Do you think that possibly if it was as easy as asking five questions in a survey, none of which address sound, that someone would have already designed the ultimate speaker?
Your survey method of style and size might work as far as marketing and sales are concerned, but have nothing to do with building a better speaker.

www.legacyaudio.com Bill D. has done it already check out his HELIX....and he doesn't sell it at more than 10X what some other's could try to do...they may also be pricey, but the do everything a loudspeaker for hi fi should. And LOOK beautiful, fit and finish it's all there. Withc the science to back it up, not nonsense, .

Bill D. has it all covered, he's done it, and continues to make it teh best there is. http://legacyaudio.com/index.php?option=com_content&task=blogsection&id=9&Itemid=235

Just avoid becoming one of the 3 types of designers I giggle at.

Hi-Fi Fashion designer
Over Engineer
Wheel re-inventor

Innovative crossover design to improve efficiency and performance would be a worthwhile activity.

Quote:
Just avoid becoming one of the 3 types of designers I giggle at.

Hi-Fi Fashion designer
Over Engineer
Wheel re-inventor

what about two other types?
The under engineer
The completely not engineered, but lots of hype

Quote:

Innovative crossover design to improve efficiency and performance would be a worthwhile activity.

I read lots of ad speak about crossover design but have seen nothing technical to back it up (proprietary of course). Asking as a real question and not being critical, why do you think that improving crossover efficiency would be useful? With the availability of amplifier power I would think that crossover losses would be a non issue.

Quote:

I read lots of ad speak about crossover design but have seen nothing technical to back it up (proprietary of course). Asking as a real question and not being critical, why do you think that improving crossover efficiency would be useful? With the availability of amplifier power I would think that crossover losses would be a non issue.

Hi

Firstly the "Underengineers" just get ignored and people tend to vote with their wallets - thus those claiming Watts in "Volts Squared" just get laughed at by us as they tend to be the low-life designers yet to get out of nappies. (Dypers - or how ever Americans spell it) - Look at most Car Audio. Eeeeuuuch!

The next bit is the best question I have had aimed at me since my first post in Stereophile and I would like to back it up with a ton of technical papers from Universities both UK and USA (Cal-Tech and MIT) relating to passive Chebyshev and Butterworth + all the other possible filters.

if the problems were uniform spectrally, power would be the answer as the losses and anomalies would also be uniform

They, however are not -- as the problems are nuanced and different with respect to "f"

power is only one cure -

The need to research in - better materials - better electrolytes - better coil forming - better circuit design - to make better components - to apply those components properly etc etc.

Coils and Capacitors are used in crossover circuits as "Frequency dependent resistors" - the key here is "Frequency dependent" and they all leave their own perculiar fingerprint on the Audio pathway due to tolerances.

The Ideal would be to have a group of "Frequency dependent" pieces of wire - neither capacitive nor inductive - In current componets there is still a long way we can go to make improvements.

"I'm beginning to spel like DUP" -- please have me put down by the first Vetinary

What we need you to design is a speaker that sounds as good as a Wilson Watt Puppy(yes it's a real speaker), Looks like a Dodge Ram to my guy friends, looks like a fine piece of furniture to my wifes friends (as in " That is So cool, You can put it anywhere you want") And, AND, Costs Lots but is always worth more than you paid. "Honey, I'm Soo glad you invested in those speakers, now my mother can go live with her new boyfriend in Florida!

That's all there is to it. Half of the design challenge is setting your goals.

Quote:
Coils and Capacitors are used in crossover circuits as "Frequency dependent resistors" - the key here is "Frequency dependent" and they all leave their own perculiar fingerprint on the Audio pathway due to tolerances.

The Ideal would be to have a group of "Frequency dependent" pieces of wire - neither capacitive nor inductive - In current componets there is still a long way we can go to make improvements.

"I'm beginning to spel like DUP" -- please have me put down by the first Vetinary

Ahh, the straight wire with selective frequency gain but has no phase value.
We do have such an animal, but it exists in the digital domain.

Quote:
Ahh, the straight wire with selective frequency gain but has no phase value.
We do have such an animal, but it exists in the digital domain.

My understanding that even in the digital domain filters still exhibit phase shift, and to minimize phase shift other trade-offs must be realized. Thus, the ability to switch filters on some CD players, etc.

While I deal with the unprocesed analog/digital side of synthetic aperture radar and don't have a detailed understanding of what the processing guys are doing, I do believe that in the digital processing domain most anything can be compensated for. Phase compensation is not difficult.
Of course all you need is computing power and speed.

Hi Cycle

If I had the technology and the brain capacity to design digital filters -- I'd find it fun to have a go and see how close we humans can get to ideal. This means one would shift the crossover into the amplifier and the amplifier would become "Amplifiers".

However, when it comes to digital design and writing code into Proms to run FPGA and DSP - my brain caves in and my eyes glaze over. YaaaaaaaaaaaaaaaaaaaaaaaaWWWWWWWWN !

Quote:
. . . I do believe that in the digital processing domain most anything can be compensated for. Phase compensation is not difficult.
Of course all you need is computing power and speed.

This had been my assumption as well. However I later learned that there are tradeoffs that are inherent in the process - a little like Heisenberg data uncertainty.

For example, phase shift is a big issue in sample rate conversion software as anti-aliasing filters are employed. Those with better phase characteristics have more issue with the passband. Wadia elected better phase and gave up a little high end frequency response as its solution to the issue.

SRC's also have issue with aliasing artifacts. If one converts only a 1 kHz sine wave from 96kHz to 44.1kHz for example, and then look at the audio spectrum of the result you will usually see all sorts of spurious frquencies appearing on either side of the resulting 1kHz 44.1kHz tone. The spectrum looks like JA's jitter tests.

The quality of the conversion varies greatly depending on the software. Thus, I am in full agreement that more sophistication and more processing power makes a tremendous difference.

As a culture we like to assume that digital/math is perfect and that science can already measure everything. If only it were so. We are getting closer however.

I wish I understood enough of the math to appreciate the issues and to explain them. Perhaps someone here does.

Quote:
For example, phase shift is a big issue in sample rate conversion software as anti-aliasing filters are employed. Those with better phase characteristics have more issue with the passband.

First I need to admit that I know just enough about the subject of digital proccessing to toss around some buzz words. So feel free to disagree.
I thought that part of the purpose of SRC was to move the sampling frequency higher so that the low pass filter cut-off frequency would be high enough to minimize its effect in the audio band.

Quote:

As a culture we like to assume that digital/math is perfect and that science can already measure everything. If only it were so. We are getting closer however.

I wish I understood enough of the math to appreciate the issues and to explain them. Perhaps someone here does.

If math was so perfect then it wouldn't need to use imaginary numbers.

Quote:
I thought that part of the purpose of SRC was to move the sampling frequency higher so that the low pass filter cut-off frequency would be high enough to minimize its effect in the audio band.

Yes, some CD players employ upsampling for this purpose.

My example is SRC from a higher sampling rate, such as 96kHz, to 44.1 for release on a CD. This process occurs solely in the digital domain and phase shift is an issue.

Quote:
If math was so perfect then it wouldn't need to use imaginary numbers.

Good one!

Oh, downsampling. While I understand why it's done, I can't deal with the concept of throwing out perfectly good bits.
What a waste.

Quote:
Oh, downsampling. While I understand why it's done, I can't deal with the concept of throwing out perfectly good bits.
What a waste.

Agreed.

But it isn't as useless as burning a 24/96 disk and trying to play it in a CD Player.

If you follow the late John Bowers saying at B&W Loudspeakers Ltd, ""We are not trying to give the most in a Loudspeaker, we are trying to loose the least", you can't go wrong IMHO.

Especially given the quality of his speakers.