World's Best Midranges - Shocking Results & Conclusions.

I actually think what you're suggesting is interesting. I'm simply pointing out that if you want to present your ideas in something approaching a scientifically informed manner then these are considerations you might want to include in your study, and if you can't do that then you might at least consider tempering your pronouncements by noting the omission of these these kinds of qualifying considerations. After all these are precisely the kinds of points to which 'audiophiles' and other interested parties are likely to appeal if challenging the soundness of your claims.

Best,
 
The most important is to start with some SPL target (peak) also in relation with the room properties, and then choose the components (drivers) to achieve a flat 10 octaves frequency response.

95db peak target for a 30m³ room is not the same as a 120db peak target for a 200m³ room.

In the first case, you can probably use a fullrange but not in the second scenario.

Of course, some can argue that a full 10 octaves is not mandatory ''for good sounding speakers'' but if you're looking for anything that can be called ''hifi'' i don't know how it could be anything less than 9 octaves (30hz-15khz)...
 
so I guess its safe to say that the most important thing in a speaker is the crossover and cabinet?
If you designed a speaker for outside use, the considerations would largely be practical. A clean omnidirectional (point source like to side-step diffraction) pattern would be difficult for practical drivers with regard to coil/excursion at higher levels. Narrow makes for a complex cabinet.

At least a room is a starting point. It has to be. It isn't necessarily a good thing, but it can be. It isn't a regular shape in the bigger picture but it is workable. Room reflection control is fundamental to the design goal and the cabinet and configuration is the way to do that.

The cabinet represents the way to hold and to diffract the sound.

A crossover fundamentally arranges this configuration but also encompasses the necessary equalising, as I see it.
 
Well, ''time'' is actually a very important factor here. Please refer to this article: https://en.wikipedia.org/wiki/Short-term_memory

So far, it seems that 10-20 seconds music excerpts gives the best chance for the participants to spot differences.

I will speak mostly based on my hands-on experience in differentiating sounds in an ABX... (You will learn that STM is not the end of the road)

Short-term memory (STM) is indeed needed to detect several types of differences, mainly small (0.1dB) level difference. The shorter the better. Half to one second is needed for very small differences. But STM is not the only one...

Comparing 80dB with 80.1dB requires STM because none of it relates in any way to a fixed information in a memory that is stored as long term memory (LTM).

The logic is like this (I have explained this long time ago):

(1) We listen to DUT1 and store it as STM1, then one second later (STM1 is still in memory) we listen to DUT2 and compare it to STM1. If DUT2=STM1 it means DUT2=DUT1.

(2) We listen to DUT1 and compare it to LTM (a benchmark stored as long term memory), then one day later (DUT1 is not in memory but LTM is still in memory) we listen to DUT2 and compare it to LTM. If DUT1<LTM and DUT2>LTM it means we know how DUT1 differs from DUT2.

Benchmark points (the LTM) can be in the form of (a) enjoyment (b) fatigue/disturbance (c) wrong pitch/rhythm (d) rare tonality, etc.

In an ABX, we can listen for a song and repeating the song 4 times. During this long listening we can just wait till one thing happen, i.e. a "positive experience" or a "negative experience" to register in to our sense (these experiences are LTM). The more experienced you are, the faster you can do this.

I know this is not a bias because I have been able to shorten these processes. This is why I know how weak human ability is, just by comparing what I can do now compared to what I can do 10 years ago (same skill item), or what I can do in 15 minutes compared to 10 seconds... (Of course also by comparing what I can do with others).
 
Who me? I'm nobody to most and to a few I mean the world.
Typical mistakes yield false conclusions of the young amateur.
All things being equal, better parts are not supposed to sound better they are made to work better.
At the end of the day if you can hear Andres Segovia click his teeth as he draws on his pipe as he attacks a chromatic run, all the better. Electricity is like water and through time will work it's way through any barrier it comes across. A brick wall and a pile of sand bags will do the job but fail in the end. One sooner, the other later. Can you tell which one will go first? A ton of lead or a ton of feathers, which is heavier? Get the picture? No? Are you standing too close? Did you pay too much and now you have to justify your decision? Did you go cheap and now you think they all sound the same? Did you buy your way into the club with a gift coupon or did you take your time and truly pay your dues. 99% of you will find this all so droll maybe even bothersome. The other 1% ...........well, you know what I mean. You would be surprised how poorly a $100,000
Sonus Faber can sound while sitting in front of an unhealthy ego.
 
So you're philosophy is eq everything out that the engineer intended you to hear??

It's not a philosophy. But flat FR is what every engineers would want since the first century. It was not possible then, but it is now (well, almost).

Similarity between what the engineers hear (in studio) and what the listeners hear (at home) can only be ensured if the FRs are at least similar. And the only sure way to make the FRs (close to) similar is to flatten them.

Of course, there are other things beside FR, but as you can see from the test, most people cannot hear it easily.
 
The most important is to start with some SPL target (peak) also in relation with the room properties, and then choose the components (drivers) to achieve a flat 10 octaves frequency response.

95db peak target for a 30m³ room is not the same as a 120db peak target for a 200m³ room.

In the first case, you can probably use a fullrange but not in the second scenario.

Of course, some can argue that a full 10 octaves is not mandatory ''for good sounding speakers'' but if you're looking for anything that can be called ''hifi'' i don't know how it could be anything less than 9 octaves (30hz-15khz)...
so, most drivers eq'ed flat sound the same. do you think that a driver that dont need eq is better then a driver that require extra processng to be flat?

what is the way to build the best speaker? DSP/EQ is absolutely necessary for you? I know you can eq for on axis response, but what about off axis?

Do you think that going passive with very simple crossover with flat drivers with bitperfect source (if you use dsp you are not bitperfect= problems) would likely be the best possible way? or do you think dsp/eq is always necessary no matter the driver and then think that the trade off of dsp are worth it?
 
so, most drivers eq'ed flat sound the same. do you think that a driver that dont need eq is better then a driver that require extra processng to be flat?

what is the way to build the best speaker? DSP/EQ is absolutely necessary for you? I know you can eq for on axis response, but what about off axis?

Do you think that going passive with very simple crossover with flat drivers with bitperfect source (if you use dsp you are not bitperfect= problems) would likely be the best possible way? or do you think dsp/eq is always necessary no matter the driver and then think that the trade off of dsp are worth it?

I never stumbled upon any driver that do not require EQ. And i'm talking 1/3 octave resolution...

and then, you have room corrections...

DSP/EQ is the single most important breakthrough for 21st audiophiles. And yet, big names in industry will continue making passive uncorrected speakers and stereo amps/DACs, etc... Because that is where the money is and this market is far too conservative to move the smart way.
 
what is the way to build the best speaker? DSP/EQ is absolutely necessary for you? I know you can eq for on axis response, but what about off axis?

You start with the room. What are the dimensions of the room.

Then, based on room's properties you ask yourself about the target SPL and way to obtain it with enough headroom so you're not at the limits all the time.

Finally, i think you should question yourself about point-of-source V.S. large diffused ''sonic presentation''... That can go from a 4'' fullrange to a full height line array.

... i'm assuming everyone will want a 10-octave bandwith here... And even if that is too much for you, well, you have an EQ to do whatever you wish...
 
I never stumbled upon any driver that do not require EQ. And i'm talking 1/3 octave resolution...

and then, you have room corrections...

DSP/EQ is the single most important breakthrough for 21st audiophiles. And yet, big names in industry will continue making passive uncorrected speakers and stereo amps/DACs, etc... Because that is where the money is and this market is far too conservative to move the smart way.
I personally take passive room treatment before dsp.

in the headphone world, many also consider EQ/dsp is the way to go. I personally have tried those method, and there area trade off.

I really have not enough experience with active system, you may be right!