Check out this video session from the Hottinger Bruel & Kjaer EA virtual conference ten days ago noting that lower order harmonics are masked and higher order are not. And a effective loudness algorithm in frequency and time domains might make a better judge of driver quality.
-
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!
Are we looking at the wrong distortion components in loudspeakers.
- Thread starter JimWeir
- Start date
- Joined
- Feb 23, 2016
- Messages
- 20,761
- Likes
- 37,616
Pretty interesting presentation. I've known for a couple decades at least that 2nd and 3rd harmonics are about all most speakers put out. And that such are heavily masked. I've also been aware that hum, buzz and squeak was something of an unexplored issue. I also wondered if that was part of why ESL panels sound cleaner because usually they likely don't have much hum, buzz or squeak.
It also is why I've doubted tube amp have a sound from low order distortion. We just don't hear it very well. Plus I've found tube preamps to sound clean while transformer coupled amps have a signature it seems to me. Often that would be related to FR changes with transformer coupling.
I have noticed speakers on slow sweeps putting out sound like the BSR demo with tape over the test device. Some panels and some cone and box speakers.
It also is why I've doubted tube amp have a sound from low order distortion. We just don't hear it very well. Plus I've found tube preamps to sound clean while transformer coupled amps have a signature it seems to me. Often that would be related to FR changes with transformer coupling.
I have noticed speakers on slow sweeps putting out sound like the BSR demo with tape over the test device. Some panels and some cone and box speakers.
Ja, the wisdom of the past ;-) replicated over and over, now B&K again. What about a breakthrough?
Connecting to the past: "with music ...". With music or any noise to speak of, the spectrum comprises harmonics in itself. Question is, if the harmonics additionally generated by the speaker for instance, contribute so much, that the overall impression changes.
So, how are the harmonics in music distributed (spectrum)? Bassguitar, played 'clean': 400% (!!!) second, 100% (!!) third harmonic plus lots of higher components. How far would like 10% of HD from the bass-speaker distract the listener? Especially as the spectral content of the instrument depends strongly on how the chords are picked by the musician ... . Another is, again 'clean', the electric organ with just 50% 2nd, 10% 3rd, and little higher components.
More crucial is intermodulation, that originates in the same non-linearities. But worse, because in regular musical presentations, the instruments generate their own spectrum, and do not interfere with each other. But a wideband speaker logically would generate IM from a bass tone combined with all the sweet 'female voices', reproduced at the same time. These IM come not natural to the music. They distract as an additional phenomenon in the sound field. Ony that people got used to it. So much so, that a clean speaker sounds 'thin', 'clinical', what have You. The thickening, the muddiness, maybe even heft became a quality parameter in its own right, a matter of taste. How much speaker sound do You prefer?
Plus, IM can be generated not only in the motor. A very prominent source is the surround of bass/midrange speakers. It may raise the IM of an otherwise o/k-ish speaker to 30% with just 3mm of an excursion. (That's why the Purify's surround looks that weird.)
Thing is: tja, what is it? Breakthrough needed, me thinks. Discard the two-way concept for starters. Quality begins with a four-way, may it be with a 5" bass ... especially with smaller basses because of the high excursion needed ... the cheapskates we are ;-)
Last edited:
- Thread Starter
- #4
The “breakthrough” is applying non-stationary loudness measurements (something we have use in
Product and automotive testing for decades) in the stimulus - response loudspeaker test.
I didn't follow that. Non-stationary in the statistical sense?
- Thread Starter
- #6
No, how loudness varies with time.
Loudness incorporates the critical bands or sound, so if the excitation signal (sine wave) is steady the algorithm for loudness looks for masking.
However, there is also temporal masking, if a sound shows up a little before and a bit more after (say a sine pulse) it is also masked. But if it lingers, like with an undamped resonance is shows up as something audible.
Bu incorporating both algorithms we can sweep a sine wave while sampling quickly Overtime and assess audibility of frequency and time anomalies and ignore anomalies that are not audible.
Me too. Only that the acceptable limits for IM are (implicitly) well overestimated. With some training (don't do that for obvious reasons: https://www.klippel.de/listeningtest/?v=3) IM as low as 0,3% become perceivable when playing pop-music. So, a speaker @3% IM (-30db) I would consider overloaded, used out of specs.
I still don't understand the relation to audio-quality. It sounds (pun intended) as if it were more suitible to identify shattering noises, rattling the like. With music, say a foundational bass note plus singing on top would generate IM in a two-way speaker. A 3-way is way better (pun not intended). What I'm after with a breakthrough is a new design paradigm.
versus @500Hz
- Joined
- Dec 17, 2021
- Messages
- 647
- Likes
- 1,417
So at 7:40 in that video, they play a tone that is supposedly masked. However, I can hear the tone failry easily the instant it starts (miniDSP Flex > Topping PA5 > Revel M105). Maybe you can't hear it on crappy computer speakers, but I can hear it fine on my desktop system. Am I the only one? Am I misunderstanding that I shouldn't hear the tone, and intead it's normal to hear it, but it's just quiet under the other tone? The fact that I can clearly hear it makes it hard for me to buy into ideas based on the fact that I should not hear it. Maybe he used a bad example.
I can hear the "masked" tone on my MacBook Pro speakers also so maybe a bad example (or an artefact of Youtube??)
Yes I have the same experience, the masked tone is audible immediately. Not dramatically, but nonetheless unmistakably. I don't know the reasons for or implications of this (perhaps YT adds higher harmonics). The theory looks good otherwise.
The concept of the masking curve isn't new of course, but nicely described and applied here. And I've always liked the representation we see in the 'human domain' spectral graphic starting around 8:50 when it is used in speaker reviews. I don't expect human domain from ASR (surprise me) but more people should do it, certainly
Also, I want that acoustic camera.
Last edited:
