Harbeth Monitor 30 Speaker Review

[napilopez]

The Olive formula seems far then perfect in this case, as this loudspeaker isn't either smooth on axis nor has smooth directivity, but only (like most of those designs) a clever chosen frequency on axis which compensates the directivity problems to some degree on the sound power / far field, but the psychoacoustically important direct sound above the transition frequency remains coloured.

I agree and disagree. I disagree in the sense that the smoothness is determined by the formula, therefore the speaker is smooth enough. Even if the on-axis is coloured, there's nothing to suggest the speaker wouldn't still be preferred as often as a speaker that achieved the same score but had smoother on-axis. Though academically, I would believe that to be the case.

I think there are some discontinuities that look bad to the eye but are not that bad in practice. The on-axis curve is probably the spekaer's weakest part, but the Predicted in room curve - which almost always looks just like the early reflections curve, is actually quite smooth.

Compare with, say, the Revel Ultima Studio2:

Obviously better on-axis, but the directivity isn't that much better.

But yes, I agree with you in the sense that it's worth noting the formula doesn't actually include smooth directivity as part of the calculation. The only variables are operations on the predicted in-room response, on-axis curve, and low frequency extension. Because both the predicted in room response and on-axis curve should be smooth, that generally implies smooth directivity. For the perfect speaker, these would both be perfectly smooth. But for anything less, the formula does not necessarily require it. The fomula also doesn't seem to account for tilt. So you can "cheat" the formula a bit

That's something the careful observer should watch for. A smooth directivity curve will tell you how much you can "correct" a speaker via EQ (Distortion and built-in EQ such notwithstanding).

As you note, in this case there appears some compensation for the directivity issues in the on-axis curve. That's a legitimate way of designing a decent speaker methinks, even if not the ideal.

 

[Deleted member 2944]

re: rough impedance plot, are we sure this speaker is in proper condition? I was going to suggest checking that all of the fasteners are tight, then I saw the rear-panel picture. Not 100% that I'm seeing it clearly, but it appears that someone attacked them with more enthusiasm than mechanical skill.

I'm having trouble with the impedance plot as well. To me, that indicates something wrong with the speaker.

Can somebody outline the Klippel impedance measurement routine? Is this a stepped measurement with thousands of data points that takes many minutes to complete??

Dave.

 

[Phorize]

Sorry, a bit dramatic. They did censor it from (an active and enthusiastic contributor on) their own forum, but certainly never tried anywhere else.

I posted it last night and confirmed it was awaiting moderation. Today I look and it is nowhere in the thread or my content - clearly deleted.

I see what you mean, it would be nice if AS felt able to contribute to this discussion as he did design the speakers in question!

 

[hardisj]

I'm having trouble with the impedance plot as well. To me, that indicates something wrong with the speaker.

Can somebody outline the Klippel impedance measurement routine? Is this a stepped measurement with thousands of data points that takes many minutes to complete??

Dave.

https://www.klippel.de/fileadmin/klippel/Bilder/Our_Products/R-D_System/PDF/S2_LPM.pdf

"Excitation Signal:
The stimulus used during the measurement is a sparse multi-tone complex spaced logarithmically over frequency. This signal is optimal for the parameter identification at small amplitudes because the transducer is only excited at frequencies of interest. The user may specify the amplitude and the frequency range covered by the tones and their distance (relative resolution). Furthermore, either the voltage at the output connector (OUT 1) or the voltage at the terminals of the speaker connected to output SPEAKER 1 (SPEAKER 2) may be specified. In the latter case the amplifier gain is determined at 750 Hz without load prior to the main measurement and the excitation level is adjusted accordingly. Also, the amplifier low frequency roll-off is determined and compensated for the two lowest frequency lines. "

 

[hardisj]

and, yes, I agree the impedance looks very, very odd. The woofer's impedance indicates something is wrong or there's a very poor woofer design/implementation. The only way to know for sure is to measure the raw driver itself; outside of the enclosure.

 

[Frank Dernie]

Isn’t It effectively a 70 year old design , BBC started to investigate speakers when late 50’s?
Keith

I wouldn't say so, not the design.
The BBC started their research a long time ago but multiple tests and prototypes were done before the first ideas of what seemed to be important emerged and then designs focused on achieving the goal.
One of the key things was that paper cones couldn't be good enough but alternatives were a long time coming and manufacturing techniques were limiting, early vacuum formed plastic cones were thin where they ideally should have been thick and vice versa because of the manufacturing process. The first fruits of their work started to be available outside the BBC in the 1970s and have been developed ever since. The plastic, in particular, is more suitable today than the early bextrene (which is still pretty good IME) and people can afford to injection mould it so the cones behave much better and are more consistent.
Sound is still the same thing today as it was 70 years ago but manufacturing techniques, materials and analysis methods are much more readily available to far more people and much cheaper.
Apart from the out of control deep bass on the bigger models I have found the BBC derived speakers I have heard and owned to sound extremely realistic, FWIW.

 

[Krunok]

I agree and disagree. I disagree in the sense that the smoothness is determined by the formula, therefore the speaker is smooth enough. Even if the on-axis is coloured, there's nothing to suggest the speaker wouldn't still be preferred as often as a speaker that achieved the same score but had smoother on-axis. Though academically, I would believe that to be the case.

I think there are some discontinuities that look bad to the eye but are not that bad in practice. The on-axis curve is probably the spekaer's weakest part, but the Predicted in room curve - which almost always looks just like the early reflections curve, is actually quite smooth.

Compare with, say, the Revel Ultima Studio2:

Obviously better on-axis, but the directivity isn't that much better.

But yes, I agree with you in the sense that it's worth noting the formula doesn't actually include smooth directivity as part of the calculation. The only variables are operations on the predicted in-room response, on-axis curve, and low frequency extension. Because both the predicted in room response and on-axis curve should be smooth, that generally implies smooth directivity. For the perfect speaker, these would both be perfectly smooth. But for anything less, the formula does not necessarily require it. The fomula also doesn't seem to account for tilt. So you can "cheat" the formula a bit

That's something the careful observer should watch for. A smooth directivity curve will tell you how much you can "correct" a speaker via EQ (Distortion and built-in EQ such notwithstanding).

As you note, in this case there appears some compensation for the directivity issues in the on-axis curve. That's a legitimate way of designing a decent speaker methinks, even if not the ideal.

Something is telling me F208 could get better score than Ultima2 you posted..

 

[amirm]

I'm having trouble with the impedance plot as well. To me, that indicates something wrong with the speaker.

Can somebody outline the Klippel impedance measurement routine? Is this a stepped measurement with thousands of data points that takes many minutes to complete??

Dave.

No. It uses the same swept log sweep as the rest. The little side peaks only show up in highest resolution sweeps. Just measured another speaker to same effect.

 

[thewas]

Compare with, say, the Revel Ultima Studio2:

Obviously better on-axis, but the directivity isn't that much better.

While I agree with the rest of your post (which I therefore didn't repost here) I have to disagree a bit on this, the directivity of above Revel ist quite better than that of the Harbeth, it may also have its peak around 2 kHz but after it doesn't fall to a significantly lower level than before this like the Harbeth does, which is to be expected if you see its smaller midrange driver, lower crossover frequency and (unfortunately too) small waveguide at the tweeter. Newer Revels like the F208 posted from Krunok are of course even better in regard to this discipline.

 

[dshreter]

and, yes, I agree the impedance looks very, very odd. The woofer's impedance indicates something is wrong or there's a very poor woofer design/implementation. The only way to know for sure is to measure the raw driver itself; outside of the enclosure.

Unfortunately for the forum, I would prefer the woofer stay exactly where it is inside the enclosure.

 

[hyperknot]

One thing to note here I think should be that Alan Shaw mentioned multiple times, that the biggest change in the recent models is especially that their response is flat. He even wrote about mentioning that older models, like the M30 were not flat and newer models became flat and the sales went up exactly at the same time. So I think this is not a bad review, and would make it even more important to test an M30.2 to see if his notes can be measured or not.

BTW, Alan Shaw is also one of the speaker designers who said he wouldn't do anything without his measurement equipment and also that he doesn't trust his ears. So I don't think it's that far from ASR philosophy, the only thing what happened here was that a 20 year old design was tested and not a recent one.

---


From Designer's Blog

Do I trust my ears? Good question. Trust them to do what? Be the final arbiter between A and B augmenting the technical measurements, or to be reliable objective tools in themselves? I would never trust my ears unless I had access to test equipment to establish whether I was being self-deceived or not. So, even if I was washed up on a mythical desert island with an audio source, a crate of inductors, capacitors and resistors, some drive units and planks of wood I would not commence speaker design unless my test equipment has been washed up too. It would to too frustrating not to be able to cross-correlate what I heard with what I measured: an endless, maddening circular tweaking process without end as my ears and mood changed from day to day.

 

[Krunok]

Here they are, neck to neck: with Early reflections DI Harbeth seems to be dropping from 4 to 1 and Ultima2 from 7 to 3.

 

[dukanvadet]

That impedance measurement is bad. Every bass reflex speaker have two pronounced peaks, one on either side of the tuning frequency, i have made some crappy test boxes that propably were not completely air tight but i have not seen that kind of inpedance result before. That there are some minor irregularities is normal but that the top of the first peak is completely cut of like that looks crazy.

 

[617]

That impedance measurement is bad. Every bass reflex speaker have two pronounced peaks, one on either side of the tuning frequency, i have made some crappy test boxes that propably were not completely air tight but i have not seen that kind of inpedance result before. That there are some minor irregularities is normal but that the top of the first peak is completely cut of like that looks crazy.

We've hashed that out. It is due to the impedance sweep being done at high enough levels that room reflections are influencing it.

 

[dukanvadet]

We've hashed that out. It is due to the impedance sweep being done at high enough levels that room reflections are influencing it.

Ok. I tought i read the thread trough but obviously not I was trying to work and read at the same time, maybe not the best idea...

 

[MattHooper]

I want to quote MattHooper: I was dumbstruck by hearing the spookiest sense of a woman standing way behind the speakers singing and a full trumpet just "appearing" 10 feet off to her right to play. Not so much that wispy super sharp, clean, artifical and electronic sounding "imaging" from most audio systems, but the sense of a full sized person singing with a throat, a body, and a full sized corporeal trumpet just being "there." Such a surprise to have that experience from such unassuming old-school looking speakers!

I think it is very funny, that designer Shaw is so proud of these flimsy British resonator boxes!

Hmm. It seems odd to me to think it "funny" (or odd?) that Shaw is proud of his speaker design, while quoting a listener report on the apparent success of the design in producing natural sound.

Shaw is proud because he's created a speaker brand that is quite successful and appeals to many people, to the degree of having a particularly devoted following for decades. What's "funny" about that? I'd be proud too!

 

[Deleted member 2944]

Unfortunately for the forum, I would prefer the woofer stay exactly where it is inside the enclosure.

Yes, understood. Thank you for providing the speakers to Amir.
Removing the woofer would likely invalidate the issue anyways, since that would remove the box and port effects from the results seen.

Dave.

 

[dshreter]

Yes, understood. Thank you for providing the speakers to Amir.
Removing the woofer would likely invalidate the issue anyways, since that would remove the box and port effects from the results seen.

Dave.

It was my pleasure! I was only poking fun at the discussion. I’m thrilled to have learned more about these speakers, and that it has prompted so much discussion.

I think it’s been well suited for a discussion of where listener preference or application fits into speaker design. These speakers have flaws and also a lot of fans. To me, the measurements are showing that these speakers likely do fit some people’s true individual listening preferences. There’s also a healthy dose of appealing visual design and construction along with heritage that helps make the sale too, but I’d like to think that’s not the exclusive reason these are enjoyed by many.

I have two systems I listen to, and I’d like to put some objectively superior speakers in the theater and keep these in my living space. I think there will be qualities to appreciate about each in their respective places!

 

[q3cpma]

To everybody saying that these sound great in their specific use case (jazz, rock 'n' roll, etc...), one of the points of well engineered speakers is that they sound good with everything. These would probably shatter if trying to play some Godflesh at a Godflesh suited volume.

I think I'll stay with my "toy speakers" costing toy money for toy performances.

 

[MattHooper]

These speakers have flaws and also a lot of fans.

Which suggests certain "flaws" can be re-considered as "features" not a "bug."

I'm in a position I think most audio-enthusiasts are in: I don't have a high level of technical knowledge about speaker design so I have to sit on the side-lines and watch the issues being hashed out among people who do have more knowledge. But among those people who build speakers, be they manufacturers, DIYers or technically knowledgeable enthusiasts, including on this site, there are still battles over "which approach is best/right." Sometimes I read from these more-knowledgeable types that certain speakers represent "just a wrong, wrong, wrong way to do things!" (Harbeth has often a target). But when I actually hear the results from said speakers they sound "right, right right" and just wonderful to listen to...to me. (No I don't always like "badly designed" speakers BTW). So, while it's fascinating to see the debates and graphs, listening to a speaker and satisfying my personal preference remains my method for selecting speakers. The Harbeths seem to tickle something in my personal preferences as they do in the rest of their audience. (I also quite like aspects of a speaker I now own that are if anything an opposite to the Harbeths).