[Piotr Michalak]

Ever wondered how does M20 and M15 really affect the sound? And what happens when you take it off?

 

[buonassi]

If all else fails, you've always got shift+command+4?

and here I've been launching 'grab' and capturing screens the super hard way all this time! You've got to be kidding me, this is awesome. Took me a while to realize the file was saved to desktop. thank you sir.

 

[Otto Motor]

Quantity vs. Quality: An Indirect Approach!

The Brainwavz B200 v1., so it was speculated based on its FR graph, may lack midrange and sound dark, but it does not…voices are rather intimate…this is no surprise as the ideal, neutral sound features a somewhat linear response from sub-bass to the upper midrange.

When comparing the FR responses of the B100 and the much better sounding B200 v1., one is truly surprised.

Strictly speaking, the curves only give information on quantity but not on quality. But indirectly, these 2-5 kHz peaks point to a lack of quality. This statement can be tested by looking at FR graphs of really expensive items.

Where do I get this "subversive? information from - while adding my own interpretations? Mainly from the earphone wizards and master modders who use measurements to guide their modding steps. It is great to see how a V is straightened out to a flat line in the process while listening to the individual steps.

Another example of an upper midrange peak contrast. See how the ibasso IT01 gets away with less. I conclude, some of these low-budget multis have drivers just to compensate for other drivers. The ibasso is a single DD.

In the case of the KZ ZSN, the upper midrange “tower” again accentuates the otherwise rather thin voices. Next time, If I see such a curve, I won’t buy it.

Here a weird graph of a $400 earphone (sometimes $179 on massdrop). Against any expectations, this one sounds good with its four Knowles BAs. But, again, it gets away without the broad upper midrange peak.

Last but not least the universally loved Brainwavz B400 (with a healthy channel imbalance). I found the mid-bass unnaturally strong and the treble was rolling off early. It has a wee little upper midrange spike, but much less pronounced than the low-budget Chifi multis. The graphs also show that the B100/200 and the B400 are totally different beasts.

In summary, there are clearly observable FR trends between earphones of different qualities (at least good as a “rule of thumb”).

 

[Shma]

Thanks

 

[csglinux]

Wow, you've been busy @Otto Motor - nice job

Strictly speaking, the curves only give information on quantity but not on quality. But indirectly, these 2-5 kHz peaks point to a lack of quality. This statement can be tested by looking at FR graphs of really expensive items.

There are some expensive IEMs that intentionally add a lot of energy in that 2-5 kHz region, e.g.:

https://www.head-fi.org/threads/rha-cl2-tuning-preferences.894595/

There appears to be some proportion of folks who actually love this kind of tuning. My guess is it depends what kind of standing wave frequencies/amplitudes you end up with, given your own unique ear anatomy.

Mainly from the earphone wizards and master modders who use measurements to guide their modding steps. It is great to see how a V is straightened out to a flat line in the process while listening to the individual steps.

My personal preference would be to take that flat line and squish it back into a V

 

[castleofargh]

But indirectly, these 2-5 kHz peaks point to a lack of quality

I entirely disagree with this premise. a bump somewhere around 3khz is normal on raw measurements. it's the ear canal resonance that we get normally from sounds around us, and the various IEC models do their best to recreate it as an attempt to measure what a statistically average human ear would get. objectively of course, not subjectively. if you want a flat graph to sound flat to you, create your own compensation or try some of those already available.
now because different people are different, the actual resonance you get could be different in frequency and amplitude from what you measure(usually between 2.5 and 4khz apparently, and I'm not sure about how much it can vary in amplitude but it most certainly does). so when a midrange bump is very local(and that could be a bad sign), it's possible that sometimes it will align with what you get with your placement in your ear, and sometimes it's possible that it will show enough offset to feel really crappy and unnatural to you. but you have to consider that this could be as much about you as it about the IEM.
I'm not trying to say that the IEMs you don't like sound amazing, and I understand how easy it can be to find the UE900 nice as it's most certainly not an aggressive IEM. but if all IEMs were tuned that way, I'd cry over the loss of my favorite female voices.

 

[Otto Motor]

I entirely disagree with this premise. a bump somewhere around 3khz is normal on raw measurements. it's the ear canal resonance that we get normally from sounds around us, and the various IEC models do their best to recreate it as an attempt to measure what a statistically average human ear would get. objectively of course, not subjectively. if you want a flat graph to sound flat to you, create your own compensation or try some of those already available.
now because different people are different, the actual resonance you get could be different in frequency and amplitude from what you measure(usually between 2.5 and 4khz apparently, and I'm not sure about how much it can vary in amplitude but it most certainly does). so when a midrange bump is very local(and that could be a bad sign), it's possible that sometimes it will align with what you get with your placement in your ear, and sometimes it's possible that it will show enough offset to feel really crappy and unnatural to you. but you have to consider that this could be as much about you as it about the IEM.
I'm not trying to say that the IEMs you don't like sound amazing, and I understand how easy it can be to find the UE900 nice as it's most certainly not an aggressive IEM. but if all IEMs were tuned that way, I'd cry over the loss of my favorite female voices.

Interesting discussion - I am certainly listening! To me, the quality of an earphone is mainly defined on the accuracy of natural reproduction. My examples of "unwanted peaks" all related to low-budget earphones.

 

[azel831]

is there anything in here that sounds maybe 90% - 100% of my beloved pfe232 i can no longer find?

 

[EagleWings]

Various attempts have been made to derive a headphone-target-curve and regardless of the approach, the common trait you would see on all these curves is a upper-mid bump (~15dB), centred at around 3kHz. But the quantity of the bump and the centre of the bump can vary considerably between individuals, as the target curves are derived based on an average ear as @castleofargh mentioned.

Apart from variations of ear geometry between individuals, what the target curves don't take into account is, the difference in how this bump is perceived by the human ear, from a source that is placed 3 metres away in an open field (speakers), versus, a source placed at the ear-canal opening in an enclosed space (IEMs). But I will argue that this bump is definitely necessary for the overtones to be reproduces correctly. But depending on your ear-geometry and preference, the quantity and centre of this upper-mids bump will vary. I myself prefer a 10-12dB bump for closed back IEMs and a 11-14dB bump for open-back headphones.

If a manufacturer wants to add this upper-mid bump as the headphone target curves compel the existence, then the manufacturer also needs to make sure they tune the headphone or IEM to have sufficient lower mids (300Hz-700Hz) and centre-mids (700Hz-1kHz). Its when the lower-mids and centre-mids are not in sufficient quantity, the upper-mids bump start to sound too prominent in the form of making the notes sound thin and dry, or making the instruments and vocals sound shouty. But I can also imagine some folks preferring only a gentle bump or even a no-bump, even though the headphone might have sufficient lower and centre mids.

It is because of all these reasons, many successful IEM manufacturers, trust their ears to tune the IEMs than strictly following curves derived through science. Of course, the baseline reference they would still use is, how the music sounds on a good set of speakers.

 

[crinacle]

On the the topic of target curves, I've been exploring the idea of coming up with my own based on my experiences with hundreds of IEMs and having the data for all of them as well.


Assuming dynamic driver IEM, separate curve probably required for BAs

Additional explanations, context and applied examples back in my subreddit: https://www.reddit.com/r/inearfidel...mpt_at_a_target_curve_based_on_my_subjective/

Never really felt like Harman or DF sounded "neutral" so I thought maybe I'll throw my hat into the ring.

 

[james444]

On the the topic of target curves, I've been exploring the idea of coming up with my own based on my experiences with hundreds of IEMs and having the data for all of them as well.


Assuming dynamic driver IEM, separate curve probably required for BAs

Additional explanations, context and applied examples back in my subreddit: https://www.reddit.com/r/inearfidel...mpt_at_a_target_curve_based_on_my_subjective/

Never really felt like Harman or DF sounded "neutral" so I thought maybe I'll throw my hat into the ring.

Looks pretty good to me. I agree that the Harman bass bump is not needed for DD IEMs. Perhaps a slight bump for BAs, to compensate for faster transients and lower tactility, as compared to DDs.

I also agree that the 2-3kHz centered bump should be moderate for IEMs with a shallow fit (= most DDs). A larger bump is likely appropriate for deep fitting IEMs like the Etys.

Just my 2c...

 

[hakuzen]

On the the topic of target curves, I've been exploring the idea of coming up with my own based on my experiences with hundreds of IEMs and having the data for all of them as well.


Assuming dynamic driver IEM, separate curve probably required for BAs

Additional explanations, context and applied examples back in my subreddit: https://www.reddit.com/r/inearfidel...mpt_at_a_target_curve_based_on_my_subjective/

Never really felt like Harman or DF sounded "neutral" so I thought maybe I'll throw my hat into the ring.

agree. all the iems which sound to me nearer to flat analytical response of studio monitors, don't have +10-14dB bump at 3-4kHz. about +5-7dB instead. nor they have that bass boost. and have slight bump at 1.5kHz.
so my subjective experience matches your curve. keep on it!

 

[CoiL]

In the case of the KZ ZSN, the upper midrange “tower” again accentuates the otherwise rather thin voices. Next time, If I see such a curve, I won’t buy it.

I`m with You on that one. ZSN bump is too much for me.

 

[Otto Motor]

I`m with You on that one. ZSN bump is too much for me.

I was really trying to address the principles citing the ZSN as an example. It applies to many other multis in this price category. I am presently very very happy with the Pioneer SE-CH3, a $22 single DD.

 

[SilverEars]

I'm looking to put together a rig like Crin's. Where can you purchase IEC60318-4 compliant inner ear simulator for a good price? Or where can you purchase the same ear simulator like Crin's?