If you don't think there is any difference, or cant hear any, in electronics I'm not sure there is a so much value in your test. Are you saying that you never heard any difference between two power amplifiers?
You have to secure the best possible signal chain to the DUT. Get a second equal channel and set up a pair of good speakers - is what you hear a really good/superb reproduction system (in absolute terms!), then commence test... if not - rethink strategy.
//
1.+4. A 3-4" compression driver is to the liking to me, so I would call it a winner, which I did. I would never just use a regular speaker for midrange. The cone is to heavy and slow and lacking detail.
2. Then why use it in a midrange shootout? It is clearly a tweeter horn.
3. Tell that to the horn enthusiast on this forum.
5. Again, if you haven't sold it yet, try it with a proper horn, unless the criteria is to be able to fit in that little box. Then the test should be called:
"Midrange Driver Shootout for Little Box." 😛
Last edited:
Your whole point is irrelevant, unless you'd prove a bandpass of, say, 1000hz-4000hz would lead to a completely different result (positive identification), which i'm sure you couldnt.
The Dayton woofer, the Radian w/short horn and the ATC are all ALIENS in this test. By that, i mean they shouldnt be there. At all.
Why i did put 'em anyway ? Simply because we're still looking for an audibility threshold.
We can ask Tomahawk to bring his measurement gears.
But then again, what if his EQing is not as good as mine ? Will we identify the drivers because they are different or because the EQ's are uneven ?
In fact, what kind of outcomes we expect ?
A positive identifiation because of the alternate measurement/EQing ?
Based on what ground ?
If the room is that much of a problem then we should go directly to an anechoic room and get it done with it for good.
But if the room is that much of a problem, how do we all, audiophiles, live with that ?
The room (and common acoustic effect) are not PART of the final result we all hear ?
The way that the midrange drivers interact with the room ie the power response will be the biggest percieved difference once all are eq'd flat near field.
But your questions are good ones. Maybe i have missed it but what are you trying to get out of this test? If it is prove that once different drivers are eq'd flat at the listening position with room treatment they all sound very similar / the same within a chosen pass band and SPL limit then I think you have succeded
But your questions are good ones. Maybe i have missed it but what are you trying to get out of this test? If it is prove that once different drivers are eq'd flat at the listening position with room treatment they all sound very similar / the same within a chosen pass band and SPL limit then I think you have succeded
In an ABX where we can use our short-term memory by switching between drivers very very quickly, 0.5-1dB is audible, especially at LF where some instruments will have different "impact".
360 Hz (cut-off) is low enough, either for discernible impact or room effect. The difference you heard might creep in from/during off-axis listening (which is not EQed).
Original goal was to make an appreciation blind test (...World's best midrange)
''which one do you prefer''
But first, we have to make an identification blind test to be sure participants can actually notice differences.
So, that's where we're stucked!
I see you still don't "get" what I'm saying....
Here's a new idea: EQ like you did in previous tests, but move the listening position 30 degrees off axis. Listen there, don't EQ that spot. Listen for differences. The driver that wins that test has the best power response.
Or: find the driver that needs the least EQ and declare that one the winner.
Like it or not, your EQ scheme only works at that single (sweet) spot. You are EQ-ing the room along with the driver, that makes it harder to find/hear the differences. Moving off axis without changing the EQ spot should present those differences.
Again: you don't need the anechoic chamber, we actually like hearing the room in our listening. But once you check out more than one position your way of EQ won't be the best anymore.
For most people on here, the best driver will be the one that behaves best by itself without the need of EQ. For others it will be the one that still has good response outside the pass band for about an octave, on and off axis.
Obviously, this test will not provide all answers for everyone.
Here's a new idea: EQ like you did in previous tests, but move the listening position 30 degrees off axis. Listen there, don't EQ that spot. Listen for differences. The driver that wins that test has the best power response.
Or: find the driver that needs the least EQ and declare that one the winner.
Like it or not, your EQ scheme only works at that single (sweet) spot. You are EQ-ing the room along with the driver, that makes it harder to find/hear the differences. Moving off axis without changing the EQ spot should present those differences.
Again: you don't need the anechoic chamber, we actually like hearing the room in our listening. But once you check out more than one position your way of EQ won't be the best anymore.
For most people on here, the best driver will be the one that behaves best by itself without the need of EQ. For others it will be the one that still has good response outside the pass band for about an octave, on and off axis.
Obviously, this test will not provide all answers for everyone.
I understand your concerns wesayso, and power response is important, but surely listening on axis as he is, should include the power response in it, as he isn't doing this in an anechoic room..
If his eq technique is so flawed that it is not taking into account the room/power response or factors such as driver dispersion/beaming/off axis frequency inconsistency etc then the power response of each driver will be very different.
By your argument because he is doing it in a reflective non-anechoic environment, it should make the drivers easier to tell apart because these variables will be so different from driver to driver.
As I see it, that fact he is only eq'ing on axis response demonstrates how on axis response massively dominates everything else.
His methodology may not be the best for 'best sounding' but actually, should show the greatest differences between the different driver shapes/sizes if power response and room effects are important...
If his eq technique is so flawed that it is not taking into account the room/power response or factors such as driver dispersion/beaming/off axis frequency inconsistency etc then the power response of each driver will be very different.
By your argument because he is doing it in a reflective non-anechoic environment, it should make the drivers easier to tell apart because these variables will be so different from driver to driver.
As I see it, that fact he is only eq'ing on axis response demonstrates how on axis response massively dominates everything else.
His methodology may not be the best for 'best sounding' but actually, should show the greatest differences between the different driver shapes/sizes if power response and room effects are important...
Last edited:
Not if you EQ with pink noise. You basically average all of the sound at the listening spot only. Once you move out of that spot you should be able to detect the differences in power response.
The room certainly isn't bad, the listening position is pretty close but it isn't anechoic. From the pictures posted, even the listening chair close to the mic will influence the EQ.
Yes, he's EQ-ing on axis, but without any gating or window. Just averaging out the response at that spot. So that means direct sound + indirect sound. Where should the differences in power response come from EQ-ing with pink noise? It will only be valid at that spot though.
What I propose isn't a hard test, just place the chair to the side and listen again. That will show the differences between driver/room integration.
Of coarse the direct sound will be dominant, but due to the Haas effect all response after that won't be heard as separate events but added and it will change tonality.
EQ-ing without gating, as in using pink noise, will average the direct plus any added room sound.
EQ-in with a (relatively short) frequency dependent window will still EQ some room at the low end, but it will avoid a lot more of the added "room stamp" in the end result. To avoid any room sound you'd need the anechoic room. At that point there's no influence of power response in what you hear.
Without seeing what the room is doing we can't state anything about the added part, that's why I requested to "see" an IR and preferably a filtered IR as well, showing the levels of reflections. I'd rather have the IR itself to be able to dissect it further as an IR in a measurement suite is just a chosen rendering of what happens over time. With the actual IR, we could dissect what's happening with the sound over time at each frequency.
The standard plots only show part of the story I'm afraid. But it will tell us way more than: "EQ-ed flat at the listening spot using pink noise". That's the level of information we have now.
The room certainly isn't bad, the listening position is pretty close but it isn't anechoic. From the pictures posted, even the listening chair close to the mic will influence the EQ.
Yes, he's EQ-ing on axis, but without any gating or window. Just averaging out the response at that spot. So that means direct sound + indirect sound. Where should the differences in power response come from EQ-ing with pink noise? It will only be valid at that spot though.
What I propose isn't a hard test, just place the chair to the side and listen again. That will show the differences between driver/room integration.
Of coarse the direct sound will be dominant, but due to the Haas effect all response after that won't be heard as separate events but added and it will change tonality.
EQ-ing without gating, as in using pink noise, will average the direct plus any added room sound.
EQ-in with a (relatively short) frequency dependent window will still EQ some room at the low end, but it will avoid a lot more of the added "room stamp" in the end result. To avoid any room sound you'd need the anechoic room. At that point there's no influence of power response in what you hear.
Without seeing what the room is doing we can't state anything about the added part, that's why I requested to "see" an IR and preferably a filtered IR as well, showing the levels of reflections. I'd rather have the IR itself to be able to dissect it further as an IR in a measurement suite is just a chosen rendering of what happens over time. With the actual IR, we could dissect what's happening with the sound over time at each frequency.
The standard plots only show part of the story I'm afraid. But it will tell us way more than: "EQ-ed flat at the listening spot using pink noise". That's the level of information we have now.
Last edited:
So a high performing measurement equipment chain is for you "totally irrelevant" when used to measure something. OK.
//
Ok I get that his technique is averaging room and direct sound, but surely this means some drivers will have much more reflected sound and some much more direct, depending on directivity etc. It is interesting to me that this hasn't made them sound different. No?
I still feel flawed or not, the fact they are not distinguishable is very interesting.
Last edited:
He did state he found someone that could pick them apart. You can actually train that behaviour. There will be differences. It's just not that easy for most people. I've done a lot of timing experiments. I can't always tell immediately. But I do notice I feel differences. Easier to tell by looking at unsuspecting listeners. Do they start to move along to the music? A lot harder to ABX using short fragments though. And much easier with an actual bottom end.
I bet listening to ~370 Hz to 7.2 Hz isn't that comfortable to start with. I've listened that way, only midrange or only a tweeter, it's definitely a hard test, harder than listening to a full spectrum.
For its purpose, I don't think it is flawed. You can say that the opposite is even truer. You want to hear only REAL differences (in quality) between the drivers.
By removing the variables of differences that has less effect on quality, we are more focus on real quality differences, not taste or expectation biases...
Like I said before, distortion and detail/speed, are the only 2 things that I want to listen/look for in a midrange test like this. (When the test is flawed, even when the drivers are 'theoretically' producing the same frequency band, one that has more bass will win).
Well, seems that i'm not the only one who don't get it 🙂
There is no ''winner'' possible because there is no test yet about any winning drivers... At this moment it's an identification test (A+B+X... which one is it?)
Wesayso,
I understand perfectly well your point of view about the EQ. I do get it.
Problem is: i don't think the unavoidable room effect in my measures method would be affecting the end results.
The very fact that i'm able to EQ completey different drivers in a way they are indistinguishable for the high majority of people (if not all)... that alone is something unexpected and totally shocking.
I understand perfectly well your point of view about the EQ. I do get it.
Problem is: i don't think the unavoidable room effect in my measures method would be affecting the end results.
The very fact that i'm able to EQ completey different drivers in a way they are indistinguishable for the high majority of people (if not all)... that alone is something unexpected and totally shocking.
And this isn't even mentioning different driver distortion, materials, transient response, efficiency, etc, etc. i.e. supposedly completely different sounding drivers........
Exactly!!!!!!! 😱😱😱