• 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!

Ideas for more meaningful speaker measurements

tuga

Major Contributor
Forum Donor
Joined
Feb 5, 2020
Messages
3,343
Likes
3,422
Location
Oxford, England
They are. Check Apple Music. Unfortunately the whole thing is more of a marketing thing than an attempt in advancing the art. As long as the music industry isn't implementing any meaningful standards that's mainly all we get.

I'm perfectly happy with my 2-channel rig.
 

markus

Addicted to Fun and Learning
Joined
Sep 8, 2019
Messages
619
Likes
623

tuga

Major Contributor
Forum Donor
Joined
Feb 5, 2020
Messages
3,343
Likes
3,422
Location
Oxford, England
Well, you can't miss what you've never experienced ;)

Indeed, life is to short not to enjoy what I have. FOMO is not my thing.
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
2,268
Likes
1,632
Location
Canada
This post requires more information before it can be responded to thoughtfully.

Am interested, but please create another thread as is clearly off topic.

Thanks!

Rick

Sure, I still have to confirm this so will have to recheck what exactly is it I’m seeing in the LZpeak value.
 

dshreter

Addicted to Fun and Learning
Joined
Dec 31, 2019
Messages
534
Likes
775
I admit I don't really understand the fixation on listening before measuring. Like, I guess vice versa creates some bias? But so do 100 other things. You're still doing sighted tests, rendering your listening test questionable AT BEST...

If a reviewer tells me they listen before measuring I just think "So what? You eliminated 1 bias, what about the other 99?". it's just totally irrelevant.
I think the point is that if observations determined prior to measurement are consistent with the measurements it lends credibility to the reviewers critical listening abilities.
 

markus

Addicted to Fun and Learning
Joined
Sep 8, 2019
Messages
619
Likes
623
Indeed, life is to short not to enjoy what I have. FOMO is not my thing.
That's perfectly fine but it doesn't negate there might be something out there you might enjoy even more :)
 

gnarly

Senior Member
Joined
Jun 15, 2021
Messages
418
Likes
467
My understanding of M-Noise is that it's similar to pink noise up to 700hz or so. I don't know of any case in which any normal home/studio speaker would be SPL limited by frequencies above that. I think this is solving for an issue with how the dynamic capacity of extremely large PA/cinema speakers is measured. Not really a problem that we have.

I have the same understanding about the spectral content on M-Noise.

I do think dynamic capability is a big problem for home audio...for two basic reasons.

First, at the same SPL level, most home gear sounds quite stained compared to live sound or install speakers to me. Home gear gets plain harsh quick, when the volume is turned up. By up i mean anything approaching 90 dB or greater at listening position.

Second, a small amount of home audio number crunching usually indicates trouble imo.
It seems to me a system needs 24 dB of headroom above its rated sensitivity, just to linearly reproduce regular pink noise at listening position.
That's with regular pink having +12 dB crest, and another +12 dB needed for 4m listening position.

So an amp needs 24 dB, 256x the power, just to play pink at the speakers rated 1 watt sensitivity.
And music often has a higher crest than pink...hence the case for M-Noise.
Clipping has to occur it seems.

Especially for home audio speakers, where sensitivity often is in the 80's dB range.
Plus, most sensitivity specs are overstatements, coming from the higher part of the frequency response curve.
We know the lowest part of the sensitivity curve is what determines the maximum linear SPL output....for the entire speaker...(if we want flat or downward sloping response.)

I can only repeat, I think for any kind of realistic SPL, either waveform clipping, or thermal compression of some frequency range is likely to occur with many home speakers....if they are trying to play as loud as their rated sensitivity spec at a moderate listening position.

I like that Erin provides his dynamic range test. I'm not sure how to think about it in the context of how, or if, it is showing average compression and/or any peak suppression. Need to look into it a little i guess.

Admittedly, I'm talking about measuring an amp and speaker together as a system, not just the speaker.
But i do think this is crucial for linear preproduction, after all is factored in.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
2,314
Likes
4,230
Location
Stow, Ohio USA
I think that the 50dB-wide vertical scale is informative enough, it is more or less standard too.
Too much smoothing is common though, but that is not a problem with Amir's measurements.

Since we can supposedly hear differences as little as 1 dB, why should we settle for visual representation that is +/- 25 dB? Agree it is a standard, but does why allow tolerances way beyond components comprising other parts of the audio system?
 
Last edited:

Sancus

Major Contributor
Forum Donor
Joined
Nov 30, 2018
Messages
2,432
Likes
6,125
Location
Canada
I do think dynamic capability is a big problem for home audio...for two basic reasons.
I think you misunderstood my post. I was saying dynamic capability above 700hz(or 1khz really, theres little divergence by that point anyways) is not an issue, so M-Noise doesn't help much. You generally have to get into crazy multi-12" or 15" woofer designs before midrange and tweeter output become an issue. Those types of speakers are more like PA and commercial cinema designs, and only the absolute tiniest percentage of people really need them.

For the average cone and dome domestic loudspeaker(or comparable design), the woofer is going to fall apart long before any other driver. This is even more true with home theatre, where bass requirements are higher than with music. I would say most HT content requires at least +20dB in bass relative to 1khz output.

So I agree dynamic capapability in general is an issue, but that's primarily because we don't have any good conditions for test failure that have been shown to correlate with audible problems. The way we currently measure and depict distortion has no correlation with audible problems. M-Noise doesn't help with that problem, as it's just a test signal. We still need some kind of measurement(s) to judge failure that actually correlates with audible issues in a clear way.
 

HooStat

Addicted to Fun and Learning
Joined
May 11, 2020
Messages
751
Likes
779
Location
Calabasas, CA
Along these lines, if a subwoofer is going to be used, how useful are some of distortion and dynamic capability measures based on full range signals? I would like to see some tests done with an appropriate cross-over for the speaker (or have every speaker use 80 Hz if easier). Particularly for the common 2-way bookshelf, but it would also apply to a lot of other speakers.
 

gnarly

Senior Member
Joined
Jun 15, 2021
Messages
418
Likes
467
I think you misunderstood my post. I was saying dynamic capability above 700hz(or 1khz really, theres little divergence by that point anyways) is not an issue, so M-Noise doesn't help much. You generally have to get into crazy multi-12" or 15" woofer designs before midrange and tweeter output become an issue. Those types of speakers are more like PA and commercial cinema designs, and only the absolute tiniest percentage of people really need them.

For the average cone and dome domestic loudspeaker(or comparable design), the woofer is going to fall apart long before any other driver. This is even more true with home theatre, where bass requirements are higher than with music. I would say most HT content requires at least +20dB in bass relative to 1khz output.

I think we are probably more in agreement than not, but coming at it from different angles.
I totally agree that woofers will fall apart first, and that high frequency drivers which get so much less power, hang in just fine from a SPL perspective (providing they get the full unclipped signal).

The problem as i see it, is that the high frequency drivers don't get the full signal if an amp is undersized, because of summed waveform clipping at high SPL.
High frequency wave forms ride on top of low frequency waveforms, giving an overall higher voltage than either alone. We know how voltages stack on an o-scope. (Relevant for passive speakers.)

So the +24 dB (or +30dB) headroom i think that is needed for an amp to provide above a sensitivity spec, is about being able to maintain the voltage implied by that gain, along with the power required by that gain, to keep from clipping summed waveform voltage peaks.

The fact that the HF driver isn't using much power doesn't matter for a passive speaker. It can still get clipped signal pretty easy, when a amp has used all its power capacity to drive the woofers. Amp rail voltage sags, and hello clipping tweeters.

And of course with multi-amping, the issue largely goes away....just get the right sized bigger amp for woofers, and the right sized smaller amp for tweeters. etc
So I agree dynamic capapability in general is an issue, but that's primarily because we don't have any good conditions for test failure that have been shown to correlate with audible problems. The way we currently measure and depict distortion has no correlation with audible problems. M-Noise doesn't help with that problem, as it's just a test signal. We still need some kind of measurement(s) to judge failure that actually correlates with audible issues in a clear way.

I agree, we really do need a better standardized way of testing large signal performance. Both average SPL and peak, along with whatever correlates with degrading audibility as SPL rises.
I've been trying a few measurements off the beaten track...might post one in a bit.
 

gnarly

Senior Member
Joined
Jun 15, 2021
Messages
418
Likes
467
Along these lines, if a subwoofer is going to be used, how useful are some of distortion and dynamic capability measures based on full range signals? I would like to see some tests done with an appropriate cross-over for the speaker (or have every speaker use 80 Hz if easier). Particularly for the common 2-way bookshelf, but it would also apply to a lot of other speakers.
I just start the measurement sweep for distortion, no more than a 1/2 octave below the speaker's intended xover point. Keeps there from being excessive garbage distortion down low.

Have you seen Josh Ricci's tests for subwoofers? https://data-bass.com/#/articles/5cc0bc36a75a260004255c88?_k=8xshia
CEA2010 seems a decent test for sub distortion. I've been trying it for mids and highs....seems to work there too.
 
D

Deleted member 46664

Guest
Since we can supposedly hear differences as little as 1 dB, why should we settle for visual representaion that is +/- 25 dB? Agree it is a standard, but does why allow tolerances way beyond component comprising other parts of the audio system?

Yes, many people can pick a +-1db level change in a steady tone. That changes rather drastically when you play two, non harmonically related tones. Most people won't pick out a much larger difference between the levels in that case. Thus while we hear the level change on a steady signal we really aren't very good at judging the relative loudness of differing parts of the audio spectrum.

That is to say that in yer average living room, frequency response variations are far less critical than most people think.

Here is a plot of my speakers in my living room taken with REW ... The scaling is set to 1db and 1/12 octave smoothing.
Notice how disastrously bad it looks... Really... how could anyone listen to THAT?
Rew_at_1db.png


Well, when we adjust the same plot to more closely resemble what people actually hear ... with 0db at the bottom of the chart and the active trace at about half of full scale and then use psychoacoustic smoothing... we get this... and, that is pretty much what I hear in the room ... nice even sound, a bit of punch to the bass, some sub bass activity and gently rolled down highs... perfect for my ears!
rew_at_0_base.png


Keep in mind it's the same measurement... all I did was adjust the scale and smoothing to more closely resemble human hearing.

If I did not tell you those are the same measurement of the same speakers ... which would you prefer?

Here are some test tones you can play with... all are equal volume at -10dbfs...
 
Last edited by a moderator:

sigbergaudio

Major Contributor
Manufacturer
Forum Donor
Joined
Aug 21, 2020
Messages
1,214
Likes
2,128
Location
Norway
I think it would be a mistake to shrink the scale and make the deviations more detailed. People are already struggling to understand what they're seeing, and assuming this and that without knowing how to interpret the graphs.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
2,314
Likes
4,230
Location
Stow, Ohio USA
Yes, many people can pick a +-1db level change in a steady tone. That changes rather drastically when you play two, non harmonically related tones. Most people won't pick out a much larger difference between the levels in that case. Thus while we hear the level change on a steady signal we really aren't very good at judging the relative loudness of differing parts of the audio spectrum.

That is to say that in yer average living room, frequency response variations are far less critical than most people think.

Here is a plot of my speakers in my living room taken with REW ... The scaling is set to 1db and 1/12 octave smoothing.
Notice how disastrously bad it looks... Really... how could anyone listen to THAT?
View attachment 205147

Well, when we adjust the same plot to more closely resemble what people actually hear ... with 0db at the bottom of the chart and the active trace at about half of full scale and then use psychoacoustic smoothing... we get this... and, that is pretty much what I hear in the room ... nice even sound, a bit of punch to the bass, some sub bass activity and gently rolled down highs... perfect for my ears!
View attachment 205149

Keep in mind it's the same measurement... all I did was adjust the scale and smoothing to more closely resemble human hearing.

If I did not tell you those are the same measurement of the same speakers ... which would you prefer?

Here are some test tones you can play with... all are equal volume at -10dbfs...

Thanks, this is comparable to my original point (from a different perspective).

The question is what value is the visual representation for comparison? They clearly do no sound any different. I suggest both have value for different reasons.
 
D

Deleted member 46664

Guest
I think it would be a mistake to shrink the scale and make the deviations more detailed. People are already struggling to understand what they're seeing, and assuming this and that without knowing how to interpret the graphs.
Your average customer can follow a line on a graph .... but they have no way of equating that with what they are actually hearing. If you present them with a "loudness curve" instead, most will look at it and go 'Wow... that's better than I expected" and most can follow the tones to confirm that's pretty much what they are hearing.

Misinterpretations like this are always a risk when engineering level data is presented to layman level listeners.
 
D

Deleted member 46664

Guest
Thanks, this is comparable to my original point (from a different perspective).

The question is what value is the visual representation for comparison? They clearly do no sound any different. I suggest both have value for different reasons.

That's the thing ... I mean no offence to anyone... but sometimes the engineers do get lost in detail and minutia that the average listener can't discern and probably shouldn't care about.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
2,314
Likes
4,230
Location
Stow, Ohio USA
That's the thing ... I mean no offence to anyone... but sometimes the engineers do get lost in detail and minutia that the average listener can't discern and probably shouldn't care about.

Thanks again and I appreciate that you chose to discuss here. My more recent post is more for fun. It may say something about measurement perceptions within the membership too but was not my intent. Most seem to be fine with the graph scaling but may be just because ASR attracts more technical folk than other open forums.

Am not in favor of dumbing down for the lowest common denominator but can see why the manufacturers seem to go there. IMO, it should not be an excuse to not have some way for more technical folk to get more detailed information. :confused:
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
2,314
Likes
4,230
Location
Stow, Ohio USA
I think it would be a mistake to shrink the scale and make the deviations more detailed. People are already struggling to understand what they're seeing, and assuming this and that without knowing how to interpret the graphs.

Maybe but suggest it really depends on who the "People" are. ;)
 
Top Bottom