ASR Getting Into Measuring Headphones!

[m8o]

Some what's left of innerfidelity is found here:

Thankfully the way back machine archive has snapshots of all the measurement pdf resources and other stuff.

 

[solderdude]

Not all of the pdf's though.
These can be found here and in alfabetical order.

 

[or965]

I bought many earphones that has been measured with different head simulator and found out that the RA0401/02 give very good result with IEM measurement. It so much better then the RA0045 or similer 60318-4, I think that it much more accurate over 6 khz . I equalize earphones and found out that the harman target without the bass boost and treble tilt is quite good but I prefer a modified Etymotic target which very similer to the Etymotic er4sr. I know that the 0db.co.kr also using the GRAS 45CA with the RA0401/02 and I equalize my kz as10, kz zs10 pro and sonicast dirac+ mk2 with their measurements .
@amirm I saw that you measured the earstudio he100 with the BK 5128, can you measure the earstudio he100 with your GRAS 45CA .It will be very interesting to see the diffreance. I anticipate that there will be a difference over 6 khz . I will get the earstudio he100 very soon and want to equalize it using the BK 5128 frequency response graph and GRAS 45CA and see if one is more accurate then the other .
Also I was very surprised to know that the GRAS 45CA cost 15,000 USD after seeing that the BK 5128 cost 41,000 USD , they do not post prices, it is much cheaper then the BK 5128 and give very good IEM measurement.

 

[nj75f]

Leaving the note, to contemplate, that amirm now becomes one of the subjectivist reviewers himself, since he wants to start subjectivist headphone reviews mixed with a tiny bit of objective data (measurements).

Measuring only frequency response and distortion is not nearly enough to give a fair comparison of headphones.
Especially the notion to review and mix up different headphone drives types (dynamic, electrostatic, planar magnetic, air motion transformer, ribbon, tesla) is inherently wrong in itself, since head/earphones with different driver technology does not sound the same, even when the frequency graph is basically very similar.
You can crank up the lower frequencies of a sennheiser hd650 as much as you like. Besides from distorting very fast, it will never give the same punch in the lower frequencies, as the audeze lcd-2 does, a planar magnetic type.

There is also no "bad or good" response, no matter the targeting curve (harman or whatever).
Amirm can give as much recommendations and decapitated or golf playing panthers as he wants.
It is mainly subjective at this point, not enough measurements to justify the science in adiosciencereview.
The problem arises that just the recommendation by him/giving a panther, makes people believe that this headphone now is "not worth buying/or is", even if he writes "suvbjective review" in size 50 big and red letters.

 

[solderdude]

Well that's a bit negative ... Amir will provide more measurement points and his opinion.
Isn't that what all reviewers do ?
Have you seen other headphone reviewers providing more info that is more 'sciency' and what more 'science numbers/graphs/plots' do you want Amir to show that will make it different from others ?

You could have just mentioned what other measurements you would like to see from Amir that other sites don't show and you feel are needed to make it more 'sciency'.

 

[amirm]

@amirm I saw that you measured the earstudio he100 with the BK 5128, can you measure the earstudio he100 with your GRAS 45CA .

Yes, I plan on repeating all those measurements with 45CA and more in depth.

 

[jazzendapus]

You can crank up the lower frequencies of a sennheiser hd650 as much as you like. Besides from distorting very fast, it will never give the same punch in the lower frequencies, as the audeze lcd-2 does, a planar magnetic type.

That's easily measurable.

There is also no "bad or good" response, no matter the targeting curve (harman or whatever).

You need to brush up on your ABC (aka, Harman/Sean Olive research regarding headphone preferences) before barging in with those claims.

I don't like when Amir gets into subjectivism and replaces things that need to be measured with "vast experience" and/or "exceptional training" and so, but you're way off target here IMO.

 

[MayaTlab]

There is also no "bad or good" response, no matter the targeting curve (harman or whatever).

So following that logic, this would be an appropriate frequency response curve ?

If the answer to that is yes, then please be my guest I'm launching my new headphones company and will happily sell you $3000 headphones made of my new material, a special brand of acoustically enhanced carbon fibre composite with newly developed extra-high excursion drivers made of Unobtainium and special, exclusive dampening material made with a carefully selected fibre obtained from specially raised silk worms listening to Mozart all day to enhance the quality of their silk.
If the answer to that is no, then it means that there exist at the very minimum objective criteria that bound what is acceptable or not in terms of frequency response curve. So we can definitely assert that there is such a thing as a "bad" frequency response curve.

 

[Vini darko]

 

[DualTriode]

Hello,

What does Impulse Response tell us? Something to be tested?

This is HE400i.

Thanks DT

 

[solderdude]

for starters... how high is the pulse supposed to go ? What's the reference ? Why use a 24ms scale and not 5ms ? How to tell driver resonance apart from that of the measurement system (ear canal/pinna) ?

 

[DualTriode]

for starters... how high is the pulse supposed to go ? What's the reference ? Why use a 24ms scale and not 5ms ? How to tell driver resonance apart from that of the measurement system (ear canal/pinna) ?

Al good questions.

I set it all up pressed the start button and selected auto scale.

This is the first Impulse Response with the new 45CA-9 test hammer.

I will run a few others and compare, adjust the scale to see what falls out.

What do you say?

Thanks DT

 

[amirm]

What do you say?

Tests like this are only useful for designers searching for something specific. Ditto for CSD/waterfall/time domain response, etc. For end-customers, they only serve to confuse and add noise to otherwise useful measurement.

We may occasionally use them on digging deep on "why" something is the way it is. But as a routine measure, they are not useful. This is why I don't post waterfalls for speakers even though I secretly run them all the time .

 

[solderdude]

What do you say?

Thanks DT

My HE400i impulse response.

It tells me the amplitude doesn't reach the amplitude of the stimulus (green trace)
My impulse is wider than a needle (positive half of a 5kHz square-wave).
The driver likes to vibrate at around 4kHz which is also easily seen in the CSD.
As can be seen in the CSD the membrane is fast and goes beyond 30kHz so speed is not a problem but there are other issues.
Not surprisingly as this was entry level planar when it came out.

There is also a harmonic frequency (I suppose because of the tensioned membrane as many planars with stretched membranes do this) around 8kHz. Luckily the ear canal also resonates a bit so the brain doesn't really care about certain resonances it has by itself.
A well damped driver doesn't have those resonances (or are much shorter)

Below the HE-6 (4 screw)

This one does reach the target but is less well damped at lower frequencies.

The CSD above shows the left channel's membrane has less tension than the right one and likes to vibrate at a lower frequency.
Also pretty well extended FR.

But... as Amir states. Mostly there is little extra info in it. Not in the cases above it gives some extra info here.
I have much more time to do a write up and measure than Amir who has a very busy schedule and pumps out reviews in a pace I could never do nor keep on going for that long.

 

[MZKM]

Can sensitivity and max SPL (however defined) be measured?

For instance, I believe it’s Æon, they can’t get loud at all, so no matter how good the FR is, if they can’t get sufficiently loud, many people would strike them off their list.

 

[amirm]

Yes, that is the plan. I need to disassemble the coupler and pinna unfortunately to calibrate it.

 

[Cahudson42]

Tests like this are only useful for designers searching for something specific. Ditto for CSD/waterfall/time domain response, etc. For end-customers, they only serve to confuse and add noise to otherwise useful measurement.

Yes, CSDs confused me when I first saw them. Then I read @solderdude 's explanation - a couple times I admit - and they started to make sense. I would suggest that most ASR 'regulars' are at least as capable as I am, and can understand them should they take the time to do so.

And while speakers, except for a few cases like Maggies, are all dynamic/voice coil - headphones are planar - electrostatic and magnetic, and dynamic. Much more of a differentiator. So until shown otherwise that CSD/time domain is not useful, could we keep an open mind that they might be, and show them?

Personally, I seem to be shifting in my 'music listening' to basically two modes - speakers for HT and casual background, often familiar stuff, I simple enjoy while I am moving around, cooking, whatever. And 'serious listening' - new stuff, or new renditions of old stuff - on Headphones. These days we can compensate for fr'relatively easily'..But that still leaves Distortion, and IMO time domain response, as two areas of difference. Why not examine them closely until shown they make little difference?

Looking forward to the new HP reviews!

 

[amirm]

I would suggest that most ASR 'regulars' are at least as capable as I am, and can understand them should they take the time to do so.

They show you redundant data though. Note how the frequency response were responsible for elongation in time:

So until shown otherwise that CSD/time domain is not useful, could we keep an open mind that they might be, and show them?

Unfortunately not. While Klippel software produces these, the AP software cannot. You have to export the data and plot it externally which is time consuming and a pain in the neck.

 

[Illtrick]

Comfort is such a critical component of the overall headphone experience. Many aspects are subjective but some values can still be expressed objectively: clamping pressure/force and earpad dimensions are two major factors in comfort that never get expressed. What do you guys think of adding data like that to the reviews?

 

[tvrgeek]

Darn y'all. Now I am going to read and follow this and second guess my headphones. I have a set of Grados SR 80I use for critical listening for transcribing disks or preamp tweaking, and my antique Yamaha blacks. The most musical I have ever heard. Circa late 70's.