As Earl says, it shouldn't be the speaker that seduces you but the music - the speaker should get out of the way, be transparent to & accurately replay the electrical signals it is receiving.
The seduction only occurs when nothing is distracting one's attention from the soundfield being replayed
It's possible but it appears that we are sensitive to accurate reproduction - it's an internal QA we all seem to have. Where I find it gets complicated is that we are often listening to a studio creation, not a recording of a live event so this is very much a combination of the musician's art with the recording engineers art played back over a 2 channel system which is itself not capable of reproducing the full soundfield, just an illusion that has proven sufficiently satisfying. So the question, "what is accuracy", does get complicated when judged for the perspective of auditory perception.
It may be possible that some of the shortcomings in this whole setup can be compensated by certain types of distortions in some circumstances/setups? Again, without knowledge of auditory perception & how it works, what it expects, we are just guessing at these issues
Can all of this lead to situations where our preference for accuracy is over-ridden? Sure it can.
You are misinterpreting why they don't emotionally connect you to the music - they are missing whatever it is which allows your auditory perception to seamlessly communicate the with the emotional centers of the brain - it may be that there are some elements in the soundfield which contravene it's internal model of realistic sound i.e we are drawn more towards the playback itself than to the music being played.
Audio forum ABX blind tests are certainly not the way to measure this but not for the reason you state. Olive's blind AB preference tests show that people from all walks of life & experience gravitate towards accuracy & that these obviously are more enjoyable. The limitations of the test have to be borne in mind but in this particular test, accuracy wins the day, preference-wise.
for further clarification so i understand things a little better if THD and IMD numbers don't serve as an indication of sound quality what does? (i sincerely hope that doesn't seem like a troll question as i would like to explore that topic!?) i clearly see linear response as king and that if sufficiently reduced distortion is irrelevant but where does the divide lay?
I can count on my fingers how many times I've ever been anywhere near passing the "down the hall" criterion. I wouldn't even mistake my cell phone electronic ringer for a loudspeaker reproduction of it on my electrostats even from down the hall. Would you... if tested blind?
So if by distortion we mean passing the "down the hall" criterion as a true goal, we sure have a lot of distortion to address. Finding a yet more perfect Rice-Kellogg driver would not seem to be helpful, as Earl has argued. Adding motional feedback to address ringing, would be helpful.
So we need to think what are at the sources of "distortion" in getting to the true goal. That includes speakers, locations, rooms, and their interactions. If we can't reproduce a cellphone ringer, we have a ways to go.
(BTW, music is full of harmonics and intermodulations. I find it easy to believe you can achieve a sub-threshold levels of HD and IMD for music reproduction.)
B.
Last edited:
As I said before, there is no established means of quantifying the level of nonlinear distortion that is objectionable. It has not been studied to any great extent. Why? you may ask. Because the little work that was done has shown that there just isn't enough nonlinear distortion in a good loudspeaker system to worry about, so no one studies it. I started out to solve this problem, but quickly learned that there was nothing to be gained from such studies. That's pretty much what everyone (with a few exceptions) has concluded.
So its not that exceptions to the rule don't exist, where what we know to be true doesn't apply, its that these exceptions are not sufficiently interesting to warrant the effort to study them. Studying nonlinear distortion in loudspeakers is just never going to bear much fruit. So here we stand not able to define what "sufficiently reduced distortion" means in quantifiable terms, except to say that it probably only matters in rare cases.
Let me say my post was meant to stoke the conversation, and yes I do believe there are certain assumptions about audio that may ultimately prove untrue, but haven't yet. A lot of audiophilia seems like chasing ones tail since no one really knows.
Sure the accurate speaker is better but with rapidly diminishing returns. The music only exists in the creative sense. There isn't perfect accuracy when the music is recorded and mixed. For example, much music will be too bass heavy if the speaker is flat to 20 hz.
Consider this, what is more accurate, a line source or point source? They certainly sound different. It seems common sense that some speakers excel with different rooms and with different music and so there are many combinations.
Why doesn't everyone listen to electrostatic headphones that are equalized flat? Very accurate. I am suggesting that there could be a performance aspect to a speaker, that involves source, the look of the speaker (gasp), dispersion, room, types of distortion, etc.
Further its entirely possible that there is such a thing as euphonic distortion that sounds good. Maybe its just dumb luck. Could something like a painting be improved upon? Its reasonable to think that a speaker and room can better express some music by errors of commission or omission.
The seduction occurs because it occurs. Generally distracting discordant sounds get in the way. The issue is that if we experience music this way, you can't really say that perfection is achievable or that there is a perfect speaker, even if it transduces perfectly.
No disagreement here,
The pistonic speakers are lacking that nice rippy break up sound that paper drivers have. I also know thinking critically about the speaker itself kills the enjoyment faster than anything, so maybe that's the lesson
Sure the accurate speaker is better but with rapidly diminishing returns. The music only exists in the creative sense. There isn't perfect accuracy when the music is recorded and mixed. For example, much music will be too bass heavy if the speaker is flat to 20 hz.
Consider this, what is more accurate, a line source or point source? They certainly sound different. It seems common sense that some speakers excel with different rooms and with different music and so there are many combinations.
Why doesn't everyone listen to electrostatic headphones that are equalized flat? Very accurate. I am suggesting that there could be a performance aspect to a speaker, that involves source, the look of the speaker (gasp), dispersion, room, types of distortion, etc.
Further its entirely possible that there is such a thing as euphonic distortion that sounds good. Maybe its just dumb luck. Could something like a painting be improved upon? Its reasonable to think that a speaker and room can better express some music by errors of commission or omission.
The seduction occurs because it occurs. Generally distracting discordant sounds get in the way. The issue is that if we experience music this way, you can't really say that perfection is achievable or that there is a perfect speaker, even if it transduces perfectly.
No disagreement here,
The pistonic speakers are lacking that nice rippy break up sound that paper drivers have. I also know thinking critically about the speaker itself kills the enjoyment faster than anything, so maybe that's the lesson
Last edited:
i made no reference to what type of distortion. i find it interesting that you insert "non linear".
can i ask why?
i'm trying to understand your perspective better.
can i ask why?
i'm trying to understand your perspective better.
Last edited:
I'm pretty sure that's been covered before too, but I'll answer. The surest way to see if I really know is to post it on the Internet and see what corrections I get.
When something is Non-linear means it causes harmonic and intermodulation distortion. A speaker example might be a spider or surround that takes more force to move it one direction than it does to move the same distance the other direction, thus causing a flattening of the peaks in that direction. This is roughly analogous to a tube/valve amplifier stage going into saturation.
"Linear" distortion is caused by resonances, delays, diffraction, dispersion, a crossover with a less-than-perfect frequency response, or anything that could vary the frequency response, on-axis or off-axis. These don't add harmonics or cause intermodulation distortion, but they do change the sound of the speaker, if I understand what Mr. Geddess is saying, these are responsible for almost all of the differences between high-end speakers.
I recall a fellow engineering student in college saying "You know why they never give THD specs for speakers, don't you? Because they're so high!" Looking back now, I can see that even as this is true, a driver's nonlinear distortion is much more like that of a tube/valve amp, generating only lower harmonics that are easily masked in most music.
I certainly would like to listen to some of the speakers anyone here has designed. Lots of repeating arguments which I have no way of really understanding how it will improve sound. So let’s see if someone can suggest a measurement method for the case below.
I was long using normal stuffing material inside speaker enclosures, but quite often find that same stuffing and same weight will sound different with each speaker no matter how hard you try to stuff them the same way.
So my question would be, how do you measure them to ensure that the differences are inaudible?
I was long using normal stuffing material inside speaker enclosures, but quite often find that same stuffing and same weight will sound different with each speaker no matter how hard you try to stuff them the same way.
So my question would be, how do you measure them to ensure that the differences are inaudible?
I would start by measuring the impedance curve of that driver in it's enclosure.
You'd be surprised what you can get out of that for starters.
You'd be surprised what you can get out of that for starters.
Last edited:
Generally the impedance does not show differences unless the stuffing quantity changes. How much change would be meaningful? Generally the lines are overlapping, while not absolutely (some edges of different colors will show). The difference is quite within the range that would exist in a batch of divers.
Which raises another question. Even if you do accept the published specs, the actual drivers are going to be off. Say you like a specific design, when you put together a pair yourself, there is probably no way to ensure you are going to get the sound which you originally heard, so is there anything one can do with all this knowledge floating around?
I made some enclosure measurements (not speakers) with medical grade pressure transducers, by drilling a small hole and inserting it so it samples the internal pressure. They respond to DC and the linearity and calibration are excellent the ones I had went out to 3.5kHz even though they are not specified for AC save rise time.
Have you noticed different results depending on where the sensor actually is located in the enclosure? It would be almost impossible to get it in the same location due to stuffing.
Last edited:
One needs to be willing to zoom in to the graphs to get to actual differences, but it would make sense to start with a set of well build drivers and get their free air curves first, both impedance and FR.
What are you willing to do. How far do you want to go. Get into a scientific mode and question everything and measure every step to confirm.
Plus there's always FIR filtering for those that trust enough in that. Lots of ways to get that FIR part wrong, resulting in even more that could be interpreted wrong. Don't question the means, question your own actions. Dig a little deeper.
I'll openly admit that I've had cases of expectation bias influence me.
Dissect an IR to find what it's made of. If you hear a difference, chances are you can find and measure them, if you're willing to invest the time.
Start a thread and show or share the measurements. Let's see what we can dig up.
What are you willing to do. How far do you want to go. Get into a scientific mode and question everything and measure every step to confirm.
Plus there's always FIR filtering for those that trust enough in that. Lots of ways to get that FIR part wrong, resulting in even more that could be interpreted wrong. Don't question the means, question your own actions. Dig a little deeper.
I'll openly admit that I've had cases of expectation bias influence me.
Dissect an IR to find what it's made of. If you hear a difference, chances are you can find and measure them, if you're willing to invest the time.
Start a thread and show or share the measurements. Let's see what we can dig up.
Last edited:
I was wondering if anyone has done this before. To me, it would seem quite annoying if I paid for something that I liked and actually get something that only looks the same. Does this not bother anyone here? Nobody has discovered this?
Putting sensors through holes in your speaker isn't needed at all. Maybe looking at the impedance wouldn't hurt.
An overstuffed speaker sounds like someone has there finger on the cone, and in bass reflex the bass will be less (though the very bottom end won't be affected as much.
Drivers that aren't broken in can also sound like a stuffing issue so rule that out first. Pretty much go by ear.
Though if you are a manufacturer you might want to do some serious investigation if you think that there is a consistency problem
An overstuffed speaker sounds like someone has there finger on the cone, and in bass reflex the bass will be less (though the very bottom end won't be affected as much.
Drivers that aren't broken in can also sound like a stuffing issue so rule that out first. Pretty much go by ear.
Though if you are a manufacturer you might want to do some serious investigation if you think that there is a consistency problem
Last edited:
Why would you be surprised? The better the engineering behind the driver the more likely it is they will closely resemble each other.
High dollar product isn't a guarantee for this.
Go by ear would be my last resort. If I think I hear a difference I'll grab the measuring gear and find out.
If I can't find out I'll rethink, dig deeper and try to learn more about it, in other words, question my own competence.
I cannot completely trust my ears, as said, I have proven to myself to be susceptible to expectation bias. So I'll need to use more steps to rule that out too. I used to think something like that would be impossible to happen to me, fooling myself? Nehh.. that couldn't be it, could it?
Last edited:
Because one must be clear about what type of distortion we are talking about. Making sweeping claims about "distortion" is misleading at best and completely wrong worst case.
The more pertinent question is: "How do you know that what you think is an audible difference actually is?" You will never measure a difference in things when none exists even though you think it must.