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

as a rule of thumb, i try not to exceed +/-6db of corrections in the wideband range (8 middle octaves)

That means i can have a +6db @ 2200hz and a -6db @1300hz but if it exceeds 12db total, i will surely question the driver/integration/xover points.
 
yes of course, it will be done with an old school AudioControl RTA, an Earthworks M50 mic, a pink noise. All in a large sound isolated room, with moderate sound treatments.

If Tomahawk is not happy with the Audiocontrol RTA, he will bring his own gears. 😉
 
Eldam,
I know what model Troels used as I talked with him and I also talked with Per Skaaning about what Troels used. Did I try a set of his, no I didn't.

I outlined to Skaaning what I was had for my other drivers, my cabinet, the enclosure size for the existing mid etc etc and Skaaning suggested the one I ended up getting. Do they sound better than the previous Dynaudio M560D mids I was using, yes.

And whay about the medium dome from Proac in a face to face benchmark ?

For the conversation it would be interressant to know the range of filtering on the medium and the slopes !

Today made an experiment by ading a very low amount of airy medical cotton behind my aluminium tweeters (filtered at 2100 LR4) and it's changing subjectivly the amount of détails for more but mainly how the mid sounds ! Divided by two the little amount was even more better and of course better with than without this micro amount of airy medical cotton !

Which makes me think than maybe the compact amount of coton damping pad than some add in the internal aera of the voice coil (so behind the dome) is not the good way... at least with my aluminium tweeter cones.

I'm very curious to know if it's easy this time with a mid to listen a difference with a kapton voice coil versus an aluminium-kapton one ???

My subejective but limited experience is metal is sounding more precise and less colored after some home works than polypro or paper... But never tried at iso perimeter (same Fhz range and XO and near T/S parameters) !

Very hard to judge a driver alone... But the source remains an essential factor, that's why I have a doubt with the layout of main active filter like MiniDSP ; poor and poor power supply as well ! Just my subjective opinion which is not the last one of course... many things to discover...:xmasman:
 
The mic will be placed exactly where the head of the participants will be.

Room will be quiet as much as possible (i will give the noise floor level once measured) with ventilation off while participants runs the test (and on in-between, to keep the temp stable!)
 
One thing that is not yet determined is the distance from the listening position to the driver.

Also, the driver's set-up is at fixed height. So maybe an adjustable chair would be needed, to keep the same 0deg/ears level, and avoid some differences because of different dispersion/beaming, etc...
 
Time to move to this century Jon. EQ is as good as the measurement. 1/3oct and pink noise no longer cut it.

well, i sure don't trust your iPad app and mic any more than my gears... I sure can tell you that.

..and if we are going to limit the # of EQ bands, then anything with more resolution than 1/3oct is completely moot.

I far prefer reliability than resolution.
Unreliable resolution is completely useless and, even worse: misleading.
 
An IR could show if you're right or not. Just because you're above 400 Hz doesn't mean you won't include any reflections, especially with pink noise. No matter how big the room is, there's going to be a floor right? Not sounding very F1 to me at the moment...

Read up on this site, there's bound to be several projects that use higher tech than what you are proposing right now.
 
Last edited:
An IR could show if you're right or not. Just because you're above 400 Hz doesn't mean you won't include any reflections, especially with pink noise. No matter how big the room is, there's going to be a floor right? Not sounding very F1 to me at the moment...

No it's not F1 at all and it's about the maximum F1-ish as we will get. In fact, as i see it today, we should consider ourselves lucky to be allowed of using electronic xovers and few bands of EQ!
 
well, i sure don't trust your iPad app and mic any more than my gears... I sure can tell you that.

..and if we are going to limit the # of EQ bands, then anything with more resolution than 1/3oct is completely moot.

I far prefer reliability than resolution.
Unreliable resolution is completely useless and, even worse: misleading.

I'm talking about REW. Reliable, precise, easy and accurate.
 
I think you're missing the point of this test, Wesayso.

The main goal of this (part 1) blind test, is to determine whether or not it's possible for people to identify drivers from each others.

Then, if part 1 succeeds (positive identification possible), we can move to part 2, which would be about ''Which is the best''.

-------

My own educated-guess as of now ?
I think, EVEN UN-EQd, some drivers would be difficult to distinguish. And any EQ will make this just even harder... Especially if we get ''F1'' about it...

So, the idea of this part 1 is to level-match carefully (important) and to EQ moderately without going over the top about it.
Samething about room's acoustics... I see no point of making the test in an anechoic chamber or anything of that sort. It's not a real-life situation and, anyway, won't really help the case.
 
I think you're missing the point of this test, Wesayso.

The main goal of this (part 1) blind test, is to determine whether or not it's possible for people to identify drivers from each others.

Then, if part 1 succeeds (positive identification possible), we can move to part 2, which would be about ''Which is the best''.

-------

My own educated-guess as of now ?
I think, EVEN UN-EQd, some drivers would be difficult to distinguish. And any EQ will make this just even harder... Especially if we get ''F1'' about it...

So, the idea of this part 1 is to level-match carefully (important) and to EQ moderately without going over the top about it.
Samething about room's acoustics... I see no point of making the test in an anechoic chamber or anything of that sort. It's not a real-life situation and, anyway, won't really help the case.

There's a middle of the road way to EQ drivers and a few different technical ways to do it. You're still using PEQ, that's fine, but for me the test will be uninteresting if done with pink noise. You'd include the room if you do. Just Measure with a sweep in REW and look at the IR...

I take it this is a single driver on test? I'm just proposing to use a more valid way of EQ. Take the room out as much as you can. With REW you can easily do that. Very common technology on this forum 🙂. Even more high tech is available (and often used on this forum).

I hope tomahack will bring you up to speed.