• 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.

Speaker Max SPL Criteria?

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,338
Likes
89,967
Location
Seattle Area
#1
I like to pursue adding max SPL capabilities to my speaker tests (to the extent they are speakers I own at least). I know of the IEC standards for them but I don't know of any real research behind the thresholds they have picked. Limits are in bass where there are multiple mechanisms for distortion. In an active speaker for example, there is the amp clipping and driver becoming non-linear. The latter has multiple mechanisms, each with their own distortion profiles.

Have there been formal studies/papers that we/I should be reading on this? Any of you have done any tests to see what your threshold of distortion is?

Let's discuss what makes sense to do. Note that I don't care about following the crowd. I like to find the right thing to do. We will have enough of our own data as to not need to compare to others doing this.

Open to any and all feedback.
 

Vini darko

Major Contributor
Joined
Jun 1, 2020
Messages
1,112
Likes
1,300
Location
Dorset England
#3
Hi amirm since you test apmlifier max @1% distortion. Maybe that would be a good line in the sand for speakers too. Though it would need to be in the say 500-5khz range where hearing is most sensitive. Below and above that I have no idea. Most speakers tested have what looks like hidious bass distortion , picking a % will be tricky.
 

pozz

Data Ordinator
Forum Donor
Editor
Joined
May 21, 2019
Messages
2,960
Likes
4,498
#5
Also this graph for the audibility and masking of harmonic distortion vs. level and frequency can be helpful:

View attachment 94929

Source: https://www.hifi-selbstbau.de/grund...nmenu-70/239-klirrfaktor-wie-viel-ist-zu-viel
Sorry, again German.
That presents H2 only. The thresholds for higher harmonics will be lower.

However, I'm glad you posted it since there is a reference to a program (or perhaps instructions) for modelling distortion audibility according to Zwicker. Perhaps our Deutschsprachig members can investigate:)
 
Last edited:

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,476
Likes
8,523
#9
@amirm I guess you've already read/watched Klippel's own material on this topic? This is the most pertinent one here.

You may also want to consider using m-noise instead of IEC-prescribed band-limited pink noise. I'm not sure it's clear yet whether the industry will pick on on m-noise as the new de facto standard for SPL testing, but it does seem to be getting talked about a lot lately.
 

pozz

Data Ordinator
Forum Donor
Editor
Joined
May 21, 2019
Messages
2,960
Likes
4,498
#11
You may also want to consider using m-noise instead of IEC-prescribed band-limited pink noise. I'm not sure it's clear yet whether the industry will pick on on m-noise as the new de facto standard for SPL testing, but it does seem to be getting talked about a lot lately.
Wasn't m-noise developed specifically for Meyer's new(ish) analyzer? It's a general purpose test signal?
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,338
Likes
89,967
Location
Seattle Area
Thread Starter #12
@amirm I guess you've already read/watched Klippel's own material on this topic?
I have. They are following specs to the letter. I like to find out if they are relevant.
 

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,476
Likes
8,523
#13
I have. They are following specs to the letter. I like to find out if they are relevant.
I see, thanks.

Unfortunately (as you no doubt know), the research findings are extremely mixed in this area (mostly because the metrics chosen, e.g. THD, are fairly useless when it comes to human hearing.

Of all the existing tests I've looked into, it seems that the distortion metrics with the best correlation to audibility are the DS and R(nonlin) metrics developed by de Santis and Henin. These are the only distortion metrics I'm aware of that research has demonstrated to correlate well to audibility.

However, I don't know if either metric is publicly available. Nor am I aware of attempts to reproduce the findings (which were undertaken by the creators of the metrics).

Geddes' metric has shown some promise (though not as much as the metrics developed by de Santis, as I understand it), and there are now online calculators available for it using mathlab or similar IIRC (let me know if you'd like links).

Is this answering your question or am I still not quite understanding where you're coming from?
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,338
Likes
89,967
Location
Seattle Area
Thread Starter #15
Is this answering your question or am I still not quite understanding where you're coming from?
No, it is a very good start. Thanks. I am also looking to see if anyone has experimented on their own to find the thresholds.
 

detlev24

Senior Member
Joined
Dec 3, 2019
Messages
305
Likes
272
#16

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,476
Likes
8,523
#17
No, it is a very good start. Thanks. I am also looking to see if anyone has experimented on their own to find the thresholds.
I'm sure plenty here have done online tests like the Klippel one (I have too, ages ago). But I'd think more rigorous studies would be the most reliable path forward, especially given the online tests are quite limited.

And especially because finding a threshold for one form of distortion won't necessarily tell us much about relevant thresholds for a DUT.

Sorry if I'm overstepping here and being overly opinionated about it, but f I were you, I'd contact de Santis and co to see if they will share with you the formulae their metrics are based on (which are derived from standard IMD measurements, and therefore should not be too difficult to apply to standard measurements).

It's their metrics that appear, at this point, to offer by far the best correlation to subjective experience.
 

andreasmaaan

Major Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,476
Likes
8,523
#19
I’ve mentioned it before, but Axiom Audio did a small trial (wouldn’t call it a study), and this is what they got:


Of course using sine waves or whatnot would have tighter thresholds.
One little note: it needs to be kept in mind when interpreting Axiom's findings that the masking "signal" in this study was music, which has a far-from-flat spectrum itself!
 

AnalogSteph

Addicted to Fun and Learning
Joined
Nov 6, 2018
Messages
937
Likes
772
Location
.de
#20
It's a bit perplexing that nobody seems to have established audibility thresholds for a few standard IMD tests so far. We can superbly measure multitone distortion these days, which basically could tell us everything if conducted at a range of levels, but relating it to the speaker audibly "falling apart" seems to be a different kettle of fish.

I rather like @hardisj's IMD Test 2 (80 Hz + 200-6000 Hz sweep). It may have to be complemented by something to annoy 3-ways with (maybe 800 Hz + 1100 Hz, 800 Hz + 1500 Hz, 800 Hz + 1750 Hz or similar). Or maybe one could go 3-tone (80 Hz + 825 Hz + sweep or similar so there's something for everyone, with relative levels to match pink noise). This seems a promising kind of test signal to correlate measurements and hearing, unfortunately without any existing data.

EDIT: I just generated a tritone in Audacity. Generate sines 80 Hz, 825 Hz, 1505 Hz @ 0.316 each, make EQ preset with two points connecting 20 Hz @+10 dB to 20 kHz @-20 dB and apply that to all tracks, export. It definitely seems suitable to annoy my K+Hs at higher levels.
EDIT ^2: Test tone uploaded here. I cut it to 20 seconds and the forum software still complained, so I uploaded this to my personal webspace. It's just 5 minutes worth of using Audacity anyway. I can definitely get my K+H O110s to get "raspy" at elevated but not at all inordinate volumes (easily 20 dB below the 80 Hz tone by itself giving them trouble). They would seem to be doing the business as-is but I think a sub is going on the x-mas wishlist...
 
Last edited:
Top Bottom