• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Harbeth coming out with a “digital” speaker

ahofer

Major Contributor
Forum Donor
Joined
Jun 3, 2019
Messages
4,952
Likes
8,698
Location
New York City

Interesting that in a separate conversation I had (and posted here) he poo-poo’d the idea that a prospect’s in-home experience might be improved with measurement and DSP. Alas, such are the marketing imperatives of high end - it’s no good until it is (re)invented here.
We will be following-up this announcement with a formal Press Release in due course, but I thought that I owe it to you to give you a heads-up.

For the last 18 months we have been developing digital equivalents of the existing Harbeth speakers. We first set ourselves a challenge with the most complex model, the three-way Monitor 40.3, and worked our way through the range to the P3ESR and to other concepts. At the Bristol show next week in one of our three adjacent demo rooms (230, 232, 234) we will be demonstrating the Monitor 40-D amongst other speakers, on open frame studio-style stands.

Over the coming weeks we plan to discuss the development of these digital Harbeths in as much detail as we can in a public forum, hint at the sort of technical challenges we encountered and overcame, and consider together where this digital technology takes our brand. Digitising the existing models is a logical but hardly radical move: the M40-D is our 'concept car' that hints at a journey which is already significantly mapped-out. But first, for the avoidance of any doubt, I want to make it absolutely clear that there has been no change in ownership of the Harbeth company, no buy-ins or buy-outs, all R&D has been self-funded from our reserves and most importantly, there have been no tie-ups or commercial or contractual arrangements with any third parties with the exception of expert guidance in PCB physical layout of our prototype circuits to avoid hum and noise problems from digital circuitry sharing a PCB with analogue circuitry.

In short, the entire design has been undertaken in-house with my continuous guidance as to what sounds 'right' as my colleagues experiments with different potential DSP implementations. The team collaboration has worked effectively because we have given each other space and time to experiment and respect each other's contribution, and no timescales of any sort were set. And when they invite me to listen to alternative candidate DSP strategies, I'm confident that they have carefully pre-selected potential winners from a myriad of possibilities. My colleagues know that I'm going to listen carefully and give them constructively feedback in meaningful language that they can understand, absorb and build in to DSP code. Absolutely no hi-if mumbo jumbo guff. No audio poetry. Comments on frequency, level, balance, distortion and integration. They can work with that sort of feedback.

The power of even a moderately sophisticated DSP engine is simply incredible. The smallest lump and bump in the already world-class frequency responses of our Harbeth-made drive units can be completely eliminated. The energy in sub-sub-sub octaves can be minutely adjusted on a Hz by Hz basis to achieve any desired sonic experience. The relative timing between drive units can be adjusted to subsample accuracy (a few microseconds): none of this would be remotely achievable using passive crossover components at any cost or complexity. That said, the investment that we've made in designing sophisticated passive crossovers over the past near 50 years has stood the test of time: the passive solutions have set a high benchmark, and will continue to be honed and refined as our digital solutions trickle down to new passive implementations. Both can and will coincide - forever - as they solve different problems.

There are and were many unknowns at the start of our journey, not all technology matters, many are pragmatic consumer-led issues. Will enough consumers abandon their love-affair with hi-fi amplifiers and cables and opt for an all-in-one, buy-and-forget digital solution? How would such a package make business sense to a hi-fi retailer whose income depends upon continuous upgrades? What sort of connectivity is required at a minimum and how would that allow the continuing use of existing sources (incl. vinyl)?

Then there is the issue of how much access to give to the user to tinker with the Harbeth DSP. This is really a serious matter and our opinion has much altered as we have purchased and analysed many competitive DSP speakers and familiarised ourselves with how they have addressed that question. In every case, the graphical user interface (GUI) that sits between the DSP chip and the user's plugged-in laptop both (reasonably) protects the DSP and speaker but locks-out more than a superficial sub-set of the available DSP facilities. In some cases, the speaker manufacturer's default ex-factory (reset) DSP settings are so bizarre, that the motivated and able user could indeed sort-out and dramatically improve the speaker if only he had deep access to the DSP. So the question is, do you give the user complete access to the DSP, for which he will need a minimum level of acoustics/DSP/loudspeaker design knowledge as the 'real' DSP code is not managed through a pretty consumer GUI but in a much less familiar engineering language - or none at all? Or somewhere in between?

For me, the experience of taking the M40.3 that I designed and know well and digitising it has confirmed, above all, the excellence of the 'raw' electro-mechanical-acoustic properties of the Harbeth-made woofers/midrange drivers and their exclusive material science and injection moulded cones. I assumed, wrongly, that the 40.3 was as about as good at it got for transparency, neutrality, imagery: the 'being there' sensation. But I was wrong.

We'd love to see you at Bristol new week. If you do come, please introduce yourself and allow us to take you on a new sort of journey into sound.
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
Welcome to the XXI Century!
Bristol is only an hour's drive away, but it's been years since I last felt like enduring an audio show...

It doesn't sound like whoever wrote that has a grasp on the DSP side of things.
 
OP
ahofer

ahofer

Major Contributor
Forum Donor
Joined
Jun 3, 2019
Messages
4,952
Likes
8,698
Location
New York City
Welcome to the XXI Century!
Bristol is only an hour's drive away, but it's been years since I last felt like enduring an audio show...

It doesn't sound like whoever wrote that has a grasp on the DSP side of things.
It was Alan himself. You should go over there. I’m listening to my Harbeths now, and I do like them (and I have Revel, KEF, and Genelec in other systems), but it’s frustrating how Alan has one foot firmly in objective science (and is one of the few high end CEOs to do so), but keeps the other foot suspended to stomp on objectively defensible loudspeaker advances not invented at Harbeth or the BBC.

I’d also say he’s not saying much about DSP for in-room application. I mean, what’s the point of tinkering “Hz by Hz” with bass response that gets swamped by room characteristics?
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
It was Alan himself. You should go over there. I’m listening to my Harbeths now, and I do like them (and I have Revel, KEF, and Genelec in other systems), but it’s frustrating how Alan has one foot firmly in objective science (and is one of the few high end CEOs to do so), but keeps the other foot suspended to stomp on objectively defensible loudspeaker advances not invented at Harbeth or the BBC.

I’d also say he’s not saying much about DSP for in-room application. I mean, what’s the point of tinkering “Hz by Hz” with bass response that gets swamped by room characteristics?

Shaw is not an engineer (he has a degree in marketing) and that shows.
 

dualazmak

Major Contributor
Forum Donor
Joined
Feb 29, 2020
Messages
2,820
Likes
2,950
Location
Ichihara City, Chiba Prefecture, Japan
They wrote;
>The relative timing between drive units can be adjusted to subsample accuracy (a few microseconds): none of this would be remotely achievable using passive crossover components at any cost or complexity.

Yes, this is exactly the "DSP time alignment between the SP drivers in sub-msec precision" which I (and many Multichannel-DSP DIY people) have already established in multichannel multi-driver multi-amplifier DSP-base fully active stereo audio project; nothing new for me (please refer to my summary post here) , but I assume the Harbeth concept would be putting all the DSP+multi-amplifier in their multi-driver SP box eliminating passive LCR-XO network (and attenuators). You can find my latest system setup here.

My post here would be also of your reference;
- Not only the precision (0.1 msec level) time alignment over all the SP drivers but also SP facing directions and sound-deadening space behind the SPs plus behind our listening position (i.e. our room acoustic) would be critically important for effective (perfect?) disappearance of speakers:

BTW, I am very much interested in what would be their fully "validated" measurement and (flexible?) tuning method for sub-msec precision time alignment in Harbeth's multi-driver SP(s); hopefully at the listening position in listener's home acoustic environments.
 
Last edited:

FeddyLost

Addicted to Fun and Learning
Joined
May 24, 2020
Messages
752
Likes
542
I think they just wanted to hide obvious fact that they are completely incompetent in digital domain and even cloning transfer function of passive Xover with DSP is tough.
In case of active DSP speaker you don't have any external source of error, everything is up to your design.
And if you are really good "listener" and voicing your speakers with some intent, you have a lot of options to tweak.
It's really hard.
Like you was riding horses for 50 years and now you need to design a rocket engine.
 

Robovox

Member
Forum Donor
Joined
Nov 19, 2018
Messages
27
Likes
59
Location
London
Good to see Harbeth responding to market trends. I heard the 40.D at the Bristol show yesterday and spoke to the engineer who is developing the active versions of all of the speakers. They are essentially cloning the passive crossover, but adding time alignment between the drivers and adding a few "small corrections to the driver response" but not extending the bass response with the DSP / amplifiers beyond what the passive versions do.

I enjoyed the sound, probably as much as anything in the show. Unlike almost every other manufacturer, they did play a variety of material and not just breathy vocal & guitar hifi show music. The speakers were plausibly realistic and un-fatiguing with a variety of material. The sort of speakers that I could enjoy listening to all day long. By "plausibly realistic" I don't damn with faint praise, I just think that's the best you can really hope for listening in stereo in a domestic environment. Or a Hotel bedroom.

Without a side by side comparison active vs passive it's impossible to say how much of an improvement there is.
 
Last edited:

sergeauckland

Major Contributor
Forum Donor
Joined
Mar 16, 2016
Messages
3,440
Likes
9,100
Location
Suffolk UK
I think they just wanted to hide obvious fact that they are completely incompetent in digital domain and even cloning transfer function of passive Xover with DSP is tough.
In case of active DSP speaker you don't have any external source of error, everything is up to your design.
And if you are really good "listener" and voicing your speakers with some intent, you have a lot of options to tweak.
It's really hard.
Like you was riding horses for 50 years and now you need to design a rocket engine.
Except that audio DSP chips like Sharc, which have been around for a good few years, have a whole series of libraries that do all the 'routine' stuff like audio crossovers, equalisers, time delay, none of which is particularly difficult. This massively simplifies software development.

S.
 

Bjorn

Major Contributor
Audio Company
Forum Donor
Joined
Feb 22, 2017
Messages
1,286
Likes
2,562
Location
Norway
I assumed, wrongly, that the 40.3 was as about as good at it got for transparency, neutrality, imagery: the 'being there' sensation. But I was wrong.
What a surprise....
 
OP
ahofer

ahofer

Major Contributor
Forum Donor
Joined
Jun 3, 2019
Messages
4,952
Likes
8,698
Location
New York City
Good to see Harbeth responding to market trends. I heard the 40.D at the Bristol show yesterday and spoke to the engineer who is developing the active versions of all of the speakers. They are essentially cloning the passive crossover, but adding time alignment between the drivers and adding a few "small corrections to the driver response" but not extending the bass response with the DSP / amplifiers beyond what the passive versions do.

I enjoyed the sound, probably as much as anything in the show. Unlike almost every other manufacturer, they did play a variety of material and not just breathy vocal & guitar hifi show music. The speakers were plausibly realistic and un-fatiguing with a variety of material. The sort of speakers that I could enjoy listening to all day long. By "plausibly realistic" I don't damn with faint praise, I just think that's the best you can really hope for listening in stereo in a domestic environment. Or a Hotel bedroom.

Without a side by side comparison active vs passive it's impossible to say how much of an improvement there is.
Any word on amplification & DAC choices?
 

Robovox

Member
Forum Donor
Joined
Nov 19, 2018
Messages
27
Likes
59
Location
London
IMG_2434.jpeg


The power amps were hidden in the black boxes between the speakers. When asked they didn't want to say too much - it's a work in progress so fair enough. Sources and amps in all three Harbeth rooms were nothing exotic; the P3ESRs were being driven with a modestly powered Hegal and a Marantz CD player.
 
OP
ahofer

ahofer

Major Contributor
Forum Donor
Joined
Jun 3, 2019
Messages
4,952
Likes
8,698
Location
New York City
Sources and amps in all three Harbeth rooms were nothing exotic; the P3ESRs were being driven with a modestly powered Hegal and a Marantz CD player.
Yes, He's made a point on the forum of choosing fairly simple amplification, and complains that he can't go even cheaper because some customers would write off the demo if he had, say, a Behringer.
 
OP
ahofer

ahofer

Major Contributor
Forum Donor
Joined
Jun 3, 2019
Messages
4,952
Likes
8,698
Location
New York City
In the thread over at the forum, Alan offers this about the future of active speakers and what he was streaming:
We continuously engaged with visitors about their expectations for a modern quality hifi system at home. Repeatedly I heard words like ‘simplification’, ‘de-clutter’, ‘basic essentials’, ‘getting back to the music not the tweaks…’ and the like.

We left with a definite feeling that there will be a squeeze on accessory and cable manufacturers as the future system seems to be coalescing on active speakers and a streamimg box, and that’s all.

Although we had fixed pre-selected music lists for two of our rooms, in the M40-D room my colleague used a Mac and Spotify, both of which he is comfortable with. It allowed us to take requests. And boy! Did we have some fabulous tracks offered. So I’m a complete convert to streaming now as I can see endless possibilities to expand my musical horizons.

The marketing director of Qobuz took a seat and commented favourably on the sound. He then noticed that it was being streamed and was horrified to discover that it was Spotify at 320kb.

So he sees all-in-one speaker appliances as the future. And this on not using DSP in the demo and a criticism of a listener that “the low end needed more weight”

No, no, no! Not only was the low end unconstrained, any more bass quantity, regardless of the size of the driver, would be bitterly regretted. Deep, tight bass is so subtle that it does not draw attention to itself.

As you might expect, we took appropriate measuring equipment with us and actually checked the frequency response in-room. We found, absolutely predictably from the physical dimensions of the room that the ever-present several dB tall standing waves were there, but chose to leave them uncorrected (we could have notched them out in the DSP, indeed tried it) and felt that given the terrible environment that the bass was just right.

We have no interest whatever in floor standing loudspeakers using conventional wooden boxes.

And

We deliberately did not even consider making communications between the source (laptop with streaming audio) a digital or Bluetooth stream in this prototype because that added needless complexity to the project, and would make it difficult off-site and without test equipment to cope with unexpected problems in the Bristol show such as radio-frequency BT interference from other exhibitors, glitches, splats, distortion, drop-outs and so on. In fact, the option to use analogue was respected by those who asked as a prudent move for a prototype system and indeed, many visitors were actually comforted by our decision as their experience (or their perceived experience) of BT in other exhibitor's rooms gave them concerns.

Given that we're not discussing the detailed implementation of the electronics in these laboratory prototype systems I can talk in general terms about what we did. Perhaps most surprising is that the DSP code to implement a stereo-pair of three-way Monitor 40s only used about 2% of the potential DSP engine in the DSP chips. That means that should we wish to add DSP room correction, there is more than enough capacity to do so.

To be clear: the DSP code in the M40-D as auditioned at Bristol did not include room correction: they represented the lab design as at November 2022 when we gave a small presentation to test their functionality in an unfamiliar room well away from HQ.

BTW, he says there was also analogue crossover hardware in the boxes. It seems like the boxes perform ADC-DSP-(crossover?)-DAC-crossover process, if I’m reading him right. As usual he’s being vague but then accusing others of being unable to understand. And then this marketing assertion that takes a swipe at nearly everyone:

One Harbeth dealer who attended said that his personal frustration was his inability to persuade (using words, not physical violence!) customers who actually were prime potential Harbeth users from not taking the biased advice of various Internet forums and buying other speaker brands when he could predict with certainty that they were wasting their time and money and would be back ‘as sure as the sun follows the moon’ for another expensive bite. Their time and money wasting journey could be foreshortened if they a) ceased to trawl the internet looking for tips and guidance b) explained what sort of personal sound experience that sought but hadn’t yet found c) actually listened to a dealer’s experience and above all, listened for themselves and trusted their own ears.

As you can imagine, we recognise familiar faces at our rooms, and of stories of their sometimes serendipitous acquisition of Harbeth between annual shows. One regular said that last year he was walking past a hifi shop announcing its demise, and there in the window a pair of Harbeth. The planets were in alignment and the credit card quickly sealed the deal to the delight of all!

And when we compared notes over dinner we toasted those CEOs of our competitors who, arriving at our rooms, sat down and to the amusement of all present and said, as they do every year “Sanity!”, returning wearily to their ghetto where loud and inappropriate noise is played endlessly by their sales staff.
 

Willem

Major Contributor
Joined
Jan 8, 2019
Messages
3,658
Likes
5,276
I did not attend but the description of the set up was pretty clear. For convenience sake, since this is a prototype, all electronics are in the external box. This means three class D (chip?) amplifiers, plus the dsp electronics. To allow direct comparison with the old crossover design, the traditional crossovers were also available externally, together with a traditional amplifier. Of course, the final units will have the electronics built in. I take it (not sure) that the speakers will have both analogue and digital inputs.
Alan Shaw is aware that the dsp system allows for programming of room eq filters etc, and he is fully aware of the impact of the room, but thus far he hesitates because he fears consumers may baulk at the complexity involved or mess up the sound quality of his designs. I think he should integrate something like Audyssey or Dirac, preferably only for the lower frequencies, but it is a complex marketing decision. The alternative would be to simply leave it to consumers to use either, e.g., the DSpeaker X2, or add subs and use MSO.
 

Phorize

Major Contributor
Forum Donor
Joined
Apr 26, 2019
Messages
1,533
Likes
2,060
Location
U.K
Yes, He's made a point on the forum of choosing fairly simple amplification, and complains that he can't go even cheaper because some customers would write off the demo if he had, say, a Behringer.
I think there is some sensitivity around Harbeth’s dealer network too. Cheap appliances need to be sold in high volumes to be more than marginal. I’m aware that at least one Harbeth dealer in the U.K reports that he doesn’t shift the volumes required to become a Yamaha distributor for example.
 

DSJR

Major Contributor
Joined
Jan 27, 2020
Messages
3,314
Likes
4,427
Location
Suffolk Coastal, UK
Shaw is not an engineer (he has a degree in marketing) and that shows.
Late reply here as the usual negative vibes come out..

Alan has hearing attuned to driver blending and he seems to have done this very well over the decades. His marketing skills have grown the business very well too AND, he's made it obvious on the HUG that if he needs a skill-set to do a particular development task such as the radial cone material or in this case DSP techniques, he'll hire the brains to do the job!!! On a more practical development level, Spendor's long-standing designer jumped ship a couple of years ago to join Harbeth and as far as I know, he's still there... Spendor, Rogers, Graham and Stirling do seem stuck into resurrecting 1960's, 1970's and 1980's relics to make a bob or two and only use new materials 'cos old ones can't be obtained readily these days. If you'd ever compared LS5/8 and 5/9 to say, a M30-XD or M40-XD, you'd hear how far this brand have come with the concepts. Yes I know, other monitor makers leapfrogged them years ago, but KH420's look so ugly in a domestic environment despite the excellent performance and keen pricing in comparison. The rest of us should give some serious thought to the Dynaudio Focus 50 which impressed so much (in the absence of a full tech appraisal) and cost £8600 or so complete with wireless, A-D, dac and active amps built in if I read the spec right.

These new models will suspect be significant 'baby steps' into a new world of active speakers using DSP. I'm not expecting space age styling or anything hugely signioficant externally (and do please remember Alan was making the HHB Circle actives decades ago and apparently can still service them if needs be, albeit at a price), but I believe him when I read of his subjective views on the music's presentation through such models. It's a given they'll be expensive... Let's hope the DSP side won't be obsolescent in a very few years as rumoured in the ATC thread here - but Harbeth do seem financially secure compared to many so it may not be an issue should future servicing be required.
 

tuga

Major Contributor
Joined
Feb 5, 2020
Messages
3,984
Likes
4,281
Location
Oxford, England
Late reply here as the usual negative vibes come out..

Alan has hearing attuned to driver blending and he seems to have done this very well over the decades. His marketing skills have grown the business very well too AND, he's made it obvious on the HUG that if he needs a skill-set to do a particular development task such as the radial cone material or in this case DSP techniques, he'll hire the brains to do the job!!! On a more practical development level, Spendor's long-standing designer jumped ship a couple of years ago to join Harbeth and as far as I know, he's still there... Spendor, Rogers, Graham and Stirling do seem stuck into resurrecting 1960's, 1970's and 1980's relics to make a bob or two and only use new materials 'cos old ones can't be obtained readily these days. If you'd ever compared LS5/8 and 5/9 to say, a M30-XD or M40-XD, you'd hear how far this brand have come with the concepts. Yes I know, other monitor makers leapfrogged them years ago, but KH420's look so ugly in a domestic environment despite the excellent performance and keen pricing in comparison. The rest of us should give some serious thought to the Dynaudio Focus 50 which impressed so much (in the absence of a full tech appraisal) and cost £8600 or so complete with wireless, A-D, dac and active amps built in if I read the spec right.

These new models will suspect be significant 'baby steps' into a new world of active speakers using DSP. I'm not expecting space age styling or anything hugely signioficant externally (and do please remember Alan was making the HHB Circle actives decades ago and apparently can still service them if needs be, albeit at a price), but I believe him when I read of his subjective views on the music's presentation through such models. It's a given they'll be expensive... Let's hope the DSP side won't be obsolescent in a very few years as rumoured in the ATC thread here - but Harbeth do seem financially secure compared to many so it may not be an issue should future servicing be required.

The HUG is one of greatest achievements.
 
OP
ahofer

ahofer

Major Contributor
Forum Donor
Joined
Jun 3, 2019
Messages
4,952
Likes
8,698
Location
New York City
Let's hope the DSP side won't be obsolescent in a very few years as rumoured in the ATC thread here
Missed this. Where?
 

DSJR

Major Contributor
Joined
Jan 27, 2020
Messages
3,314
Likes
4,427
Location
Suffolk Coastal, UK
Missed this. Where?
Tales of chips being replaced by the manufacturers every few years and so on. Maybe just scaremongering in that thread.
 

kemmler3D

Major Contributor
Forum Donor
Joined
Aug 25, 2022
Messages
3,013
Likes
5,614
Location
San Francisco
he sees all-in-one speaker appliances as the future.
I mean, not exactly a dramatic prediction since they have been dominating the market (outside of the audiophile world) for more than 10 years now. :facepalm:
 
Top Bottom