You do realize that some are voiced, right? There is a popular hifi brand who rolls the top end up, giving a different sound and also making them terrible to be used in a measurement rig...
As for this thread, if the drivers aren't EQ'd to have the same response and SPL level throughout, it's kind of silly as who listens to midrange drivers without any filters?
First of all, thanks for chiming in. I appreciate both your posts and your devotion to the <strike>hobby</strike> quest for better audio (calling it "hobby" seemed quaint, or, at least somehow less than appropriate).
I wouldn't seriously consider any DAC which did not, at the very least, include a "neutral" or "direct" mode, or simply have only one "mode" which was as flat and neutral and low-noise as possible.
From the OP: "...all drivers will be on an baffle, electronic xover from about 400hz to 7000hz (somewhat the minimum for comfortable listening), EQed and all SPL matched for fair comparisons."
So, while it may not be an ideal test, at least it seems worthwhile, to me.
Last edited:
Wow, is this debate/search still going on. I hadn't heard anything for so long I thought you had found your golden midrange and left it at that.
Can't wait to hear what you've selected as your fav.
I found mine, it's the Audio Technology 15H5220613(special order, as recommended by Skanning himself), love it. Takes a while to break it in but after that it's glorious.
Can't wait to hear what you've selected as your fav.
I found mine, it's the Audio Technology 15H5220613(special order, as recommended by Skanning himself), love it. Takes a while to break it in but after that it's glorious.
Well, here's the thing: we made preliminary tests and... Breaking news:
It's not that easy to identify drivers from each others once EQd and SPL-matched.
So i had to re-think the methodology and now we have something much better for the next sessions. It's too soon to determine if it's the lack of practice or the lack of auditive capacities but one thing for sure: it's not that easy. At least, for some drivers...
Next sessions will use very different driver's types, with many participants with various ages, background, etc... (audiophiles and non-audiophiles)
Oh, and no i didnt find my favorite yet.
In fact, it's not even my goal. At all.
Even once the identification blind test is completed, i'd be more interested about ''People's favorites'' than what I prefer 😉
But for obvious reasons, identification test must be made before.
In fact, it's not even my goal. At all.
Even once the identification blind test is completed, i'd be more interested about ''People's favorites'' than what I prefer 😉
But for obvious reasons, identification test must be made before.
I'm almost finished regarding the (new) methodology that i'll use, but i'm kind of stucked with one thing:
After the first run (music excerpt no.01, playback driver A, then B, C and D) i'll ask the question:
''Which one do you prefer?''
OR
''Which one you think sound the most realistic ?''
Both questions could lead to interesting answers but probably not the same. Which question would be the most relevant ?
Keep in mind that most people who will pass the test won't be hardcore audiophiles!
After the first run (music excerpt no.01, playback driver A, then B, C and D) i'll ask the question:
''Which one do you prefer?''
OR
''Which one you think sound the most realistic ?''
Both questions could lead to interesting answers but probably not the same. Which question would be the most relevant ?
Keep in mind that most people who will pass the test won't be hardcore audiophiles!
HiFi appreciation ("realistic") is very subjective. It doesn't correlate with HiFi.
If you're interested in "people's favorite" then "Which one do you prefer?" is the better question.
If you're interested in "people's favorite" then "Which one do you prefer?" is the better question.
I'd have to agree with Jay as I doubt very few people know what "realistic" sounds like.
Unless you play an instrument or go to the live acoustic concerts regularly you'd have no idea what real instruments should sound like, even the basic ride cymbals or acoustic guitars. For that matter all acoustic guitars sound different too, I know the two different brands I have sound totally different, not to mention different strings(brand) sound different on different guitars. Even different cymbals produce different sound signatures as well as different brands of cymbals sound different.
I'm not sure what you are doing when you say you are EQ'ing them, are you making them all flat over a certain band. I would think just level matching them would give the truest test. EQing would only leave the signature of each driver and not how they are performing as a mid driver in the mid band to which they are meant to operate.
So when will this test begin /or end so you can reveal the drivers which ones came out in what order of preference.
Unless you play an instrument or go to the live acoustic concerts regularly you'd have no idea what real instruments should sound like, even the basic ride cymbals or acoustic guitars. For that matter all acoustic guitars sound different too, I know the two different brands I have sound totally different, not to mention different strings(brand) sound different on different guitars. Even different cymbals produce different sound signatures as well as different brands of cymbals sound different.
I'm not sure what you are doing when you say you are EQ'ing them, are you making them all flat over a certain band. I would think just level matching them would give the truest test. EQing would only leave the signature of each driver and not how they are performing as a mid driver in the mid band to which they are meant to operate.
So when will this test begin /or end so you can reveal the drivers which ones came out in what order of preference.
Last edited:
Yes i agree with that.
Regarding music instruments.
That's why i think if there is a question about realism, it should be paired with a non-musical playback or a musical playback that is easy for everybody to evaluate the level of realism (i.e. acapella voice)
Yes you could use that for testing, or something like clapping or bells or any percussion instruments. Of course you can throw in some well recorded jazz or classical music just to round things out. For excellent female singing try some Eva Cassidy.
I believe it's irrelevant to test drivers in 2016 without the use of DSP/EQ possibilities.
Would be like testing new sportscars with tires from the 80's...
48db/oct or steeper crossovers and EQ to get something flat within the 400-7khz or so, and all level-matched.
The subjective questionning is just a bonus, the first goal of the test is the identification part: Will people be able to identify each drivers from the others ?
fyi: the 400-7khz band is not decided yet and will vary based on the 4 drivers chosed.
I can very well use a 500hz-5.5khz for a group of 4 drivers and then next group will use different band. The idea is to keep the ''weakest'' driver(s) within his comfort zone, otherwise the test won't be as accurate.
But so far, 400-7khz works pretty well and it's surprisingly listenable.
I can very well use a 500hz-5.5khz for a group of 4 drivers and then next group will use different band. The idea is to keep the ''weakest'' driver(s) within his comfort zone, otherwise the test won't be as accurate.
But so far, 400-7khz works pretty well and it's surprisingly listenable.
I think EQ'ing each driver is not the way to test, level matching yes. Sure use steep xovers, but all drivers are not flat in their response curves so why make then flat. That's the whole point of testing them.
Yes of course you need to keep(adjust) each driver within the manufactures recommended operating range of freqs. But on the other hand you can't be smoothing out peaks and valleys using eq as that's not showing the driver's true capabilities.
The bad news is we'll have to conduct the first session (let's call it part 1) with only 4 drivers.
I takes about 25-35minutes for each participants to go through the 5 different excerpts (on each drivers) for a total of 20 playbacks, including the pauses, notes, etc...
People get kind of tired/unfocused after a while, so the next group of 4 drivers must be tested in a different session.
I believe it would be possible to make a semi-final/final kind of thing, with the 2 best drivers of each sessions together in the final test. However, we're still far from taking a decision about that. For starters we're not even sure how many people will actually be able to identify the drivers correctly, and then the ''what is your favorite'' and ''which is the most realistic'' questions will have to decide the finalists. If we're lucky, there will be a consensus, but what happen if ''favorites'' and ''realistics'' gives opposite answers ?
I takes about 25-35minutes for each participants to go through the 5 different excerpts (on each drivers) for a total of 20 playbacks, including the pauses, notes, etc...
People get kind of tired/unfocused after a while, so the next group of 4 drivers must be tested in a different session.
I believe it would be possible to make a semi-final/final kind of thing, with the 2 best drivers of each sessions together in the final test. However, we're still far from taking a decision about that. For starters we're not even sure how many people will actually be able to identify the drivers correctly, and then the ''what is your favorite'' and ''which is the most realistic'' questions will have to decide the finalists. If we're lucky, there will be a consensus, but what happen if ''favorites'' and ''realistics'' gives opposite answers ?
Well, i have to strongly disagree with that.
I think it's the complete opposite: if you want to show the driver's true capabilities, you need to EQ it!
You can discard a driver because of some nasty peak or dip, but once the flaw is fixed, the driver is great.
Visaton B200 is a good example. Lowther DX4 as well, even though hardcore fans likes it ''natural''.... until they listen to it EQd 😉