ZMF Caldera Headphone Review

[amirm]

Did you read my posts or have you missed the part where I spoke about tuning and trying Harman on each headphone via different methods before deciding on a final tuning? What do you think I've been doing for the past 12 years?

I don't know what you have been doing. I can only go by what you have post here which so far, has no data on listener preference in any kind of controlled testing to justify the frequency response of Caldera. If you have such a test, let's see it. What you have tried or not tried is of little value if you are talking subjective, sighted tests.

I am reminded of the first time I attended blind speaker testing at Harman. It forever shattered what I thought I knew about my own hearing, preferences and speaker performance. Until you tell me that you have been through this, I am not inclined to accept opinions.

 

[amirm]

I'll let you respond and be out because it seems you're not willing to explore anything I offer from my apparently invalid experience which you are continually assuming.

Why do you keep doing this "I am out" business? If you want to be out, then be out and not post. Or post without constantly saying this.

That aside, I have experienced your headphone here. Its response is wrong to my ears. And wrong per research. You can't keep answering with I know what I am doing. I know my customers, etc., etc. That is not how we assess performance here and discuss science and technology. You see from my posts where I bring in the research. I expect the same in a discussion, not any kind of "take my word for it." We are not disputing that you have been in business for N number of years. And know how to build headphones. We only care about the final outcome which is what we buy.

 

[usern]

A lot of the resistors and capacitors used to tune these other Harman tuned headphones on the market have negative affects on the impulse response, and yes I've tested this. It leads to compressed dynamics and sound.

How do you test this? Take impulse response with and without the caps & resistors?

What do you mean with compressed dynamics? More distortion?

 

[kemmler3D]

What do you mean with compressed dynamics? More distortion?

Just interpreting here, but for any given transient signal, if you move the frequencies out of phase, the peak of the transient will be lower. One way to describe that is decreased or perhaps compressed dynamics. How that shows up in terms of audibility would vary a lot.

 

[kemmler3D]

there's never going to be a way to completely refute varied needs with 64/36 statistics regardless of the accuracy of the group.

True... someone employed at another large chicago-area audio brand (you've heard of them) told me they had tried to characterize / group inner ear shapes so they could come up with categories of standard tunings for IEMs. After taking 50 or so impressions, apparently they gave up because there was no discernable pattern. This is hearsay and I can't corroborate it... but it seems totally plausible to me.

I think that if we're thinking about a "headphone tuning is a solved problem at the industry level" scenario, it will inevitably involve personal scans / HRTF custom tuning to achieve the "last mile" of tuning. The situation we're in with headphones is as if everyone had slightly different color perception when they put on VR goggles. At some point you need to just cut through the confusion with brute force, I think.

 

[Rusty Shackleford]

It can be stated that, according to the study, a randomly selected individual with a 64 percent probability prefers a headphone that matches the Harman-curve. But it is determined with equal objectivity that a minority of 36 percent prefer a different profile (some prefer less bass, some prefer more)

I do not think any of the studies were random samples. So they cannot be generalized to the general public.

 

[IAtaman]

Sure! I took these with a random off the shelf unit and with the AECM 206 711 coupled Larson Davis to match to the Harman curve. I also changed the y-axis to more approximate what Amir was doing with it. Normalized around 700 HZ which is closest to how I hear the Caldera sounding to me if using Harman as a comparison.

Here's the 1/12:
View attachment 331715

Thank you. Anecdotally, I think this graph better aligns with what most people say about this headphone. General consensus seems to be that it is a bit bright. You would not be able to tell if from the measurements Amir shared, but the graph you shared does give that perception. In that sense maybe normalizing at 700 is not a bad idea in this instance.

 

[IAtaman]

I think that if we're thinking about a "headphone tuning is a solved problem at the industry level" scenario, it will inevitably involve personal scans / HRTF custom tuning to achieve the "last mile" of tuning. The situation we're in with headphones is as if everyone had slightly different color perception when they put on VR goggles. At some point you need to just cut through the confusion with brute force, I think.

Agreed. I do appreciate the engineering and the art that goes into designing a headphone but ultimately this will most likely be solved with electronics and software.

 

[zach915m]

Just interpreting here, but for any given transient signal, if you move the frequencies out of phase, the peak of the transient will be lower. One way to describe that is decreased or perhaps compressed dynamics. How that shows up in terms of audibility would vary a lot.

Yeah I mean there's a lot of factors at play and in the end it's the owners of the headphone that have to decide the priority, as I've said numerous times my approach is that there's no right or wrong as long as all the factors and measurements are being taken into consideration for the final tuning. For me I'd rather have low distortion, better transient response, impulse and CSD's rather than give those up just for FR. I personally find it more pleasing and many of our owners do as well.

Here's a recently measured THD of a headphone on this site that does measure to harman:

Here's the Caldera from page 1 of this thread:

Now the question for the potential owner who absolutely has to have Harman tuning for their particular HRTF (which as you know I don't always but sometimes ascribe to based on the headphone format) is whether they should EQ the headphone with the better THD measurement or take the other stock tuning? There's no right or wrong for them as long as they are happy right? BUT, various users will have a greater chance of hearing distortion and how it manifests in the headphone with higher THD, and also potentially have to deal with worse impulse response, sensitivity, and other variables in the headphone that gave up those areas for the cost of adhering to harman target.

Again, no absolute argument on what is right here, as you know ZMF makes options that are closer and further to harman - BUT - just that there isn't a clean way to get to harman without sacrifice in some situations and we as manufacturers have to make individual choices for what we feel is best, just like the owners who utilize our headphones. It's a good thing there's a lot of great choices these days!

 

[kemmler3D]

Yeah I mean there's a lot of factors at play and in the end it's the owners of the headphone that have to decide the priority, as I've said numerous times my approach is that there's no right or wrong as long as all the factors and measurements are being taken into consideration for the final tuning. For me I'd rather have low distortion, better transient response, impulse and CSD's rather than give those up just for FR. I personally find it more pleasing and many of our owners do as well.

Here's a recently measured THD of a headphone on this site that does measure to harman:

View attachment 331783

Here's the Caldera from page 1 of this thread:
View attachment 331784

Now the question for the potential owner who absolutely has to have Harman tuning for their particular HRTF (which as you know I don't always but sometimes ascribe to based on the headphone format) is whether they should EQ the headphone with the better THD measurement or take the other stock tuning? There's no right or wrong for them as long as they are happy right? BUT, various users will have a greater chance of hearing distortion and how it manifests in the headphone with higher THD, and also potentially have to deal with worse impulse response and other variables in the headphone that gave up those areas for the cost of adhering to harman target.

Again, no absolute argument on what is right here, as you know ZMF makes options that are closer and further to harman - BUT - just that there isn't a clean way to get to harman without sacrifice in some situations and we as manufacturers have to make individual choices for what we feel is best, just like the owners who utilize our headphones. It's a good thing there's a lot of great choices these days!

Personally I'm a "low THD, then EQ" guy myself since I do almost all of my headphones listening on a computer where EQ is easy - how I ended up with some LCD-XCs. So this makes sense to me.

I know Amir would like to see every manufacturer back up their design approach with robust data, (which is far from easy to get) or go with Harman, because at least that's got data behind it.

Personally, I don't think it's crazy to prioritize distortion or phase instead long as you're selling units, and they sound good and measure reasonably on the other metrics. It might not comport with the Harman data, but at least it's logical. But I would also not expect Amir to cut slack for non-harman tuned cans or endorse other paradigms for headphone tuning. I think even Dan Clark got a mixed review on here for a bit of deviation...

I'm personally well acquainted with the feeling one gets when a reviewer hasn't appreciated the merits of a product you put out. With Amir at least you can say he's consistent even if you don't like his point of view. I wouldn't say that about every audio reviewer...

 

[solderdude]

I think a valid alternative point of view is - we should have lots of different tunings because the Harman curve is A) not actually universal, B) doesn't guarantee perfection so much as a good starting point, and C) lots of people can't or don't want to use EQ.

It is a valid point of view. At least when we are talking about tonality.
BUT considerable dips and peaks (and I am talking several dB) is not desirable even when it does add some character.

The response may well be skewed in a certain way but should be 'smooth' across the frequency range.
Lets face it up to 1kHz the response is excellent.
From 1kHz to 6kHz the -7dB dips here and there are not desirable. One can EQ those of course but I am always looking for headphones that don't do that (so do not need to be corrected) and can be used without EQ or just very little. These type of headphones sound best to me.
Of course, overall a little recessed 1-5kHz range can make some recordings sound 'more pleasant' / less 'edgy' and overall this is what's happening.
Dips usually are less detrimental to sound than peaks.

Also bass should have low distortion at higher levels (linearity of drivers).

 

[kemmler3D]

It is a valid point of view. At least when we are talking about tonality.
BUT considerable dips and peaks (and I am talking several dB) is not desirable even when it does add some character.

The response may well be skewed in a certain way but should be 'smooth' across the frequency range.
Lets face it up to 1kHz the response is excellent.
From 1kHz to 6kHz the -7dB dips here and there are not desirable. One can EQ those of course but I am always looking for headphones that don't do that (so do not need to be corrected) and can be used without EQ or just very little. These type of headphones sound best to me.
Of course, overall a little recessed 1-5kHz range can make some recordings sound 'more pleasant' / less 'edgy' and overall this is what's happening.
Dips usually are less detrimental to sound than peaks.

Also bass should have low distortion at higher levels (linearity of drivers).

Sure, I am not talking about this headphone per se but just the general philosophy behind it.

 

[zach915m]

I know Amir would like to see every manufacturer back up their design approach with robust data, (which is far from easy to get) or go with Harman, because at least that's got data behind it.

Personally, I don't think it's crazy to prioritize distortion or phase instead long as you're selling units, and they sound good and measure reasonably on the other metrics. It might not comport with the Harman data, but at least it's logical.

Yeah and I totally get Amir's stance and opinion. I just am not liking the low key insults being thrown my way assuming that I haven't done my work just because it hasn't be studied. An easier route would certainly be to tune to harman and not think about other aspects, but it's not what I found works best in all formats (yes some). I know our audience and the sound I'm after and how to get there, and yes it's not always Harman but sometimes is. I don't take 6+ years to make the Caldera and not consider every facet, I literally don't do anything else with my life except think about headphones.

 

[zach915m]

Also bass should have low distortion at higher levels (linearity of drivers).

Especially planar drivers.

 

[kemmler3D]

Yeah and I totally get Amir's stance and opinion. I just am not liking the low key insults being thrown my way assuming that I haven't done my work just because it hasn't be studied. An easier route would certainly be to tune to harman and not think about other aspects, but it's not what I found works best in all formats (yes some). I don't take 6+ years to make the Caldera and not consider every facet, I literally don't do anything else with my life except think about headphones.

ASR is very dismissive of any sighted / unscientific listening tests of any kind performed by anybody. Not just Amir, that's the vibe around here. Frankly, for me it's a breath of fresh air compared to some of the venues online where seemingly every subjective report and superstitious opinion is taken seriously no matter what.

Having tuned headphones once myself though, I think I can understand what this attitude feels like in the other direction... not so good when you've put in serious hard work on something. I would just say don't take it personally, ASR is actually serious about the science thing, or at least the "anecdotes don't count" thing.

 

[zach915m]

ASR is very dismissive of any sighted / unscientific listening tests of any kind performed by anybody. Not just Amir, that's the vibe around here. Frankly, for me it's a breath of fresh air compared to some of the venues online where seemingly every subjective report and superstitious opinion is taken seriously no matter what.

Having tuned headphones once myself though, I think I can understand what this attitude feels like in the other direction... not so good when you've put in serious hard work on something. I would just say don't take it personally, ASR is actually serious about the science thing, or at least the "anecdotes don't count" thing.

Yeah agreed - and I've really enjoyed the healthy and open-minded/learning parts of this, it's fun and is good for everyone because we all care so much about these silly headphones. It's all about moving the hobby forward.

 

[solderdude]

Especially planar drivers.

Planars usually are low (except for a few) it is usually more a problem for open dynamics though, and less for closed ones (in general)

 

[Mulder]

And there is no medicine that is researched that helps the people not helped by Harman's. The other targets are simply much less effective. They don't cure what Harman's solution doesn't.

Here is the thing: I know how to get to near 100% with Harman's curve (as far as tonality). Simply adjust the level of bass and treble to taste if you don't like the default. Small variations can do a lot here. I do that all the time as I develop my EQ. If the sound is a bit bright, I pull things down. If the bass is too much (or causes distortion), I pull that down a bit. But the key is that I am starting with a very good recipe. One that is faithfully reproducing the sound of a proper speaker in a room.

Right. For that very reason, I wrote in a previous post in this thread that how well a headphone can be adjusted with eq is perhaps one of the most important characteristics of a headphone.

 

[Thomas_A]

It is a valid point of view. At least when we are talking about tonality.
BUT considerable dips and peaks (and I am talking several dB) is not desirable even when it does add some character.

The response may well be skewed in a certain way but should be 'smooth' across the frequency range.
Lets face it up to 1kHz the response is excellent.
From 1kHz to 6kHz the -7dB dips here and there are not desirable. One can EQ those of course but I am always looking for headphones that don't do that (so do not need to be corrected) and can be used without EQ or just very little. These type of headphones sound best to me.
Of course, overall a little recessed 1-5kHz range can make some recordings sound 'more pleasant' / less 'edgy' and overall this is what's happening.
Dips usually are less detrimental to sound than peaks.

Also bass should have low distortion at higher levels (linearity of drivers).

I agree about the variations > 1 kHz of this headphone. This is not anything desirable if you want high fidelity. There is nothing in this thread or anywhere else that supports such variation. I don’t need or want to EQ headphones.

 

[Mulder]

I do not think any of the studies were random samples. So they cannot be generalized to the general public.

If this was going to be determined with a random sample, the sample have to be much larger. The possibilities of carrying out a random selection are probably very limited in a study like this, in practice probably not feasible. However, just because it is not based upon a random sample do not mean it can’t be used for conclusions about the general public, but it requires relying on methods other than those used in statistical sampling surveys with large samples. I am sure Sean Olive are familiar with these methods. By the way, he himself do mention some uncertainties in the paper.