• 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!

Neumann KH 80 DSP Monitor Review

Status
Not open for further replies.
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
The book is published recently but it's based primarily on research from the 80's and 90's.
The third edition of the book is quite updated including developments in the last few years.

That aside, objective modelling of audible non-linearities is a hard problem. People love to invent new metrics they hope will make them famous. :) But ultimately, it is not easy to solve. To wit, every lossy encoder has a perceptual model of hearing to control its encoding parameters yet listening tests rule to determine level of audibility. For a first level estimation, PSNR and such are used as quick shortcuts.

The other challenge for us risk of damaging speakers that are loaned. Problems arise at very high SPLs and Klippel tests have inputs for voice coil temps and such to avoid destruction.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,736
Likes
16,158
Why would you listen to any speaker indoor since that massively changes the low frequency response? Which is detrimental to sound quality?

Why the heck do you think I am measuring speakers using anechoic techniques? It is so that we have objective, room independent analysis of the speaker sound.

Plenty of people slap these things on their desk and use them that way. That is by far more common than the people who use any type of EQ, much less that 150 Hz correction based on potentially faulty evaluations years ago.
Yes, some people do so, it's their problem but not serious conditions for a listening comparsion tests due to the several times repeated reason.
And no, it's not the same like placing them on a well treated room far away from the walls on stands on a well treated room, the 150 Hz peak of the close surface gives a very strong coloration of the sound, also because its already in the root tones region.
And I didn't complain about the measuring speakers using anechoic techniques, only such listening tests which also Harman woudn't agree with.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,736
Likes
16,158
Current research doesn't back your point of view. As I mentioned read the AES paper, "The Correlation Between Distortion Audibility and Listener Preference in Headphones"

Steve Temme , Sean E. Olive , Steve Tatarunis , Todd Welti , and Elisabeth McMullin, Listen, Inc., Boston, MA, 02118, USA
Harman International Industries, USA

The study involved equalizing four headphone to the same frequency response and then emulating that on a neutral headphone. Result was that distortion was not a predictor of preference even in that situation. With respect to your IMD comment, here is the paper:

View attachment 46992

View attachment 46993

View attachment 46994

This is to be expected because these traditional methods of distortion are not psychoacoustically aware. This is why all of my electronics measurements include full spectrum of the distortion so that we can apply psychoacoustics to them. A blunt distortion metric of IMD/THD, etc. is not that useful other than something to chew on when you are bored. :)
I know that paper, the psychoacoustics of distortions are not fully understood yet, but measuring them like with your electronics tests (which I really like) are also important as you know. Just for an example, do a listening a simple compact 2 way like the 308 cranked relatively high (90dB average in a meter which can have small peaks till 110dB) with bass heavy music that has at the same time voices piano etc. and then the same with a very good 3-way monitor like the KH310, trust me, anyone hears such a large difference.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,736
Likes
16,158
And opposite of my experience. I suggest having a blind test done to make sure you can even hear that effect let alone have it be the worst offender. Room modes substantially destroy fidelity of music. No way this narrowband peak has the same effect. Floor reflections which is an analog of this has been shown to not be as relevant as one may think. See Dr. Toole and my response a bit later: https://www.audiosciencereview.com/...y-without-measurement.7127/page-3#post-162809
I can directly switch on and off my 150 Hz filter on my desktop setup via my EQ Apo, the difference as said is large. Floor reflections from a floorstander or bookshelf on stands are usually in different frequency region and not as sounding pronounced/dominant.
 

Andreas007

Active Member
Joined
Mar 11, 2019
Messages
137
Likes
353
Location
Germany, Bavaria
They measure based on certain assumptions (distance from listener, height, surface height, etc.). In this day and age when we have easy and free tools for measuring, and much more flexible EQ than just a few dumb switches, I see no need to use them at all.

I like the switches and also use them, despite having an ADI 2 Pro. o_O I don‘t want to measure my living room and fiddle around with EQ.

Thanks for the measurement of KH80, though.

Regarding listening test: I would suggest not to directly compare two speakers in a review (measurement yes, informal listening no). Like with DACs.

I want to add something about the higher price: Overall quality seems clearly above average to me. Small form factor / performance ratio, too. Noise performance...
And „Made in China“ vs „Made in Ireland“ - which is important for some people and their decision what to buy.
 
Last edited:

Juhazi

Major Contributor
Joined
Sep 15, 2018
Messages
1,717
Likes
2,897
Location
Finland
I know that paper, the psychoacoustics of distortions are not fully understood yet, but measuring them like with your electronics tests (which I really like) are also important as you know. Just for an example, do a listening a simple compact 2 way like the 308 cranked relatively high (90dB average in a meter which can have small peaks till 110dB) with bass heavy music that has at the same time voices piano etc. and then the same with a very good 3-way monitor like the KH310, trust me, anyone hears such a large difference.

I must support this. And in other words, "we" want to see distortion spectrum of a loudspeaker at high spl (like 95dB/1m) to see the when and how the speaker gets stressed with hig spl. This is very important for small speakers, not that much for hifi 3-ways. A small woofer or a tweeter crossed too low will show high distortion in low freq or around crossover. High spl is very common with "hifi" listening and may even lead to damage of driver units with poor design.

http://www.klippel.de/cn/know-how/measurements/nonlinear-distortion/harmonic-distortion.html
csm_dis_maximal_spl_level_8545243c8d.jpg


Audibility of distortion in loudspeaker is poor. Giving some single % is meaningless. I have seen/heard Dr. Klippel to demonstrate it "live" at an AES Congress in 2013!

http://www.klippel.de/listeningtest/

Klippel 2013.jpg
 

Attachments

  • siggi juha 2.jpg
    siggi juha 2.jpg
    1 MB · Views: 133
Last edited:

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
Agreed, distortion, especially LF distortion of those small speakers accelerates at low frequencies with rocket speed. One must not be surprised if more than 100% distortion is measured - because fundamental frequency would be lower in amplitude than H2. These small speakers are nothing to be used for serious listening.

LF distortion is to be measured in nearfield, close to the membrane, otherwise room modes and reflections would mask the result.
 

DDF

Addicted to Fun and Learning
Joined
Dec 31, 2018
Messages
617
Likes
1,355
And opposite of my experience. I suggest having a blind test done to make sure you can even hear that effect let alone have it be the worst offender. Room modes substantially destroy fidelity of music. No way this narrowband peak has the same effect. Floor reflections which is an analog of this has been shown to not be as relevant as one may think. See Dr. Toole and my response a bit later: https://www.audiosciencereview.com/...y-without-measurement.7127/page-3#post-162809

Other standing waves could affect the frequency region of this 150 Hz peak, agreed. In my set up, they dont, and EQing the peak is very obvious.

However, this situation is analogous to the difference between pushing your speakers against the wall or listening to them pulled into the room. While each room will have different standing wave patterns, against wall will almost always have more exagerated bass regardless of the other room modes. Similar holds for on desk effect.

Regarding the floor reflection, I spent 4 hours pouring over Bechs papers this weekend, and 1st Edition of Dr Tooles chapter on reflections. Dr Tooles conclusion about the floor reflection is not supported by Bechs DBTs. But I'll save that for another thread for some other day. First I'd like to see if Dr Toole brings forth more data or changes his conclusions in the 3rd edition, which I dont have
 

MZKM

Major Contributor
Forum Donor
Joined
Dec 1, 2018
Messages
4,240
Likes
11,462
Location
Land O’ Lakes, FL
I must support this. An in another words, "we" want to see distortion spectrum of a loudspeaker at high spl (like 95dB/1m) to see the when and how the speaker gets stressed with hig spl. This is very important for small speakers, not that much for hifi 3-ways. A small woofer or a tweeter crossed too low will show high distortion in low freq or around crossover. High spl is very common with "hifi" listening and may even lead to damage of driver units with poor design.

http://www.klippel.de/cn/know-how/measurements/nonlinear-distortion/harmonic-distortion.html
csm_dis_maximal_spl_level_8545243c8d.jpg


Audibility of distortion in loudspeaker is poor. Giving some single % is meaningless. I have seen/heard Dr. Klippel to demonstrate it "live" at an AES Congress in 2013!

http://www.klippel.de/listeningtest/

View attachment 47053
Showing max SPL at given distortion % is cool, but too many people will argue whether 1% is too high for treble and 10% is too high/low for bass (some tests have shown near 100% THD at 20Hz to be the threshold for typical music).
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
I don‘t want to measure my living room and fiddle around with EQ.

Revel Performa3 F208, definitely a very good speaker.

Here is how it measures anechoically:

F208.JPG


And here is how it measures in a treated dealer's room with no EQ:

Capture.JPG


As you can see peak-to-peak variations are a whooping 28dB (43Hz vs 95Hz), and that is what you are really getting from your speaker which anechoically measures very well when you don‘t want to measure your living room and fiddle around with EQ.
 
Last edited:

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
I must support this. An in another words, "we" want to see distortion spectrum of a loudspeaker at high spl (like 95dB/1m) to see the when and how the speaker gets stressed with hig spl. This is very important for small speakers, not that much for hifi 3-ways. A small woofer or a tweeter crossed too low will show high distortion in low freq or around crossover. High spl is very common with "hifi" listening and may even lead to damage of driver units with poor design.

http://www.klippel.de/cn/know-how/measurements/nonlinear-distortion/harmonic-distortion.html
csm_dis_maximal_spl_level_8545243c8d.jpg


Audibility of distortion in loudspeaker is poor. Giving some single % is meaningless. I have seen/heard Dr. Klippel to demonstrate it "live" at an AES Congress in 2013!

http://www.klippel.de/listeningtest/

View attachment 47053

I'm guessing most of us are aware that when listening music at SPL which equals 95dB/1m MF drivers and especially tweeters are not really producing that SPL, so maybe a realistic test should apply some "averaged" music frequency spectrum when measuring THD over the entire frequency range.
 

Absolute

Major Contributor
Forum Donor
Joined
Feb 5, 2017
Messages
1,084
Likes
2,125
Regarding listening test: I would suggest not to directly compare two speakers in a review (measurement yes, informal listening no). Like with DACs.
I disagree. Listening tests without comparison is worthless imo.

Quality is relative to something, so it only makes sense to use a known reference so that both Amir and all readers have the same reference point to judge respective qualities from. It's very good that this reference point is cheap, have good objective sound quality and is easy for most people to borrow an ear to so that we can understand where Amir is coming from and thus know what it means when he describes certain aspects of speakers relative qualities.

Without a reference, it's all fantasy talk.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,736
Likes
16,158
I'm guessing most of us are aware that when listening music at SPL which equals 95dB/1m MF drivers and especially tweeters are not really producing that SPL, so maybe a realistic test should apply some "averaged" music frequency spectrum when measuring THD over the entire frequency range.
Yes, that's why Prof. Goertz and many others use for multitone distortion measurements (IMD and other) a signal according to the EIA-426B standard which is decreasing from 1 kHz and thus representing more realistically the spectrum of most music which can be seen also in this article
https://www.hifi-selbstbau.de/grund...6/81-musik-qvergleichenq-mit-dem-waveanalyzer

MLT-096.png TrackAllAvg.png
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,736
Likes
16,158
I must support this. An in another words, "we" want to see distortion spectrum of a loudspeaker at high spl (like 95dB/1m) to see the when and how the speaker gets stressed with hig spl. This is very important for small speakers, not that much for hifi 3-ways. A small woofer or a tweeter crossed too low will show high distortion in low freq or around crossover. High spl is very common with "hifi" listening and may even lead to damage of driver units with poor design.

http://www.klippel.de/cn/know-how/measurements/nonlinear-distortion/harmonic-distortion.html
csm_dis_maximal_spl_level_8545243c8d.jpg


Audibility of distortion in loudspeaker is poor. Giving some single % is meaningless. I have seen/heard Dr. Klippel to demonstrate it "live" at an AES Congress in 2013!

http://www.klippel.de/listeningtest/

View attachment 47053
Yes and we must keep in mind that while the harmonic distortion itself might not be hearable, a driver with high such distortion can also have high intermodulation distortion which is especially problematic at midwoofers which play also the psychoacoustically sensitive mids.

Here is a nice example with simulations and sound samples of 2 drivers with almost the same harmonic distortion where one exhibits though much higher IMD due to differences in their BL function.
https://purifi-audio.com/tech/ (click on the 3rd article from the top from Dec 12th, 2019 with the title Low frequency harmonic distortion is almost inaudible. So what’s the point of low distortion drivers?)

Their conlusions are:
  • Similar distortion figures can hide very different manifestations.
  • Intermodulation measurements are indispensable for evaluating loudspeaker drivers.
  • BL as a function of position is a critical performance specification that should never be left out of a data sheet.
  • The same goes for BL as a function of current or its equivalent, position dependent self-inductance (see AES paper).
  • The stiffness curve of the suspension of a woofer can be designed purely based on practical requirements of robustness and power handling.
  • For pure subwoofers, low distortion is optional.
  • With their limited control bandwidth (200Hz or so), motional feedback (MFB) systems can only improve distortion in the subwoofer frequency range but are powerless against midrange distortion, including IMD. This explains the limited take-up of this technique.
  • Anyone asking “can we ever hear distortion mechanism x when elsewhere in the signal chain distortion from mechanism y is so much greater” should look at what types of distortion either mechanism can and cannot produce. Only then can they know if one distortion mechanism is likely to mask the other.
(and finally)

The science was right: even gross bass distortion is remarkably difficult to hear. The fine print says yes, provided it stays in the bass.
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,779
Location
Oxfordshire
Wankels are not flawed. They just have different tradeoffs versus other ICEs. If I were to build a small, light sports car, a Wankel would be a great choice since it's so small, light, and powerful.
In as much as no ICE has no flaws you are slightly correct.
The Wankel flaws:-
Combustion chamber shape - shite.
Gas sealing, super difficult to achieve because its shape and thermal expansion are incompatible. A complex system with a relatively short life is slightly less shite than the original.
Open ports - look good when drawn fully open, as they usually are, but at all times other that than around that instant they are a an orifice with sharp edged opening - which is shite for gas flow.
At the moment of ignition the difference in flame path distance to the nearest chamber wall and to the furthest is massive = shite combustion.
Because the combustion and sealing are poor the pollution and specific fuel consumption are both shite.
Yes, it is light. Yes it is mechanically smooth.

If it was any good lots of people would use it. NSU (Audi) did the sensible thing (I really loved the Ro80 but I wasn't an experienced car engineer then...). Mazda probably have some old senior executive with an interest of some sort, maybe to to have a unique product.
I really like the RX8 but would be unwilling to pay for all that fuel...
 

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
Sorry @DjBonoBobo that's exactly what I meant! Brain fart. And @amirm - that would probably explain the 2K dip then, at least. Per manual:

View attachment 46848

Small difference, not saying it's necessarily worth remeasuring, but neumann is one of the few that provide an exact acoustic reference, so it's worth noting, especialy going forard. Lots of speakers, especially among those with waveguides, have their reference axes somewhere around where the waveguide and woofer meet. The KH80 is well controlled enough that it's a small difference, but others, like the Buchardt S400, showed pretty massive deviations just a few degrees off (the S400 then straightens out at steeper off axis angles, but just giving an example).


Was this post addressed? In any case when measuring in the nearfield not starting at the proper acoustic axis would affect the spinorama performance. Any chance to remeasure the loudspeaker in accordance with this information?
 

bobbooo

Major Contributor
Joined
Aug 30, 2019
Messages
1,479
Likes
2,079
I saw that and agree it's a good idea but it's only half the equation, which is the measurement side. When doing actual listening tests, you have to decide if you want to run every speaker full range or normalize the bass to not give the bigger one an advantage. This is a personal choice and depends on how you're using them but generally people buying a speaker with a 4" driver are going to be using a sub.

True, but as Amir has said multiple times, these are informal listening tests. Unless he does controlled, exactly volume-matched, double-blind ABX tests with multiple listeners and a large number of speakers, in my opinion they're only useful at the moment for flagging up obvious problems that the Klippel system doesn't show, like the audible hissing from the JBL 305P. There are just way too many variables at the moment to draw any firm conclusions from such informal listening tests, like speaker placement / listener position; room / desk acoustics; frequency response accustomization; volume imbalance; listening material; sighted, choice-supported and expectation bias etc. Not to mention sub integration as you said.

I think it's also necessary though to have some kind of listening tests just as a token gesture to placate subjectivists, otherwise they may be put off from reading the reviews entirely - this way they could actually learn something about the science and measurement of speakers they might not have if it was a purely objective review.

But instead of arguing over informal listening tests that were never intended or claimed to be scientific, we should be concentrating on the actual measurements and preference ratings here, and getting those to be as precise, detailed, and scientific as we can, to further this site's reputation as one of real, genuine audio science (reviews), which differentiates it from the sea of subjectivist snake-oil infested websites, 'hi-fi' magazines and YouTube channels that have dominated the industry for years. These thrive off endless opinion-based arguments that don't actually reveal any useful truths about audio reproduction, instead of the real academic research and measuring techniques like Klippel's NFS and Sean Olive's predicted preference ratings we're using here, which have been proven to highly correlate with subjective reported speaker preferences, and are based on decades of solid science and psychoacoustics in the field.
 
Last edited:

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
True, but as Amir has said multiple times, these are informal listening tests.

If he was to do "formal" listening tests in a non-anechoic environment the speakers would have to be properly room EQ-ed to the same response precison and aligned to the same target curve otherwise there would be not much use of a listening tests.
 

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,109
Likes
8,420
Location
NYC
The best theory was mentioned earlier in the thread: it's possibly the DSP compressor/limiter controlling bass distortion by limiting the bass response at high SPLs. A good illustration of this effect is shown in the response vs level of a bluetooth speaker, for example:
http://www.oluvsgadgets.net/2015/09/review-jbl-flip-3-ue-boom-killer.html
View attachment 46964

FWIW I authored a patent on this: https://patents.google.com/patent/US6201873
and this response accurately represents what a compressor with that function does.

Thanks @DDF, I did think this was the most likely explanation. 90dB @1m (a realistic distance for a nearfield monitor) is quite loud =]

But the above comment made me take a second glance at the OP, so question @amirm. You said you did the measurement at 106dB @ 0.3m, equivalent to 90dB at 1m. On second glance that doesn't seem right. 106dB at 0.3m would be 95.5 dB, which would be very loud indeed and certainly explain the bass dip due to compression on a small monitor.

Snag_1bb51875.png


Typo, or am I missing something? Sorry, not trying to be a pain... I know speaker measurements and data presentation are a lot of work, even with the Klippel.

Microphone curve based the five numbers variation in high frequencies amirm shared including a 3,5Hz 1st order high pass knee based datasheet of MK 255, on a 50dB scale it doesn't mean much calibrate for response or not :) it has beatiful low end reach where NFS seems to run over anechoic data in resolution, for @napilopez puzzle of variations in the various KH 80 curves added one standard UMIK-1 into graph :).

View attachment 47028

Thanks for sharing =] I assumed the mic was fine. Umik-1 is good enough for EQ in home theater purposes, but for measurements I got one calibrated from Cross Spectrum labs - which I believe calibrates the mic to an ACO pacific 7052E. Anyway, the mics definitely aren't a significant reason for the different results.

If he was to do "formal" listening tests in a non-anechoic environment the speakers would have to be properly room EQ-ed to the same response precison and aligned to the same target curve otherwise it would be not much use of a listening test.

I definitely think its best to accept the listening tests as informal for now now. This is something researchers spend a lot of effort doing.
 
Last edited:

bobbooo

Major Contributor
Joined
Aug 30, 2019
Messages
1,479
Likes
2,079
If he was to do "formal" listening tests in a non-anechoic environment the speakers would have to be properly room EQ-ed to the same response precison and aligned to the same target curve otherwise there would be not much use of a listening tests.

As I mentioned in my post, one of the many variables that would need to controlled for is room acoustics. So all this fervent talk about the listening tests in their current form is pointless. Let's move on now and focus on the measurements and preference ratings, as the whole point of Sean Olive's preference formula is that he and other academic researchers have already done the large, controlled listening tests which correlate subjective preferences to objective measurements with high accuracy, so all we have to do is get the measurements and rating calculations correct.
 
Last edited:
Status
Not open for further replies.
Top Bottom