• 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). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Neumann KH80 DSP Monitor Measurements #3

hyperplanar

Active Member
Joined
Jan 31, 2020
Messages
177
Likes
304
Location
Los Angeles
#21
Ah forgot to mention that the final measurements are for reference axis and not tweeter. This makes some difference here because speaker is so small.
Thank you for taking the time to measure this speaker (yet) again! I get why it may have felt like you were coming under fire over this whole saga, but honestly as an observer I think most people just wanted to resolve why such differences, however small, were occurring in the first place with what is supposed to be a state-of-the-art measurement system that can be more accurate than an anechoic chamber. Determining and/or resolving these sources of variation only serve to increase the trustworthiness and reliability of ASR’s measurements which is a great thing in my opinion. We also now have some data on the effect of temperature on speaker performance as well, which seems to be a tad bit more than expected in some cases.

Going forward, can we expect the correct reference axis specified by each manufacturer to be used in calculating the measurements, or is it too much of a hassle? I understand the practical reasons for the mic protection cage being left in place but maybe a disclaimer should be included in the reviews regarding potential error sources if people are not to interpret the measurements as absolute.
 

digicidal

Major Contributor
Joined
Jul 6, 2019
Messages
1,372
Likes
2,117
Location
Sin City, NV
#22
To me the real "gem" of this review is the fact that Neumann runs and stores the FR of every monitor they produce and is willing to reference them in order to troubleshoot a nearly meaningless difference in response under completely different conditions (the largest of which is constantly variable). Sure most of us find similar temperatures comfortable... but even that difference might be statistically significant.

On the other hand... this is more time that @amirm did not spend reviewing the SDP-55... which to me is a far greater tragedy than if all 3 pairs of the KH80's had caught fire and burned up on the Klippel during testing. One man's ceiling... ;)
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
29,101
Likes
80,513
Location
Seattle Area
Thread Starter #23
Going forward, can we expect the correct reference axis specified by each manufacturer to be used in calculating the measurements, or is it too much of a hassle? I understand the practical reasons for the mic protection cage being left in place but maybe a disclaimer should be included in the reviews regarding potential error sources if people are not to interpret the measurements as absolute.
The cage was shown to not make a difference so don't think we need a notice for that. On reference axis, it is easy to do but I need to know about it in advance. Something about reading the manual for a speaker seems odd to me to find that info. :)
 

napilopez

Major Contributor
Joined
Oct 17, 2018
Messages
1,385
Likes
4,445
Location
NYC
#24
Seems that if you want to create a riot in the streets of ASR, all you have to do is show a non-ruler flat response of Neumann KH80 DSP professional monitor (speaker). Not knowing that, that is precisely what I did in my first review of KH80 DSP:



This was one of the earliest speakers I measured with the Klippel Near-field scanner. So naturally (or not), there were a lot of suspicions that it could actually recreate anechoic chamber measurements such as provided by the company:



After what seemed like a hundred pages and numerous attacks from every corner on validity of measurements including me not even being dressed right to measure speakers, I decided to test a second KH80 DSP monitor sample that another member was kind enough to offer. Was that the end of it? Of course not. As I said, if the measurements were not ruler flat, folks were not satisfied and this is what we had:



A sin was committed once more with the same dip around 80 Hz or so. Protests ensued and library books burned in streets with assertion that ASR measurements were flat wrong.

After some 25 pages of posts, folks seemed to forget about the fight and thread went dormant. Until... we heard back from Neumann which had asked Klippel Germany to measure an identical sample to what they had measured in anechoic chamber. While there were some differences between the two measurements, they were closer to each other than mine was. So folks went for the jugular this time. Nothing short of repenting on my part and accepting that my measurements were bogus would satisfy. The proof as they said, was right there.

When someone kindly suggested to fund the purchase of a third sample, I decided to not make it the community's problem and purchased that myself. Meanwhile @GuyLayfield reached out privately to me offering that we work together to see what is going on. Suspicion quickly went after the temperature difference for when I tested mine (in winter) and theirs (temperature controlled at 21 degrees C). Guy showed me a couple of useful slides from Klippel indicating changes in speaker parameters based on temperature:

View attachment 72743

View attachment 72744

Meanwhile my sample arrived and measured it. And what do you know, the low frequency dip had disappeard:

View attachment 72746

We still had to verify the problem was temperature. So Guy put a unit in the fridge overnight and then made measurements in the morning every 3 minutes as the speaker and its components warmed up. Here are his results, normalized to 21 degree C (black line):

View attachment 72747

As we see in the bottom blue line, lower temp most definitely created the dip around 80 Hz and gradually disappeared as the speaker warmed up. This then showed that my measurements are indeed accurate but were simply impacted by the lower temperature at the time.

Guy, knowing you all, wanted to still investigate some variations in measurements, some of which were also in Klippel Germany measurements. As it happens, they have measurements for each speaker sold and that included my sample. So we set out to try to reduces differences.

The guilty party here appeared to be the microphone protection cage that I use. This is a sensor shroud around the mic with micrswitches that detect any impact to any object (e.g. speaker) and instantly shuts down the system. It avoids banging an expensive microphone against an expensive speaker! Quite a while ago I had noticed that it was causing some comb filtering due to reflections. I experimented with absorption around it and got it under control. Still, we thought maybe it is at fault.

So I took out the cage and sadly as soon as you do that, all alignment of the system is gone. So while I was at it, I re-adjusted everything better than ever before, and even worked on optimizing the signal processing for least amount of error. Here is that measurement:

View attachment 72749

Yes, it looks the same! It is a bit flatter in bass but I think it has tiny variations here and there. According to Guy, we are now with +-0.6 dB from their measurements.

We don't know if the variations are because my measurements are wrong, or theirs. There is no gold reference here.

Frankly, I am sick of it at this point so told Guy that we are done and I should just post the results. People need to remember that every measurement system has variability and there is fair bit of that between anechoic chambers as well. For the purposes of evaluating speakers, what we have is more than good enough and far, far better than anything any reviewer is producing.

Importantly the moment the speaker is put in a room, nothing remotely flat comes out of it. So we need to keep that in perspective.

Conclusions
The main variable in our measurements not quite matching Neumann's was temperature. Whether this impacted just this speaker or the others, I don't know. If I get some time (yeh right), I will go back and remeasure a few of the other ones I still have and see what difference there may be. Come this winter, I will heat up the measurement room to get it closer to spring weather at least. Fortunately it never gets that cold here and hopefully we will find that other speakers are not affected as much.

Beside temperature, I have also fine tuned the system, using averaging for example to reduce noise level, deploying more optimal parameters to get the high frequencies right, etc. These have been in place for months so the bulk of what we have measured are accurate as they also enjoy warmer temperatures.

@MZKM has a new score for this speaker which he will post. New spin data is enclosed.

Be on notice that I am not measuring this speaker again so don't even think about asking me to do that. As it is, I have to now rip the system apart again and put the cage back on it and re-align it all.

------------
As always, questions, comments, recommendations, etc. are welcome.

As you can tell I am quite grumpy because of this ordeal. Fortunately it is not something where a few hundred dollars won't fix. So please donate what you can using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
Thank you @amirm. This is tremendous work, and it's wonderful you and Guy were able to isolate the meaningful differences. I certainly don't think anyone could reasonably ask for more - it was always the wide deviations that bothered people more than the small squiggles. Sorry for all the trouble we put you through :). But I think we all learned a ton from this process. I'd honestly had never known temperature could have such a difference. Note to self: don't listen to dupstep in a refrigerator:D
 

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
2,152
Likes
1,996
Location
SoCal
#25
I keep my HT room at 73 degrees Fahrenheit. Could you please retest this speaker and every other speaker at this temperature? Thanks in advance!
I'm going to be modest in my request, I just want to know if the 800hz dip becomes a bump at higher temperatures.
 

hyperplanar

Active Member
Joined
Jan 31, 2020
Messages
177
Likes
304
Location
Los Angeles
#26
The cage was shown to not make a difference so don't think we need a notice for that. On reference axis, it is easy to do but I need to know about it in advance. Something about reading the manual for a speaker seems odd to me to find that info. :)
Ah, I see. Different strokes for different folks, but nowadays I always try to at least take a glance at the manuals for whatever I buy. I've been burned too many times by assuming I could figure things out completely myself! :oops:

If you end up measuring the KH120s, the reference axis is also the midpoint between the centers of the woofer and tweeter.
 

ROOSKIE

Senior Member
Joined
Feb 27, 2020
Messages
473
Likes
686
#27
@amirm I don't know why this would change, but do you still prefer the JBL 3 series speaker to this one? (subjectively) Is either one more capable of low distortion+higher SPL over 100hrz (once high passed to a sub?)
 
Last edited:

phoenixdogfan

Addicted to Fun and Learning
Joined
Nov 6, 2018
Messages
850
Likes
739
#28
So all reviews going forward will not just test at 30 degrees off-axis, but at 30 degrees Celsius also.

You could put one speaker on ice while testing the other... the possibilities are endless!
In the old days you used to have to make a speaker bend over to take its temperature.
 

richard12511

Major Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
1,475
Likes
1,881
#29
For those sending in speakers for testing, I think it would be a good idea to figure out the reference axis before sending it in so that you can communicate that with amir before the test happens. Maybe shoot an email to the manufacturer(if they have good CS) or see if you can find any info in the manual(if you still have it).
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
29,101
Likes
80,513
Location
Seattle Area
Thread Starter #30
For those sending in speakers for testing, I think it would be a good idea to figure out the reference axis before sending it in so that you can communicate that with amir before the test happens. Maybe shoot an email to the manufacturer(if they have good CS) or see if you can find any info in the manual(if you still have it).
Now, there is a good idea!
 

AudioStudies

Senior Member
Joined
May 3, 2020
Messages
404
Likes
175
#31
Where and when will it all end? Will we develop computer models for the thermal conductivity of all the speaker components to determine how they relate back to speaker measurements? After all some people prefer a warm house and others a cooler environment. Or since the room will change the performance anyway, are we just spinning in circles? Are there any mandates for what temperature should be used in the chambers? With powered monitors, won't the temperature be dependent on what type of music is being played and the volume? I very much appreciate ASR and others with knowledge and equipment to investigate these issues. More than that I appreciate the dedication of those who work so hard. But is there a point of no return? Are we falling into a rabbit hole that may have no carrots at the bottom? Can science sometimes go too far into a gnat's ass?
 

richard12511

Major Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
1,475
Likes
1,881
#32
@amirm I don't know why this would change, but do you still prefer the JBL 3 series speaker to this one? (subjectively) Is either one more capable of low distortion+higher SPL over 100hrz (once high passed to a sub?)
Going by Olive's metric, using a standard deviation of 0.8(from the study), and using maths(thanks @edechamps for helping me to understand), I would say that there is a 97.44(repeating of course)% chance that amir will prefer the Neumann.
 
Joined
Feb 6, 2020
Messages
88
Likes
114
#36
Seems that if you want to create a riot in the streets of ASR, all you have to do is show a non-ruler flat response of Neumann KH80 DSP professional monitor (speaker). Not knowing that, that is precisely what I did in my first review of KH80 DSP:



This was one of the earliest speakers I measured with the Klippel Near-field scanner. So naturally (or not), there were a lot of suspicions that it could actually recreate anechoic chamber measurements such as provided by the company:



After what seemed like a hundred pages and numerous attacks from every corner on validity of measurements including me not even being dressed right to measure speakers, I decided to test a second KH80 DSP monitor sample that another member was kind enough to offer. Was that the end of it? Of course not. As I said, if the measurements were not ruler flat, folks were not satisfied and this is what we had:



A sin was committed once more with the same dip around 80 Hz or so. Protests ensued and library books burned in streets with assertion that ASR measurements were flat wrong.

After some 25 pages of posts, folks seemed to forget about the fight and thread went dormant. Until... we heard back from Neumann which had asked Klippel Germany to measure an identical sample to what they had measured in anechoic chamber. While there were some differences between the two measurements, they were closer to each other than mine was. So folks went for the jugular this time. Nothing short of repenting on my part and accepting that my measurements were bogus would satisfy. The proof as they said, was right there.

When someone kindly suggested to fund the purchase of a third sample, I decided to not make it the community's problem and purchased that myself. Meanwhile @GuyLayfield reached out privately to me offering that we work together to see what is going on. Suspicion quickly went after the temperature difference for when I tested mine (in winter) and theirs (temperature controlled at 21 degrees C). Guy showed me a couple of useful slides from Klippel indicating changes in speaker parameters based on temperature:

View attachment 72743

View attachment 72744

Meanwhile my sample arrived and measured it. And what do you know, the low frequency dip had disappeard:

View attachment 72746

We still had to verify the problem was temperature. So Guy put a unit in the fridge overnight and then made measurements in the morning every 3 minutes as the speaker and its components warmed up. Here are his results, normalized to 21 degree C (black line):

View attachment 72747

As we see in the bottom blue line, lower temp most definitely created the dip around 80 Hz and gradually disappeared as the speaker warmed up. This then showed that my measurements are indeed accurate but were simply impacted by the lower temperature at the time.

Guy, knowing you all, wanted to still investigate some variations in measurements, some of which were also in Klippel Germany measurements. As it happens, they have measurements for each speaker sold and that included my sample. So we set out to try to reduces differences.

The guilty party here appeared to be the microphone protection cage that I use. This is a sensor shroud around the mic with micrswitches that detect any impact to any object (e.g. speaker) and instantly shuts down the system. It avoids banging an expensive microphone against an expensive speaker! Quite a while ago I had noticed that it was causing some comb filtering due to reflections. I experimented with absorption around it and got it under control. Still, we thought maybe it is at fault.

So I took out the cage and sadly as soon as you do that, all alignment of the system is gone. So while I was at it, I re-adjusted everything better than ever before, and even worked on optimizing the signal processing for least amount of error. Here is that measurement:

View attachment 72749

Yes, it looks the same! It is a bit flatter in bass but I think it has tiny variations here and there. According to Guy, we are now with +-0.6 dB from their measurements.

We don't know if the variations are because my measurements are wrong, or theirs. There is no gold reference here.

Frankly, I am sick of it at this point so told Guy that we are done and I should just post the results. People need to remember that every measurement system has variability and there is fair bit of that between anechoic chambers as well. For the purposes of evaluating speakers, what we have is more than good enough and far, far better than anything any reviewer is producing.

Importantly the moment the speaker is put in a room, nothing remotely flat comes out of it. So we need to keep that in perspective.

Conclusions
The main variable in our measurements not quite matching Neumann's was temperature. Whether this impacted just this speaker or the others, I don't know. If I get some time (yeh right), I will go back and remeasure a few of the other ones I still have and see what difference there may be. Come this winter, I will heat up the measurement room to get it closer to spring weather at least. Fortunately it never gets that cold here and hopefully we will find that other speakers are not affected as much.

Beside temperature, I have also fine tuned the system, using averaging for example to reduce noise level, deploying more optimal parameters to get the high frequencies right, etc. These have been in place for months so the bulk of what we have measured are accurate as they also enjoy warmer temperatures.

@MZKM has a new score for this speaker which he will post. New spin data is enclosed.

Be on notice that I am not measuring this speaker again so don't even think about asking me to do that. As it is, I have to now rip the system apart again and put the cage back on it and re-align it all.

------------
As always, questions, comments, recommendations, etc. are welcome.

As you can tell I am quite grumpy because of this ordeal. Fortunately it is not something where a few hundred dollars won't fix. So please donate what you can using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
I believe a long time ago on Zaph's site he mentioned something about "Why car audio sucks in freezing/blazing states like Wisconsin" ...
 

phoenixdogfan

Addicted to Fun and Learning
Joined
Nov 6, 2018
Messages
850
Likes
739
#37
When I see how high these things rate, I really have my reservations. They're just too small for use in anything other than a small desktop/studio monitor, yet their preference rating is above all the Revels, Kefs, and on a par with the Genelecs, and I just don't see it.

Probably highlights that the preference ratings don't take into consideration the power handling/dynamic range of speakers, which certainly makes them incomplete as a predictive model.
 

JohnYang1997

Major Contributor
Technical Expert
Manufacturer
Joined
Dec 28, 2018
Messages
4,741
Likes
7,456
Location
China
#38
When I see how high these things rate, I really have my reservations. They're just too small for use in anything other than a small desktop/studio monitor, yet their preference rating is above all the Revels, Kefs, and on a par with the Genelecs, and I just don't see it.

Probably highlights that the preference ratings don't take into consideration the power handling/dynamic range of speakers, which certainly makes them incomplete as a predictive model.
It's already measured at 105db. What more power handling do you need?
 

AudioStudies

Senior Member
Joined
May 3, 2020
Messages
404
Likes
175
#40
Science is valid, even when it isn't about effects that are practically significant. It is important to value small values.
In the purest sense, yes, I agree. In a practical sense, if we spend too much time and resources on things too small, it takes away from investigating bigger things in other areas.
 
Last edited:
Top Bottom