Agreed 100%, but I'm still skeptical about your test because:
1 Psychological bias. Your goal is obviously "no one can tell the difference", and in that atmosfere, it's questionable if the participants have zero mental bias, both consciously and subconsciously.
2 Room. I don't think I can tell the difference in this test room. Noise floor is high, and RT60 is high for critical listening. I can't find the picture of the room, but I guess the early reflection of the room is pretty high also.
???
@scottjoplin
"Perhaps an indication of how important FR is or maybe the importance of direct sound and minimising the influence of reflections, I get the most enjoyable sound from my speakers listening near field but I'm not convinced it's due to a flat FR
It doesn't matter what you think is enjoyable. In the blind test, different drivers were not detectable. Btw, have you measured your room response/speakers?
"It's a real shame there are no measurements of the FRs, a gaping hole in the test in my view"
It would be a gaping hole in all drivers at both ends if not equalised...
And yet when equalised, Jon claims not detected...
When listening to music over that bandwidth and nearfield I'm not surprised there was no difference in FR heard.
Jon,
You answered most of the questions raised in posts # 15 and #21 (you skipped my questions #2 and 3 😉) which gives me enough information to respond to your test, and conclusions.
Your on axis, near field, equalized test effectively nearly eliminated any driver’s major differences to the ears due to :
A) Power response- though power (off-axis) response of each transducer was different, the single listening position relatively near the DUT (driver under test) reduced off-axis room contribution cues. The fixed size of the baffle made LF power response similar for all the DUT.
B) Potential SPL- all DUT were below the SPL where distortion issues would become apparent in the bandwidth chosen.
C) Frequency response-the drivers on-axis response was corrected, and no test of the resulting off-axis response was auditioned.
If the “King” was allowed to move from the location of the “Throne”, the results would have been quite different.
Your test protocol eliminated what you regard to be “that's about it” regarding the DUT, with predictable results.
At least, predictable to me, having done my own
High Frequency Compression Driver Evaluation
8 years ago, which also allows for people to hear recordings of each DUT.
That said, though I couldn’t reliably tell the difference between the equalized response of the six drivers tested on the same horn when used at “home stereo” levels, still wouldn’t bet my car that others couldn’t.
Cheers,
Art
Hmm.... I have 4 presets on my preamp/DSP - so that I can turn on and off different settings that I made in the DSP.
For me it's clear that I have less listening fatique, when I clean up issues in the FR. When I try to ignore them and simply make a more simple filter - I get tired of my speakers. When I go into a more in dept EQ'ing - I like my speakers for both youtube, netflix, music and simply voice in documentry's, the news and so on. Sometimes it's just a notch at maybe 1,5dB. But when my speaker is most pleasing to the ear - it measures very flat. BUT - it has to measure pretty flat off-axis too. I tried to ignore this - but then it sounds annoying over time.
It would be interesting to see the varying off-axis responses, that's for sure. However, the test conditions clearly preclude that from having any significant influence, and were also presumably manipulated to ensure that HD could not have any influence either. Which pretty much prevents any findings (still can't find any data) from being carried over to most real world conditions, and why I'm struggling a little to understand the purpose of the exercise.
Be that as it may -assuming sufficiently low levels of HD, and nearfield on-axis listening, with baffle dimensions appropriately sized, & in a suitable room acoustic, then the on-axis first arrival behaviour dominates perception. Plenty of studies on that for years. What is preferable, flat or some other form of curve e.g. a Bruel / Kjaer derived declining trend is another question though, & is one of practical useage, especially for multiways, and also affected by the progamme material & mixing. I've no fixed approach on that myself, other than varying as appropriate to a given set of conditions. YMMV as ever of course.
Last edited:
JonBocani, BTW, how loudly was the music played?
If the test was done @80dB RMS, there is only 27.5db dynamic range in your test room. Do you really think someone can hear subtle difference in that environment? It's far from quiet. The noise floor of the good professional music studio's control room is probably 20dB, and the dynamic range is 60dB. I would not perform a critical listening test in the running bus, and you should not. If your other tests were done in the same room, I consider your tests are all invalid just because of this single reason.
Your tests are interesting and I do not want to be offensive, so I would say your conclusion is only valid for domestic average listening environment, but still, my very average suburban living room's dynamic range is better than your test room's.
The noise floor of my critical listening space is 20dB or so (don't remember exactly), and I can hear the difference between DAC on ABX test in this room, and I'm pretty sure that you will. (Well, I always make some mistakes here and there, not 100%, but perfect listening ABX score is as hard as SAT perfect score. 😉)
If the test was done @80dB RMS, there is only 27.5db dynamic range in your test room. Do you really think someone can hear subtle difference in that environment? It's far from quiet. The noise floor of the good professional music studio's control room is probably 20dB, and the dynamic range is 60dB. I would not perform a critical listening test in the running bus, and you should not. If your other tests were done in the same room, I consider your tests are all invalid just because of this single reason.
Your tests are interesting and I do not want to be offensive, so I would say your conclusion is only valid for domestic average listening environment, but still, my very average suburban living room's dynamic range is better than your test room's.
The noise floor of my critical listening space is 20dB or so (don't remember exactly), and I can hear the difference between DAC on ABX test in this room, and I'm pretty sure that you will. (Well, I always make some mistakes here and there, not 100%, but perfect listening ABX score is as hard as SAT perfect score. 😉)
Last edited:
JonBocani,
Let’s get down the very simple basics in testing. For this test to have any meaning you would have to conduct the identical test, with exactly the same people, but without the bandpass filter implemented in the miniDSP. In other words, the same test, same people, same equipment, but wide open frequency response.
Then you could make a before and after comparison of the data to see the effect, if any, of attenuating the top and bottom end of the frequency range.
If you did the test without filtering, we have not yet heard about it.
If you didn’t do it, then any conclusion you come to is totally without merit and pointless to discuss. Your test proves absolutely nothing without a before and after comparison.
Let’s get down the very simple basics in testing. For this test to have any meaning you would have to conduct the identical test, with exactly the same people, but without the bandpass filter implemented in the miniDSP. In other words, the same test, same people, same equipment, but wide open frequency response.
Then you could make a before and after comparison of the data to see the effect, if any, of attenuating the top and bottom end of the frequency range.
If you did the test without filtering, we have not yet heard about it.
If you didn’t do it, then any conclusion you come to is totally without merit and pointless to discuss. Your test proves absolutely nothing without a before and after comparison.
Last edited:
I'm not, I'm questioning the conclusion that it's down to FR. Jon says himself they aren't EQ'd that precisely and implies they don't have to be to sound the same. The question then becomes, how precise does the EQ need to be for the same result?
I have no idea what that comment means. Do you want try to say something without being cutesy about it.
Yes, and I believe it's also been shown that we are not particularly sensitive to FR variations when listening to music? Listening to tones and sweeps could show differences that aren't heard with music, better still, looking at graphs of the FRs measured at the listening position. 😉
Haven't read every post but if you want to test drivers to see if they sound different but you EQ them to make them the same?? Then you have an ahah moment that there is no difference in sound. WTH yeah you made them the same to start with. Wouldn't you just make sure the SPL @ like 1kHZ is set the same and then listen to the drivers not EQ'd. I'm missing something!
I have been thinking why audio ABX test is so hard. I think it is because the brain is comparing the sound in the memory to the sound currently listening. They are not on the same table. Human memory is altered by later experiences, and it is scientifically proved.
Audio ABX is fundamentally different from medical ABX in this regard. I don't think audio ABX is totally nonsense, though.
Audio ABX is fundamentally different from medical ABX in this regard. I don't think audio ABX is totally nonsense, though.
Last edited:
Yes, you're missing the point.
So let's go back to numbers, shall we?
A 10$ driver, once EQd, can sound the same as a 2000$ driver.
Enough said. 😉
As said before, the human ear/brains combo is unprecise.
My tests and many others have found that minimum 0.3db and as much as 1.5db is necessary to spot a difference in SPL.
Also, about frequencies, we are talking between 1/6 and 1/2 octave.
So, basically, EQing 1/12octave with 0.2db precision is useless.
- Status
- Not open for further replies.