Jump to content

Review and Discussion of Toole's Book, Third Ed.


Chris A

Recommended Posts

One thing that I should mention:

 

In general, loudspeaker frequency response is largely correctable using EQ--that's why I don't consider it in the same light as the other measures of performance (distortion, directivity, etc). 

 

Only in those cases where the room is quite live, rather small in its internal dimensions, and the loudspeakers unequal in terms of their directivity vs. frequency...does EQ fail to return the perceived frequency response to flat.  Toole talks about this in section 18.2.3 in his first edition.

 

Chris

  • Like 2
Link to comment
Share on other sites

4 hours ago, adam2434 said:

Dr. Toole has been posting on an AVS thread, and I have been thinking about posing the following questions on that thread.  Basically, I’m wondering if there are situations where flat frequency response is trumped by other factors.  I have not read his books, so perhaps these areas are covered in the books.  I should read a book.  :)

 

1) Distortion vs. SPL:

As SPL increases, is there a point at which a lower distortion speaker with less flat frequency response is preferred to a higher distortion speaker with a flatter frequency response?  For example, let’s say that at 80 dB, the flatter frequency is preferred, but at 95 dB, could lower distortion dominate preference vs. frequency response?

 

2) The listening room – dead vs. live:

Can preference swing away from flat frequency response depending on whether one’s room is at the extreme of reflective vs. absorptive for high frequencies?  For example, in a large, relatively dead room, could one prefer a speaker that has more high frequency output vs. a flatter speaker?

 

3) One’s hearing:

If one’s high frequency hearing is diminished due to age or other factors, could one prefer a speaker with more high frequency output vs. a flatter speaker?  Could this factor plus the dead room factor be additive in terms of one preferring a speaker with more high frequency energy?

This is where imo he is no longer as relevant.  On avsforum and gearslutz there are true experts in applied design that discuss at length these questions as they apply to a room or a particular situation.

 

I view him as a pioneer and still relevant for the basics but when i read the advanced design principles being applied in todays recording studios, home theaters and commercial venues.. ..these pros are solving these problems constantly and very successfully.   No dedicated studio or theater would use the minimalist approach he tends to suggest.  

 

Same for what is important....go to gearslutz where they build studios they will send away someonenwho only discusses FR....they spend all their time on decay and discuss endlessly the balance of decay vs response.

Link to comment
Share on other sites

6 hours ago, Chris A said:

As I stated above, I believe that Toole has shied away from discussions on nonlinear distortion (which is proportional, albeit non-linear in loudspeaker output SPL) and frequency response flatness at the listener's ears at higher than 80 dB.  This is the same problem that Geddes ran into when he published his report on the audibility of modulation distortion--he wasn't able to get data for listening conditions higher than 80 dBA at the listener's eardrums (OSHA).  Geddes later acknowledged that there are indications that his statements of the inaudibility of modulation distortion might not hold for SPL greater than 80 dB (a bit sheepishly, I might add). I have reason to believe that Toole listens to his loudspeakers typically at very low levels compared to what most home theater enthusiasts typically do. 

 

And, given that orchestras at peak levels or fff are about 105 dB (according to Toole), or higher  (according to PWK), and an "average" level (meaning a subjective "medium" level, I presume) is 80 dB (Toole), and 85 dB (PWK), it would make sense to conduct distortion testing well above 80 dB.  The complex beauty of a great orchestral climax, or a prolonged, triumphant ending a la Beethoven, Mahler, or Stravinsky must be defended.  PWK did defend them.  I have measured 110 dB during many high SPL passages, that were probably marked fff, but there are markings of ffff to be seen from time to time (the end of the Firebird suite), with Ligeti marking as high as 6 or 8 fs.   At such moments, I want to hear everything, including what PWK was fond of calling "inner voices of the orchestra" without detail being masked by distortion. 

 

  • Like 2
Link to comment
Share on other sites

7 hours ago, Chris A said:

One thing that I should mention:

 

In general, loudspeaker frequency response is largely correctable using EQ--that's why I don't consider it in the same light as the other measures of performance (distortion, directivity, etc). 

 

Only in those cases where the room is quite live, rather small in its internal dimensions, and the loudspeakers unequal in terms of their directivity vs. frequency...does EQ fail to return the perceived frequency response to flat.  Toole talks about this in section 18.2.3 in his first edition.

 

Chris

Indirectly, this is another vote for Danley SH-50's and other synergy horns, where the sound is not allowed to splash on the walls, but if they do (much later in the time domain), all frequencies are represented equally above 300 Hz. or so.

Edited by ClaudeJ1
INcomplete
  • Like 2
Link to comment
Share on other sites

48 minutes ago, garyrc said:

and 85 dB (PWK), it would make sense to conduct distortion testing well above 80 dB.  The complex beauty of a great orchestral climax, or a prolonged, triumphant ending a la Beethoven, Mahler, or Stravinsky must be defended.  PWK did defend them.  I have measured 110 dB during many high SPL passages, that were probably marked fff, but there are markings of ffff to be seen from time to time (the end of the Firebird suite), with Ligeti marking as high as 6 or 8 fs.   At such moments, I want to hear everything, including what PWK was fond of calling "inner voices of the orchestra" without detail being masked by distortion. 

Indeed. There are more of us than you may think!!

  • Like 1
Link to comment
Share on other sites

  • 1 month later...

I've finished reading the 3rd Edition book.  As I had guessed beforehand, it's chocked full of subtleties of the author's statements on loudspeakers, room acoustics, and psychoacoustics, and is multi-faceted, sometimes quite subtle, and filled with many implied assumptions about consumer needs/wants, basic design choices/trades, and how people listen to music and movies. 

 

So in order to give the book and the author the benefit of the doubt for these implied assumptions, I will need to continue the chapter-by-chapter reviews that I began above, because if any of the author's implied assumptions turn out not to be true, the answers that the author provides change, and not by a little bit, it seems. However, I do think that an overall, 50,000-foot viewpoint is needed at this point...

 

An Overall Review of Toole's Third Edition of Sound Reproduction

 

First and foremost, Toole's books spend most of their time on how people hear...the psychoacoustics of human hearing...and almost zero time on "how to make a better loudspeaker or room".  One great strength of the books (1st-3rd editions) is in their weaving of their extensive technical references on how people hear into a narrative that dominates the informed consumer viewpoints on "how to achieve better hi-fi reproduction based on how people hear"--but not so much "how to build a better loudspeaker".   When the book departs from that strength area (i.e., summarizing relevant technical reports), I believe that it gets much less entertaining and educational, and I find myself disagreeing with the author, sometimes quite sharply.

 

I'll attempt to identify those areas that I thought were most important, those that were basically unchanged from the 1st/2nd editions (not very much original material remains in the 3rd ed.), those areas that I disagree with, and why I disagree...

 

Chris

  • Like 1
Link to comment
Share on other sites

By the way--a very important comment:  Toole's books look more like a collection of his white papers weaved together from over the years while still at Harman International.  While the 3rd edition definitely improves the book's organization and train of thought, honestly it has a long way to go before being what I'd call "well organized", like for instance Philip Newell's book of Recording Studio Design (2nd ed), or even Beranek's or Olson's textbooks on Acoustics/Acoustical Engineering--albeit these last two texts are much older but still relevant due to the material covered there--and nowhere else. 

 

Chris

Link to comment
Share on other sites

  • 1 month later...

When I locate my book, I'll continue with my chapter-by-chapter comments.  I believe that I left my copy at my physician's office (and I only go there once a year on average...).  It may be a while, in other words.

Link to comment
Share on other sites

  • 1 year later...

Well, I never found my original copy of Toole's 3rd Ed., so my wife gave me another copy as a Christmas gift.  Still it's a loss since I wrote many comments/glosses in the margins of that prior copy as I read along the first time.  I'll try to re-create my original impressions, but the fact remains that over a year has passed and these "original impressions" now are actually "second impressions".  C'est la vie!

 

This edition of the book has 18 chapters--all of which are relevant to the subject at hand (i.e., loudspeakers in home-sized listening rooms), so this will likely be an interesting (and extended) journey to discuss the most prominent aspects of this important book for home audio enthusiasts. I'll start where I left off...but note that I'll be hitting only the high points.  If you want to know more in-depth views about each topic discussed here, you certainly know where to get a copy.

 

Loudspeakers in personal listening rooms seems like it should be a fairly closed and constrained subject, but it's not--in my experience.  In fact, it still amazes me how much information exists, but how much further we can take that information to learn even more that's relevant to better home hi-fi.  I actually believe that we are presently living in a "golden age" of hi-fi audio reproduction.  YMMV.

_________________________________________________________________________________________________________________

 

Chapter 4 of Toole's 3rd edition book is billed as a "tutorial" (as I stated above)--but I think it's a lot more than that, for it sets a real foundation for what is to come.  It primarily discusses how we hear (psychoacoustics), not how loudspeakers and listening rooms work. 

 

Why is psychoacoustics important to loudspeakers and listening rooms?

 

For me, the answer to that question is that all the engineering trades of loudspeakers (and the acoustics of listening rooms) depend on the way that we hear, not some abstract notion of "what is linear and faithful to the original source recording".  

 

I could repeat that again, probably several times, but I doubt that those folks that lie on either side of the subjective/objective audiophile argument will take it to heart.  But it's the essence of the source of those arguments: how we hear.  Many people would like to think that psychoacoustics is unquantifiable (sort of like typical romantic notions of the "boundless human spirit")...and others that want to believe that what is important is basically already all known and easy to quantify.  It's neither, in my experience.  We actually know a lot about that psychoacoustics...not all, but a great deal in order to make engineering trades that results in excellent loudspeakers...and listening rooms.  It used to be a black art...and nowadays, not really.  It's that last 5% or less that is difficult to convey, (but is probably still largely understood) within existing engineering state of the art.

 

A couple of observations that I would like to add before we get further into Toole's writings:

 

1) There seems to be an issue with "accommodation" with some listeners in that they seem to prefer the distortions typical of some direct radiating loudspeakers over actual live/real performance.  I also can't tell you why those people that seem to hold these views aren't immediately apologetic and defer to real (acoustic) performance as the only standard by which we can mutually judge and agree on what's hi-fi and what's not.  [This is a source of disappointment and, frankly, perhaps a bit of intemperance in my observations.] 

 

2) A second observation is that some seem to think that psychoacoustics and acoustical engineering knowledge of 50-75 years ago can be somehow become "old/outmoded".  I.e., that there are better experts today whose thoughts today somehow eclipse those original writers, and that all of what the physics and engineering giants of the past wrote and posted data on--is now "not worth the time to read, understand, and discuss" as if it is not just as important today as the day when it was first written. I will be making observations of Gunness, Geddes, et al. in the same breath as discussing Beranek and Olson writings.

 

Chris

  • Like 1
Link to comment
Share on other sites

31 minutes ago, Chris A said:

some seem to think that psychoacoustics and acoustical engineering knowledge of 50-75 years ago can be somehow become "old/outmoded".  I.e., that there are better experts today whose thoughts today somehow eclipse those original writers, and that all of what the physics and engineering giants of the past wrote and posted data on--is now "not worth the time to read, understand, and discuss" as if it is not just as important today as the day when it was first written

"My theories on audio and audio reproduction  will change when the laws of physics change" -PWK

"Come on guys it's all done with ball bearings now"- Fletch

:)

  • Like 1
  • Haha 1
Link to comment
Share on other sites

For anyone interested, there is a thread over at AVS where the Doctor is actively sharing observations and responding to questions including one of mine.  I asked him if he has listened to point source speakers designs such as the Danley SH50's. he said, he never did.    The topic at the time was whether a speaker design can express a bigger presentation or sense of scale.  

 

https://www.avsforum.com/forum/89-speakers/3038828-how-choose-loudspeaker-what-science-shows.html

 

Link to comment
Share on other sites

Continuing on with the book chapter-by-chapter review, Toole, 3rd ed., after a long hiatus (and re-discovering this thread of mine)...

 

We're now at Chapter 5: "Characterizing Loudspeakers - Can We Describe What is Good?"

 

In many ways, this is the beginning of "where the rubber meets the road" for Toole.  He is plowing new ground for a lot of readers, and the discussion on his "four schools of thought of loudspeaker performance" over time:

 

1) the flat anechoic on-axis frequency response crowd

2) The on- and off-axis anechoic response (weighted assessment of on- vs. off-axis) crowd

3) The total radiated power crowd

4) The in-room, listening position , steady-state SPL response crowd (i.e., room and loudspeaker together)

 

I wasn't aware of these "schools of thought" in the past, and I still don't believe that it accurately reflects the different factions that I've witnessed over time (especially since 1986 when Toole references his own JAES article).  I would say that the schools of thought that I've been acquainted with include mostly from group #1 (flat on-axis SPL) to the exclusion of almost all other factors. 

 

Of the folks that take up with the "Klipsch sound" (i.e., fully horn-loaded loudspeakers), I would say that none of the groups listed really encapsulate the loudspeaker preferences.  I believe this should be noted that Toole could have formulated four basic groups that do not actually capture the most important elements of loudspeaker design, as evidenced by the Klipsch (PWK in particular) experience. 

 

To summarize PWK's points, the first requirement is low distortion...not harmonic distortion but rather modulation distortion (AM and FM types) that increase the "mud factor" of the loudspeakers.  In general, direct radiating loudspeakers have lots of modulation distortion.  Having multiple loudspeaker "ways" can cut down on the severity of the mud factor, but can't minimize the effects of modulation distortion. One of the ways to achieve low modulation distortion is to design a very highly efficient loudspeaker, greater than 95 dB/m at 2.73v broadband input.  This is the so-called "Klipsch's Law": highly efficient loudspeakers = low distortion.

 

Secondary to low modulation distortion, at least good horizontal dispersion is probably next on the list, with consistent polars all the way up and down the audible spectrum, at least until you reach down to the Schroeder frequency of your listening room (generally 100-200 Hz).  I think Roy has extended the horizontal dispersion requirement also to the vertical axis, too, as this is a major point in the K-402 and K-510 horns over predecessor horns used by the Heritage family of loudspeakers (Khorn, La Scala/Belle, Cornwall, Heresy). This implies a certain minimum size to the mouth dimensions of the horns used.

 

It's interesting to note that PWK himself stated that SPL response (frequency response) flatness was "almost omitted from the list". This is like a polar opposite to what we're seeing with Toole's loudspeaker design requirements. 

____________________________________________________________

 

So what gives?  Well, one is an observation based on the ease by which loudspeakers can be EQed to flat response, especially nowadays.  Similarly, phase/group delay response is equally amenable to correction if the passive crossover is replaced by a DSP crossover.  It's even easier if "overlay" FIR correction filters are applied to the loudspeaker upstream of the DSP crossover, using such as the better performing "room correction software" packages as JRiver, Dirac, rePhase, etc.

 

Apparently, we really can't agree on describing what is good about loudspeakers and what is not... 💥

 

Chris

Link to comment
Share on other sites

The discussion in the above post ^^^ explains why the "reviewers" on other forums simply miss the point with Klipsch loudspeakers, and why so many people prefer them to "correctly designed and implemented" loudspeakers (in their opinion), usually little monkey-coffin boxes on sticks that measure well, sound okay at low SPL, but simply cannot reproduce the music like fully horn-loaded loudspeakers (i.e., including horn-loaded bass bins). 

 

Why is it that many people still prefer the sound of the "badly designed" Klipsch loudspeakers, even after they've been shown just how bad they measure?  Because these "reviewers" are simply not measuring the things that matter to the people that prefer the Klipsch sound, and are "sticking to their story" of what's important to measure.  They have to, because if they questioned what they were measuring, then they would have to question all their assumptions and the factors that people other than PWK and all the other Klipsch loudspeaker design engineers already know: 

 

these reviewers and the pundits that write books and propose to be experts in how loudspeakers should be designed and implemented simply don't really understand how to make loudspeakers that customers actually enjoy, instead they rely on "drinking their own bathwater", even going to the extreme measure of training and selecting people to listen for them that exhibit the traits that match what they think is the "right" things to listen for.  People don't need to be selected and trained to know what sounds better--and it's clear that many people do not follow what these others believe is "the right answer".

 

So when we read the "reviews" from this school of loudspeaker design/implementation, they will certainly not want to include those that value the sound of loudspeakers having much different capability traits than they proscribe.  A case in point: the Heresy IV review on the forum professing to be using "science".  The reviewer completely panned the Heresy IV's performance and encouraged the readers to join in bashing the design.  What's clear to me is that the review in question simply doesn't know how measurements correlate to different listeners' preferences, and doesn't know when to use which measurement (no matter how well the measurements are performed) because they are basically not measuring the factors that make a difference to the people that buy and listen to Klipsch designs, or understanding how to read the measurements they have taken, in terms of their precedence of importance. What does he do to justify what he's done?  He denigrates the listeners that prefer the Klipsch sound over the "correct" sound. 

 

What has he done that's actually useful?  He's given us some data, but he hasn't published all that he's measured (I had to cajole him into publishing the full transfer function plots, and when he did, he didn't understand why I asked for the data).  In addition, he apparently is mainly motivated to make money in order to pay for the measurement tools that he's bought (and wants to buy) to continue making the same kind of "reviews" that he has been doing.  I don't recommend encouraging him, because I don't think that he's done his homework yet.

 

Chris

Link to comment
Share on other sites

Now that I've shot holes through the idea that "one size fits all" in loudspeaker requirements...

 

51nnVSc+8HL.jpg

 

...or more importantly, the precedence of requirements...

 

2076287492_RequirementsPrioritization.jpg.dee713efac84314773bf05d78235fb7a.jpg

 

I'll comment on the Toole's identified performance characteristics related to "transfer function", namely SPL and phase response, below:

___________________________________________________________________

Curiously, most "hi-fi loudspeaker reviewers" fail to recognize that DSP and other SPL and phase correction packages widely abound in today's hi-fi marketplace (and only anti-digital mossbacks choose not to use them; everyone else that actually hears good transfer correction packages never go back to the "bad old days"). 

 

The other factors (modulation distortion, compression distortion, polar coverage/power) cannot be corrected via DSP.  If the loudspeaker has defects in these performance areas, then you have to physically alter the design of the loudspeakers to correct them.  These factors should be the focus of the reviewer's time and testing--in light of today's hi-fi marketplace (and this is what Klipsch itself does...).  Add-on DSP correction packages for loudspeakers/rooms should be marketed for each loudspeaker that doesn't already come with DSP and their own amplifiers (i.e., powered loudspeakers)...

 

https://audiophilestyle.com/uploads/monthly_2020_09/412219948_Top-LevelLoudspeakerCapabilitiesHierarchy.GIF.332775aad1110fad926bf52c616bfdb8.GIF

 

A Hierarchy of Loudspeaker Performance Factors (all of which are non-correctable except SPL and phase response)

 

If these add-on DSP transfer function packages are not offered for a loudspeaker by its manufacturer, these are simply business opportunities for third party providers.  (It's not like hi-fi enthusiasts don't know how to add on extra components or to trade among their upstream electronics.)  There are already many of these--some performing better than others: Audyssey, ARC, Dirac Live, Accourate, Sonarworks, Waves TRACT, Neumann Control, and even DIY applications with low-cost calibrated microphones.  DSP crossovers paired with a calibrated measurement microphone and an in-room acoustic measurement app (REW being the best example and the software is free and quite powerful, but there are others) offer even more powerful and controllable application of transfer function correction capabilities to today's hi-fi enthusiast.

 

images?q=tbn:ANd9GcQbwcEEC3z73fHRqWQE6SM        image.png.5f305172a379242064ca28b02b193c7a.png , etc.

 

This is what almost all reviewers have missed, and it is such a simple idea to think in terms of "correctable transfer function performance" that it leads me to believe that these reviewers are either stuck in a rut of thinking about "flawed transfer function" or just "uneven SPL response" (as almost all loudspeaker reviewers concentrate on to the exclusion of the phase response).  Almost 75% of what Toole/Olive have concentrated on (summarized in the pie chart below) is correctable, but the portions that aren't correctable aren't even complete in their requirements hierarchy.  Where's modulation, compression, and harmonic distortion?  Everything in Olive's requirements hierarchy (the pie chart below consolidates his factors) is typically fully correctable, except smoothness of sound power and low frequency extension, and the biggest factors (low distortion at higher SPLs) is totally missing:

 

https://audiophilestyle.com/uploads/monthly_2020_09/SeanOlivePreferencesofLoudspeakerPerfFactorsPieChart.thumb.gif.affc1ec1de1084a147e45641b88928a7.gif

 

Chris

  • Like 1
Link to comment
Share on other sites

So to continue with the discussion of the factors found in Toole's book (3rd ED, chapter 5 - Characterizing Loudspeakers), most of what he focuses on are correctable--one time corrections that are even automated in some instances with some of the packages identified above.  Toole showed a figure in chapter 5 (figure 5.2) showing the phase growth response of passive crossover loudspeakers from the mid-1980s.  I was stunned to see these plots (the black lines--more on the red lines in a moment):

 

505356491_TooleLoudspeakerPreferences-PhaseResponsegroups7_0to7.9.GIF.a1c4fac3f58722ba46a6cadb857d4764.GIF

 

These are terrible.   The average phase growth of a typical passive-crossover loudspeaker is something between one and a half total wavelengths (200-->20k Hz 540 degrees ) and three wavelengths (1080 degrees).  He doesn't show the phase growth below 200 Hz, which is where the really large phase growth actually occurs due to the drivers themselves (i.e., minimum phase...and is compounded by the phase shifts found in the recorded music chain--which is typically never published). 

 

No wonder why our reproduced music has historically never sounded very much like the real thing. 

 

Now for the red traces: if you guessed that these are the phase responses that I measured in my listening room on my Jubilees (TAD TD-4002s), you can pat yourself on the back, because that is exactly what they are.  Do you think that there is a difference in sound quality between those passive crossover loudspeakers that Toole measured and the Klipsch Jubilees in my listening room?  (A rhetorical question.) There is a massive difference in sound quality--and almost no one is paying attention to this performance factor--which represents one half of the transfer function of the loudspeaker.  Additionally, the directivities vs. frequency of those passive crossover loudspeakers are nothing like those of the Jubilee (except around the crossover region, which I mitigate by shortening down the K-402s and use zero phase growth crossover filters):

 

175008980_KPT-Jubilee_535-BHorMap.thumb.png.ea0d2d117960566446942bcb70bcee95.png

 

Here's the directivity map of the Danley SH-96, which is almost a perfect twin for the K-402-MEH in terms of horn directivity vs. frequency:

 

817087489_SH96Hormap.thumb.png.08ba08ba56d5225eac9a3e874024db90.png

 

Chris

  • Like 1
Link to comment
Share on other sites

On 12/10/2020 at 3:10 PM, Chris A said:

Continuing on with the book chapter-by-chapter review, Toole, 3rd ed., after a long hiatus (and re-discovering this thread of mine)...

 

We're now at Chapter 5: "Characterizing Loudspeakers - Can We Describe What is Good?"

 

It's interesting to note that PWK himself stated that SPL response (frequency response) flatness was "almost omitted from the list". This is like a polar opposites to what we're seeing with Toole's loudspeaker design requirements. 

____________________________________________________________

 

So what gives?  Well, one is an observation based on the ease by which loudspeakers can be EQed to flat response, especially nowadays.  Similarly, phase/group delay response is equally amenable to correction if the passive crossover is replaced by a DSP crossover.  It's even easier if "overlay" FIR correction filters are applied to the loudspeaker upstream of the DSP crossover, using such as the better performing "room correction software" packages as JRiver, Dirac, rePhase, etc.

 

Apparently, we really can't agree on describing what is good about loudspeakers and what is not... 💥

 

Chris

I think I'd be easily diverted from reading literature that doesn't directly talk about horns, since they make up the entirety of my preference, and the majority of my experience. I'd just end up back here reading resurrected book review threads... 😉 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...