• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

The quest for my hyper speaker - Very Large room dilemma

Bjorn

Major Contributor
Audio Company
Forum Donor
Joined
Feb 22, 2017
Messages
1,254
Likes
2,482
Location
Norway
Unfortunately, these are all generic specifications that you would get with any manufacturer. And we all understand line source tech. A lot of other speakers also meet my requirements, that's not really the point. The point is to have objective data for all the systems that meet requirements, and to find the best one for my needs. These audio companies (wisdom et al) do not seem to provide anything other than some specs and blurbs. Thus apart from listening (which is notoriously unreliable even in your own room much less a showroom) they make it impossible to evaluate and compare with other speakers. Spins are needed, and failing that, at minimum on and off axis responses, as well as directivity charts for the speakers in question.
As you say, these are simply claims/marketing with no proper measurements. However, it's possible to see how traditional line arrays perform in several of the papers in the links below. Bottom line is that line arrays suffer from polar lobing, comb filtering and poor directivity.

But I want to point out that complete measurements is what's needed, not spinoramas. Parts of spinorama is an interpretation of data and some are highly discussable. To put in another way: Would you rather see the complete results from a medical study where you can draw your own conclusion or only see the interpreted results?
 

Sokel

Master Contributor
Joined
Sep 8, 2021
Messages
5,550
Likes
5,467
As you say, these are simply claims/marketing with no proper measurements. However, it's possible to see how traditional line arrays perform in several of the papers in the links below. Bottom line is that line arrays suffer from polar lobing, comb filtering and poor directivity.

But I want to point out that complete measurements is what's needed, not spinoramas. Parts of spinorama is an interpretation of data and some are highly discussable. To put in another way: Would you rather see the complete results from a medical study where you can draw your own conclusion or only see the interpreted results?
I would prefer a doctor or two interpret them for me,there's people jump from cliffs because they made their own research,specially on internet.
 

Keith_W

Major Contributor
Joined
Jun 26, 2016
Messages
2,349
Likes
5,409
Location
Melbourne, Australia
As you say, these are simply claims/marketing with no proper measurements. However, it's possible to see how traditional line arrays perform in several of the papers in the links below. Bottom line is that line arrays suffer from polar lobing, comb filtering and poor directivity.

But I want to point out that complete measurements is what's needed, not spinoramas. Parts of spinorama is an interpretation of data and some are highly discussable. To put in another way: Would you rather see the complete results from a medical study where you can draw your own conclusion or only see the interpreted results?

You make an excellent point. Measurements are objective, but interpretation of measurements is subjective. When I read medical papers, I place different weighting on the data to the authors, and the weighting I give changes depending on the patient sitting in front of me. This is why I consider the "conclusion" of a paper to be fun reading only, usually contaminated by whatever spin the author is trying to put on the data. What is of real importance is the actual data, provided it has been collected properly. I read the data, I make my own conclusions.
 
OP
A

aliqaz

Member
Joined
Nov 25, 2021
Messages
96
Likes
130
@aliqaz Side note: thank you for expecting manufacturers to be accountable with data at this level of expense, so it doesn't look like just some cynicism from reverse snobs driven by envy for systems they can't afford, which is how many megabuck boutiques tend to handwave and dismiss requests for data.
Absolutely, transparency and accountability are essential at every level for any endeavour, otherwise you enter the realm of Veblen goods where the value of something is unrelated to performance.
 
OP
A

aliqaz

Member
Joined
Nov 25, 2021
Messages
96
Likes
130
As you say, these are simply claims/marketing with no proper measurements. However, it's possible to see how traditional line arrays perform in several of the papers in the links below. Bottom line is that line arrays suffer from polar lobing, comb filtering and poor directivity.

But I want to point out that complete measurements is what's needed, not spinoramas. Parts of spinorama is an interpretation of data and some are highly discussable. To put in another way: Would you rather see the complete results from a medical study where you can draw your own conclusion or only see the interpreted results?
I see what you're saying. From where I am coming from, I simply don't have the expertise to interpret raw audio data. So I have to rely on the spinorama pioneered by Dr. Toole (my "doctor") which is an accepted international standard.
 

Impossible

Active Member
Joined
Aug 2, 2020
Messages
136
Likes
47
You want to make a data driven decision... But the data quality isn't there to give you an answer as accurate as you would like, other then point you in the right direction.

You don't want to do the most accurate test and gather that data, which is listen with your own ears, the thing you are actually going to be doing. Having the least variance and also being the most reliable and trust worthy source. The closest thing to a real world test. The only thing being better would be to put the system in your room and listen.

It's like you go for ice cream and ask for the best one. They go here you can sample them and you say no instead describe them to me in as much detail as possible, only then will I know.
 
OP
A

aliqaz

Member
Joined
Nov 25, 2021
Messages
96
Likes
130
You want to make a data driven decision... But the data quality isn't there to give you an answer as accurate as you would like, other then point you in the right direction.

You don't want to do the most accurate test and gather that data, which is listen with your own ears, the thing you are actually going to be doing. Having the least variance and also being the most reliable and trust worthy source. The closest thing to a real world test. The only thing being better would be to put the system in your room and listen.

It's like you go for ice cream and ask for the best one. They go here you can sample them and you say no instead describe them to me in as much detail as possible, only then will I know.
Silly. Of course the data is there for a large number of products. For those that don't provide, there's no issue, they get to be out of contention.

The most accurate test is absolutely not listening with your own ears, unless the conditions are very strict and direct comparisons are made level matched, absolutely any reasonably impressive system has the potential to sound great in almost every room. In the same token, almost any system has a potential to sound crappy in almost any room depending on circumstances.

This is why we have data, objectivism and psychoacoustics. Since we know that our ears are highly unreliable, you rely on data to help us figure out which system has the highest potential to sound best under a variety of different conditions.

Also in ideal circumstances I would absolutely listen to them in my room. But honestly, logistically, that is not worth my time and effort. Listening in small showrooms is worse than garbage data because the variability, the setup, the room absolutely everything is incredibly confounding to making accurate judgments
 

Impossible

Active Member
Joined
Aug 2, 2020
Messages
136
Likes
47
Listening with your own ears is absolutely the most accurate test as thats the use case.

So your saying you want to buy something that sounds good to your ears.
But your ears can not tell if something sounds good?
 
OP
A

aliqaz

Member
Joined
Nov 25, 2021
Messages
96
Likes
130
Listening with your own ears is absolutely the most accurate test as thats the use case.

So your saying you want to buy something that sounds good to your ears.
But your ears can not tell if something sounds good?
I think you might be on the wrong website.
 

Impossible

Active Member
Joined
Aug 2, 2020
Messages
136
Likes
47
I think there is great value in sound testing and data, mostly for telling if something is broken and what to expect. But ultimately music is for enjoyment and the only test for that is to listen with your own ears.
 

Keith_W

Major Contributor
Joined
Jun 26, 2016
Messages
2,349
Likes
5,409
Location
Melbourne, Australia
Are you planning to use DSP, Ali?

After using a DSP based digital crossover for many years, it makes me realize that a traditional analogue crossover is crude by comparison. Furthermore, any speaker designer no matter how talented has no idea how it will perform in your room. Well designed DSP will perfectly integrate your speaker into your room. I don't think of speakers any more, rather it would be more accurate to think of it as a speaker-room system. Getting a good speaker is important, but the speaker is only half of the story.
 
OP
A

aliqaz

Member
Joined
Nov 25, 2021
Messages
96
Likes
130
Are you planning to use DSP, Ali?
Yes. In fact I am quite interested in the application of the upcoming Dirac active room treatment. Regardless, all my systems use room correction and this one will be no different.
 

Keith_W

Major Contributor
Joined
Jun 26, 2016
Messages
2,349
Likes
5,409
Location
Melbourne, Australia
Sorry, I was editing my post while you replied ;) The problem with Dirac is that it is a closed system, meaning that you use Dirac software to generate the filters, and then you must use Dirac's convolver or supported third party hardware (like MiniDSP, and some Yamaha and Denon receivers) to do the convolution. Dirac is also fully automated which can be a good thing (easier for beginners) but might also be bad (you don't know what assumptions it is making).

I too am quite interested in Dirac's active room treatment, but I would need to learn more about how it works, and most importantly - whether it can be integrated into a non-Dirac system like mine.

Acourate is the choice of doctors :D I use it, and so does Archimago*. The downside of using PC based DSP is ... you need a PC. Not a problem if you play files or stream. But using an external source like HDMI, vinyl, etc. requires additional hardware and inconvenience.

IF you are planning to go the "ultimate" route, meaning individual DSP control of every driver, then you will need to think ahead. Because every driver has its own DAC channel and amplifier, you need a speaker where the crossover can be bypassed. This would be a true hyperspeaker - one where every driver is linearized digitally with measurements, time aligned, phase aligned, overall room correction and target curve applied, etc. and all done in 0.7Hz resolution with 65k taps. I suggest you read this blog for a description of how it all works, and read @dualazmak thread here.

Of course, you don't need to jump all in. You will get a good result if you buy a well sorted speaker and apply some overall DSP to it. But if you wish to think ahead, you should consider a speaker that one day can be upgraded to DSP every individual driver.

* EDIT it just occurred to me that Archimago is also in Canada. Vancouver, I believe. Whistler isn't too far from Vancouver, perhaps you should meet him. I would love to meet him, but I am too far away (Australia).
 
Last edited:

Impossible

Active Member
Joined
Aug 2, 2020
Messages
136
Likes
47
Haha
Sorry, I was editing my post while you replied ;) The problem with Dirac is that it is a closed system, meaning that you use Dirac software to generate the filters, and then you must use Dirac's convolver or supported third party hardware (like MiniDSP, and some Yamaha and Denon receivers) to do the convolution. Dirac is also fully automated which can be a good thing (easier for beginners) but might also be bad (you don't know what assumptions it is making).

I too am quite interested in Dirac's active room treatment, but I would need to learn more about how it works, and most importantly - whether it can be integrated into a non-Dirac system like mine.

Acourate is the choice of doctors :D I use it, and so does Archimago. The downside of using PC based DSP is ... you need a PC. Not a problem if you play files or stream. But using an external source like HDMI, vinyl, etc. requires additional hardware and inconvenience.

IF you are planning to go the "ultimate" route, meaning individual DSP control of every driver, then you will need to think ahead. Because every driver has its own DAC channel and amplifier, you need a speaker where the crossover can be bypassed. This would be a true hyperspeaker - one where every driver is linearized digitally with measurements, time aligned, phase aligned, overall room correction and target curve applied, etc. and all done in 0.7Hz resolution with 65k taps. I suggest you read this blog for a description of how it all works, and read @dualazmak thread here.

Of course, you don't need to jump all in. You will get a good result if you buy a well sorted speaker and apply some overall DSP to it. But if you wish to think ahead, you should consider a speaker that one day can be upgraded to DSP every individual driver
Is this something like a Trinnov?
and is dirac on a storm audio receiver no good, it does active crossovers etc but not pc based like the Trinnov?
 
OP
A

aliqaz

Member
Joined
Nov 25, 2021
Messages
96
Likes
130
Sorry, I was editing my post while you replied ;) The problem with Dirac is that it is a closed system, meaning that you use Dirac software to generate the filters, and then you must use Dirac's convolver or supported third party hardware (like MiniDSP, and some Yamaha and Denon receivers) to do the convolution. Dirac is also fully automated which can be a good thing (easier for beginners) but might also be bad (you don't know what assumptions it is making).

I too am quite interested in Dirac's active room treatment, but I would need to learn more about how it works, and most importantly - whether it can be integrated into a non-Dirac system like mine.

Acourate is the choice of doctors :D I use it, and so does Archimago*. The downside of using PC based DSP is ... you need a PC. Not a problem if you play files or stream. But using an external source like HDMI, vinyl, etc. requires additional hardware and inconvenience.

IF you are planning to go the "ultimate" route, meaning individual DSP control of every driver, then you will need to think ahead. Because every driver has its own DAC channel and amplifier, you need a speaker where the crossover can be bypassed. This would be a true hyperspeaker - one where every driver is linearized digitally with measurements, time aligned, phase aligned, overall room correction and target curve applied, etc. and all done in 0.7Hz resolution with 65k taps. I suggest you read this blog for a description of how it all works, and read @dualazmak thread here.

Of course, you don't need to jump all in. You will get a good result if you buy a well sorted speaker and apply some overall DSP to it. But if you wish to think ahead, you should consider a speaker that one day can be upgraded to DSP every individual driver.

* EDIT it just occurred to me that Archimago is also in Canada. Vancouver, I believe. Whistler isn't too far from Vancouver, perhaps you should meet him. I would love to meet him, but I am too far away (Australia).
I actually have houses both in Whistler as well as Vancouver as I live in both places. Also mitchco lives in Vancouver I believe.

Minidsp allows time and phase alignment if individual drivers as well as the implementation of custom crossovers. This would definitely be on the docket if I were to choose going the DIY route, which may be the case if I can't find something suitable.
 
Last edited:
OP
A

aliqaz

Member
Joined
Nov 25, 2021
Messages
96
Likes
130
Sorry, I was editing my post while you replied ;) The problem with Dirac is that it is a closed system, meaning that you use Dirac software to generate the filters, and then you must use Dirac's convolver or supported third party hardware (like MiniDSP, and some Yamaha and Denon receivers) to do the convolution. Dirac is also fully automated which can be a good thing (easier for beginners) but might also be bad (you don't know what assumptions it is making).

I too am quite interested in Dirac's active room treatment, but I would need to learn more about how it works, and most importantly - whether it can be integrated into a non-Dirac system like mine.

Acourate is the choice of doctors :D I use it, and so does Archimago*. The downside of using PC based DSP is ... you need a PC. Not a problem if you play files or stream. But using an external source like HDMI, vinyl, etc. requires additional hardware and inconvenience.

IF you are planning to go the "ultimate" route, meaning individual DSP control of every driver, then you will need to think ahead. Because every driver has its own DAC channel and amplifier, you need a speaker where the crossover can be bypassed. This would be a true hyperspeaker - one where every driver is linearized digitally with measurements, time aligned, phase aligned, overall room correction and target curve applied, etc. and all done in 0.7Hz resolution with 65k taps. I suggest you read this blog for a description of how it all works, and read @dualazmak thread here.

Of course, you don't need to jump all in. You will get a good result if you buy a well sorted speaker and apply some overall DSP to it. But if you wish to think ahead, you should consider a speaker that one day can be upgraded to DSP every individual driver.

* EDIT it just occurred to me that Archimago is also in Canada. Vancouver, I believe. Whistler isn't too far from Vancouver, perhaps you should meet him. I would love to meet him, but I am too far away (Australia).

Also in regards to Archimago he has a great new article on his blog but in it I was disappointed to see a fellow physician espouse outdated and invalidated ideas (cathexis from discredited psychoanalysis?) to postulate why there are less women in the hobby rather than the preponderance of evidence of 60 years of targeted marketing and gatekeeping:facepalm:.

I had erased this from my memory but going back into his archives I remember the article he wrote near the beginning of the pandemic minimizing impact and decrying lockdowns, and then further comments re hydroxyhloroquine and the minimization of morbidity. I had shared it with my colleagues in the trenches ( public health doctors and ID specialists) and we were all profoundly disappointed at the dangerous psuedoscience that was being peddled by a fellow Vancouver physician. Thankfully opinions like his were in the vast minority: https://archimago.blogspot.com/2020/03/covid-19-few-thoughts-finding.html#more

So my desire to meet him has withered on the vine so to speak. I still respect him as an audiophile objectivist of course.
 
Last edited:

Keith_W

Major Contributor
Joined
Jun 26, 2016
Messages
2,349
Likes
5,409
Location
Melbourne, Australia
The problem with MiniDSP and other hardware based convolvers is that they have too few taps. Mitch points out that they typically have 4096 FIR taps across 4 channels, or 1024 taps per channel. The resolution is sample frequency / taps, so at 48kHz it is 48000/1024 = 46.875Hz resolution. Mitch argues that this is insufficient for bass correction, because you only have two bins below 100Hz at 46.9Hz, and 93.8Hz. Compare this to Acourate or Audiolense, which both can do 128K taps at 48kHz - that's a resolution of 0.37Hz. You do not need a resolution of 0.37Hz for bass, that wastes computing power for no good reason. But you do need more resolution than 1024 taps provides.

MiniDSP's website does not mention how many taps their products offer, but they do say that they use Analog Devices SHARC ADSP21489 processors. Those have 4096 taps. Even if you put all those taps into 1 channel, that's only a resolution of 11Hz.

So, I would say that a MiniDSP is good for time alignment, good for phase alignment, good for target curves in the mids/highs (where you don't want to do fine corrections anyway), but not useful for any correction below the Schroder frequency. Still .. any speaker that can be MiniDSP'ed can also be DSP'ed by other methods. It would be best to find a speaker that does not have any major physical issues out of the box, and never mind if the crossover is a bit wonky because you will fix it anyway :)

I don't have enough knowledge about the speakers on the market to recommend something that fits your needs, and like you I am similarly hobbled by lack of any decent measured data on directivity, distortion, etc. on ANY high end "hyperspeaker". Perhaps you could narrow down a list of speakers you are interested in and write to the manufacturers requesting measurements, or make the dealer really do their work to earn a sale.

I used to go to Whistler every few years until I injured my knee and then got too fat to go skiing. I tried skiing again a few years ago but aggravated my old knee injury again. My orthopaedic surgeon said that I need to lose weight, and even if I lose weight I have to give up skiing, or I could have an operation and continue skiing. I decided to go cycling and swimming instead. So sadly I don't think I will be visiting Whistler again :(

Re: Archi, I did not know he said those things about the pandemic. I was firmly on the vaccinate + lockdown camp until the new mutations appeared and it became evident that it would not be a viable strategy any more. Even during the height of the pandemic my physician group was vigorously debating public health measures, and there were a few dissenters who gave well made arguments. I don't have to agree with someone to respect them, I just want to see that they are making reasonable arguments. Even if I disagree with them.
 
OP
A

aliqaz

Member
Joined
Nov 25, 2021
Messages
96
Likes
130
The problem with MiniDSP and other hardware based convolvers is that they have too few taps. Mitch points out that they typically have 4096 FIR taps across 4 channels, or 1024 taps per channel. The resolution is sample frequency / taps, so at 48kHz it is 48000/1024 = 46.875Hz resolution. Mitch argues that this is insufficient for bass correction, because you only have two bins below 100Hz at 46.9Hz, and 93.8Hz. Compare this to Acourate or Audiolense, which both can do 128K taps at 48kHz - that's a resolution of 0.37Hz. You do not need a resolution of 0.37Hz for bass, that wastes computing power for no good reason. But you do need more resolution than 1024 taps provides.

MiniDSP's website does not mention how many taps their products offer, but they do say that they use Analog Devices SHARC ADSP21489 processors. Those have 4096 taps. Even if you put all those taps into 1 channel, that's only a resolution of 11Hz.

So, I would say that a MiniDSP is good for time alignment, good for phase alignment, good for target curves in the mids/highs (where you don't want to do fine corrections anyway), but not useful for any correction below the Schroder frequency. Still .. any speaker that can be MiniDSP'ed can also be DSP'ed by other methods. It would be best to find a speaker that does not have any major physical issues out of the box, and never mind if the crossover is a bit wonky because you will fix it anyway :)

I don't have enough knowledge about the speakers on the market to recommend something that fits your needs, and like you I am similarly hobbled by lack of any decent measured data on directivity, distortion, etc. on ANY high end "hyperspeaker". Perhaps you could narrow down a list of speakers you are interested in and write to the manufacturers requesting measurements, or make the dealer really do their work to earn a sale.

I used to go to Whistler every few years until I injured my knee and then got too fat to go skiing. I tried skiing again a few years ago but aggravated my old knee injury again. My orthopaedic surgeon said that I need to lose weight, and even if I lose weight I have to give up skiing, or I could have an operation and continue skiing. I decided to go cycling and swimming instead. So sadly I don't think I will be visiting Whistler again :(

Re: Archi, I did not know he said those things about the pandemic. I was firmly on the vaccinate + lockdown camp until the new mutations appeared and it became evident that it would not be a viable strategy any more. Even during the height of the pandemic my physician group was vigorously debating public health measures, and there were a few dissenters who gave well made arguments. I don't have to agree with someone to respect them, I just want to see that they are making reasonable arguments. Even if I disagree with them.
Thank you so much for the info re Mini DSP, I did not know that at all. I use their products in 2 of my systems so perhaps I hay have to rethink it for bass correction. As for the hyperspeaker I am currently in the process of emailing various manufacturers to get the requisite data. If they provide and allow me, I will share on this forum.
It is a real privilage for me to live where I do. I am still in my mid 30's so the knees are holding well enough for 30ish days of snowboarding a year.
Also, I too respect people I disagree with but I tend to draw the line where there is actual societal harm involved. In regards to his views on the pandemic (and his comments on gender in his most recent article) as a physician and a blogger he really ought to know better due to his reach and the weight his words carry.
 

Axo1989

Major Contributor
Joined
Jan 9, 2022
Messages
2,656
Likes
2,670
Location
Sydney
I think there is great value in sound testing and data, mostly for telling if something is broken and what to expect. But ultimately music is for enjoyment and the only test for that is to listen with your own ears.

I would say it is the final test, but self-evidently not the only test.
 

Mr. Widget

Major Contributor
Forum Donor
Joined
Oct 11, 2022
Messages
1,120
Likes
1,611
Location
SF Bay Area
This would definitely be on the docket if I were to choose going the DIY route, which may be the case if I can't find something suitable.
I am still in my mid 30's so the knees are holding well enough for 30ish days of snowboarding a year.
Being that young, I think you should definitely give the DIY route serious consideration. The journey is as much fun as the destination... and besides considering your fairly unusual problem set, i.e. wanting to be able to play back at realistic levels in an unusually large space for a domestic system, looking at DIY makes a lot of sense!
 
Top Bottom