes the win slightly in terms of musical reproduction.  
 
I think a lot of the "musicality" aspect has to do with the bass extension depths which the Etymotics lack a bit (and a lot of BAs) [the last HK research backs this up]. 

 

[tinyman392]

Quote:

 

 
 
That first part makes no sense to me. Take for example, the 10k region, these are "sparkle" regions because they are fundamental regions which will cymbal crashes lie in. There is no subjective in what the 10k region does (boost it and cymbal crashes do indeed sparke more), if there's somehow a disagreement on what "sparkle" is, it's the terms themselves that need to be fixed, the effect of the 10k region remains objective. Sparkle is fine but terms like "sweetness" are useless..imo and just cause confusion. If used/read correctly, FR graphs are an amazing tool, just note of transients/decay extension and you've got a pretty concrete idea of what to expect. 
 
 
 
 
Compensations are based on resonants of the ear canal and concha, make a an IEM flat purely on it's raw data and it will sound veiled because of it [ML Mikros/Etymotic ER2] and it be a horrible idea to use a db changes of volume to tune and IEM, read this from Toole. 
[size=x-small]"The curves tell us that different frequencies at the same sound level may be perceived as having different loudness. This is not a message that anything needs correcting. We live with these characteristics from birth, and they are a part of everything we hear, whether it is live or reproduced. That is why audio equipment must exhibit flat-frequency responses—uniform output at all audible frequencies—so the sounds we perceive have the correct relative loudness at all frequencies, assuming they are reproduced at realistic sound levels."[/size]
 
 
 
Not exclusive to IEMs, headphones in general. No matter what, they can't be perceived as speakers, best a manufactuer could to is maximize extension just as Sony did with EX1000 and Audeze in the LCD2. 
 
 
 
In terms of IEMs, there are some that are just as neutral as the ER4 series, modded UE9000-to ER4S, RE272 to ER4P. 
 
 
 
I think a lot of the "musicality" aspect has to do with the bass extension depths which the Etymotics lack a bit (and a lot of BAs) [the last HK research backs this up]. 

 
The definitions states that a bump in the x Hz range will cause property Y (in this case, x = 10k, Y = sparkle which isn't defined in the glossary yet IDK why)?  The question is, in relativity to what?  We need an objective definition, not a generalization that may or may not hold true in all aspects.  That is what I'm getting at.  With respect to a flat frequency, the statements all hold true.  Take a look at the graph below.  There is a major bump at the 10k range (from ~6.5k and up; bump is about 6 dB), I'm more than certain you won't get too much "sparkle" with this though as the rest of the frequency spectrum will overrun it... Then the question becomes, how much can it be overrun before the property doesn't hold (The answer to this shouldn't be a single number for all frequencies; rather a function in terms of frequencies...).  
 
 

​

 
OK, maybe this was a bit of a cheap shot at the definition.  But a lot is left open for discussion with the definition (this is a legal counter-example).  What is a bump specifically?  Is there a bump at the ~16 kHz range?  How can we reform the definition of "sparkle" to not include this example?  This is what I meant when I said the definitions need to be reworked to account for every case.  When the definitions in the glossary were created, they were with respect to a flat response.  Headphones generally aren't always flat.  The definition breaks down with anomalies like the one above.  I've ran into anomalies all the time with multiple IEMs for different descriptors.  Creating an objective definition for audio terms is very difficult, you have to make it broad enough to include all headphones that would fall into it, but not too broad where it lets some in (that shouldn't be).    
 
And you state if they are read/used correctly...  Key word is if.  The way they are read is 100% subjective (graph above, by definition, it should have the sparkle property; it can be read that way).  It's your interpretation of the results, you can read it one way, another person as experienced can read it another.  It remains subjective (for that reason).  
 
You didn't understand my argument regarding compensation.  My argument is not what sounds flat to the ear, it's how the ear responds as the frequency trends away from what is considered flat.  Is a 2 dB divergence at the 1 k range the same as a 2 dB divergence at the 20 Hz range (or take any frequency, 10 kHz for example)?  Yet, with compensation, they are still treated the same...  It'll take away 2 dB regardless to form the compensated.  It shouldn't be like this.  Actual divergence and perceived divergence hasn't been studied, so it can't be assumed that it is in fact the same.  This is the reason why I only look at raw graphs for data anymore, the compensated ones are highly misleading much of the time IMO.  I truly believe that more research needs to be done before we can have any accurate compensation (there just simply isn't enough taken into account).  To sum this up, I'm saying compensation shouldn't be the same compensation per IEM and not be linear with respect to frequency as none of that has been confirmed yet (but it still is).
 
You know what I meant with I said the ER4's neutrality was second to none... This was an ER4 and W4 statement really, why bring the UE 9000 (NOT CLOSE TO NEUTRAL AT ALL.  The UE 9000 has a slight v-shape to my ears...  You actually don't have the graph for the UE 9000 but Tyll does...  Sorry, I had to throw this cheap shot ) and RE272 into the story?
 
Musicality (being musical) deals with a lot more than just bass extension and depth.  If it was, every bass-focused IEM would get 5 star reviews from me...  It's not the case.  You are overgeneralizing this aspect.  It deals with the entire audio spectrum, not the first 100 or so Hz of it.  Although you are correct (to a point), the bass extension has something to do with it, but it's not the whole story.  It's not the only difference separating musicality from neutrality (there are others that can differ as well).  I don't agree with your standpoint that everyone will find neutral (to the ear with compensation) to be the best (personally, I do find neutral to be the best; the majority of the public does not), but I understand the idea completely (which is why I've stopped attacking it; you rate one way, I rate another, I've accepted that).  
 
I know I'm being a bit nit-picky right now.  However, I feel the statements of stuff being "objective" still aren't objective.  Are they closer to being objective than a subjective description?  Yes.  They still aren't objective though IMO.  This is the last I'll say about the objective chart statements in this thread, I've said enough.  You can refute, but there will be no one to answer.  

 

[mark_h]

Quote:

Many neutral IEMs are not "accurate" in this sense, but can be.  In the case of the Etymotic (HF2, I don't have an ER4) vs W4, the W4 takes the win slightly in terms of musical reproduction.  
 
 

 
Given that there is a clearly audible difference in neutrality between the 4P and the 4S then I think the HF2 will further emphasise this deviation from the target curve. 

 

[castleofargh]

this is me years after years:
"wow this is super good, listen to those powerfull ugly bass! is it bose? "  ----->  i know hifi, it's flat to get the sound like it was played. super interesting. I'm not listening to music I'm working. where is my fun?  -------> found a headphone with reasonable audio quality,but what i really love is its sound signature. i enjoy it very much and couldn't care less if it's flat or not as long as it makes my ass shake --------> i try to get the same signature with better imaging and will die doing so. always happy with music"
 
Forthwall, welcome to step 2. 
take your "flat" headphone, try 3 DAPs  get 3 sound signatures.
add an amp get 6 results!!! good luck on the flat road mate. not an easy one.
 
 
 
about er4:
everybody feels like the er4 is sounding real(except bass). some love it, and some really feel like they're with the band!!!!!! getting a headache from percussions. so real!
getting er4 is pretty much like flipping a coin. hard to know the result until you try.
i love them balanced for classical and .... that's it.

 

[fourthwall]

Thanks guys, useful information here

 

[gnarlsagan]

Good points made all around. The ER4S has been the most enjoyable iem for me. You'll have to try it to find out for yourself how you'll perceive it. If you take technical ability conveyed through graphs and measurements seriously, and you find flat or neutral to be the most desirable goal for audio reproduction, then you will likely prefer the ER4S. Many people don't understand or care about those things and consequently prefer other iems. But even people who do understand or care about measurements can prefer other iems, |joker| being one of them. Personal preference is the overriding theme here.

 

[tinyman392]

Quote:

Good points made all around. The ER4S has been the most enjoyable iem for me. You'll have to try it to find out for yourself how you'll perceive it. If you take technical ability conveyed through graphs and measurements seriously, and you find flat or neutral to be the most desirable goal for audio reproduction, then you will likely prefer the ER4S. Many people don't understand or care about those things and consequently prefer other iems. But even people who do understand or care about measurements can prefer other iems, |joker| being one of them. Personal preference is the overriding theme here.

 
I thought |Joker| preferred neutral IEMs like the RE272 (though he stated his preference as analytical).  Unless I'm thinking of ClieOS...    

 

[bobeau]

Quote:

 But even people who do understand or care about measurements can prefer other iems, |joker| being one of them. Personal preference is the overriding theme here.

 
The ER4P was my first IEM back in 2k4 or so, got the S adaptor about 6 mos later when I got an amp.  I remember at the time really liking them, being totally awed by the detail, but I moved onto the Shure E500s once those were released a couple years later.  I wasn't a fan of the treble in those, but overall my attitude was... well these sound more like headphones (or consequently speakers if you think of it on that continuum).  
 
Fast forward to today and my preferences have pretty much kept to this path.  I've had some HF2s for a couple years now and can't stand them, to the point where I'll listen to any cheap piece of gear I have in preference to those.  My attitude on neutral sources as well is mixed... couldn't stand the Anedio D1, oddly enough loved the O2/ODAC combo I had.  
 
All that said, at one point in time I really did love the Ety thing.  It opened my ears to something I hadn't really experienced before.  I'm not sure exactly how other IEMs that do the thing they do compare but it's an experience I would recommend to any budding audiophile.