3rd edition of Floyd E. Toole's "Sound Reproduction"

[cooky1257]

Some background, it's a bit of PR but lots of interesting info;

https://www.mixonline.com/technolog...oped-m2-and-re-ignited-its-studio-line-383733

 

[tuga]

The fewer drivers and crossover points the better IMO. I have certainly come to the conclusion I am a fan of point-sources or very close to them. The less distance between the fewest number of drivers the better anyway.

If I was designing a speaker myself I’d want to use the highest quality widest ranging full-range driver I could find and cross it as low and as high as possible to a sub and supertweeter as I could. The closest I have to this are the ancient La Scalas which have many flaws, but do allow the range from 400 Hz to 6kHz to be handled by a single very high efficiency and low distortion horn unit. This makes for a remarkably clean and articulate vocal range, it is so easy to follow dialogue through these speakers (I use them in the TV rig). I’d love to hear a really modern 3-way with the crossover points even wider apart giving an even more uninterrupted mid-band as I’m convinced this is the best way fr a multi-driver speaker to function. I’d like to hear it properly time-aligned too (La Scalas are just shocking in this respect, there is about a foot and a half between the three drivers front to back!).

With modern drivers it may well be able to cover all that matters with a large bass unit and a single horn driver, i.e. a modern take on an Altec VOTT, Model 19 or whatever. I’d like to hear that.

PS I don’t really care about extension below about 35Hz as long as the driver is big and can move air like real instruments. Efficiency, transient ability and scale is way more important than extension.

As you can see I agree with everything you said:

I think that for acoustic instruments it would be best to have a single driver cover the lower- and core-midrange (~150-250Hz to ~2KHz).

Even with a steep filter you are mixing the sound from two different transducers. This disrupts the continuity of tone and timbre of instruments producing continuous or legato (as opposed to staccato) sound like a bowed cello or a violin, or human voice.

The audibility of a well executed transition may be questioned but the theoretical advantage of not having two drivers cross at a crucial frequency is in my view undeniable. A single wide-band driver would be the best option if it weren't for the innumerous shortcomings.

3-ways is fine for quasi-fullrange and one can easily add a pair of subs for the lower octave.

 

[tuga]

Some background, it's a bit of PR but lots of interesting info;

https://www.mixonline.com/technolog...oped-m2-and-re-ignited-its-studio-line-383733

Really?
Has a lot of information about the company, which is of little interest to all but the JBL fan, but very little about the speaker design.
This sentence sums it up nicely:

The end result is a free-standing M2 monitor that is 49.5 inches tall, 20 inches wide, and a mere 14 inches deep, with an in-room frequency response of 20 Hz to 40 kHz and 129 dB maximum SPL.

 

[cooky1257]

Really?
Has a lot of information about the company, which is of little interest to all but the JBL fan, but very little about the speaker design.
This sentence sums it up nicely:

The end result is a free-standing M2 monitor that is 49.5 inches tall, 20 inches wide, and a mere 14 inches deep, with an in-room frequency response of 20 Hz to 40 kHz and 129 dB maximum SPL.

The link was posted in the spirit of shared info, enlightenment, for the curious and was qualified with 'PR'. It goes into the need for a new waveguide to match the 2216 dispersion, the novel design ideas in the D2 and quite a bit more-it gives context to where Toole sits in the scheme of things and how it was the quality of the concert line array transducers that prompted the monitor application-its just background but am getting a bit pissed off now. Anyone who reads the piece will see this is trolling.

 

[DANOFDANGER]

If a driver is free from distortion and can support a wide frequency response, then its likely ideal because of the prevention of crossover distortion. However i think its because of these physical limitations that we address the problems with higher order systems. I have heard and owned a few two way systems and have usually been dissapointed. Dynaudio were one of the few that could provide a good system and it was likely because of their motor design.

How audible is the level of IMD introduced with lower order systems ? I would think frequency response sweep tests wont reveal true distortion levels as the distortions develop with complex wide band signals. I think the discussion should be focused on the audibility of these distortions, not which has more. As far as i can see, theres no question which has more. Lower order systems have to compromise more to reduce these distortions, higher order systems less so.

 

[John Phillips]

If a driver is free from distortion and can support a wide frequency response, then its likely ideal because of the prevention of crossover distortion.

Indeed. But just to clarify, wide on-axis frequency response is necessary but not sufficient to prevent audible frequency response problems at crossover points.

Off-axis, driver dispersion normally narrows as frequency goes up. As audio wavelength goes down and becomes comparable to the driver's size the driver "beams" at the top end of its range. So typically a large woofer will see a droop in it's off-axis frequency response at a woofer/tweeter crossover. Conversely a small tweeter will have a wide dispersion at the bottom of its frequency range. So off-axis there is a sudden discontinuity in frequency response.

In real rooms you hear both direct (on-axis) sound and reflected (off-axis) sound. So a good loudspeaker has to match the off-axis dispersions (the directivities) of its drivers at the crossover frequencies. That's one of Floyd Toole's principles for good loudspeaker design. It's easier to achieve this in a three-way design than in a two-way design. The steps between driver sizes are smaller and off-axis discontinuities are easier to correct. However, if you look at the M2 graphs you will see a good directivity index curve with no serious discontinuities. This will help the loudspeaker to sound good in real rooms and will have been achieved via the rather interesting tweeter waveguide.

 

[DANOFDANGER]

Indeed. But just to clarify, wide on-axis frequency response is necessary but not sufficient to prevent audible frequency response problems at crossover points.

Off-axis, driver dispersion normally narrows as frequency goes up. As audio wavelength goes down and becomes comparable to the driver's size the driver "beams" at the top end of its range. So typically a large woofer will see a droop in it's off-axis frequency response at a woofer/tweeter crossover. Conversely a small tweeter will have a wide dispersion at the bottom of its frequency range. So off-axis there is a sudden discontinuity in frequency response.

In real rooms you hear both direct (on-axis) sound and reflected (off-axis) sound. So a good loudspeaker has to match the off-axis dispersions (the directivities) of its drivers at the crossover frequencies. That's one of Floyd Toole's principles for good loudspeaker design. It's easier to achieve this in a three-way design than in a two-way design. The steps between driver sizes are smaller and off-axis discontinuities are easier to correct. However, if you look at the M2 graphs you will see a good directivity index curve with no serious discontinuities. This will help the loudspeaker to sound good in real rooms and will have been achieved via the rather interesting tweeter waveguide.

Yes thats a great point, i hadnt even thought of that. Thats another aspect that lower order crossovers have to compromise with. But the M2 in question does have as you say a unique throat. Lower order systems can definitely compensate to get close to higher order, but whether the benefits of less crossover distortions outweigh the compromises of lower order crossovers is hard to say. Interesting none the less.

 

[tuga]

Indeed. But just to clarify, wide on-axis frequency response is necessary but not sufficient to prevent audible frequency response problems at crossover points.

Off-axis, driver dispersion normally narrows as frequency goes up. As audio wavelength goes down and becomes comparable to the driver's size the driver "beams" at the top end of its range. So typically a large woofer will see a droop in it's off-axis frequency response at a woofer/tweeter crossover. Conversely a small tweeter will have a wide dispersion at the bottom of its frequency range. So off-axis there is a sudden discontinuity in frequency response.

In real rooms you hear both direct (on-axis) sound and reflected (off-axis) sound. So a good loudspeaker has to match the off-axis dispersions (the directivities) of its drivers at the crossover frequencies. That's one of Floyd Toole's principles for good loudspeaker design. It's easier to achieve this in a three-way design than in a two-way design. The steps between driver sizes are smaller and off-axis discontinuities are easier to correct. However, if you look at the M2 graphs you will see a good directivity index curve with no serious discontinuities. This will help the loudspeaker to sound good in real rooms and will have been achieved via the rather interesting tweeter waveguide.

I agree that the power response of the M2 is stellar.

And I can easily imagine that the drivers are healthy performers.

All that I am questioning is the choice of topology and of the crossover frequency.

And if I were being picky I would also point out the use of a ported cabinet.

These design choices, the stress on high SPL abilities jumping at you in ever piece of literature and the lack of measurements makes me wonder which prospective buyer JBL has in mind: pop music and movies studios?

 

[cooky1257]

Indeed. But just to clarify, wide on-axis frequency response is necessary but not sufficient to prevent audible frequency response problems at crossover points.

Off-axis, driver dispersion normally narrows as frequency goes up. As audio wavelength goes down and becomes comparable to the driver's size the driver "beams" at the top end of its range. So typically a large woofer will see a droop in it's off-axis frequency response at a woofer/tweeter crossover. Conversely a small tweeter will have a wide dispersion at the bottom of its frequency range. So off-axis there is a sudden discontinuity in frequency response.

In real rooms you hear both direct (on-axis) sound and reflected (off-axis) sound. So a good loudspeaker has to match the off-axis dispersions (the directivities) of its drivers at the crossover frequencies. That's one of Floyd Toole's principles for good loudspeaker design. It's easier to achieve this in a three-way design than in a two-way design. The steps between driver sizes are smaller and off-axis discontinuities are easier to correct. However, if you look at the M2 graphs you will see a good directivity index curve with no serious discontinuities. This will help the loudspeaker to sound good in real rooms and will have been achieved via the rather interesting tweeter waveguide.

Thanks, I said as much in post #130, another thing to bear in mind is the bass driver is crossing to a large MF/HF horn mouth roughly equal to the 15”.Fewer crossover points/ drivers mean fewer potential discontinuities and fewer steps away from a point source. Some leading monitor manufactures such as Geithian cluster the Mf/ hf in front of the bass cone( they don’t seem too concerned with a 550/800 hz crossover point range either), while others such as Westlake offer 5 ways (some with an 800hz crossover)on a baffle the size of Wales. FWIW a common rule of thumb mistake people make when estimating the wavelength/ cone diameter / max crossover freq is that for example Tannoy the 15” is actually 12.5”and the JBLs are approx 13”. Cone profile also has a bearing on the beaming ‘issue’. Loudspeakers present a myriad of issues and an equal number of compromises and design decisions/approaches often impacted upon by the transducers that are available to the designer. Porting is another chestnut where people claim they can hear group delay at 25hz or, because they've heard bookshelf sized boxes ported to give fake bass all ports are bad.Honestly do you think ATC(for example) would invest so much R&D in improving their bass units then throw the benefits way in a 'sub standard' cabinet loading/tuning method?

 

[cooky1257]

4367 essentially a tweaked passive version of the M2(includes some distortion measurements);
https://hometheaterhifi.com/reviews...g/jbl-4367-studio-monitor-loudspeaker-review/

No one is forcing you to read it lol.

 

[tuga]

Thanks, I said as much in post #130, another thing to bear in mind is the bass driver is crossing to a large MF/HF horn mouth roughly equal to the 15”.Fewer crossover points/ drivers mean fewer potential discontinuities and fewer steps away from a point source. Some leading monitor manufactures such as Geithian cluster the Mf/ hf in front of the bass cone( they don’t seem too concerned with a 550/800 hz crossover point range either), while others such as Westlake offer 5 ways (some with an 800hz crossover)on a baffle the size of Wales. FWIW a common rule of thumb mistake people make when estimating the wavelength/ cone diameter / max crossover freq is that for example Tannoy the 15” is actually 12.5”and the JBLs are approx 13”. Cone profile also has a bearing on the beaming ‘issue’. Loudspeakers present a myriad of issues and an equal number of compromises and design decisions/approaches often impacted upon by the transducers that are available to the designer. Porting is another chestnut where people claim they can hear group delay at 25hz or because they've heard bookshelf sized boxes ported to give fake bass all ports are bad.Honestly do you think ATC(for example) would invest so much R&D in improving their bass units then throw the benefits way in a 'sub standard' cabinet loading/tuning method?

The crossover frequencies are determined by the number of ways and by the intrinsic characteristics of the drivers.
With horns you need more ways due to their narrow operating range but the M2 uses a waveguide not a horn.

TAD floorstanders are 3-way and cross over at 250Hz and 2kHz,
The largest Vivids are 4-way and cross at 250Hz, 1000Hz, 4000Hz.
The LSR6332 crosses at 250Hz and 2.2kHz and needs to be partnered with a sub to achieve almost full-range (26Hz @ -10dB), crossed at 80Hz.
The ATC SCM150ASL Pro is a 3-way and crosses at 380Hz and 3.5kHz.
All use "assisted" low frequency extension (ports).

As far as I know, there are four major benefits for using bass reflex: cabinet size, load/efficiency, reduced excursion at tuning frequency, SPL.
The downside is transient response, although this can probably be mitigated to inaudibility by tuning the system to a frequency below 30Hz.

And as anywhere else in he chain the need for more accuracy in the sub-bass range depends on the listener's expectations as well as the type of music that he listens to.

 

[cooky1257]

With distortion figures; http://atcloudspeakers.co.uk/wp-content/uploads/2018/01/Hi-Fi_Critic_SCM50PSL_WEB.pdf

JBL 4365, H9800 type diffraction throat biradial HORN, 4" mg comp driver flat out to 20khz, augmented with super tweeter rolled in at 15khz-40khz.

In this context they M2/H9800 are still horns and operate as such though with more pattern control contouring in the M2, a more correct term would be horn-waveguide..

 

[tuga]

4367 essentially a tweaked passive version of the M2(includes some distortion measurements);
https://hometheaterhifi.com/reviews...g/jbl-4367-studio-monitor-loudspeaker-review/

Thanks for the link.

I wish he were using a narrower amplitude scale for the frequency response measurements (Stereophile, Soundstage and most other magazines use 50dB); a 160dB wide scale is worthless as it makes any curve look flat.
It would also help to know how much smoothing he applied.


Here's a review of the 4365 showing distortion figures (page 51 for noise measurements, page 52 for THD @ 90dB '3rd graph from top'):

http://www.cieri.net/temp/JBL_4365_-_Test_di_Audio_Review_(05-2012).pdf

 

[tuga]

With distortion figures; http://atcloudspeakers.co.uk/wp-content/uploads/2018/01/Hi-Fi_Critic_SCM50PSL_WEB.pdf

JBL 4365, H9800 type diffraction throat biradial HORN, 4" mg comp driver flat out to 20khz, augmented with super tweeter rolled in at 15khz-40khz.

In this context they M2/H9800 are still horns and operate as such though with more pattern control contouring in the M2, a more correct term would be horn-waveguide..

At what frequency is the waveguide supposed to be working as a horn?

 

[tuga]

Thanks for the link.

I wish he were using a narrower amplitude scale for the frequency response measurements (Stereophile, Soundstage and most other magazines use 50dB); a 160dB wide scale is worthless as it makes any curve look flat.
It would also help to know how much smoothing he applied.


Here's a review of the 4365 showing distortion figures (page 51 for noise measurements, page 52 for THD @ 90dB '3rd graph from top'):

http://www.cieri.net/temp/JBL_4365_-_Test_di_Audio_Review_(05-2012).pdf

P.S.: here's how Audio Review measures TND (Total Noise Distortion):

http://www.audioreview.it/tecnica/total-noise-distortion-seconda-parte.html

 

[cooky1257]

Thanks for the link.



Here's a review of the 4365 showing distortion figures (page 51 for noise measurements, page 52 for THD @ 90dB '3rd graph from top'):

http://www.cieri.net/temp/JBL_4365_-_Test_di_Audio_Review_(05-2012).pdf

Thanks, for the link.I wish all reviews were that thorough. I love the phrase 'mid-range distortion desert'.

 

[cooky1257]

At what frequency is the waveguide supposed to be working as a horn?

Any horn that controls dispersion is now more commonly termed a waveguide so I'd say over its entire bandwidth. Theres a bit of an amusing squabble over who coined the term, with Earl Geddes using it in a 1990 patent(and Tannoy launched their Tulip waveguide drivers in that same year) but Altec also used the phrase much earlier.

 

[merlin]

2 points about this thread.

1) I'm staggered that Tony has to make a repeat an argument about the specific demands of panel speakers, and more so that people as knowlegeable as Werner don't appear to have sufficient experience with electrostats to realise how crucial the argument about positioning is.

I'm staggered that people on here (including Tony) think that the Logan was in some way disadvantaged given the image of the positioning and the room in question. Those, allied to the plots published suggest it is an incorrect assertion from a scientific POV.

I lived with Logans for something approaching three years. I now own JBL horn hybrids and have obviously had speakers similar to the 800's in the interim.

I've previously set all three types up using Tact software and measurements taken by the engineers at Proac in my listening room. I'd say I have a fairly good idea of how to get the optimum out of the designs.

 

[tuga]

I don't

Any horn that controls dispersion is now more commonly termed a waveguide so I'd say over its entire bandwidth. Theres a bit of an amusing squabble over who coined the term, with Earl Geddes using it in a 1990 patent(and Tannoy launched their Tulip waveguide drivers in that same year) but Altec also used the phrase much earlier.

Horns increase the efficiency through beaming. Waveguides merely control dispersion.

If at all, the M2 will using the waveguide to EQ response by amplifying the lower part of the D2 passband.

 

[cooky1257]

See post 95 for the DSP settings for the D2 on the M2 waveguide that looks like a cut to me. Some guys over at Lansing Heritage have bolted their 4" comps onto the same M2 and the way it loads the driver is different from a normal Bi radial, the eq on the standard horn is a pretty bog standard +5dB/oct above 3khz or -5db/oct from 24khz ish depending on if you want to pad/cut the comp down instead..