KeithR's "Dream Speaker" Search

[tima]

Horns are best for flow and continuity of music (similar to between good vinyl and bad digital), inflections on vocals, micro nuance and micro details, high dynamic range due to ease of drive. The soundstage is large for big horns spaced out well. The macro dynamics are great and tone is best

While small speakers might disappear easier in the average room as compared to horns, in most rooms the non-horn speaker will be oversized and interact too much with the room. Almost no one except for a handful have their speaker room interaction sorted. with horns this is less of an issue so they work better.

I like you both better when you stick to specific models and avoid sweeping generalizations that cause your message to be less credible.

 

[morricab]

Just put Cheever in the search box of WBF - KeithR, 853Guy, Amir , myself, we all explained what KeithR said briefly in post https://www.whatsbestforum.com/threads/stereophile-january-2017-issue.22040/post-439171

His experimental data is not statistically valid. It kills any conclusive value of his work.

So now WBF brain trust is the source for debunking scientific work with what? Opinions? Did any one of you conduct an experiment to invalidate anything in Cheever's thesis? How does KeithR know that the amps used by Cheever were "crappy"? Did he listen to them? They served the purpose of his thesis well enough: one was low distortion with high order harmonics and the other was relatively high distortion but with exponential decay of harmonics. He limits other things to try to make the comparison between the amps more fair. They fit opposite ends of his TAD metric and then the small listening panel decided and not surprisingly came down in favor of the amp with lower TAD. Probably if he had used Shorter's simpler equation he would have gotten a similar result.

I once made a spreadsheet using Shorter's equation and data that I took from Soundstage.com for harmonic distortion. When different powers were used (1 watt mostly but sometimes 10 watts) I corrected the absolute levels based on the THD vs. power curve. What I found is that, once corrected distortion values were used, the SET amps did very well compared to much lower distortion (summed raw harmonic distortion) amps. Cheever's sharper non-linearity would of course put this more in favor of the amps without feedback and/or low order dominated harmonics, but to use Cheever, you need to set an SPL. The problem then is that speakers with all different sensitivities will require different power levels to get the same SPL, so the problem quickly gets very complicated. Sadly, the hard drive died and i lost the spreadsheet...I haven't had the energy to recreate it as it took some time. One could do the same from Stereophile IMD data, which would highlight what I call the "high frequecy" problem. That being a lot of amps that measure with very low THD low levels or harmonics at 50 or even 1Khhz will perform sometimes up to 100 times worse for higher frequecies. Then their IMD plots look quite poor and I think this hurts sound quality quite a lot (so does Nelson Pass as he notes in his white paper on distortion).

Psychology studies are always plagued by numbers of participants and drawing general conclusions from low study power (it is impractical to get hundreds or even thousands of listeners to your test to make it statistically "relevant")...that doesn't make it invalid. It just means that drawing too strong conclusions from it are problematic. I also never use Cheever's study in isolation, it is one piece in a puzzle of evidence that THD is meaningless (Geddes found this as well in two papers he published for the AES) and that seemingly inconsequential distortion may very well be not benign. The weighting is actually exponential and this means the weight falls heavily on high orders. Humans are very bad at understanding exponential behavior. They make the mistake of saying, "But its so small it CAN'T be audible" not understanding the power of the exponential function. Probably Shorter's paper is not much more rigorous than Cheever's but since he did it for the BBC it will have more respect.

The funny thing is that you don't seem to have a problem with Shorter coming up with an equation that shows the perception of harmonic distortion is HIGHLY non-linear and this was echoed by others such as Crowhearst but when Cheever says essentially the same thing, expands on why he thinks it is so (ear distortion, masking, SPL dependence etc.) and then comes up with a more refined equation that has a stronger non-linearity and is SPL dependent and does a small study to show he is on the right track. If you want to claim it is invalid then prove it...that is the scientific way. Appeals to authority and non-expert opinions do not invalidate. Data does.

 

[bonzo75]

I like you both better when you stick to specific models and avoid sweeping generalizations that cause your message to be less credible.

That was only because Ron generalized on only two factors, dynamics and jump factor

 

[tima]

Sure, make a bad example worse to assure disbelief. Once upon a time the Hyperbole Police would write you up. Did he go to the dark side or just debadged?

 

[the sound of Tao]

Sure, make a bad example worse to assure disbelief. Once upon a time the Hyperbole Police would write you up. Did he go to the dark side or just debadged?

I hear the wildly improved and spectacularly unrelentingly best ever new new Hyperbole Bureau now also has just got in a new super duper chief trooper, his most magificent supremeness the vastly superior and most absolute and much loved General I. Zation.

 

[morricab]

I like you both better when you stick to specific models and avoid sweeping generalizations that cause your message to be less credible.

I do think it is interesting though that of all the speakers that Keith has tried, the Duos seemed to tickle his fancy the most.

Perhaps an overall summary from him would be useful to put all his thoughts on all the speakers he heard in his own context... and will he keep his Devores or not??

 

[microstrip]

So now WBF brain trust is the source for debunking scientific work with what? Opinions? (...)

Yes, the fundamental faults were so flagrant that anyone could see them ... The reasons were spread and discussed in many posts. BTW, what I am objecting is just mixing obscure work with serious work of known audio scientists and developers.

For me the interesting subject now is how time seems to have challenged all those studies - they were carried at a period where typical distortions were one or two order higher than current best measuring and/or sounding amplifiers. IMHO we can not generalize on the subjective experience and comparisons carried many decades ago in completely different conditions. Subject for a separate thread?

 

[Al M.]

While small speakers might disappear easier in the average room as compared to horns, in most rooms the non-horn speaker will be oversized and interact too much with the room.

If it's not "most", then it's at least "many", yes.

Almost no one except for a handful have their speaker room interaction sorted.

Possibly true. It takes the right decisions, and often a lot of work.

Even though I have "easy" speakers (monitor/sub combo) it took me years to figure it all out, before now I can reasonably say that I am (mostly -- never discount possible future insights) done with it.

with horns this is less of an issue so they work better.

Interesting argument, you may be onto something.

 

[morricab]

If it's not "most", then it's at least "many", yes.



Possibly true. It takes the right decisions, and often a lot of work.

Even though I have "easy" speakers (monitor/sub combo) it took me years to figure it all out, before now I can reasonably say that I am (mostly -- never discount possible future insights) done with it.



Interesting argument, you may be onto something.

Planars also in many ways have less room interaction, with very little first sidewall reflection due figure 8 dispersion and little ceiling or floor bounce. Seems the normal box speaker has the most room interaction to deal with...

 

[KeithR]

I do think it is interesting though that of all the speakers that Keith has tried, the Duos seemed to tickle his fancy the most.

Perhaps an overall summary from him would be useful to put all his thoughts on all the speakers he heard in his own context... and will he keep his Devores or not??

I just received my amp back today, so will have music for the first time in 3 weeks - should be eye opening. I am going to move into my new room with Devores in hand before making any final decision. On that front, found some intriguing options yesterday. I will post up photos when appropriate!

 

[Al M.]

Planars also in many ways have less room interaction, with very little first sidewall reflection due figure 8 dispersion and little ceiling or floor bounce.

Does that also hold for electrostats?

 

[Folsom]

Cheever was trying to prove why he liked SET's with high sensitivity speakers. All of his thoughts can be accounted for with other things than his conclusions. Also it wasn't the best time for SS amps either. Modern (not stereo-poop amps) amplifiers produce almost nothing past the 3rd harmonic. And besides that transistors are logarithmic devices, so unless you converted the graphs logarithmicly, Brad, they probably aren't remotely accurate.

I just received my amp back today, so will have music for the first time in 3 weeks - should be eye opening. I am going to move into my new room with Devores in hand before making any final decision. On that front, found some intriguing options yesterday. I will post up photos when appropriate!

You keep me glued to this topic waiting for these replies.

 

[bonzo75]

Does that also hold for electrostats?

For planars you need space from the front wall. You also need space for the seat from the planar, ideally.
Some planars the back wave can cancel front and create a bass null in some rooms.
Ideally I like planars to be at the right width to disappear and play as one coherent wide plane with sound starting behind coming in front as of the speaker is not there.

 

[Folsom]

For planars you need space from the front wall. You also need space for the seat from the planar, ideally.
Some planars the back wave can cancel front and create a bass null in some rooms.
Ideally I like planars to be at the right width to disappear and play as one coherent wide plane with sound starting behind coming in front as of the speaker is not there.

You just say it like this, "they're dipoles"

 

[morricab]

Does that also hold for electrostats?

All dipole speakers will have figure 8 pattern but only a line source dipole like a full range ribbon or electrostat will have both features.

 

[morricab]

Cheever was trying to prove why he liked SET's with high sensitivity speakers. All of his thoughts can be accounted for with other things than his conclusions. Also it wasn't the best time for SS amps either. Modern (not stereo-poop amps) amplifiers produce almost nothing past the 3rd harmonic. And besides that transistors are logarithmic devices, so unless you converted the graphs logarithmicly, Brad, they probably aren't remotely accurate.



You keep me glued to this topic waiting for these replies.

Again, many modern examples defy what you claim and there is plenty of data to prove it. Besides, if what you were claiming was true they would all sound VERY similar...and yet funny enough...they don’t .

Actually mosfets follow a quadratic transfer function while bipolars are even less linear with some kind of exponential function. See Boyk and Sussmann for a mathematical modeling of these devices on simple circuits.

As to reading the graphs accurately, I didn’t get a Ph.D in Analytical chemistry by not being able to accurately read and analyze data ...it’s what I do for a living.

 

[Folsom]

Again, many modern examples defy what you claim and there is plenty of data to prove it. Besides, if what you were claiming was true they would all sound VERY similar...and yet funny enough...they don’t .

Actually mosfets follow a quadratic transfer function while bipolars are even less linear with some kind of exponential function. See Boyk and Sussmann for a mathematical modeling of these devices on simple circuits.

As to reading the graphs accurately, I didn’t get a Ph.D in Analytical chemistry by not being able to accurately read and analyze data ...it’s what I do for a living.

I'm not including modern amps that aren't modern at all. I specifically, every time, disclude that stuff. And no, there's a lot of things that make amplifiers sound different. If what you're claiming I'm claiming was true, all SET would sound the same within the same distortion range, too. They don't. There's so much more going on.

See Boyk and Sussmann? No, see Folsom. There's hardly anyone out there that models transistors as well as we do. I'm not going to dive into the math in this thread.

And yes, you have a Ph. D in Analytical chemistry, with a penance for memorizing audio white papers with enough comprehension to play weekend warrior non-designer. You don't actually work with, design, or anything of the sort so you have no idea when the information you've read really applies, was accurate, or anything much besides arguing on here. Also if the accuracy of every published paper was so complete, new people to the industry wouldn't suffer so much and we'd all be enjoying perfection - but we aren't anywhere near that as we continually find out lots of stuff - like Cheever - was optimistic bias.

 

[KeithR]

How did the Cheever "thesis" get on my speaker thread

 

[KeithR]

thanks to Micro's repost of that Cheever debate, rediscovered my amp selection since 2010:

Since 2010:
BAT VK300se
McIntosh MA6600
McIntosh MC601s
Almarro 318B
Mastersound Due Venti
Shindo Haut Brion
Shindo Montille (EL84 version)
DeHavilland 845Gs
Audion Black Shadows
FirstWatt SIT 1 and 2
Sophia Electric 845s
Ayre VX-5
Valvet A3 Class A monos
DarTZeel CTH-8550
Vac Phi 200
Vac Phi Beta (integrated)
Melody AN845
Quad ii Jubilees (yes, the gold plated beauties not the cheap Chinese crap)
Line Magnetic 219 (not my room, but on my speaker at dealer's)
Luxman 590ax
Audio Research Ref75SE

 

[morricab]

thanks to Micro's repost of that Cheever debate, rediscovered my amp selection since 2010:

Since 2010:
BAT VK300se
McIntosh MA6600
McIntosh MC601s
Almarro 318B
Mastersound Due Venti
Shindo Haut Brion
Shindo Montille (EL84 version)
DeHavilland 845Gs
Audion Black Shadows
FirstWatt SIT 1 and 2
Sophia Electric 845s
Ayre VX-5
Valvet A3 Class A monos
DarTZeel CTH-8550
Vac Phi 200
Vac Phi Beta (integrated)
Melody AN845
Quad ii Jubilees (yes, the gold plated beauties not the cheap Chinese crap)
Line Magnetic 219 (not my room, but on my speaker at dealer's)
Luxman 590ax
Audio Research Ref75SE

That's a fair number . Do you really think the ARC Ref 75SE is the best of this lot or just for the Devores? A few of those would sound pretty good on the Duos as long as they are quiet enough.