New white paper of the PEQdB tool with interesting results out

[JSmith]

Founder and CEO, conclusion;

5 CONCLUSION
The PEQdB over-ear, in-ear, and generic magnitude frequency response target curves are the most statistically optimal headphone target curves created and should be the industry standard for tonality.

JSmith

 

[RogerSmith]

Yes, I have published many peer-reviewed papers in medical and scientific journals and co-authored national guidance documents. I have been a peer-reviewer for many papers and national guidance documents. I served as an Editorial Board Member for a medical journal.

Not sure why any of that is relevant for identifying shortcomings in a technical document.

Yes, this white paper superficially follows a journal format in the naming of the various sections, but the text contained within those sections do not. The paper gives the appearance of self-promotion aided by criticism of existing work in the form of a diatribe. It is not presented as a well designed and executed study intended to explore a hypothesis and contribute to scientific literature in a defensible and reproducible manner.

Good, but not ever scientific work has to explore a hypothesis. This work presents finding, based on the analysis of several hundreds of listeners which resulted in a new target curve. I wouldn't know what is not scientific about it. Often in scientific publications, you describe the current status of things and then go on to point out some shortcomings or limitations and present your improvement. At least in theoretical physics, where I have published, it is like this and I guess in medicine as well. And this is exactly what is done in this paper.

 

[amirm]

This work presents finding, based on the analysis of several hundreds of listeners which resulted in a new target curve.

What new target curve? This?

I see no new curve here, just noise around the general curve. Since audio production has no standards for tonality, and Dr. Olive's research has proven that there is high variability in bass and treble preference, attempting to get exact curves is silly.

I have always said that what we want out of this research is a standard. Something that is close or at preference level of many listeners. Then the listeners must be given tone controls to finalize to their preference. This work and that of headphones.com, etc. pollutes this idea with nonsense. We don't need 10 takes at target curve when a single curve cannot possible work for everyone.

I frequently deviate from target in my own reviews. I start with what the difference to target is, but often fine tune across a dozen or so clips, and balancing things like bass distortion and such. Reminds me of what a Mercedes salesman told me when I mentioned the steering in Benz took higher effort: "you have to drive the car and not have the car drive you." A bit of effort is needed here on behalf of the user to drive the tonality they want. This is what the author's tool was supposed to do. Him pretending to be an audio researcher and jumping the gun to show off while matching his personality, is as usual shows appetite for recognition beyond one's knowledge and wisdom.

 

[RogerSmith]

What takes me with quite some surprice is the resistence, which borders on ignorance, which most commenters present here. Science is a process and always open and improves and changes. So how many of you have actually used the tool or at least EQed to the presented target and tested if they actually might like it? Being stubborn, narrow-minded and not even try anything new for sure is not very scientific.

 

[amirm]

What takes me with quite some surprice is the resistence, which borders on ignorance, which most commenters present here. Science is a process and always open and improves and changes. So how many of you have actually used the tool or at least EQed to the presented target and tested if they actually might like it? Being stubborn, narrow-minded and not even try anything bew for sure is not very scientific.

He reached out to me when he was getting started. I tried it and gave him feedback. There is a lot of knowledge in this forum. Please refrain acting like the author and speaking in these terms.

 

[wwenze]

Maybe the true lesson behind this is how one piece of data can give two different conclusions depending on the person

 

[Lion♡]

I don't follow many threads, but I will follow this one.
This makes me crave popcorn more than watching a Netflix series.

 

[RogerSmith]

What new target curve? This?

View attachment 453396

I see no new curve here, just noise around the general curve. Since audio production has no standards for tonality, and Dr. Olive's research has proven that there is high variability in bass and treble preference, attempting to get exact curves is silly.

You always critize deviation from the target curve in your reviews, as a measure of good sound. So suddently this is not important anymore?

I have always said that what we want out of this research is a standard. Something that is close or at preference level of many listeners. Then the listeners must be given tone controls to finalize to their preference. This work and that of headphones.com, etc. pollutes this idea with nonsense. We don't need 10 takes at target curve when a single curve cannot possible work for everyone.

Maybe you did not understand the concept completely. It is exactly about that everyone can, through AB test, by the way as many as you like, 10 or 33 or any number, find their own specific target curve. And this not only with the very limited filters for bass and treble applied by the Harman research.

 

[Lion♡]

Maybe you did not understand the concept completely. It is exactly about that everyone can, through AB test, by the way as many as you like, 10 or 33 or any number, find their own specific target curve. And this not only with the very limited filters for bass and treble applied by the Harman research.

To what specific "their own specific target curve" are you referring?

 

[RogerSmith]

To what specific "their own specific target curve" are you referring?

The one that will come out as a result after you run the AB algorithm. You should give it a try. I suspect that many, especially strong defenders of the Harman curve, will be surprised by the curve that they really prefer.

 

[Lion♡]

The one the will come out as a result after you run the AB algorithm You should give it a try.

I know my own curve: for each azimuth, elevation, and distance, and even my personal DF.
That's why I'm asking you.
You can always apply EQ as much as you want.
So what do you mean by "their own specific target curve" here? Just a simple preference?
"Their own specific target curve" cannot be an industry standard.
"My curve" also cannot be an industry standard.

Also, Amir kindly wrote this just a few minutes ago. Please read what he wrote slowly, and then reconsider what I asked you.

I see no new curve here, just noise around the general curve. Since audio production has no standards for tonality, and Dr. Olive's research has proven that there is high variability in bass and treble preference, attempting to get exact curves is silly.

I have always said that what we want out of this research is a standard. Something that is close or at preference level of many listeners. Then the listeners must be given tone controls to finalize to their preference. This work and that of headphones.com, etc. pollutes this idea with nonsense. We don't need 10 takes at target curve when a single curve cannot possible work for everyone.

I frequently deviate from target in my own reviews. I start with what the difference to target is, but often fine tune across a dozen or so clips, and balancing things like bass distortion and such. Reminds me of what a Mercedes salesman told me when I mentioned the steering in Benz took higher effort: "you have to drive the car and not have the car drive you." A bit of effort is needed here on behalf of the user to drive the tonality they want. This is what the author's tool was supposed to do. Him pretending to be an audio researcher and jumping the gun to show off while matching his personality, is as usual shows appetite for recognition beyond one's knowledge and wisdom.

 

[RogerSmith]

I know my own curve: for each azimuth, elevation, and distance, and even my personal DF.
That's why I'm asking you.

How do you know it? Did you try all possible targets (which are infinitely many) ?

You can always apply EQ as much as you want.

Yes, but that is specifically what is done with the tool, together with machine learning. After a few trials, you will converge or your personal most prefered curve. Doing this by hand is impossible.

So what do you mean by "their own specific target curve" here? Just a simple preference?

Yes, your target curve is your most prefered one, how could it not be?

"Their own specific target curve" cannot be an industry standard.

True, the one presented in the paper is some average of all user's target. But the good thing is, you don't need to care about this and just find your own one.

"My curve" also cannot be an industry standard.

It could be, but will probably not.

 

[Lion♡]

How do you know it? Did you try all possible targets (which are infinitely many) ?

I have my own HRTF.

Yes, your target curve is your most prefered one, how could it not be?

My curve isn't just my preference; it reflects how I actually hear. I naturally find the way I hear in reality to be the most natural.

Yes, but that is specifically what is done with the tool, together with machine learning. After a few trials, you will converge or your personal most prefered curve. Doing this by hand is impossible.

True, the one presented in the paper is some average of all user's target. But the good thing is, you don't need to care about this and just find your own one.

It could be, but will probably not.

Is an individual's curve an industry standard?
YOU CAN EQ ANYTIME AS YOU WANT.

I know it seems like I'm repeating myself, and I am. You can EQ as much as you like based on your preferences.
However, what "your EQ" cannot become an industry standard.
If their A/B results and samples become more numerous and validated, other people's opinions might change. But even if a new curve is created that way, it still won't reflect the interaction of your preferences, eardrum, pinna, and every individual, including yourself. In that case, someone else will claim that new curve is also wrong. Will they then create another A/B curve?
You can EQ to your heart's content. There's absolutely nothing wrong with that.
Just consider what measurements and industry standards truly mean.
I shared the link to Amir's writing, but it seems you didn't look at it closely. Please review the text below one more time. His short statement contains everything you need.

We don't need 10 takes at target curve when a single curve cannot possible work for everyone.

 

[amirm]

You always critize deviation from the target curve in your reviews, as a measure of good sound. So suddently this is not important anymore?

I am telling you, as I explained, that the target curve cannot ever be 100% precise or even 80% precise especially in certain region. That doesn't mean we throw out the target. If I see a headphone be 1 or 2 dB above or below the target in bass, you don't see me complaining. I will however in some cases apply EQ to see if it matches my taste or not. And I routinely mention that the vendor may be right in choosing to do that.

Maybe you did not understand the concept completely. It is exactly about that everyone can, through AB test, by the way as many as you like, 10 or 33 or any number, find their own specific target curve. And this not only with the very limited filters for bass and treble applied by the Harman research.

No one is determining any "target curve." They play content and mess with tonality to see what they like. No different than when I apply EQ in every review I do and adjust. I find that depending on content, I may have to still mess with these knobs. This is one thing. Taking these ad-hoc, uncontrolled results and claiming to have found a precise target curve has nothing to do with this.

Finally, I am going to strongly caution you on your tone. You are starting to talk like the author of the paper which we showed the door to.

 

[Old_School_Brad]

This is one ad that slipped past my ad-blockers..

 

[RogerSmith]

I have my own HRTF.

Even if you knew your HRTF exactly, this is not identical to knowing your preferred target. There are additional questions like if you base it on a flat speaker in an anechoic chamber, possible crosstalk ajustments, etc. I assume you calculated your headphone target curve from your HRTF, so it would be interesting if you really would approach at the same curve when running the PEQdB AB test.

 

[Lion♡]

Even if you knew your HRTF exactly, this is not identical to knowing your preferred target. There are additional questions like if you base it on a flat speaker in an anechoic chamber, possible crosstalk ajustments, etc. I assume you calculated your headphone target curve from your HRTF, so it would be interesting if you really would approach at the same curve when running the PEQdB AB test.

I have anechoic responses, and I also have responses that include reflections.Of course, there's also crosstalk.
I even have responses with crosstalk canceled out
I'm not calculating a target curve from HRTF; that's just my REAL response. At this point, both headphones and IEMs should be HPCF-equalized, allowing me to accurately hear my body's response and the space.
Preferred target? I'm not fixated on the word "target." I just used "curve" to make it easier for you to understand.
I don't want any "target curve".
You still don't understand the fundamental point that I, or other users, are making.
I can somewhat grasp the intent of what you're trying to say, but other users have already kindly provided answers and explanations.
Your understanding of their points should come first. If you simply keep talking about "target curves" without that understanding, my interest in this thread will likely end here.

 

[RogerSmith]

I have anechoic responses, and I also have responses that include reflections.Of course, there's also crosstalk.
I even have responses with crosstalk canceled out
I'm not calculating a target curve from HRTF; that's just my REAL response. At this point, both headphones and IEMs should be HPCF-equalized, allowing me to accurately hear my body's response and the space.
Preferred target? I'm not fixated on the word "target." I just used "curve" to make it easier for you to understand.
I don't want any "target curve".
You still don't understand the fundamental point that I, or other users, are making.
I can somewhat grasp the intent of what you're trying to say, but other users have already kindly provided answers and explanations.
Your understanding of their points should come first. If you simply keep talking about "target curves" without that understanding, my interest in this thread will likely end here.

Thanks, with such a condescending answer, I have no intention to go any further as well.

 

[Lion♡]

Thanks, with such a condescending answer, I have no intention to go any further as well.

I have never made an arrogant statement on the ASR community.
As I said, with more data, validation, and further research, a new A/B target curve could be created. Many people might welcome it, and I clearly stated that. But is that the case right now? And can a single curve apply to every individual?
Your target might be preferred over the Harman target, and that's perfectly understandable. But what if others find your target less preferable? Would we then create a new curve again?
Neither Harman's research, many posts on ASR, nor any other audio community opposes EQing based on individual preference (which includes how an individual actually interacts with devices).
Don't you agree that the idea of the Harman curve being wrong and a new target being better, simply due to differences in EQ-applied preference curves, can be equally negated in reverse? Even if it's individual or average data.
That's why I initially asked what you meant by "their own specific target curve" and what your thoughts were on it. Amir and other users have already said it many times. So, I completely agree with and respect your idea that a different "curve" reflecting "preferences" can be good.
However, saying that the Harman curve, or a specific curve, or a new curve, is "wrong" or "doesn't match my preference" suggests to me that you're focusing too much on the term "target curve" itself. Please reconsider what that curve signifies and the purpose behind various contributors providing data through measurements.

 

[JSmith]

starting to talk like the author of the paper

Agree, getting that impression too... possibly at this point the onus should be on @RogerSmith to prove otherwise to continue posting in this thread?

Seems to be becoming rather belligerent in replies to members.


JSmith