[flinkenick]

Fair enough. The ideas that are put forth are definitely similar but that's the facts of the situation, whether due to housing resonance or whatever mumbo-jumbo Audeze's portaplanar line have major issues in tuning. The magnitude of the 2-3k bump is up for debate but I'm pretty confident that most established ears would agree that a bump at that region is definitely necessary and has nothing to do with the Harman or whatever else target curves (which I kinda sorta agree with and not at the same time. Well, I agree with Harman more than HRTF in IEMs anyways, which is what Tyll uses in his IEM measurements for some reason...).

I wasn't disputing the necessity of the 2-3 KHz bump. In fact, the argument was that there was a large dip in this region. In my first response I posted this:

I know an iem should ideally have a 2-3 KHz bump.

This is among others reflected in my reviews of Zeus, W900, and Flamenco from the top of my head.

I don't recall having much of an issue with bass and treble on raw i4 either. My main gripes were with the midrange which caused this veiled-ish masking effect mainly in vocals, but bass and treble were pretty clean for the most part. That's what I think you might be hearing when you referring to that veiled-yet-clear effect.

I don't recall mentioning any issues with the cleanliness of the bass or treble. The only thing I mentioned about the treble in my messages to Vishnu was that the treble was prominent, which will occur if the midrange is relatively attenuated by comparison. However, in the same message I said it was very transparent.

 

[Mimouille]

Don't have the jazzy oBravo amp, but they sound damn fine out of my LG V30 or the Echobox Explorer. I saw they've gone back up to full price - that deal was an absolute steal for the quality (even managed to avoid import fees on mine)!

Not heard any other AAW stuff - is the tuning of the Q in line with their general house sound, does anyone know?

Well they sound nothing like the AAW900, apart from being more on the warm side.

 

[flinkenick]

Back on topic, I too did hear "serious issues" on raw i4, though I still maintain it's probably due to auditory masking from the 1.5k peak. I didn't really hear a hit in technicalities either, apart from it it just sounding really really odd.

I missed this part; but again, a simple bump in the 1.5 KHz does not mask the vocal performance. In fact, an iem that has a bump here will have a better vocal reproduction than a curve that is completely flat; it will increase the solidity of its formation and even transparency. Bumping the range between 500 Hz to roughly 4 KHz will affect selective components of the vocal density and forwardness. The issues I'm reporting are not minor variations that are common to iems with similar bumps in the midrange. It will only result if the fundamentals of the midrange as a whole are completely decreased.

With the non-cipher cable, the iSine 20's vocals sounded off(which has the same frequency response as the i4). Only source that sounded proper was out of the Lotoo Paw Gold since it had a proper sounding parametric EQ. I tried everything(I was so ****ing pissed with the fact that they sold something that the mids sounded like doodoo without EQ), it was such a headache!!

It's the 1.5khz bump, and even if you EQ with most of what is out there, it still doesn't sound right.

They don't deserve to be on the top!!

You seem to be agreeing with Crinacle, against my argument. But if what you were posting was true, then reducing the 1.5 KHz bump would solve the issue, which you yourself agree it didn't. The solution is not reducing the 1.5 KHz, it is EQ'ing up the 2-5 KHz range, and reducing the following treble.

 

[SilverEars]

The best EQ I've used was from a member here named Headwacker. He saw Crinicle's graph with the 1.5khz peak and reduced the band around it by 7dB, and made 4 additional filters to his EQ, and the mids, vocals issues was fixed.

It sounds very very odd with the 1.5khz bump. Once again, I will emphasize the 1.5KHZ bump nick if you have not registered with your ears.

 

[flinkenick]

The best EQ I've used was from a member her named Headwacker. He saw Crinicle's graph with the 1.5khz peak and reduced the band around it by 7dB, and made 4 additional filters to his EQ, and the mids, vocals issues was fixed.

It sounds very very odd with the 1.5khz bump. Once again, I will emphasize the 1.5KHZ bump nick if you have not registered with your ears.

See post before.

Basically you are saying in two posts: the issue is the 1.5 KHz bump, you need to fix this. But even if you fix this, it still sounds off.

 

[crinacle]

You seem to be agreeing with Crinacle, against my argument. But if what you were posting was true, then reducing the 1.5 KHz bump would solve the issue, which you yourself agree it didn't. The solution is not reducing the 1.5 KHz, it is EQ'ing up the 2-5 KHz range, and reducing the following treble.

Ah I think you've been misunderstanding me. It's not just the 1.5k peak but also the lack of 2-3k emphasis. There needs to be both; simply EQing down the 1.5k peak creates a somewhat linear response which isn't fully optimal. The final effect of Audeze's DSP is just that anyways; from a 1.5k bump to a "proper" 2-3k one. That combo whammy of early peak and null is the real killer.

 

[SilverEars]

The vocal, mids issue was fixed with the 1.5Khz drop(there were additional tweaks in EQ, bit this was causing the major issue). It still didn't sound right after EQ.

The 1.5Khz bump was the major issue. Keep playing with it, you'll see it is.

 

[flinkenick]

Ah I think you've been misunderstanding me. It's not just the 1.5k peak but also the lack of 2-3k emphasis. There needs to be both; simply EQing down the 1.5k peak creates a somewhat linear response which isn't fully optimal. The final effect of Audeze's DSP is just that anyways; from a 1.5k bump to a "proper" 2-3k one. That combo whammy of early peak and null is the real killer.

The vocal, mids issue was fixed with the 1.5Khz drop(there were additional tweaks in EQ, bit this was causing the major issue). It still didn't sound right after EQ.

The 1.5Khz bump was the major issue. Keep playing with it, you'll see it is.

The issues I mentioned with the i4 are the note weight, imaging, and transparency. These are fundamental issues that have no relation to a 1.5 KHz bump; a bump at 1.5 KHz by itself does not have such a negative effect on the vocal transparency. But of course if you were not able to pick up on these differences, there is little left to discuss. If you have a bump at 1.5 KHz instead of 2-3 Khz you still have a fairly linear signature; there are plenty iems that do this without sounding veiled. Examples of iems with a bump around 1-2 KHz are the 8.2, NT6, ES80. What causes a lack of transparency is huge fluctuations in the frequency response, whether it is a large dip or peak.

Look at the raw graphs of the iSine before and after DSP. You see the same rise at 1 KHz continuing through 1.5 KHz, followed by a 20 dB difference between 2-4 KHz. The large discrepancy difference between the dip and bump here is is the primary reason of the issues I mentioned.

SilverEars, if it didn't sound right after your EQ, it clearly wasn't the solution. So I'm failing to see how this is a strong argument.

 

[crinacle]

The issues I mentioned with the i4 are the note weight, imaging, and transparency. These are fundamental issues that have no relation to a 1.5 KHz bump; a bump at 1.5 KHz by itself does not have such a negative effect on the vocal transparency. But of course if you were not able to pick up on these differences, there is little left to discuss. If you have a bump at 1.5 KHz instead of 2-3 Khz you still have a fairly linear signature; there are plenty iems that do this without sounding veiled. Examples of iems with a bump around 1-2 KHz are the 8.2, NT6, ES80. What causes a lack of transparency is huge fluctuations in the frequency response, whether it is a large dip or peak.

Look at the raw graphs of the iSine before and after DSP. You see the same rise at 1 KHz continuing through 1.5 KHz, followed by a 20 dB difference between 2-4 KHz. The large discrepancy difference between the dip and bump here is what is causing 95% of the issue.

SilverEars, if it didn't sound right after your EQ, it clearly wasn't the solution. So I'm failing to see how this is a strong argument.

I believe the 8.2 and NT6 have a typical 2-3k bump. Not sure about the ES80 but the ES60 is also the same in that regard. The highest point of the bump for those IEMs are nowhere close to 1.5k. The Audeze portaplanars are unique in that regard in that the apex of the peak is at 1.5k which I've never seen in any other IEM.

I should reiterate my point, I'm not saying the 2-3k null is not the issue in an absolute sense but rather the 10/20/i4 problems are unique due to that peak. The Jupiter for instance (graph here, if anyone's interested) is the most prominent example of an IEM having that 2-3k null but it doesn't sound as odd as the portaplanar trio. The main reason for that? That peak.

 

[flinkenick]

I believe the 8.2 and NT6 have a typical 2-3k bump. Not sure about the ES80 but the ES60 is also the same in that regard. The highest point of the bump for those IEMs are nowhere close to 1.5k. The Audeze portaplanars are unique in that regard in that the apex of the peak is at 1.5k which I've never seen in any other IEM.

This is Piotr's own provided measurement of the 8.2:


And this is Rin Choi's measurement of the NT6, albeit compensated and smoothed:

However these are two more random measurements of the NT6:

I should reiterate my point, I'm not saying the 2-3k null is not the issue in an absolute sense but rather the 10/20/i4 problems are unique due to that peak. The Jupiter for instance (graph here, if anyone's interested) is the most prominent example of an IEM having that 2-3k null but it doesn't sound as odd as the portaplanar trio. The main reason for that? That peak.

I'm growing tired of repeating myself. The reason is not due to a 1.5 KHz peak. At all. Interpreting frequencies is understanding that the specific points within a frequency range are only one part of the sound. Implementing a 7 KHz peak for instance will not always mean a sound is bright. A 2-3 KHz bump does not automatically mean vocals will sound natural. And a 1.5 KHz bump certainly does not mean that vocals will sound off. Properly interpreting a frequency response involves understanding both the effects of specific frequencies, as well the balance and coherency of the frequency response as a whole.

So no, the main reason is not that peak. It is the large discrepancy between the peak and the dip what causes the effect. If the peak was instead at 3 KHz followed by a similar dip, the tuning would still be off, although it would manifest itself in a different way.

 

[crinacle]

This is Piotr's own provided measurement of the 8.2:


And this is Rin Choi's measurement of the NT6, albeit compensated and smoothed:


However these are two more random measurements of the NT6:




I'm growing tired of repeating myself. The reason is not due to a 1.5 KHz peak. At all. Interpreting frequencies is understanding that the specific points within a frequency range are only one part of the sound. Implementing a 7 KHz peak for instance will not always mean a sound is bright. A 2-3 KHz bump does not automatically mean vocals will sound natural. And a 1.5 KHz bump certainly does not mean that vocals will sound off. Properly interpreting a frequency response involves understanding both the effects of specific frequencies, as well the balance and coherency of the frequency response as a whole.

So no, the main reason is not that peak. It is the large discrepancy between the peak and the dip what causes the effect. If the peak was at 3 KHz followed by a similar dip, the tuning would still be off, although it would manifest itself in a different way.

I'm not really disagreeing with you there. It's the 1.5k peak into null (that is UNIQUE to Audeze's portaplanars) that is giving that weird tuning effect. In fact, that's I believe that's what I meant when I say "peak", in that it is the highest point after (and before) which there would be inclines. So in essence, yes, the 1.5k peak is the problem since it results in that decline afterward, as opposed to the DSP'd response that mitigates the peak by pushing the 2-3k frequencies to a higher SPL relative to the initial.

(Also off topic, and I really hope you don't respond with hostility considering our past discussions about measurements, but I was referring to peaks in a raw measurement setting. Introducing different compensation curves will of course result in different peaks and dips being shown. Comparing Olive-Welti, HRTF, Harman, Diffuse-Field and raw curves are all going to look different. The NT6 Pro curve on raw setting does not have a 1.5k apex but rather a 2k apex. That is a result of the Olive-Welti compensation and not the IEM itself. When you compare Tyll's HRTF-compensated iSine graphs with Speakerphone's Olive-Welti-compensated graphs, the differences cannot be analysed without proper understanding of such target curves. And of course, Piotr's own measurement of the 8.2 is made raw, therefore you see the 2k apex.)

Further edit: Rin Choi's raw NT6 graph (in grey) also shows a 2-3k apex.

 

[flinkenick]

I'm not really disagreeing with you there. It's the 1.5k peak into null (that is UNIQUE to Audeze's portaplanars) that is giving that weird tuning effect. In fact, that's I believe that's what I meant when I say "peak", in that it is the highest point after (and before) which there would be inclines. So in essence, yes, the 1.5k peak is the problem since it results in that decline afterward, as opposed to the DSP'd response that mitigates the peak by pushing the 2-3k frequencies to a higher SPL relative to the initial.

I think we might be converging. The main argument, is that a 1 or 2 KHz bump is not negative at all. An iem with a gradual bump between 1-2 KHz can still have a very nice tuning. There are also iems with nice vocals that peak after 2-3 KHz, around 4 KHz, or do not have a bump in the midrange at all. Besides, the distance between 1.5 and 2 or 3 KHz is small, so it would be very odd if a peak at 1.5 KHz would mask vocals, and the neighbouring frequency at 2 or 3 KHz would enhance it. The range between 500 - 4 KHz frequencies will selective affect the fundamentals of the vocal production. The main presence is between 2-3 KHz, but the surrounding frequencies still play an important role. So if there is a large drop in the fundamentals of instruments and vocals (roughly between 500 Hz - 6 KHz), they lose their presence. In the case of the i4 this results in an airy, diffuse, and veiled reproduction.

(Also off topic, and I really hope you don't respond with hostility considering our past discussions about measurements, but I was referring to peaks in a raw measurement setting. Introducing different compensation curves will of course result in different peaks and dips being shown. Comparing Olive-Welti, HRTF, Harman, Diffuse-Field and raw curves are all going to look different. The NT6 Pro curve on raw setting does not have a 1.5k apex but rather a 2k apex. That is a result of the Olive-Welti compensation and not the IEM itself. When you compare Tyll's HRTF-compensated iSine graphs with Speakerphone's Olive-Welti-compensated graphs, the differences cannot be analysed without proper understanding of such target curves. And of course, Piotr's own measurement of the 8.2 is made raw, therefore you see the 2k apex.)

Further edit: Rin Choi's raw NT6 graph (in grey) also shows a 2-3k apex.

Yes I understand it was a risky move to post a compensated graph, which is why I myself added it was smoothed and compensated, and also added other measurements to complement it. Even so, this is still a moot point if the above is clear that it doesn't matter if it is a 1.5 or 2-3 KHz peak when referring to the issues discussed.

 

[subguy812]

and at the end of the day it doesn't really matter in this thing called life.....

 

[crinacle]

The main argument, is that a 1 or 2 KHz bump is not negative at all. An iem with a gradual bump between 1-2 KHz can still have a very nice tuning. There are also iems with nice vocals that peak after 2-3 KHz, around 4 KHz, or do not have a bump in the midrange at all.

The latter I can agree with, I've heard decent IEMs with a 4k-5k bump. Gradual bumps I don't mind, but the transition for the iSine 20 was very jarring for me. Maybe it's due to different listening perceptions, who knows.

Besides, the distance between 1.5 and 2 or 3 KHz is small, so it would be very odd if a peak at 1.5 KHz would mask vocals, and the neighbouring frequency at 2 or 3 KHz would enhance it.

It's very significant. You of all people should know, 1500Hz vs 2000Hz vs 2500Hz is huge considering the scaling of frequencies. I just double-checked with a tone generator too; it's not something that can be ignored.

Anyways, doesn't seem like this discussion is going to go anywhere anytime soon. At least we can agree that the Audeze portaplanars suck when used raw, and requires DSP to truly shine whether via EQ or Cipher. Cool?

and at the end of the day it doesn't really matter in this thing called life.....

An actual discussion involving the intricacies of audio and it gets a beating-the-dead-horse comment...

 

[flinkenick]

It's very significant. You of all people should know, 1500Hz vs 2000Hz vs 2500Hz is huge considering the scaling of frequencies. I just double-checked with a tone generator too; it's not something that can be ignored.

Anyways, doesn't seem like this discussion is going to go anywhere anytime soon. At least we can agree that the Audeze portaplanars suck when used raw, and requires DSP to truly shine whether via EQ or Cipher. Cool?

The point was not whether there is a difference in perception between 1.5 and 2 or 3 KHz. Your argument was that 1.5 KHz has a huge negative effect on the vocals, while a bump at 2 or 3 KHz is desired for a positive effect. I'm saying this isn't true, and isn't logical considering they are neighbouring frequencies; i.e., if 8.2 had a bump at 1.5 KHz it would be awful, but now it's at 2 KHz it's fine. The main frequencies that will have a negative effect on vocals are for instance the upper bass or lower/mid treble; not the fundamentals in the center midrange. But I agree, let's agree to disagree here if you want to maintain this point of view.