I have been studying a lot of driver measurements from different sites, and what I can tell is: cheap drivers measures poor, midrange driver can measure excellent or poor, hi0end drivers usually measures great, but can bad too.
My main interest were midrange drivers 3-5 inches, both cone and dome. As a result I saw that:
- Scan-Speak 12MU, one of arguably best midrange on this planet, measures just OK
- SB Acoustics NBAC15 - kills the 12MU in terms of harmonic distortions, especially third, costs 20% of the former
- Accuton - despite the cost - very high distortion on 3rd harmonic (few drivers checked, 90-170mm)
Yet, most people find their sound qualist in following order: Accuton, SS, SB Acoustics. So, what else really matters to determine if one is good or great great driver? Waterfall? Time-Energy Curve? On-Axis response?
Why to pay 10 times more for Accuton, if cheap SB has literally every harmonic order better?
And, if those are telling us next to thing, why to measure it at all (besides frequency response)? Highly respected Seas drivers are nothing special in terms of harmonic distortions, yet a lot of hi-end manufacturers are using them.
I was looking for a drivers to handle something between 400-3000Hz, and while already looking on a quality manufactured drivers, why should I actually spend hefty money on Scan-Speak or Accuton? Actually why to buy Illuminator 12MU, if the old SB12NRXC measures actually the same on distortion graphs? And where I can clearly read/see the advantages of Scan-Speak over the paper SB?
Yes I know there are big variances in motor design, coil size and so on, which translates to sound, but this should be clearly visible on mesurements.
Or, we don't know what to measure, and how to combine different graphs to get proper conclusion?
If there is no measurable variance, why to pay much more for better drivers? Everyone hears, but noone can measure?
My main interest were midrange drivers 3-5 inches, both cone and dome. As a result I saw that:
- Scan-Speak 12MU, one of arguably best midrange on this planet, measures just OK
- SB Acoustics NBAC15 - kills the 12MU in terms of harmonic distortions, especially third, costs 20% of the former
- Accuton - despite the cost - very high distortion on 3rd harmonic (few drivers checked, 90-170mm)
Yet, most people find their sound qualist in following order: Accuton, SS, SB Acoustics. So, what else really matters to determine if one is good or great great driver? Waterfall? Time-Energy Curve? On-Axis response?
Why to pay 10 times more for Accuton, if cheap SB has literally every harmonic order better?
And, if those are telling us next to thing, why to measure it at all (besides frequency response)? Highly respected Seas drivers are nothing special in terms of harmonic distortions, yet a lot of hi-end manufacturers are using them.
I was looking for a drivers to handle something between 400-3000Hz, and while already looking on a quality manufactured drivers, why should I actually spend hefty money on Scan-Speak or Accuton? Actually why to buy Illuminator 12MU, if the old SB12NRXC measures actually the same on distortion graphs? And where I can clearly read/see the advantages of Scan-Speak over the paper SB?
Yes I know there are big variances in motor design, coil size and so on, which translates to sound, but this should be clearly visible on mesurements.
Or, we don't know what to measure, and how to combine different graphs to get proper conclusion?
If there is no measurable variance, why to pay much more for better drivers? Everyone hears, but noone can measure?
Another type of test that's not normally published is the shaped tone burst/energy storage one that Linkwitz liked:
https://www.linkwitzlab.com/mid_dist.htm
There are also the Klippel measurments that most manufacturers don't share. Those show more linearity/motor effects than a typical driver datasheet contains.
Of course optimally you would have great performance on all tests. With current technology, you typically have to make compromises on nearly everything. Each company/designer has their own priorities and cost constraints, which is why you see so many different types of construction still.
At the higher end of performance, you're often chasing minor improvements. Like any luxury item, practicality, cost constraints, bang for the buck, etc. often go out the window.
https://www.linkwitzlab.com/mid_dist.htm
There are also the Klippel measurments that most manufacturers don't share. Those show more linearity/motor effects than a typical driver datasheet contains.
Of course optimally you would have great performance on all tests. With current technology, you typically have to make compromises on nearly everything. Each company/designer has their own priorities and cost constraints, which is why you see so many different types of construction still.
At the higher end of performance, you're often chasing minor improvements. Like any luxury item, practicality, cost constraints, bang for the buck, etc. often go out the window.
Last edited:
To answer the thread title - Absolutely they matter.
Like anything, looking at and comparing incorrect, wrong or flawed measurements will get you nowhere (or down the wrong road).
I'm also going to suggest that if all other things are held equal, the more efficient driver will perform better (lower voice coil temperatures).
Like anything, looking at and comparing incorrect, wrong or flawed measurements will get you nowhere (or down the wrong road).
I'm also going to suggest that if all other things are held equal, the more efficient driver will perform better (lower voice coil temperatures).
Are you judging them by what they look like or what the measurements tell you they sound like?
I am a mathematician specializing in adaptive filtering. I know how to separate the linear part of a driver/loudspeaker signal from non-linear distortions for any input including music. You can excite a driver in a test enclosure with a band-pass driver emulating your crossover, and see (and hear) the distortions separately.
For example, 15 sec of Mozart piano concerto, 300-3000Hz:
and the distortions (residuals of LTI filtering) are:
Now let's see the spectrum of sine-sweep distortions:
or in the "parsed form":
Some (most of
speakers are pathetic in passing rising fronts of a signal.
As you can see, constant amplitude sine sweep measurements can not tell you the entire story, far from it. The "traditional" methods of loudspeaker measurements are antiquated and laughable ... but it's the best we have had.
Actually, I am interested to learn how many DIYers would find the examination of adaptive filtering distortion residual useful?
For example, 15 sec of Mozart piano concerto, 300-3000Hz:
and the distortions (residuals of LTI filtering) are:
Now let's see the spectrum of sine-sweep distortions:
or in the "parsed form":
Some (most of
speakers are pathetic in passing rising fronts of a signal.As you can see, constant amplitude sine sweep measurements can not tell you the entire story, far from it. The "traditional" methods of loudspeaker measurements are antiquated and laughable ... but it's the best we have had.
Actually, I am interested to learn how many DIYers would find the examination of adaptive filtering distortion residual useful?
Hi mikets42,
All we had were swept sine, and they revealed terrible speakers for sure. I would agree they won't tell the entire story, but they can show a lot. Once we had waterfall responses to examine, well that told you a lot more. We even tried measuring distortion, however the signal to noise ratio results in a high residual. Signal averaging today can drop the noise if it is random.
Passive crossovers (are never passive in that they affect things) really make controlling the drivers difficult. An active crossover with one amplifier per driver cleans everything up a great deal. So you have to approach a system with realistic expectations. A passive crossover is a "make it go" solution. Want performance? You need to use an electronic crossover and separate amplifiers.
Leading edge response will always look bad when you hit a driver with a rise time that is outside it's passband. They have mass and limited motor ability. That's why flame (ion) type drivers were investigated. Panel speakers follow the same thoughts.
All we had were swept sine, and they revealed terrible speakers for sure. I would agree they won't tell the entire story, but they can show a lot. Once we had waterfall responses to examine, well that told you a lot more. We even tried measuring distortion, however the signal to noise ratio results in a high residual. Signal averaging today can drop the noise if it is random.
Passive crossovers (are never passive in that they affect things) really make controlling the drivers difficult. An active crossover with one amplifier per driver cleans everything up a great deal. So you have to approach a system with realistic expectations. A passive crossover is a "make it go" solution. Want performance? You need to use an electronic crossover and separate amplifiers.
Leading edge response will always look bad when you hit a driver with a rise time that is outside it's passband. They have mass and limited motor ability. That's why flame (ion) type drivers were investigated. Panel speakers follow the same thoughts.
Hi Anatech,
I use 2 mics - a calibrated measurement mic (with noise in mid-30s dBA), and a low-noise condenser (12 dBA). Then I match the condenser to the measurement electret because both are min phase. That alone gives you 100x reduction in the need for averaging.
I am not talking about linear distortions. Everything is within the bandwidth. There are perfectly normal LTI waveform shaping variations due to limited bandwidth FR. They are fine. I am talking about purely non-linear distortions. Too bad I can not attach .mp3 with distortions for you to hear.
I use 2 mics - a calibrated measurement mic (with noise in mid-30s dBA), and a low-noise condenser (12 dBA). Then I match the condenser to the measurement electret because both are min phase. That alone gives you 100x reduction in the need for averaging.
I am not talking about linear distortions. Everything is within the bandwidth. There are perfectly normal LTI waveform shaping variations due to limited bandwidth FR. They are fine. I am talking about purely non-linear distortions. Too bad I can not attach .mp3 with distortions for you to hear.
Hi, morning coffee philosophy.
Big part of sound is how good you can make the whole system. If you must use only passive crossover for example, with some budget constrain and so on, another driver could give better system performance even though it wasn't technically as clean looking but if it resulted nicer system response you managed to come up with. On the other hand, you can put aside almost all of the stuff in your post and design your system not based on drivers but based on requirements for sound. Design the system with ideal drivers so that it meets your goals, and then just use best drivers that fit to the design, and budget. I mean, why wonder about some driver, if it is not even fit to your system. Design first, buy drivers last. Well, if you must use some driver and design around that, then why even wonder about this stuff?
Hi, all data is fun, especially if it's useful!
I'm sure there is lot's of people interested on what you have there. I find the residual sweep graph above very peculiar, gives insight on cone resonance. Horrible sound on the mp3, anything that helps make sure it's not audible would be very good.
Hi, look information about "current drive" and how it works, what are the mechanisms at play. Passive parts actually can reduce driver motor distortion going acoustic, as long as they increase impedance in series with the driver. Hooking a driver directly to an voltage amplifier makes basically a short circuit between driver terminals, which enables maximal acoustic emission of driver motor distortion. Controlling of driver you refer to happens only at drivers resonant frequency, where mass is cancelled by spring, which could reduce excursion and thus also reduce distortion. So, either could be nice for some application but why not utilize both! One can be clever with it, use low impedance on resonant frequency, and high impedance above. With active system one could use passive parts in addition, not to filter but to manipulate impedance. This kind of hybrid system gives relatively independent frequency response and circuit impedance control.
Last edited:
average 8' high living room, some sound absorption on the wall, area rugs throughout (I have dogs). Measurements are made in the center, 4' up, directed to avoid exciting mage resonance modes.
I use a dual mic setup for measurement mic is not good for anything besides measuring FR. A low-noise condenser mic, manually (sw) equalized to match the measurement mic, is used for actual measurements. There is also a sound level meter - just in case because the measurement mic was calibrated to have 94dB SPL at -11 dBFS. Cross-examination of a kind.
The DUT is a speaker mounted on a quite rigid test box full of damping foam.
The DUT and mics must be as close as possible without exceeding mic's {max sound pressure -25} but also to avoid letting amp noise dominate the room noise. Outside of the picture: Focusrite 2i2 G3 and 20i18 G1, i7-11700K desktop, TI's 3255 amplifier, cables.
Room's windows must be closed and all constantly blowing appliances stopped. Wait for a refrigerator to calm down and here we go....
You will need a home license of MATLAB (US$150 last time I checked) with signal processing and audio toolboxes. Register, download https://www.mathworks.com/matlabcen...peakers-for-aec-measurement-and-linearization
(GPL freeware) and start having fun! All your feedback is appreciated as I have endless blind spots.
Did I miss anything?
If this math-loaded approach gets enough customers then someone shall convert it into a more usable package,
An important thing with measurements is that don't measuring the right thing can easily lead to false conclusions and there are many things that can be measured and often even harder to interpret. Added harmonics (aka harmonic distortion) is just one of them.
Anyway, I don't think there are two different drivers/speakers that don't have any measurable difference.
Hi tmuikku,
Very familiar with current drive. It's been poked at since at least the 1970's and earlier. There is a very good reason it hasn't caught on.
You can always focus on one aspect of driving a speaker, but you have to look at the entire picture. The sound reproduction industry always experiments with different concepts, but in the end it does get it right and the market moves forward with the best technology. Absolutely nothing Joe or others have come up with is new these days. Just because someone takes off along a different path does not make them insightful or brilliant.
Hi YSDR,
I'll agree. However more measurements does increase understanding and also helps determine what is important. Audio electronics has travelled this same path and we now exist in a time when it is very well understood. However we have a large number of folks who don't understand. Old ideas are difficult to dispel.
One thing that is proved and true. Using modern measurements, the resulting speaker systems are vastly improved. PSB and others used the Canadian NRC facility and bang - world class speaker system designs. JBL, same thing as they hired Floyd Toole from the NRC, their products got better. Everyone here must be doing something right, they sure didn't design by ear!
-Chris
Very familiar with current drive. It's been poked at since at least the 1970's and earlier. There is a very good reason it hasn't caught on.
You can always focus on one aspect of driving a speaker, but you have to look at the entire picture. The sound reproduction industry always experiments with different concepts, but in the end it does get it right and the market moves forward with the best technology. Absolutely nothing Joe or others have come up with is new these days. Just because someone takes off along a different path does not make them insightful or brilliant.
Hi YSDR,
I'll agree. However more measurements does increase understanding and also helps determine what is important. Audio electronics has travelled this same path and we now exist in a time when it is very well understood. However we have a large number of folks who don't understand. Old ideas are difficult to dispel.
One thing that is proved and true. Using modern measurements, the resulting speaker systems are vastly improved. PSB and others used the Canadian NRC facility and bang - world class speaker system designs. JBL, same thing as they hired Floyd Toole from the NRC, their products got better. Everyone here must be doing something right, they sure didn't design by ear!
-Chris