• 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 Speaker Measurements: Take Two

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,111
Likes
8,438
Location
NYC
Far field is good. But that is NOT what anechoic measurements typically show. Due to size limitation and modes in the anechoic chamber, measurements are usually NOT in far field but somewhere between it, and near field. In contrast, the Klippel NFS measurements properly show the true far field sound radiation. The larger the speaker, the worse the typical anechoic measurement in this regard.

If you are performing far field measurements outdoor using gating, then you have signal to noise ratio, temperature and humidity variations.

True, but in the same vein, the smaller the speaker, the better anechoic chambers (and gating) are in this regard, and we've mostly been dealing with quite small speakers. I've seen many definitions of "far-field" in terms of wavefront integration. In his book, Dr Toole refers to a paper defining the farfield as starting at 3 to 10 times the largest dimension of a speaker, but clearly Harman is confident enough in 2m to trust that distance to measure and presumably design speakers like the 1.35m tall Salon2 - not anywhere near 3 to 10 times.

Personally, I've yet to see a single bookshelf speaker whose response meaningfully changes at 2M vs 1M within the resolution of my gated measurements. Still, with every speaker, I always repeat several measurements at 2m to see if my 1m measurements hold up. There's never been a difference, so I save the 1m data because of the better resolution. I expect towers would show a difference at that distance, but I have no plans to measure towers. :)

Anyway, at this point I'm just academically curious as to what caused that droop measuring at the stated acoustical center with the Klippel, since it uses gated measurements for the upper portion of the response. I highly doubt it's just an artifact of nearfield measurements as I'm pretty sure even at 1 foot only I wouldn't see that big of a treble drop. It's also a bit mysterious that the treble would then "fix itself" so dramatically once measured at the tweeter again, as the difference in position is definitely less than 20 degrees which was noted as looking flatter.
 
Last edited:

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
Here are the two new measurements side by side:
index.php


It is pretty clear that if Neumann says the speaker has flat response, its acoustic center can't at the same time be where they say it is. One or the other is wrong. I may have an explanation for this though. See later.

Are both measurements done at 0.3m? Maybe re-measuring at 1m should move back the audio-center between woofer and tweeter (due to tweeter's narrow freq. propagation cone) or at least will eliminate any possible and unwanted driver's vibrations or reflections that might interfere with the final measurements, and in the end maybe will get a flatter measurement @1m than @0.3m.
 

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,111
Likes
8,438
Location
NYC
Are both measurements done at 0.3m? Maybe re-measuring at 1m should move back the audio-center between woofer and tweeter (due to tweeter's narrow freq. propagation cone) or at least will eliminate any possible and unwanted driver's vibrations or reflections that might interfere with the final measurements, and in the end maybe will get a flatter measurement @1m than @0.3m.

The klippel doesn't go that far. More importantly, it's not supposed to need to.
 

LeftCoastTim

Senior Member
Forum Donor
Joined
Apr 15, 2019
Messages
375
Likes
757
How does the Klippel System with mic. at the baffle(position not critical?) correlate with mfrs. measurements with mic. at one metre, 4ft. or further, on mfrs. preferred baffle axis?

All the sound from a speaker must travel through an imaginary sphere at 0.3 meters away from the source before it reaches the sphere 1 meter away. If one takes enough measurements at the sphere at 0.3 meters, one has all the information to predict the sound at 1 meter or any other distance away.

One could not do the same with a single microphone measurement at 0.3m and at 1m.
 

Juhazi

Major Contributor
Joined
Sep 15, 2018
Messages
1,717
Likes
2,897
Location
Finland
Mic distance to two separated sources will affect summation and spl response. Klïppel NFS measures at two distances and thus can calculate 3D radiation at different distances and this.
index.php
index.php


But I am really puzzled, that negative vertical angle gives best hf response, when measurement axis is below the tweeter! Was the speaker upside down or is the NFS algorithm upside down?

Low bass foll-off in NFS measurement is real, I think. Manufacturer does the typical quasi-anechoic estimation of response based on sum of nearfield measurements of driver and port output. Stereophile's Atkinson has said that the formula he uses leaves the bass thin when the curve is flat, and I believe that the same thing applies to K&H Neumann.
 
Last edited:

LeftCoastTim

Senior Member
Forum Donor
Joined
Apr 15, 2019
Messages
375
Likes
757
View attachment 49068

View attachment 49069

These two plots alone speak (pun intended) volumes on what the Klippel can do that no other measurements can.

If "Sound und Recording" and "Stereophile" do not purchase their own Klippel NFS (and publish results), they would have to concede that ASR is a better audio publication than both of them combined.

Kudos to @amirm !
 

ZolaIII

Major Contributor
Joined
Jul 28, 2019
Messages
4,069
Likes
2,409
@amirm I think it's a good idea to mention serial number (or possibly some other ID) where possible to a measured equipment for documentary purposes and so that you can avoid future accusations how you didn't done it properly or how you got a "golden sample" rubbish.

Best regards.
 

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,111
Likes
8,438
Location
NYC
Low bass foll-off in NFS measurement is real, I think. Manufacturer does the typical quasi-anechoic estimation of response based on sum of nearfield measurements of driver and port output. Stereophile's Atkinson has said that the formula he uses leaves the bass thin when the curve is flat, and I believe that the same thing applies to K&H Neumann.

While at this point the Klippel's results look more reliable, I highly doubt Neumann uses a simple bass summation.

JA doesn't use baffle step compensation in his nearfield measurements, which I believe is because he wants to maintain consistency with his measurements over the years - that's why most of his show a bump in the bass, and as you said, speakers will sound thin if his graph is flat in the bass.

I use baffle step compensation in all my measurements, as it's usually a trivial process with some of the software out there especially with box-shaped speakers. I'm not on my measurement laptop to show it now, but doing a simple woofer and port summation without compensation on the KH80 shows a bump in the bass - as expected from a flattish speaker. With compensation, the frequency response flattens almost perfectly.

That said, it's easy to be a dB or two off depending on where you pick your splice point or due to small errors in the summation or compensation. I find it unlikely three sources did the nearfield summations wrong, but it's possible. It's also possible the Klippel is capturing something about the bass performance other measurements are missing. I might try to do some ground plane measurements, which I normally avoid because they are a pain, to see if they differs from my nearfield summation.
 

BYRTT

Addicted to Fun and Learning
Forum Donor
Joined
Nov 2, 2018
Messages
956
Likes
2,452
Location
Denmark (Jutland)
.....I wonder if Neumann now knows the bass is not actually flat, they can adjust the internal DSP of further models.

.....Currently I find myself more inclined to trust the Klippel with regards to bass.....

.....The slightly sagging bass suggests to me that there is a systemic error there, likely Neumann not having very good data in that region.....

.....Anechoic chambers also have to be "calibrated" at <100Hz, unless it is very large (and maybe even then for really low frequencies). So I would trust the math here (used by Klippel) over the "calibrated measurement" in an anechoic chamber.

My 2 cents, of course..

If they didn't know ported bass low reach response is a slow rolloff before a overshoot, they should know now, in samples scanned sounds was from two different owners, also NFS curves looks be a free calibration curve or pointer to anechoic chambers or napilopez or anyone haveing a KH 80 sample, red curve below is take 1 from 21. of January and green plus blue is both take 2 with the variable of using a different acoustic center, red is copy of NFS txt file and green/blue is traced curves if that could make a difference, they so close overlaid so a copy is offset -10/-11/-12dB.

Take_One_verse_Take_Two_low_end_reach.PNG
 
Last edited:

JIW

Senior Member
Joined
Nov 11, 2019
Messages
378
Likes
551
Location
Germany
As to bass dip, we still have that.

It is indeed still there but compared to the previous measurements at 106 dB SPL at 0.3 m or 96 dB SPL at 1 m, this measurement at 79 dB SPL at 2 m or 85 dB SPL at 1 m (according to CEA2034, I presume), has higher levels below 600 Hz: relative to 900 Hz, the peak at around 65 Hz is ca. 1 dB higher while the following dip at around 90 Hz is ca. 0.5 dB higher and the hump between 150 and 600 Hz is ca. 1 dB higher.

Thus, it seems there was between 0.5 and 1 dB of compression in the low to low-mid frequencies in the first measurement due to high SPL.
 
Last edited:

Andreas007

Active Member
Joined
Mar 11, 2019
Messages
138
Likes
353
Location
Germany, Bavaria
This is a great, great review! Thanks, we learn so much!
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,375
Likes
234,488
Location
Seattle Area
@amirm I think it's a good idea to mention serial number (or possibly some other ID) where possible to a measured equipment for documentary purposes and so that you can avoid future accusations how you didn't done it properly or how you got a "golden sample" rubbish.

Best regards.
Thanks. Serial number is actually in the photo but I black it out before posting:

index.php


If you look at the block below FCC/CE label, you can see blacked out area where the serial number is.
 

JIW

Senior Member
Joined
Nov 11, 2019
Messages
378
Likes
551
Location
Germany
Clearly this is not the acoustic center as one would define has having the flattest response. The droop in high frequencies clearly indicates we are not on-axis. We can confirm that by looking at the vertical dispersion:

index.php


The highlighted red graph is showing the "on-axis" as measured. The graphs above it are at -10 and -20 degree so they are more "correct." Translating, following the Neumann recommendation has clearly messed up the measurements.

To what distance are the measurements calculated?

The vertical difference between the acoustic axis and the tweeter axis is about 50 mm which over a distance of 1 m or 1000 mm gives an angle of about 3 degrees between them and over a distance of 2 m or 2000 mm, the angle is only about 1.5 degrees.

Looking at the vertical contour plots from the previous review which are at a distance of 10 m, giving an angle of -0.3 degrees vertical, unsurprisingly does not seem to make a significant difference in terms of frequency response.
index.php

This seems to be rather at odds with the data you present.

However, over a distance of 100 mm or 0.1 m, the angle between them is about 27 degrees which is much closer to what your measurement suggests. Further, over a distance of 300 mm or 0.3 m, the angle between them is about 10 degrees.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,755
Likes
16,208
As I have been at Neumann twice I have good contact to their senior engineer, so I just sent him the link to the new measurements and am sure he will look at them and respond in the next weeks.
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,690
Likes
6,013
Location
Berlin, Germany
The 20kHz balloons (or any other frequency, for that matter) *must* have the minutely exact same shape no matter where you put the acoustical axis, the only difference is that the whole thing is tiltet a bit. If they don't, something is wrong... at least if the assumption is that the balloons shall represent far field conditions.
My educated guess is that the measurement distance is too close, and more importantly, dynamically varying depending on choosen reference axis and that produces systematic errors despite of the technique using two distances and calculating the far field from that.

It is not for nothing that there is the recommondation, for traditional style measurements, using a distance of at least ten times the largest dimension of the DUT, and preferably larger than 20 times. Otherwise you're not in the far field.

I kindly suggest @amirm to do some experimenting to establish a baseline of repeatiblity and error bounds. Measure the same speaker (this one or any other) using a set of somewhat arbitrary reference positions. Say, the upper left corner and the lower left corner, and also a different set of the two distances the far field extraction is made from. Unless the ballons (at high frequencies) perfecty match in shape (ignoring the tilt), it's not working as intended.

EDIT: This assumes a mic that has close to no directivity (way less than 1dB) at angles up to least 30degree. A condition not met by the mic currently in use, it seems.

EDIT2: I also suggest to really think about double checking the whole measurement rig in the way I've explained here.
 
Last edited:

JIW

Senior Member
Joined
Nov 11, 2019
Messages
378
Likes
551
Location
Germany
The 20kHz balloons (or any other frequency, for that matter) *must* have the minutely exact same shape no matter where you put the acoustical axis, the only difference is that the whole thing is tiltet a bit. If they don't, something is wrong... at least if the assumption is that the balloons shall represent far field conditions.

The distance is given in the top of the figure as r. For the first one, using Neumann's specified axis, it's 2 m while for the second one using the tweeter axis, it's 1 m. The largest dimension of the KH80 is its height of 233 mm. Thus, the measurement at 2 m should give a better indication of the far field than the one at 1 m.
 

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,161
Location
Tarragona (Spain)
About the measurement distance, I thought there was a standard (1 m) but each one measures as he wishes. The logical thing, for me, is 1 meter away and at the listening point.

The last controversy was with the measurements of Klipsch RP-800F. Danny Richie detected a huge hole in his usual measurement at 1 m. Much more softened at 2 m in Audioholics measurements.



* https://www.audiocircle.com/index.php?topic=166535.msg1779874#msg1779874

index.php



* https://www.audioholics.com/tower-speaker-reviews/klipsch-rp-8000f/conclusion

image


image
 
Last edited:
Top Bottom