Rhamnetin
Active Member
- Joined
- Dec 1, 2023
- Messages
- 217
- Likes
- 447
Tardigrades are overpowering the plankton. Do I have a bad cable?
But how's your PRaT?
Tardigrades are overpowering the plankton. Do I have a bad cable?
But how's your PRaT?
Exactly. Garbage in=garbage out. A speaker or headphone shouldn‘t be editorializing either to spice up the mix beyond what was intended nor should it be self-correcting. It shouldn’t be adding more bass and treble (the dreaded bathtub or U- or V-shaped frequency response curves). A speaker or headphone shouldn’t compensate for its distortion or noise that is masking the source response’s dynamic range. Any of these approaches counterintuitively often makes the music more constrained and congested because each joule of energy that goes into something not intended to be reproduced distracts and steals energy from what it was intended to reproduce. Of course, there is the Fletcher-Munson curve (fun fact: I went to college and was friends with the great granddaughter of that Harvey Fletcher!) which says that psychoacoustically a bit more bass and treble can be good at lower listening volumes to maintain a perceived higher volume, but that goes beyond this discussion. And even so, an EQ or loudness button or control employing a Fletcher-Munson curve will sound infinitely better on an excellently objectively crafted speaker than a mediocre speaker employing the same technique in its native response to hide the shortcuts and warts of its design.I found out audio passages that like mush on my three sets of speakers also sounds like mush on a Zero Red.
In numerous discussions within the "science-oriented" headphone community and elsewhere online, I've often encountered the argument that the system is minimum phased, leading to a somewhat tribal belief that the FR curve fully describes the system. Contrary viewpoints are frequently dismissed and punished, even laughed at. This trend is even observable in YouTube videos.Right - and I don't think anyone is saying it is, after all Amir shows more than one chart per review. It is a large part of the "description" of headphones, though. If I'm looking at a headphone review here and wondering if I should get excited over it, I look at the FR first (as well as how different the left channel's FR is to the right channel's). Then distortion - after all I want my headphones to reproduce sound, I don't want them adding it or changing it in any way. Now group delay to cover some weirdness and phasing. If they are all great, even if not perfect, then it is very likely I can EQ it to my preferred target without distortion. If you only look at FR you are actively ignoring other important aspects of the headphone.
Refer to my earlier post for a basic outline of the dynamic headphones model. My expertise doesn't extend to the intricate details of acoustics or mechanical vibrations in headphones, I'll study that further in the future, when time permits.Why do you think this is not possible? I am not saying you can EQ any one specific headphone to match any other headphone, but provided the distortion of both headphones are low and the group delay is fine, you can get really close (and the more filters you can add the better). I'd even bet there are certain matches you can make where the differences would be inaudible - so tiny they wouldn't be noticeable (if that's what you mean by 'exactly like').
To step in front of your next question, after looking at the group delay for the susvara, no I would not say that you can EQ the 7Hz x Crinacle Zero:2 to sound like the susvara because there's a lot going on with the group delay, and if it's related to phase then good luck. You need more than EQ on such a headphone. But mess with some all pass filters in the highs on the Crinacle to mess with phase and maybe...
I am honestly curious if you could EQ those Crincles to sound like the Dan Clark E3. The EQ needed for the crincle in the lows is kinda weird - maybe you'd need a tilt shelf filter? And obviously you'd need to reconcile the difference in over-ear vs IEM, but this is just an example. But really, isn't this what we are trying to do? We want to take a headphone and match the FR to our preferred target. We take another headphone, and we want to match our target. Given that the headphones are of good enough quality to both reproduce sound without adding anything, and we use enough EQ to get within X% of our FR target in all frequencies, aren't we effectively trying to make one headphone sound like another?
If you don't want to use your own FR target but instead want to use the headphone creator's target, then that will vary per-headphone. Or if you can't EQ your headphones (or don't want to), this also doesn't apply, but it means you are even more tied to the FR of the headphone (that is, you have to like its FR to then like the headphone).
That isn't part of my reviews. I show no comparison of one headphone against another, or rank them in any order. The frequency response of a headphone is either matching target which based on a number of listening test trials, or not. If it is, then I confirm by listening and job is done. If it isn't, filters are developed by me by eye, and another member with numerical analysis. Listening tests are performed including controlled AB tests. If the outcome is more positive with filters -- which it almost always is -- that outcome is provided for readers to try. Many report positive results from those filters, some use them as is, others with minor adjustments.Given that headphone sound quality cannot be wholly predicted based on these measurements, it's unscientific to assert that they are sufficient for determining if one headphone sounds better than another.
Which is exactly the opposite of what the research says. Our work is praised by the Dr. Olive in his recent article,This again circles back to the notion that while the ASR measurements are extremely useful, they are not enough to predict whether a headphone will sound better to the majority or otherwise.
By your logic no headphone should be liked because their drivers do NOT exactly match each other. Nor do you wear the headphone consistently enough to give you the same response even with identical drivers.Attempting to equalize two headphones to sound identical involves aligning one model's characteristics (including intermodulation distortion, voice coil's force constant profile, etc.) with another's.
Dynamic range is orthogonal to tonality. You can have great dynamic range with poor frequency response and vice versa. As such, we can test for each factor independently. I test for dynamic capability with my three SPL tests and importantly, listening. I am unique in doing this because I recognize its importance and consider it both objectively and subjectively. So to come and complain here is non-sequitur especially since in this review, these problems were identified much to chagrin of fans of this headphone.Consider the dynamic headphones as an example. The underlying mathematics and sources of non-linearities and distortions differ significantly from say, electrostatic headphones. Each headphone type, and even individual designs within a type, has its own set of characteristics that contribute to Total Harmonic Distortion (THD). Thus, citing THD alone is insufficient to accurately describe a headphone's sound. It's incredibly opaque.
This again circles back to the notion that while the ASR measurements are extremely useful, they are not enough to predict whether a headphone will sound better to the majority or otherwise.
This again circles back to the notion that while the ASR measurements are extremely useful, they are not enough to predict whether a headphone will sound better to the majority or otherwise.
Sorry Amir I meant dynamic driver headphonesDynamic range is orthogonal to tonality. You can have great dynamic range with poor frequency response and vice versa. As such, we can test for each factor independently. I test for dynamic capability with my three SPL tests and importantly, listening. I am unique in doing this because I recognize its importance and consider it both objectively and subjectively. So to come and complain here is non-sequitur especially since in this review, these problems were identified much to chagrin of fans of this headphone.
BTW, while THD is not dispositive in directly predicting impairments, it definitely shows where there is smoke especially when it is so clear as is the case of Susvara. And in extreme cases, it absolutely predicts a problem as it did with Susvara completely bottoming out.
That isn't part of my reviews. I show no comparison of one headphone against another, or rank them in any order.
"Not some imagined modelling of a headphone". Imagined? Great.The frequency response of a headphone is either matching target which based on a number of listening test trials, or not. If it is, then I confirm by listening and job is done. If it isn't, filters are developed by me by eye, and another member with numerical analysis. Listening tests are performed including controlled AB tests. If the outcome is more positive with filters -- which it almost always is -- that outcome is provided for readers to try. Many report positive results from those filters, some use them as is, others with minor adjustments.
When you go and get your tire balanced, if it already is, then there is nothing to do. If it is not, then different weights are put on until it is balanced. That is precisely the process I follow above. No one tries to compare what weight and where they were located to another tire (unless you need huge number of in which case, something seriously is wrong). Or claim that they have created a model of a tire in order to spin it and find out if it is out of balance.
A corollary of above is that if a headphone deviates a lot from the target as is the case with Susvara, then by definition it is not properly engineered in meeting the research based target. This is proven by the frequency response measurements. Additional tests such as distortion show whether there are clear and specific design flaws which again, exists with Susvara. There is no research or engineering principle that says there is goodness in these distortions/resonances.
As you see, we are absolutely following the science/research here which is based on listening tests and not some imagined modelling of a headphone. Listening tests enabled testing of a target that was most preferred which is put to use in my reviews. The research made measurements just like I am so we are in compliant with it. You don't get to make up your own ideas of what we are doing and then complain about it.
By your logic no headphone should be liked because their drivers do NOT exactly match each other. Nor do you wear the headphone consistently enough to give you the same response even with identical drivers.
This is on top of the fact that music has no standards so no one can say anything about likability of a headphone across all music.
We aspire to still get the job done and this calls for getting things approximately right. Once there, we have made huge progress beyond the wild west of today where every headphone designer has his own idea of how their product is supposed to sound like. If this were TVs, one would show a red color as orange and the other as purple. TVs today have made massive progress there because we do have standards for color accuracy. So while they don't all match each other, they come very close such that the red is still "red" although it may be a hair more saturated or ever so lightly shifted. It is not purple like we have in audio with no standards. And people like you walking around saying if you don't have perfection, then you shouldn't have any standards. Which is of course nonsense.
Just for completeness let me reply to this one:Which is exactly the opposite of what the research says. Our work is praised by the Dr. Olive in his recent article,
The Perception and Measurement of
Headphone Sound Quality: What Do
Listeners Prefer?
Sean E. Olive
"The reaction from the headphone industry to this new
research has been largely positive. There is evidence
that the Harman target curve is widely influencing the
design, testing, and review of many headphones from
multiple manufacturers, providing a much needed
new reference or benchmark for testing and evaluating
headphones. Several headphone review sites provide
frequency response measurements of headphones showing
the extent to which they comply with the Harman
target (Vafaei, 2018; Audio Science Review, 2020); in
cases where they fall short, corrective equalizations are
often provided."
No higher validation exists than the above statement. To say otherwise means not remotely understanding what the research into headphone preference was about.
Really, this dog don't hunt. You can't make up arguments of your own and then use them to complain. Not a word is said in the research that the target response is any kind of model for a headphone. The target curve says that if you meet it, majority of people will like the sound. And those who don't, can be satisfied with some adjustments of bass and treble. So please stop repeating the same broken argument over and over again.
All this time you haven't read the research??? No wonder your comments make little sense.Isn't Harman a benchmarking tool, and not meant to definitely deduce if one headphones sound better than another? I haven't read the research but that's what it sounds like.
And they do because I use the recommended target in research. And for every review I also listen. It is the combined force of the two that you see in my review conclusions.You can compare it to a standard, but if you want to assert that a headphone will sound good or otherwise based solely on the measurements, then its predictions (all your reviews) have to hold up to scrutiny.
There are limitations in all studies of this sort. But such limitations don't stop of us from drawing highly useful conclusions and information. You have to learn to live with some ambiguity in this space due to again, lack of standardization for decades."Not some imagined modelling of a headphone". Imagined? Great.
There's so many papers in many fields on studies with test subjects, but that doesn't mean they hold through under all conditions. I'll be very surprised if Harman doesn't publish the limitations in the implications of their studies.
All this time you haven't read the research??? No wonder your comments make little sense.
Answering anyway, core of Harman research was to find the preference/target curve. They then took a shot at assigning numbers to deviations from target and using that for ranking. I don't like the results of that work (ranking) and don't use it.
There are limitations in all studies of this sort. But such limitations don't stop of us from drawing highly useful conclusions and information. You have to learn to live with some ambiguity in this space due to again, lack of standardization for decades.
And yes, the model is imagined by you. And critiqued by you. Nothing in the research attempts to model a headphone. It makes frequency response measurements and compares that to listener preference studies and correlation is established.
The part where you said the measurements don't predict listener preference. 5 years of research were performed to demonstrate that we can do that very thing and you said the opposite. And the fact that you keep thinking the research was attempting to build a model for a headphone which in no way attempted to do so.Okay, so which part of it doesn't make sense?
I am defending my work. What do you want me to do, turn the other cheek?You sound very defensive.