• 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). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Understanding How the Klippel NFS Works

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
559
Likes
102
Location
germany
I look for images in google about how a step response from klippel look. i find no step response from klippel. is it maybe klippel do not output step responses ?

maybe in arta o REW and other the step response is create wrong. to better compare the Frequency should be normalized or a HP filter of 100 hz should use to reduce the large diffrence of FR in bass so step response can better compare.

see this step response. https://www.stereo.de/hifi-test/produkt/kef-r11-1733 it is a large speaker with much low bass, but it have much faster step response as my best sounding speaker and measure with REW or arta
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
2,085
Likes
1,479
Location
Canada
I look for images in google about how a step response from klippel look. i find no step response from klippel. is it maybe klippel do not output step responses ?

maybe in arta o REW and other the step response is create wrong. to better compare the Frequency should be normalized or a HP filter of 100 hz should use to reduce the large diffrence of FR in bass so step response can better compare.

see this step response. https://www.stereo.de/hifi-test/produkt/kef-r11-1733 it is a large speaker with much low bass, but it have much faster step response as my best sounding speaker and measure with REW or arta

Amir and Erin occasionally add it in: https://www.erinsaudiocorner.com/loudspeakers/kii_three/
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,075
Likes
4,589
Location
Berlin, Germany
Amir ususally shows impulse response, not step response (which is the integral of impulse response). Impulse responses are not that useful as they bascially only highlight the high frequency region, which concides to how they sound (compared to a step).
 

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
559
Likes
102
Location
germany

thanks alot. luckily i find this speaker test in stereo magazineonline. it have much faster step response (Sprung Antwort is german word for step response) https://www.stereo.de/hifi-test/produkt/kii-three-1103

Are there more settings for step response in Klippel software ?. erin post step response with 2 diffrent settings but they look not so simular to the results from stereo page
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
2,085
Likes
1,479
Location
Canada
thanks alot. luckily i find this speaker test in stereo magazineonline. it have much faster step response (Sprung Antwort is german word for step response) https://www.stereo.de/hifi-test/produkt/kii-three-1103

Are there more settings for step response in Klippel software ?. erin post step response with 2 diffrent settings but they look not so simular to the results from stereo page

I don't read German, but visually the graph appears to be the IR. Sometimes people interchange the two even though they are different. I do not use a Klippel measurement device or software, but both Erin and Amir have posted one or the other type of graphs -- sometimes interchanging the label.
 

thewas

Major Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
4,309
Likes
9,552
Stereophile posts also step responses of the loudspeakers they measure, here from a couple of relatively KEF 3-way floorstanders which show that the above Stereo magazine one is deviating, possibly due to windowing:

1017KEF5fig09.jpg

source: https://www.stereophile.com/content/kef-reference-5-loudspeaker-measurements

914KEFfig7.jpg

source: https://www.stereophile.com/content/kef-r700-loudspeaker-measurements

Also interesting to see that the LS50 Meta one

1220KEF50fig08.jpg


compared to the one of the original LS50

1220KEF50fig09.jpg

(source of both: https://www.stereophile.com/content/kef-ls50-meta-loudspeaker-measurements)

which shows again how poorly a "nice looking" step response correlates to sound qualities.
 

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
559
Likes
102
Location
germany
I don't read German, but visually the graph appears to be the IR. Sometimes people interchange the two even though they are different. I do not use a Klippel measurement device or software, but both Erin and Amir have posted one or the other type of graphs -- sometimes interchanging the label.

I notice on impulse responses the 1. peak to lower 0 value reach alway nearly -100 % . on step responses the first peak to lower 0 should be short and is on good speakers. see the new meta.

here you can see the impulse response and you see the 1. low peak go very much in negative. this is good for impules response but bad for step response


still strange witgh which setting they measure step responses so my questions is. are there more settings in the klippel software.

In the compare from thewas the meta have clear better step response because the meta go not so much in negative. and in the meta the decay time from peak to 80% is faster. but here is not the thread to show step responses ;)
 
Last edited:

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,075
Likes
4,589
Location
Berlin, Germany
I don't read German, but visually the graph appears to be the IR
Exactly, but they mislabelled it as step response... which does not increase their credibility.

One and the same system, IR vs SR:
1639141104027.png


1639141115312.png


Therefore, the IR in @amirm's review linked to above is very hard to interpret, it is not at all easy to tell whether this Genelec has a linear-phase crossover which would be the ideal (the speaker seems to have partial linear-phase XO, for the mid/tweeter, according to S&R measurements). Only a proper frequency response plot of magnitude and phase would tell right away, or alternatively and less ideal, a proper step response display.
 

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
559
Likes
102
Location
germany
I look in stereo, they have the LS 50 meta test too. they write impulse response, but it look as a invertet but faster step response.

 

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
559
Likes
102
Location
germany
FWIW, I have never posted the IR. I only share the step response. And sometimes I forget to add it.

can you please post the step responses from klippel in minimal mode for the Kali LP 8 V2 you have test ?
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
2,075
Likes
4,589
Location
Berlin, Germany
I look in stereo, they have the LS 50 meta test too. they write impulse response, but it look as a invertet but faster step response.

Occasionally your right, Benny.
Screw Stereo magazine and their "expert journalists".
 

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,596
Location
North Alabama
can you please post the step responses from klippel in minimal mode for the Kali LP 8 V2 you have test ?

???

No such mode exists.

You misread the Kii THREE review. That particular speaker has two latency modes; minimal or phase exact. Thus, the two graphs in that review. Kali has no such mode.
 

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
559
Likes
102
Location
germany
???

No such mode exists.

then there is no settings you can choose for the klippel Step response display ?.
then maybe you can post the Kali LP8 step response. please ?

BTW: the compression tests are intresting. i dont thought that it is so much on lower frequency. when level is reduce then possible phase is shift too during level diffrences. For stereo hear below ITD max frequency of around 1400 hz the group delay change of 75 db and 85 db is also usefull to see. because when hear stereo signals the speaker play always at diffrent level and diffrent frequency. frequency or better wavelength or period time result in the time from peak to peak in non cyclic complex waveforms as music is. and in this wave display(after a 1.5 khz LP 36 filter) can see that sometimes 2 * more peaks happen as on other channel


left right diffrent frequency.jpg
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
2,085
Likes
1,479
Location
Canada
FWIW, I have never posted the IR. I only share the step response. And sometimes I forget to add it.

Now that you mentioned it, I think it was Amir who at one point mislabeled the graph — maybe in one of the early reviews, I don’t recall which one exactly. I have seen others besides that audio e-magazine mislabelling between the two.
 

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,596
Location
North Alabama
Now that you mentioned it, I think it was Amir who at one point mislabeled the graph — maybe in one of the early reviews, I don’t recall which one exactly. I have seen others besides that audio e-magazine mislabelling between the two.

Yea, definitely not me. I think the only time I've posted IR is outside the typical bounds of a review just to show where I gated the response. I've only ever shown step response in an actual review, though.
 

kyle_neuron

Active Member
Joined
Jun 18, 2021
Messages
145
Likes
241
I notice on impulse responses the 1. peak to lower 0 value reach alway nearly -100 % . on step responses the first peak to lower 0 should be short and is on good speakers. see the new meta.

here you can see the impulse response and you see the 1. low peak go very much in negative. this is good for impules response but bad for step response
This is not strange if you understand the principle of normalisation.
 

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
559
Likes
102
Location
germany
This is not strange if you understand the principle of normalisation.

normalisation set only the largest peak (doesnt matter if positiv or negative) to 100%. for example largest peak have 50% and is positvive. negative peak have -40%. then it scale by 2. largest peak get 100% and opposite peak get -80%.

In step response good speaker have less negative peak. on impulse response seem good much negative peak, this i see often.
 

bennybbbx

Addicted to Fun and Learning
Joined
May 27, 2020
Messages
559
Likes
102
Location
germany
does the klippel offer a overlay display of minimal phase impulse response ?. I do this in REW with my speakers. so the results can be better compare because the diffrence to the optimal impulse for this FR can see.

In rew manual stand this
Generate minimum phase will produce a minimum phase version of the measurement. Using this control also generates a minimum phase impulse plot and, for measurements that already had an impulse response, minimum and excess phase and group delay plots. If the measurement did not have an impulse response the minimum phase impulse is set as the IR for the measurement, this allows an IR and phase to be produced for measurements which do not have them, such as imported text files.

see here for images https://www.audiosciencereview.com/...r-itd-see-measures.28585/page-11#post-1019071
 

CtheArgie

Senior Member
Forum Donor
Joined
Jan 11, 2020
Messages
381
Likes
513
Location
Agoura Hills, CA.
I noticed today in PS Audio’s Copper magazine an article discussing spinorama where they used a graph provided by ASR! Potential for cooperation? @amirm , will you test their FR-28?
 
Top Bottom