Hi mikets42,
No they don't. There is no difference between music and two or more tones. The testing must be repeatable by anyone and at different points in time.
Hi soundbloke,
Using motional feedback, the waveform across the driver is identical whether you want to call it current or voltage amplification. We are simply playing with terminology now.
You threw in the massive wrinkle that changes any argument to an entirely new field. Motional feedback is not what is in the general market. I've heard some high end systems that really do not sound very good using motional feedback, and had the system been measured it would have showed the defects. What you are talking about isn't a magic bullet and depends massively on tuning and compensation. In effect, this is an entirely new topic that is still developmental in nature. What doesn't change one iota by bringing this in is measurement techniques and their importance. So this entire line of discussion is really off topic. That is what I tried to say earlier.
So how about we remain on topic? We are talking about measurements on loudspeaker systems and how important it may be. How you make the speaker diaphragm move is completely immaterial.
No they don't. There is no difference between music and two or more tones. The testing must be repeatable by anyone and at different points in time.
Hi soundbloke,
Using motional feedback, the waveform across the driver is identical whether you want to call it current or voltage amplification. We are simply playing with terminology now.
You threw in the massive wrinkle that changes any argument to an entirely new field. Motional feedback is not what is in the general market. I've heard some high end systems that really do not sound very good using motional feedback, and had the system been measured it would have showed the defects. What you are talking about isn't a magic bullet and depends massively on tuning and compensation. In effect, this is an entirely new topic that is still developmental in nature. What doesn't change one iota by bringing this in is measurement techniques and their importance. So this entire line of discussion is really off topic. That is what I tried to say earlier.
So how about we remain on topic? We are talking about measurements on loudspeaker systems and how important it may be. How you make the speaker diaphragm move is completely immaterial.
- Is this judged for all on the same standard baffle?
- Is this form being a part of a complete speaker?
- is this for all on a IEC baffle and EQed within +/- 0,2 dB
I think that you will find that a very good speaker that need a midrange between 400-4000 will sound the best if you choose a driver that has the best clinical performance figures (straightest FR, lowest IMD, smoothest directivity, highest efficiency (least losses), highest power handling, straightest impedance trace (least mechanical deficiencies) and the cleanest waterfall / spectrogram.
I'm quite sure it would. I would bet on it. I would choose that driver myself - unheard - no hesitation.
//
Gladly, and I happily refer you back to my first post (#24) in this thread, where you will see I merely responded to the comments of others re current drive and in specific relation to the subject of this thread.
But this does not mean misconceptions should not be corrected...
No, the terminology is important. The coil current will not be the same in both cases if the coil impedance changes. With a high impedance (current) source, the changes will be irrelevant; With a low impedance (voltage) source, added distortion will be present.
Critically, the coil impedance contains non-linear reflections of certain motor system distortions (as I have alluded to above). In converting from a low impedance (voltage) source to coil current, there will be distortion apparent in the output due to the aforementioned reflections. There is no such conversion in the signal path when a high impedance (current) source is employed, and therefore no such distortions.
(MFB is not the important factor here, it is the amplifier output impedance that is of relevance. Moreover using MFB to correct the non-linearities of which I am referring to would present a formidable challenge and likely require driver modification too. But it is just not relevant to the matter at hand).
With specific reference to this thread, if anyone is concerned why there moving-coil drivers do not sound as "transparent" or "noise free" as electrostatics, for example, these motor system non-linearities are a good place to start looking for your evidence. Certainly if you wish to correlate distortion measures with sound quality, simply driving your moving-coil loudspeakers in one particular mode will not serve well in isolating what you are looking for.
Driver impedance measures have a lot of information to offer; Current driving moving-coil drivers can confer some significant audible benefits too. And given this is a DIY forum, general market considerations are not of particular concern. I whole-heartedly recommend contributors here experiment for themselves.
How can you be sure that your test is correct and complete ?
How do you know you are doing your testing correctly and if you have the right instruments ?
If all those measurement criterias were met, then all speakers would be "Test good = Sound good". But outside the measurement laboratory schedule, we know that this is not always true...
T
Hi soundbloke,
I'm not going to get drawn into a current vs voltage drive argument. It does not apply to the thread at all. What matters is measurements and they don't care if the air is moved by force field, your mind, electrostatics or dynamic drivers. Get my point?
The voltage across the terminals is locked to the current flow. The driver is an impedance and you cannot separate the two, voltage across and current through, from each other. If you do, you have broken the laws of physics and the world as we know it ends. So however the correction mechanism acts, the end can be arrived at from either direction and arguing which is just plain silly.
No, you seem to have little grasp of the physics involved. I will repeat once more...
In current drive, coil current is fixed by the driving source on account of its high impedance.
In voltage drive, coil current is established by said voltage presented across the varying coil impedance.
Therefore in voltage drive, the output is prone to distortions reflected in the coil impedance, whereas in current drive the output is immune to such distortions.
It is that simple.
It is also relevant to this thread because of how different distortion mechanisms are perceived differently according to how they are correlated with the source that causes them. See my post #24 (or several other threads that I have contributed to in this forum) that discuss how to better relate measurements to subjective reports.
In your own words...
I have provided a very simple but sufficiently rigourous explanation as to why your understanding is flawed.
To reiterate, it absolutely matters how you generate the drive signal.
And to (try and) clarify why that is yet again, the relationship between voltage and current only matters in the voltage drive case. For current drive, the relationship is irrelevant.
I'm sorry, put a current through an impedance and you get a resulting voltage. Apply the same voltage across the same impedance and magic happens - you get the same current.
That's enough. Possibly your system or experiments are not good enough to show this, but I am afraid this relationship is true. If your correction system of either the voltage or current drive case does not correct properly, I can see why you might say this.
That's enough. Possibly your system or experiments are not good enough to show this, but I am afraid this relationship is true. If your correction system of either the voltage or current drive case does not correct properly, I can see why you might say this.
How ever you choose to express Ohm's Law or its adaptions makes no difference whatsoever!! To restate again...
In current drive, coil current is fixed by the driving source on account of its high impedance.
In voltage drive, coil current is established by said voltage presented across the varying coil impedance.
Therefore in voltage drive, the output is prone to distortions reflected in the coil impedance, whereas in current drive the output is immune to such distortions.
Thus, Ohm's Law is only relevant to the voltage drive case where it admits additional distortions due to the non-linear coil impedance.
Is there really anything in this reasoning that you believe is contrary to elementary physics?
Enough, you are incorrect.
As I said, you have the same impedance. To have the same cone motion, however you drive it through a correction network of some kind, the same current and voltages must exist.
If you drive a speaker with pure current drive, guess what? The output levels follow the impedance curve. If that isn't distorted, tell me what is? Plus resonances are not damped at all. I think you are trying to argue a corrected current drive against a standard voltage drive. You are not comparing apples to apples.
As I said, you have the same impedance. To have the same cone motion, however you drive it through a correction network of some kind, the same current and voltages must exist.
If you drive a speaker with pure current drive, guess what? The output levels follow the impedance curve. If that isn't distorted, tell me what is? Plus resonances are not damped at all. I think you are trying to argue a corrected current drive against a standard voltage drive. You are not comparing apples to apples.
Enough for who? For what? Your ego??? I am certainly not incorrect.
By contrast, you have demonstrated a fundamentally limited grasp of the subject, and yet just expose your misunderstanding more with each reply!
In current drive, the output is independent of the impedance and any non-linearities reflected therein. This is not the case for voltage drive. It could hardly be more simple.
Furthermore, the only resonance of concern is the fundamental and singular mechanical resonance, which can be better controlled via MFB than it can via voltage drive, where damping is both prone to manufacturing tolerances and dynamic changes precisely because the amplifier output resistance is so small. But this aspect (as I stated previously) is irrelevant to the distortion mechanisms under discussion here.
Everything I have stated is correct and I have made no unfair comparisons. I have yet to work out what exactly you are arguing about!
Anatech: "No they don't. There is no difference between music and two or more tones."
Please enjoy a comparison of various drivers and loudspeakers/headphones by LTI distortions. The proof is in the pudding, hear and you will see
You may find that drivers that THD-wise measure (and priced) very differently produce essentially the same distortions and a driver that measures about the same sounds quite different. Recordings of a reference-grade loudspeaker and good studio-grade headphones are also included. Hope that helps.
Anatech: "The testing must be repeatable by anyone and at different points in time." I fully agree that this is very desirable. But first, we shall find out if these tonal tests are relevant, which I have no evidence for. Then we can go on discussing the finer details of measurements, reproducibility, robustness, error margins, etc.
Markbakk: Unless you have a decently anechoic chamber, standard practices do not work. Whenever you don't, you have to play tricks. I used to have access to a fully-specified anechoic chamber, which I cross-examined with my setup. Actually, it works.
BTW, I can produce a video with distortions of music by the same driver but loaded with additional resistors, with the electric level adjusted to the same loudness. Anatech, soundbloke, will that help to resolve the disagreements?
Please enjoy a comparison of various drivers and loudspeakers/headphones by LTI distortions. The proof is in the pudding, hear and you will see
You may find that drivers that THD-wise measure (and priced) very differently produce essentially the same distortions and a driver that measures about the same sounds quite different. Recordings of a reference-grade loudspeaker and good studio-grade headphones are also included. Hope that helps.
Anatech: "The testing must be repeatable by anyone and at different points in time." I fully agree that this is very desirable. But first, we shall find out if these tonal tests are relevant, which I have no evidence for. Then we can go on discussing the finer details of measurements, reproducibility, robustness, error margins, etc.
Markbakk: Unless you have a decently anechoic chamber, standard practices do not work. Whenever you don't, you have to play tricks. I used to have access to a fully-specified anechoic chamber, which I cross-examined with my setup. Actually, it works.
BTW, I can produce a video with distortions of music by the same driver but loaded with additional resistors, with the electric level adjusted to the same loudness. Anatech, soundbloke, will that help to resolve the disagreements?
I imagine all kinds of effects can be detected with music, no argument from me on that. However we have to be certain the stimulus is all in-band for the driver. Because we have mass acceleration involved as well, you will detect issues with that that are unavoidable.
I think the trick in test stimulus is that the signal must be able to be reproduced exactly (easy with digital files), but also that the level is identical (easy with calibration tones). Once you can create a standard stimulus that anyone can reproduce, and an agreed upon test environment with equivalent microphones (in this case) you have what could be a standardized test. Great.
The next thing to determine is what is common or unavoidable, then deemphasize that portion of the results and concentrate on what will be heard as issues. After all, pointing out something all speakers do is not very useful. We are looking at things that can determine if one driver or speaker system is better than another.
I do see a lot of value in those kinds of tests.
I think the trick in test stimulus is that the signal must be able to be reproduced exactly (easy with digital files), but also that the level is identical (easy with calibration tones). Once you can create a standard stimulus that anyone can reproduce, and an agreed upon test environment with equivalent microphones (in this case) you have what could be a standardized test. Great.
The next thing to determine is what is common or unavoidable, then deemphasize that portion of the results and concentrate on what will be heard as issues. After all, pointing out something all speakers do is not very useful. We are looking at things that can determine if one driver or speaker system is better than another.
I do see a lot of value in those kinds of tests.
To answer the OP’s question, “Do measurements matter for sound?”, surely we need to start by choosing a number of components with a sound we like and then look for common features in their measurements which might help explain why we like them. As far as I’m aware, no one has ever found a direct correlation, although I’d be happy to be corrected.
Hi, it would be very cool if you did this! Also try inductor if you have one at hand. If possible, besides level matching can you match acoustic frequency response. It would be very cool to hear the difference. Thanks!