World's best midrange Blind Testing - Need your help.

Even different power amps can sound or be perceived a little bit different think its not a good point discussing here because real device under test (DUT) for research is the midrange driver not the power amp, so everything has to stay the same from test to test only exception being changed is DUT, also think come on its a band-pass of 360hz-7.2khz, what amp can not replicate that band with fair good quality.
 
Subjective nonsense is the key here. A lot of it this month :whip:

I'm questionning my whole audiophile past... No seriously. About 25 years of speakers design/DIY down the drain. Sort of.. 😱

...Or i just grew old and i'm deaf.
Audiophiles are a scary bunch. A lot of focus on what's either of little or no concern and totally oblivious to what really matters. And a total confidence in their subjective experience, while questioning any scientific test.

Ok but which outcome can we expect ?
I mean; regarding identification ABX ?
The important thing here is to be on sure side. You can't EQ the room without introducing artifacts. But you can EQ the driver because it's minimum phase. I would EQ the drivers in nearfield.

Whether this makes a difference in your ABX or not is speculation. You tell us!
Great work by the way. 🙂
 
Can't pick out the best performer if we're unable to identify first.

That's the whole problem.

NO. This isn't true. You're trying to circumvent a blind test's potential failures by running a double blind test first. You don't actually know that the problem will occur.

The problem would be a blind test where unnamed, but represented drivers, and able to switch back and forth listening to whatever music the user pleases that ends up statistically insignificant. If there is no conclusive preference (let's say 75% or higher) then people can't tell them apart even when you tell them they're different. This may or may not happen. You're trying to conclude with a double blind test it won't before moving forward. It's creating strife with lots of silly arguments.

If anything a passive X-over with the drivers might start to show some bigger differences. The way a driver interacts with them is a large part of it's character.
 
Last edited:
🙂
It would prove you are EQ-ing the room + speakers, not much more.

The difference between drivers don't surface as much if you use the pink noise EQ in one spot.

Like I said before, step out of that sweet spot and listen. Do the different drivers still sound the same?

What would be needed is to gate the measurement before the first big reflection. EQ that curve. I like to use a frequency dependent gate (or window). But in your room and the placement you should be able to get away with just setting the gating. Get rid of the chair at the mic, because it will cause reflections.

Measure to get an IR and post the first 10 or 20 ms. That will show where the reflections are. I don't want to post an example out of my own library as that would only be confusing matters more. There should be a clear peak (though a band pass will shape the IR) and hopefully a relatively flat line afterwards, followed by some small peaks. Those peaks represent the reflections. Normally one gates the measurement before those peaks (the reflections).
As you're looking at ~370 Hz you should be able to get away with gating.
Personally I'd still prefer a frequency dependent window, but have to admit REW's frequency dependent window is a bit smoothed compared to other tools.
Most measurement suites have their own variant of a frequency dependent window. I'm used to REW as a measurement suite, though I use DRC-FIR to get my preferred frequency dependent window.

We will work with you. I'm not denying your results so far. I'm not surprised either. 🙂

Even though i agree with this in general it doesnt mather in this test. You can't EQ both power response and direct sound at the same time. This is given by the driver geometry. If as you say he eq's room effects, the direct sound of the drivers will differ which is even more audible acording to Toole. And especially from such a short distance.

Double the listening distance to 3.2m and the ratio between room/driver will be more realistic and the directivity performance will be more important. Maybe that is something that could be tested. Keep the EQ from the 1.65m distance.
 
Last edited:
Ok then, it's a case of uneven FR, am i right ?
....if level-matched... what else?
Yes, FR is uneven if the amplifier's output impedance is not low enough, and that is a case for a range of amps to sound different. There are amplifiers with low output impedance and there are some with high (or moderate) one. Choosing a low output impedance amplifier, you can add a resistor in it's output and it's FR will be very similar to a high (or moderate) output impedance amplifier. After that, you should level-match again. In this case, difference nearly disappears. But not always!
What else? I don't know. I can speculate with distortion harmonics, but besides tube or low/no global negative feedback type of amplifiers, I don't know why some amplifiers sound different. Difference can be very small, or non-existent.

But we have digressed - you made a great experiment with midrange units and we both agree that amplifier sound quality is not the problem here.
 
Last edited:
Just keep everything the same (EQ position etc.) but move the listening chair 0.5 M left or right...
Wanna hear even more difference, move the chair a bit further out.

Each driver will have different frequency response off axis due to driver directivity.

This will therefore simply be a test of frequency response.

The same could be achieved with the same driver and EQ to roll off the highs at certain levels.......
 
Are we looking for the best driver or not? 😀

Dispersion - and hence off axis freq response is very design and personal preference dependent.

Some love wide dispersion omni designs - others would favor narrow dispersion designs.

Assessing dispersion characteristics via off-axis listening is not being assessed here - simply look at most driver data sheets, or the width of the driver to factor this into a design i.e.:
12" ~ 1,150Hz
10" ~ 1,380Hz
8" ~ 1,720Hz
7"~ 1,970Hz
6.5" ~ 2,120Hz
5.25" ~ 2,620Hz
4" ~ 3,440Hz
3" ~ 4,950Hz
2" ~ 6,880Hz
1" ~ 13,750Hz
Etc.
This is about determining the different between drivers when they have the same freq response at the listening position - i.e. removing the confounding factor of on-axis freq response (king of the hill for driver assessment).

The fact that off axis response or power response will be so different for these drivers in this assessment, is testament to the over-riding importance of on-axis response.

To me it also says there is probably less reflected sound in this set up than people are making out - for instance if he was doing this in an anechoic room, power response would be entirely irrelevant and it appears to be the case here too.

I actually wonder if people would be better able to tell the difference between drivers if his room was more 'live'.....
 
OK so let's summarise our arguments, observations and conclusions so far:

We have agreed DSP cannot alter a driver's physical dimensions, cone shape, radiating surface area etc.

We have agreed DSP cannot change the law of physics governing sound wave propagation, dispersion etc.

We therefore can conclude, that regardless of what form of DSP is used, different drivers will have different power responses and this cannot be changed by DSP. It is governed by physics and the physical properties of the drivers themselves.

I hope we all agree on this.

Therefore in this blind ABX testing, the drivers will have very different power responses - regardless of the DSP techniques used to give a uniform on axis frequency response.

In other words - a 2 inch full range driver would 'light up the room' and the on axis freq response measurement (ungated) would have lots of reflected sound energy in the EQ measurements and corrections.

Whilst a 8 inch driver would beam from approx 1700 hz upwards and therefore would have much less reflected sound energy in the measurements/corrections from 1700 upwards.

And yet these differences are not obvious on blind ABX testing.

So we conclude this appears to be similar to an anechoic environment, where off-axis energy or power response becomes irrelevant.

So we suggest moving the chair back, or using a more live room to get more room interaction.......

So right there, we have talked ourselves around in a circle! We appear to now be arguing that the room isn't a large enough factor!!

To me, personally, as I have said all along in this thread - the fact people cannot tell a difference between the drivers despite the wildly different power responses is very interesting.

And we still haven't touched on the greatly different distortion, transient response, driver BL/MMS , neo magnet/ferrous magnets....etc .etc. etc....
 
I'll loose ? Then make money on my back. I'm your fool. 😉

you'd be surprised how much i'd bet on that... especially since the midrange test!

I think 99% of people would NOT be able to distinguish a class A amplifier from a class D icepower, once level-matched. UNLESS there is differences in the FR, which shouldnt.

Who got the balls to make a bet ?
I'm not tryin' to make money, i just want to cover the test's costs!! 😀


I would take you one a bet without any fear given you let them play for more than 2-3 minutes and not 8 seconds each...

I mean, I change the power supply on my laptop. My laptop feed my USB/AES converter and then my 4X10HD in AES mode and both my wife and I hear a difference... ****, the laptop psu affect the sound, how crazy is that.
agreed, it may not be the laptop itself that change the sound but the noise the psu inject back in my 120V main splitter and make it's way back to the rest of the chain. But, the difference is still there.
I recently spend ~$100 to build a regulated linear powersupply on my laptop and I would never let it out. this $100 was one of the best bang for the buck I've spend lately.
My $350 nordost QRT is permanently installed in the system too. According to your belief, this is pure "snake oil" but it really work and I'm even considering buying a 2nd one.

do you really think that all amp sound alike cause the distortion specs and power are similar, this is like saying that a Corvette drive the same as a Ferrari because the horsepower and tire size(grip) are about the same.... I whish it could be the case.

Later
 
Last edited:
And we still haven't touched on the greatly different distortion, transient response, driver BL/MMS , neo magnet/ferrous magnets....etc .etc. etc....

I think that BL/MMS is extremely important here given the cone is rigid enough to avoid breakup in the working zone.
that where the ATC and CD would shine.

My 18" subwoofer have a very heavy cone to keep the fs at ~18hz, I'm considering changing them for some lighter one but it's a very expensive experiment...
 
I would take you one a bet without any fear given you let them play for more than 2-3 minutes and not 8 seconds each...

I mean, I change the power supply on my laptop. My laptop feed my USB/AES converter and then my 4X10HD in AES mode and both my wife and I hear a difference... ****, the laptop psu affect the sound, how crazy is that.
agreed, it may not be the laptop itself that change the sound but the noise the psu inject back in my 120V main splitter and make it's way back to the rest of the chain. But, the difference is still there.
I recently spend ~$100 to build a regulated linear powersupply on my laptop and I would never let it out. this $100 was one of the best bang for the buck I've spend lately.
My $350 nordost QRT is permanently installed in the system too. According to your belief, this is pure "snake oil" but it really work and I'm even considering buying a 2nd one.

do you really think that all amp sound alike cause the distortion specs and power are similar, this is like saying that a Corvette drive the same as a Ferrari because the horsepower and tire size(grip) are about the same.... I whish it could be the case.

Later

I hear differences all the time when I make a change to my system, I'm just not so convinced as you that it's real. The power of suggestion trumps all. So unless I have real evidence of what the improvement was I'm not sure it's real or just perceived, but I sit back and enjoy it non the less.

The corvette and Ferrari analogy is really reaching. The corvette and Ferrari are complete systems, a more apt comparison would be to compare them to two different 6"/1" bookshelf speakers that encompass many design decisions which effect the overall out come.