JBL 708i Monitor Review (DSP: Part 2)

[fineMen]

These are not consumer speakers!

I do not know what a commercial user would expect from Harman's JBL 708i.

If you are a pro, please explain for which other benefit you would accept
- a sharp drop in sound power between one and two kilohertz
- a bump in distortion at two kilohertz
- which presumably (!) would generate intermodulation (all relentlessly refuse to measure it)
- an uneven frequency response in critical bands
- a relatively narrow vertical dispersion at 2 kilohertz
- the hassle with the setup
- port resonances

with a speaker that is--regardless of the excellent bass driver, limited to some 100dB of clean output. To acknowledge the latter would easily lead to omitting the compression driver and solving 6 out of the 7 above mentioned problems.

From the press-kit:

"... JBL’s renowned Image Control Waveguide to deliver extraordinary output, stunning detail, an expansive soundstage, and greater accuracy in a broad range of rooms. 708P delivers several times the output of other comparably-sized studio monitors, ..."

Renowned by whom? What does that actually mean, "serveral times"?! In dB? Linear, or what?

And in regard to the high fr/ driver:

"Competitive monitors utilize radiating dome tweeters, rather than compression drivers, that require amplifier limiter circuitry to prevent failure at higher SPLs which also negatively affects the speaker’s stereo imaging and dynamic range. 7 Series innovative compression design does not need limiting circuitry and iscapable of reproducing frequencies with a smooth response beyond 36 kHz."

Did I get it right? If one speaker runs into the limit and is shut off, then stereo is negatively affected. Affirmed!

 

[GXAlan]

Correct, compression drivers are often used because of their output capabilities. A dome can also have a waveguide control its pattern.

My understanding is that even an advanced waveguides on dome tweeters aren't as good as what is possible with a compression driver?

 

[sarumbear]

I do not know what a commercial user would expect from Harman's JBL 708i.

If you are a pro, please explain for which other benefit you would accept
- a sharp drop in sound power between one and two kilohertz
- a bump in distortion at two kilohertz
- which presumably (!) would generate intermodulation (all relentlessly refuse to measure it)
- an uneven frequency response in critical bands
- a relatively narrow vertical dispersion at 2 kilohertz
- the hassle with the setup
- port resonances

with a speaker that is--regardless of the excellent bass driver, limited to some 100dB of clean output. To acknowledge the latter would easily lead to omitting the compression driver and solving 6 out of the 7 above mentioned problems.

From the press-kit:

"... JBL’s renowned Image Control Waveguide to deliver extraordinary output, stunning detail, an expansive soundstage, and greater accuracy in a broad range of rooms. 708P delivers several times the output of other comparably-sized studio monitors, ..."

Renowned by whom? What does that actually mean, "serveral times"?! In dB? Linear, or what?

And in regard to the high fr/ driver:

"Competitive monitors utilize radiating dome tweeters, rather than compression drivers, that require amplifier limiter circuitry to prevent failure at higher SPLs which also negatively affects the speaker’s stereo imaging and dynamic range. 7 Series innovative compression design does not need limiting circuitry and iscapable of reproducing frequencies with a smooth response beyond 36 kHz."

Did I get it right? If one speaker runs into the limit and is shut off, then stereo is negatively affected. Affirmed!

If you don’t know why do you stipulate?

Besides, I don’t know why you are asking me. I’m neither a JBL representative nor I advocated the speakers. You may want to read my posts before posting a tirad.

 

[pierre]

I do not know what a commercial user would expect from Harman's JBL 708i.

If you are a pro, please explain for which other benefit you would accept
- a sharp drop in sound power between one and two kilohertz
- a bump in distortion at two kilohertz
- which presumably (!) would generate intermodulation (all relentlessly refuse to measure it)
- an uneven frequency response in critical bands
- a relatively narrow vertical dispersion at 2 kilohertz
- the hassle with the setup
- port resonances

with a speaker that is--regardless of the excellent bass driver, limited to some 100dB of clean output. To acknowledge the latter would easily lead to omitting the compression driver and solving 6 out of the 7 above mentioned problems.

From the press-kit:

"... JBL’s renowned Image Control Waveguide to deliver extraordinary output, stunning detail, an expansive soundstage, and greater accuracy in a broad range of rooms. 708P delivers several times the output of other comparably-sized studio monitors, ..."

Renowned by whom? What does that actually mean, "serveral times"?! In dB? Linear, or what?

And in regard to the high fr/ driver:

"Competitive monitors utilize radiating dome tweeters, rather than compression drivers, that require amplifier limiter circuitry to prevent failure at higher SPLs which also negatively affects the speaker’s stereo imaging and dynamic range. 7 Series innovative compression design does not need limiting circuitry and iscapable of reproducing frequencies with a smooth response beyond 36 kHz."

Did I get it right? If one speaker runs into the limit and is shut off, then stereo is negatively affected. Affirmed!

I can partially answer your question (if it was not a rhetorical one): in one of the Dolby calibrated room I used, there had 3xM2 for fronts and I think 12 708i and 3 assorted JBL subwoofer. The whole system is/was eq-ed with an Intonato (a dsp box sold by JBL). The calibrated result is good enough to be certified by Dolby. The price of the total solution is a key argument when buying a configuration like this. It will be expensive anyway. The other room is full Genelec and was significantly more expensive. In terms of convenience, having all the amps in a rack is easier in case of an issue. It is super easy to replace the amp, use spare canals from another one etc. It also put the heat generated by the amp outside of the control room. Sound wise, I do not hear an issue. They only do surround duties so maybe there is that.

 

[sarumbear]

I can partially answer your question (if it was not a rhetorical one): in one the Dolby calibrated room I know, there are 3xM2 for fronts and I think 12 708i. The whole system is eq-ed with a Intonato (a dsp box sold by JBL). The calibrated result is good enough to be certified by Dolby. The price of the total solution is a key argument when buying that much. It will be expensive anyway. The other room is full Genelec and was significantly more expensive. In terms of convenience, having all the amps in a rack is easier in case of an issue. It is super easy to replace the amp, use spare canals from another one etc. It also put the heat generated by the amp outside of the control room. Sound wise, I do not hear an issue. They only do surround duties so maybe there is that.

Do you recall what were the subs in the JBL room?

 

[changer]

You have skirted the issue of (...)

No, I ignored you here.

The Twitter HP is clearly showing that there’s at best 12dB slope. If the slope was 24dB the response at 1500Hz would be 10dB lower.

I neither do know what a Twitter resonance is—does this got something to do with Elon’s latest endeavors?—nor what a Linkwitz Transport (??). But I can tell that you are making things up and do not care to memorize even most recent lessons.

In Amir’s near-field measurements of individual drivers the woofer’s response is not corrected for baffle step. You cannot deduct crossover point from this graphs. The sonogram can help those who stared at simulations for days and know how they look with different filters applied.
And, with a LR4 crossover, amplitude response of the drivers is -6 dB at crossover ...

 

[fineMen]

My understanding is that even an advanced waveguides on dome tweeters aren't as good as what is possible with a compression driver.

See https://heissmann-acoustics.de/wp-content/uploads/xd270_wg_winkel_normiert.png

Waveguide and dome match to some degree. But what is good? Most often a compression driver is used with a waveguide so that both combined show constant directivity. That means that the radiation pattern is the same for all frequencies. To design the combo this way has some advantages for public address speakers. The audience is spread out over a large aerea, and all should receive the same sound. (stereo is illusional under such circumstances btw).

For home use I would say that a smoothly narrowing, controlled directivity is preferred.

As the example shows, the latter goal is achievable; just use that combo! But I think you're right in saying that a true constant directivity is not that easy to achieve with a dome. JBL 305 comes close ... and sounds a bit bright.

you may want to read my posts before posting a tirad.

O/k, I recall a previous conversation that started the same. Sorry, I'm not an expert in this field.

 

[fineMen]

I can partially answer your question (if it was not a rhetorical one): ...

My question was too verbose, sorry. But it was serious. I don't think the 708i comes close to satisfy a consumer's demand. Thanks for clarifying!

For the intended operation mode it might be o/k, I cannot judge it. But for home use my vote was 'poor'.

 

[Cars-N-Cans]

That's interesting, but that would shift the crossover closer to the woofer resonance which is at around 2.6kHz.....but like you say a steeper crossover to help mitigate that influence. I guess you're thinking the tweeter is playing too low and causing distortion at high levels, although the higher crossover wouldn't solve the 5-10kHz distortion and it wouldn't do anything to affect the messed up 600-1000Hz region that is being affected by bass port resonances. To be honest there's multiple fronts where this speaker looks like it needs to be improved on a design level, not to even mention the poor user integration related to the analog trim-pot ridiculousness.

Interesting to note that the unwanted HF “boost” and the distortion are correlated. The two track each other. Wonder if it’s a production issue? Port resonances are also interesting in that not all of them are scaling directly with the woofer output and being seen directly in the FR. Probably not great as it points to them being of a higher Q, perhaps. Still, if I had this speaker, the tweeter part is what would be somewhat concerning since it looks like an issue with QC, or perhaps something else.

 

[sarumbear]

2. 708i in bi-amp mode does not have a passive crossover for the woofer just for the tweeter.
And the tweeter has a 1uf condensator in series and a 11ohm resistor in parallell.

Can you confirm the capacitance? 1 micro farad at 8 ohms will not pass much below 20kHz !

And… 11 ohms resistor will absorb almost the same power as the drive unit!

 

[ROOSKIE]

Did this speaker really measure higher distortion with its tweeter at 96dB than the much much cheaper A130 ? Don't get it - if you look at the A130 tweeter there is hardly anything to it Vs a sophisticated compression driver

Good question and while am not an expert give this some thought...

My answer is , do not confuse this with the type of distortion that happens when a driver is nearing limits.
It is not uncommon for 2nd order distortions to be a side effect of waveguides and horns and even the material choice of the driver.
This is actually pretty good for such a situation.
Now, look at the near ZERO 3rd (and higher order) distortions. This indicates a well designed unit to me.
I can you assure from using similar models the tweeter can play without stress to well beyond 96db@1m. The upper limits depending on the exact design are nearer 120+db@1meter.

I would be interesting to see some IMD testing here on the tweeter and compression testing, including at very high SPL - beyond what a dome tweeter can handle.

Anecdotally I did a lot of testing of the A130 vs the Studio 530. I tested both with 125hrz high passes so I could really crank them and I can assure you that the 530 can play cleaner and without strain in the treble much, much, much louder. Like not even close. At high volume the A130 starts to sound audibly compressed, strained and unhappy while the 530 is something that can still be turned up and up(with that HP @125hrz, otherwise the woofer will be sad) and I bring that up as the 530 shows some 2nd order HD that I attribute to the waveguide design and compression design. The 4309 that I have can play again much louder than even the 530 without said sense of strain. To loud for listening.

Anyway IMHO in terms of HD type distortion in general. Honestly, there is a lot of information out there on Harmonic Distortion. I really think folks ought to read as much as they can and then come back. I really don't feel it is well understood by very many hobbyist's. It really is not what most people seem to imagine it to be.

 

[ROOSKIE]

Looks poor from the perspective of a consumer. I don't know what a commercial user would expect from this. The sound power drops significantly between 1k and 2k. The vertical radiation is to narrow to be useful at 2kHz. The x-over woofer ./. tweeter is way too high as to match the 8" woofer. Alas, even if the passive x/o was circumvented, the tweeter would not allow to be used below that 2kHz, where it shows clear signs of overload.

The advertising says, that the compression driver outperforms any dome tweeter. There one may find a better alternative: https://heissmann-acoustics.de/en/test-vifa-xd-270-f4-waveguide-wg-300/ for just 50$ per side waveguide (!) and dome combined.

Astonishingly the raise in second order harmonic distortion, which is kind of typical for compression drivers, does not lead to equally increased intermodulation. To the contrary, as far as my measurements hold. But here the peak at 2kHz would, I suggest.

Even the bigger-than-life M2 shows some irregularity in the midrange, which are not originating in the ports' resonances. As if they wouldn't care.

Hmmm, while I am not exactly(V=Not Terrible) a fan of this speaker I must say
I think that tweeter is hardly working at 2k. The 3rd order is extremely low there and the bump is likely from inherent characteristics due to the compression driver +Waveguided relationship.
As you yourself mentioned these present in such designs but are not indications of limitations being reached.

Intermodulation is really best revealed by a test for it, which this is not.

I would expect this tweeter within this design to easily handle SPL levels far beyond what any typical situation will require. (100+db at @4 or more meters)

I have the Peerless tweeter you linked in one of my DIY speakers.
It is quite good for what it is.
I tried to buy the Monacore Waveguide but in the USA it was impossible to find.
I have a Dayton Tweeter in SEOS 8 waveguide and that sounds great and is crossed over at 1.1k without major issue, also a DIY speaker using an SB acoustics dimple dome loaded in a guide.
Can they compete with the compression drivers I have here? I don't think directly so but they are pretty good and I think most people would be satisfied with the dome solution.
Yes, the $50ish price reflects parts costs of the set-up you linked but there is more of course.
While I had fun and can certainly enjoy both, neither of these 2 particular DIY set-ups are my favorite speakers so really hard to justify the cost for them for most people. Zero way to 'return' them or really even sell them but I can keep tweaking them for fun and I do.(1 is my own design and 1 designed by Joseph Crow Audio)

Anyway, I suspect the tweeter and waveguide parts used here costs JBL themselves even less than $50 if you do not count R&D.
I wonder how much the screws cost ...kidding!!

 

[jhaider]

SCL line from JBL Synthesis which depend more on EQ.

FWIW, SCL-3/4 were on the market and SCL-2 was complete or nearly completed, prior to the Samsung acquisition.

SCL-3/4 were shown at CEDIA 2016.

The best new home theater gear and demos at CEDIA EXPO 2016

We attended CEDIA EXPO 2016 in Dallas, TX. EXPO is an industry tradeshow for the home electronics industry, which covers everything from home automation through networking to home theater and distributed audio. This blog post reveals some of the cool things that caught our eye (and ear) this...

acousticfrontiers.com

SCL-2 was shown at CEDIA in 2017. Samsung’s acquisition of Harman closed in Q1 or Q2 of 2017.

The other ones came out much later.

Edit: this is the label on the speaker. It clearly says only a specified Harman amp and a signal processor should be used. Not any amplifier.

It is badly, or perhaps more accurately self-servingly, worded. Anyone with any intelligence knows an amp is a damn amp. The BSS processor or Intonato do not ask what amp you have upstream, and they are not locked to some magic key that only exists in Harman amps! (I probably shouldn't give any company that idea...) I ran a pair of 708i’s for L/R with BSS processing (SCL-3 for center) powered by an ATI amp without issue. What is actually necessary is that the filters needed to complete the crossover are upstream of the amplifier. More on that below.

In which case there is no reason for either of these tests whatsoever other than to satisfy a hobbyist curiosity.

I can see why you'd say that as to Part I. I've been sitting on my own data of the raw speakers since apparently March of 2017 for that reason - I measured them for my own curiosity and had no desire to let other people free-ride off of my purchase of the correct processing equipment. But here, the speaker is measured with one of the approved processor-amps. So that is quite relevant. What looks like like either poor quality control or a running change to the tweeter not addressed by the speaker tuning is also quite relevant, and unfortunate.

The diagram I posted on the part 1 thread shows both a DSP equipped Crown amplifier and a BSS.

A cynic might say that diagram is targeted at maximizing sales of Harman components.

However, if you look closely you’ll see that the BSS processor is used there for bass management and signal distribution. In that setup all processing could be centralized in the BSS, or individual tuning files could be used in each amp. BLU-Link avoids a set of conversions in that case.

One other thing to note is, the Harman Pro software situation is a complete mess right now. The current Harman Audio Architect software does not support “macros” on BSS gear. That is how the 7-series tuning files are implemented for BSS. You have to use the older London Architect software. Audio Architect only supports the tuning files on Crown amps. To further muddle things in a way that must make sense to someone somewhere in Harman management but is completely ridiculous for anyone out of that bubble, Intonato runs a whole different software - and it has been broken on Macs now for quite some time. That is why my Intonato is sitting idle…

Do you know what is the reason why the Crown amp's DSP is not used for LP/HP filtering to form an active crossover?

I don’t. Logically it would seem the input configuration implies different components in the signal path - otherwise why not just two sets of terminals with a removable jumper between them?

 

[EJ3]

Part 1 showed us so much potential! Well: Part 2 That was disappointing!!

 

[Peafowl]

Edit: this is the label on the speaker. It clearly says only a specified Harman amp and a signal processor should be used. Not any amplifier.
View attachment 254821
In which case there is no reason for either of these tests whatsoever other than to satisfy a hobbyist curiosity.

The diagram I posted on the part 1 thread shows both a DSP equipped Crown amplifier and a BSS. Do you know what is the reason why the Crown amp's DSP is not used for LP/HP filtering to form an active crossover?

To read this, brings my mind back to the struggle i had when i changed the DSP on a full aktiv PA system and thought just swapping the units and loading the same XO charts will be okay.
At the end, it´s not that simple sometimes and i fond out with this this - why DSP, EQ isn´t ? DSP, EQ on a different Brand.
Sorry i didn't find the english version.

 

[GXAlan]

FWIW, SCL-3/4 were on the market and SCL-2 was complete or nearly completed, prior to the Samsung acquisition.
SCL-2 was shown at CEDIA in 2017. Samsung’s acquisition of Harman closed in Q1 or Q2 of 2017.

The other ones came out much later.

Thanks for the corrrection. I was thinking about the SCL-1 which I recall seeing a bad uncorrected spin, although I don’t trust my memory on the SCL series anymore.

 

[Sancus]

The 3rd order is extremely low there and the bump is likely from inherent characteristics due to the compression driver +Waveguided relationship.

Yeah, every compression driver I've seen measured comes with high 2nd harmonic. I thought it was just an endemic characteristic of them. Even the practically cost no object M2 tweeter has the same distortion profile.

Waveguided dome tweeters definitely do not typically exhibit the same thing.

Considering 2nd harmonic is described as benign or even pleasant by many I'm not even sure if it should be treated the same as other harmonics or non-harmonic distortions.

 

[Cars-N-Cans]

Yeah, every compression driver I've seen measured comes with high 2nd harmonic. I thought it was just an endemic characteristic of them. Even the practically cost no object M2 tweeter has the same distortion profile.

Waveguided dome tweeters definitely do not typically exhibit the same thing.

Considering 2nd harmonic is described as benign or even pleasant by many I'm not even sure if it should be treated the same as other harmonics or non-harmonic distortions.

Not a compression driver aficionado but I’m going to hazard and guess that it’s due to the high sound pressure levels developed at the diaphragm and the throat of the waveguide (Edit: Or probably just within the driver itself) resulting in adiabatic compression since it does seem to be something they all share and they all have the same general design. Also the harmonic distortion components will be at 2x the fundamental so at some point they are at a high enough frequency not to be heard as easily.

In that regard it makes sense since a dome tweeter doesn’t explicitly rely on any sort of impedance match like a compression driver does, so the distortion will be dominated by mechanical nonlinearities instead. I always though of the JBLs as being more towards being waveguides but I guess they can be a bit “horny”, too.

 

[witwald]

A known art to create them with electro-mechanical components, this you are completely making up. Acoustic slopes are acoustic slopes.

Yes, the resultant acoustic slope is just that. However, when designing a crossover it is likely to have been created for a particular set of driver impedance curves. Under heavy power input, a driver's voice-coil will heat up and change its resistance. This will affect the impedance of the driver, and the passive crossover components selected for a particular operating point will no longer be as optimal for the new operating point.

An active crossover's slopes won't be affected by heating of the voice-coil, so the active crossover could be expected to work more accurately across a very wide range of power inputs to the loudspeaker system. For domestic use, those resultant frequency response variations may not be all that concerning, but in professional use they would potentially become more important.

 

[Cars-N-Cans]

Yes, the resultant acoustic slope is just that. However, when designing a crossover it is likely to have been created for a particular set of driver impedance curves. Under heavy power input, a driver's voice-coil will heat up and change its resistance. This will affect the impedance of the driver, and the passive crossover components selected for a particular operating point will no longer be as optimal for the new operating point.

An active crossover's slopes won't be affected by heating of the voice-coil, so the active crossover could be expected to work more accurately across a very wide range of power inputs to the loudspeaker system. For domestic use, those resultant frequency response variations may not be all that concerning, but in professional use they would potentially become more important.

For the most part that can probably be neglected unless they are something like large PA speakers that will be driven hard, but there probably don't care as much since its not exactly "critical listening" territory. Even at 100 °C a copper 8Ω voice coil will only increase by about 2.5Ω, and this is only the resistive part. The reactive portion will remain mostly unchanged since its a function of the mechanical characteristics of the driver and the enclosure or whatever is loading it. Also it should have enough thermal inertia for transient effects to be negligible over short time scales. It is certainly a possibility, but probably degradation of the voice coil insulation will be the greater concern unless its something like PAI. I guess in a professional environment this will be more likely, but at high SPL levels it will also be less noticeable unless it impacts the actual operation of the speaker. At any rate, if it gets hot enough for that to really be an issue, its probably time for a bigger speaker.