• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

JBL 4349 Review (Studio Monitor Speaker)

OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
37,451
Likes
156,728
Location
Seattle Area
The only rational way to do a “thorough listening evaluation” which “reveals something significant” is... wait for it... blind tests. Since that didn’t happen here, it is rational to conclude it is far more likely that non-sonic factors are dominating the listening evaluation.
There is nothing rational about that. Being pessimistic about my conclusions doesn't give you any license to opine otherwise. You just get to not believe, not say, "oh the conclusion must be this ("none sonic factors")." You lack the very proof point you ask me: controlled testing. Lack of data is not data.

Like Amir replied to me yesterday, we have been over this turf a million times. And apparently, according to Amir, I don’t have to read his listening reports. But I don’t know if that means I am expected to bite my lip while readers of his reports, such as yourself, focus on the scientifically-invalid part of his report and declare it “educational, thorough and revealing”.
Your complaints are generic and have nothing to do with this specific review. We have a dedicated thread for such complaints where I have extensively addressed your concern regarding my sighted listening tests: https://www.audiosciencereview.com/...aint-thread-about-speaker-measurements.11139/

ALL your follow ups to this need to go in that thread.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
37,451
Likes
156,728
Location
Seattle Area
And to put that in perspective, a home theatre sat-sub with 4 subs is NOT a small speaker. Compared to such a setup it’s the 4349 that’s the small speaker!
Not if the dynamics are above what the subs produce. Here is the spectrum of the famous "key jangling" file:

1619137301208.png


See how the spectrum above 4 kHz or so reached and exceeds the bass response. Indeed when listening to the JBL 4349, it is the dynamics of the highs that would often startle me. Same happened when I heard Bach symphony with large horns that outperforms my Salon 2s. It was the raspy parts of the horns that stood way out to me.
 

Doodski

Grand Contributor
Forum Donor
Joined
Dec 9, 2019
Messages
11,063
Likes
10,696
Location
Canada
Same happened when I heard Bach symphony with large horns that outperforms my Salon 2s. It was the raspy parts of the horns that stood way out to me.
That's what I found with horns too and the drums where more percussive for me too. Anything with attack came through with more dynamics.
 

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,151
Likes
3,815
Location
.de, DE, DEU
That should be ‘was’, not ‘is’, for Gedlee speakers. And is 97 dB a really high efficiency? Only 6 more than 91 dB, so 1/4 power
That is already a damn high sensitivity. It is not easy to find suitable subwoofers, because without massive support from subwoofers, such a speaker is of course not suitable for listening to music - the f3 should be around 150Hz.



How does that weird shape increase directivity of woofer? Is there any study on this?
This is quite "simple" diffraction of waves at the single slit and the dependence on the slit width.
The narrower the slit, the wider the radiation.
This was often used with loudspeakers in the past, but is still used today - but can lead to problems with decay.
 

ROOSKIE

Major Contributor
Joined
Feb 27, 2020
Messages
1,165
Likes
2,072
Location
Minneapolis
Even with the inexpensive JBL 530, the dynamics at a high volume crush just about anything I have heard in the price/size class.
The 5 is struggling to keep up with the tweeter @ volume and better be highpassed.
A pair of them crossed at 150hz to dual sub/mid bass units is really nice. (For the $)
Anyway, I love dynamic playback.
Really attracted to these speakers.
 

Duke

Addicted to Fun and Learning
Manufacturer
Forum Donor
Joined
Apr 22, 2016
Messages
899
Likes
2,092
Location
Princeton, Texas
That should be ‘was’, not ‘is’, for Gedlee speakers.

My understanding is that Earl will still build the Summa on a semi-custom basis.

Why? On what rational basis? Or is it just because you like the conclusion?

It is fascinating as a (rare) example of a Harman-trained listener coming to a conclusion which is different from what the measurements predict. But remember that the measurements were never claimed to be 100% predictive.

It is educational because the implication is that THIS speaker is doing something RIGHT which was not captured by either the spin or distortion measurements. Lack of power compression seems to be that "something".

The only rational way to do a “thorough listening evaluation” which “reveals something significant” is... wait for it... blind tests.

The reason for having trained listeners is so that reliable evaluations can be made QUICKLY, without the time and expense of conducting a controlled blind listening test every time there is a fork in the road.

Amir does not have the time and resources to conduct controlled blind listening evaluations but he is a reliable enough trained listener that Microsoft used his conclusions (based on sighted listening tests) to make product development decisions. If you don't like his conclusions you are free to not read them.
 
Last edited:

richard12511

Major Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
4,289
Likes
6,445
On another note, I'm not sure I agree so much with the "subjective impressions at odds with the measurements" view that many here seem to have. Other than the crossover dip, the measurements actually look really good. Yes there is better out there (in terms of fidelity) for less, but these didn't get the golfing panther, and most of those "better" speakers did. Then you factor in dynamics. This is probably one of the most dynamic speakers tested to date.
 

Easternlethal

Member
Forum Donor
Joined
Mar 13, 2021
Messages
51
Likes
85
Not if the dynamics are above what the subs produce. Here is the spectrum of the famous "key jangling" file:

View attachment 125711

See how the spectrum above 4 kHz or so reached and exceeds the bass response. Indeed when listening to the JBL 4349, it is the dynamics of the highs that would often startle me. Same happened when I heard Bach symphony with large horns that outperforms my Salon 2s. It was the raspy parts of the horns that stood way out to me.

this is turning out to be an important lesson in interpreting measurements..

so how does the above graph demonstrate good dynamics as opposed to just having treble frequencies turned up?
 

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
1,984
Likes
7,708
Location
NYC
Not sure what you're in a huff about, he's right. Audiophiles get shit talked relentlessly on ASR for the same kind of thing that's happening in the OP. Their subjective experience is routinely dismissed in favor of objective measurements. This review is a rare role reversal, so I'm not surprised to see people calling it out.

I prefer ASR to AudioGon and other subjectivity paradises, but sometimes you gotta call a spade a spade.

For $7500 dollars (or even well below it), this is not a great showing. If this was a non-Harman speaker, it would have got skewered for very little bass , dips in FR and resonances. Justifying this with 'dynamics' and low distortion is what this forum constantly criticizes 'audiophools' for.

Therefore, I have to conclude significant bias in this review for the subjective portion.

I'm quoting these two posts because they are recent, but it's meant to generally address those who take issue with Amir's conclusions.

The problem with all of this is that the measurements aren't bad. They may not be pretty, but they are not bad. I think they are almost great with one or two flaws. Whether they are worth it in a $7,500 speaker is a somewhat separate discussion considering these are massive speakers

I normally really hate making this kind of argument but: sometimes I think you need to have some experience measuring a lot of speakers that you actually sit down and listen to get a sense of which issues really matter. (Not saying that everyone in this thread criticizing Amir's impressions has no experience measuring speakers, I don't know.)

But there have been definitely been times I've assumed some kind of measurement anomaly will completely bring down a speaker, when in fact it is barely audible -- and I spend way more time listening to review speakers than Amir does to identify these issues. And then there are very low-Q deviations and small resonances that constantly get ignored by readership so long as the graphs look pretty (also, why are we looking at CSDs again?) that people constantly ignore when interpreting measurements that, in my experience, are far more likely to be heard.

For example, I'd be willing to bet good money that with the Genelec 8030C...

1619137964182.png


the scoop from 2-4kHz would be nearly audible as the dip from 1-2kHz on the JBL.

Or the scoop from 1-2.5kHz in the 8050B (using Genelec as an example as they are the stereotypical perfect speakers):
1619140327121.png

That scoop covers a wider range of frequencies, so you're more likely to hear it, even if the effect is smaller. Not just pulling that out of nowhere; this is discussed in Toole's book, and it's what I've noticed in my own listening. And what's that at 900 Hz?? It looks like a resonance!

Even if you disagree with my interpretation of the data, measurements are meant to be a reliable predictor of preference over a large sample of listeners. That's why they're great: Measurements mean you can make a decision for yourself about whether the purchasing risk is worth it, and they help you re-evaluate your biases. It makes sense that many would feel these are not worth it at their price.

But as I've said too many times on this forum, nowhere in the literature are measurements supposed to be a rock-solid rule for predicting what an individual listener is going to like, especially over a shorter listening period. It doesn't work that way.

So if you're a listener who, like most of us, is just a single person, you provide your individual listening impressions. I mean, what's Amir supposed to say? "I tried really hard not to like this speaker because I thought the measurements weren't pretty. I truly hate the fact I liked these speakers, and that should be illegal! Don't even bother listening to these even though I liked them, because by golly, there's a dip in the spin!" (On the other hand, Amir didn't like the similarly-sized 4319, which to me is a more flawed speaker.)

And while I'm at it, I think it's important to caution that measurements should only really be used to compare speakers that either are in the same size class or are being compared within the limitations of the smaller speaker. Otherwise, why don't we all just get Neuman KH80s for living room purposes? We all know Amir likes to listen loud...
 

GDK

Senior Member
Forum Donor
Joined
Aug 8, 2020
Messages
458
Likes
1,113
Location
Toronto
It was Aries Cerat. Here is the exact room and demo:

Funny, I heard nearly the same set-up last year when I picked up a used amp off of a high end dealer here in Toronto and had the exact same impression. He had an Aries Cerat amp and some hORNS similar to these: http://horns.pl/en/speaker-sets/universum-3-way/

We were sitting chatting and he put some jazz on and the trumpet just sounded incredible. So detailed like nothing I had ever heard before. It’s an interesting discussion on these points, for sure.
 

bigjacko

Addicted to Fun and Learning
Joined
Sep 18, 2019
Messages
652
Likes
307
This was often used with loudspeakers in the past, but is still used today - but can lead to problems with decay.
Thank you for answering. Why don't people use it now and why it has problem on decay? Is the shape designed to have as much different length of gap throughout the perimeter?
 

Francis Vaughan

Addicted to Fun and Learning
Forum Donor
Joined
Dec 6, 2018
Messages
933
Likes
4,610
Location
Adelaide Australia
This was the track:

Bach BWV 565. (Better know as the Toccata and Fugue in D minor, probably the best know classical organ work ever. Even if Bach may not have actually written it.)
Pipe organs are probably one of the hardest things to cope with. Dynamics, deeper bass than any other source, wide timerbal palate, and can play seriously loud. Indeed something like the camarde horns many organs display can get you way louder than any other instrument. (The Grand Ophicleide horns on the Boardwalk Hall organ in Atlantic City New Jersey are 130dB SPL @ 1 metre). Reed stops can be almost comically harsh. And loud. Domestic sound systems than can do justice to a pipe organ are few and far between. Unless you have heard one in full flight in the flesh you are missing out on a fabulous experience.
 

Kvalsvoll

Addicted to Fun and Learning
Manufacturer
Joined
Apr 25, 2019
Messages
694
Likes
1,268
Location
Norway
Isn't most music heavily bass slanted?

If you look at the waveform of a trumpet recording, you will see extremely loud high frequency peaks. The energy level is lower at high fr, but the peak level is still very high.

Generally, music has a tilted energy distribution, which can easily lead to the faulty conclusion that spl requirements are much lower at high frequencies. But if you look at peak levels and compare that to the spectrum, the general trend is that peak-to-rms difference increases with higher freq.

Then directivity is also part of perception, horns generally has much better directivity control, resulting in a sound field with larger direct-to-diffuse ratio, which gives higher sound intensity.
 

Kvalsvoll

Addicted to Fun and Learning
Manufacturer
Joined
Apr 25, 2019
Messages
694
Likes
1,268
Location
Norway
My understanding is that there are four primary mechanisms (two thermal, two mechanical) by which power compression may occur:

1. Rapid-onset voice coil heating which causes the voice coil resistance to rise;

2. Slow-onset magnet heating which causes the magnet to temporarily lose strength;

3. Voice coil excursion beyond its linear limits; and

4. Suspension system excursion beyond its linear limits.

I can see how the two mechanical compression mechanisms would show up in distortion measurements because they change how the cone moves over some of its travel, but it is not clear to me how the two thermal compression mechanisms would show up in distortion measurements.

So I do not think that power compression can be adequately evaluated from distortion measurements. I think it would need to be measured directly, then those findings would need to be correlated with perception, and I don't think that has been really been done yet.

One of the more insidious issues thermal modulation can give rise to is described in the Floyd Toole quote in my previous post. To paraphrase, if the different drivers in a speaker system have significantly different thermal compression characteristics, then the system's tonal balance can change with power level.

In general the two best predictors of thermal compression are probably driver efficiency and voice coil mass (driver thermal power handling ratings have grown in recent years as high-melting-point formers and glues are being used in the voice coils to forestall deformation, but this does not mitigate thermal compression). X-max is probably a good predictor of the onset of significant mechanical compression, as x-max may be established either by the voice coil's linear excursion limits or the suspension system's linear excursion limits.

The motor system also has several significant dynamic nonlinearities, both inductance and field strength is affected by the current in the voice coil.
 

Rottmannash

Major Contributor
Forum Donor
Joined
Nov 11, 2020
Messages
1,922
Likes
1,534
Location
Nashville
The bigger brother D2340k is struggling below 1500 hz and really stretching it in the M2 application to reach the 792 hz crossover at a steep LR 36 acoustic slope.
I think the crossover-point is good for the tweeter, but perhaps a couple of hundred hertz above the ideal for the 12" woofer.

Choosing to make a high-capacity speaker in this format makes for some compromises that may not be perfect on paper, but it looks to be a well thought-out design overall. Not perfect, but hey. Way cooler to have next to the telly than most slim-boxes. Picture from a random dude on the internet with fugly speakers to prove my point :D


View attachment 125407
Is that the Eval 1 Purifi amp Amir tested atop that enormous power amp?
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
1,570
Likes
3,193
Location
US East
I believe in this case, the Olive formula is broken. The way PIR_SM is calculated heavily penalizes flat PIR curves, which it is not supposed to. Here is the illustration. By adding a downward slope to the PIR, the raw PIR_SM score improves by 0.453, which results in an increase of 1.05 in the final score.

The PIR_SM calculation is based on the Pearson's correlation coefficient. With a perfectly horizontal line, it is undefined. Add a just a tiny amount of noise, the x-variable and y-variable become completely uncorrelated, and will give a correlation coefficient of 0. If it is used to judge smoothness, it will falsely indicate that the curve is "not smooth", which is a wrong conclusion.
PIR_SM.png
 
Top Bottom