Hi mikets42,
Makes perfect sense. Pro audio companies typically operate on a shoestring. They may not behave professionally. I've had my own negative experiences.
Nothing wrong with making things better at all. Select an achievable goal. As for measurements, heck yes! Anything to make them easier and more repeatable!
Makes perfect sense. Pro audio companies typically operate on a shoestring. They may not behave professionally. I've had my own negative experiences.
Nothing wrong with making things better at all. Select an achievable goal. As for measurements, heck yes! Anything to make them easier and more repeatable!
Hello Lars and All,
I can not say that I have seen that, This type of noise is rarely reported. "Bark....ing". this noise is only addressed in a handful of posts primarily by @mikets42 .
Will this noise show up in a single frequency sine wave FFT or a two tone test sampled across a current sensing resistor?
What am I expecting to see on the FFT plot?
I will test a woofer with a APx555 analyzer.
I have a $10.00 12 inch woofer sitting here. mikets42 tell me the setup; how many watts, what frequencies all that setup sort of stuff?
Thanks DT
D
Deleted member 375592
Barkhausen noise will not show up as a frequency in the FFT plot, nor with dual tone. You need the latest REW beta and to choose FSAF measurements. Your mic shall be low noise (<20dBA) and linear, i.e. a decent true condenser or B&K. No artificial tones. True music only. I have observed this noise on the 80dB SPL RMS @ 1m, which is not too loud, but your mileage may vary. I did not do much detailed research on the topic, it's too boring for me.
Correct, this type of noise can not be observed (except for the extreme cases Lars reported) with "standard" measurements, which are IMHO of debatable meaning. AP staff is mostly useless, they act on assumptions that may have had some sense in the Stone Age. You need proper, full-blown, system identification and adaptive filtering, which does not make any implicit or explicit assumptions. Before my research on FSAF, these had been out of reach for most people who could not afford to lease supercomputers. Now, everyone can do these on a regular PC in less than a second, thanks to REW.
Correct, this type of noise can not be observed (except for the extreme cases Lars reported) with "standard" measurements, which are IMHO of debatable meaning. AP staff is mostly useless, they act on assumptions that may have had some sense in the Stone Age. You need proper, full-blown, system identification and adaptive filtering, which does not make any implicit or explicit assumptions. Before my research on FSAF, these had been out of reach for most people who could not afford to lease supercomputers. Now, everyone can do these on a regular PC in less than a second, thanks to REW.
This is my take on this:
Only a limited number of woofers will audibly produce this "noise" and then only the the best laboratory grade microphone can parse it from the noise floor. This noise may not be so audible and problematic as you would have us think. The solution is, buy better woofers.
If this noise is completely random it will not show up in the FFT. Testing with a sine wave I suspect that it is not so random after all and tends to be periodic.
If this noise is so low in amplitude perhaps sampling across a current sensing resistor will help.
For microphones My GRAS mics will work just fine.
Thanks DT
Last edited:
D
Deleted member 375592
An interesting take. Thank you for sharing it!
I am not trying to make you think anything. I may be completely wrong, as everybody else. I am suggesting you try the same experiments I did and tell us what you have found. IMO, a scientific theory shall be based on experiments done first, and unbiased conclusions made afterward. Also, a scientific theory can be disproven by a single experiment. So far, I have tested 2-3 dozen woofers, and each one showed Barkhausen noise. Once you hear it isolated, you start hearing it easily and understand that it was exactly what bothered you. I consulted my theoretical physicist friends, and they confirmed I was not daydreaming: that's why experimental science uses current to create a constant magnetic field, never a permanent magnet. Yet, it is absolutely normal for a novel theory to be rejected. That's how it was, is, and likely always will be. But please, argue with me based on hard data.
BTW, a non-scientific theory aka belief does not need experimentations or proofs and can not (and shall not) be argued with.
I am not trying to make you think anything. I may be completely wrong, as everybody else. I am suggesting you try the same experiments I did and tell us what you have found. IMO, a scientific theory shall be based on experiments done first, and unbiased conclusions made afterward. Also, a scientific theory can be disproven by a single experiment. So far, I have tested 2-3 dozen woofers, and each one showed Barkhausen noise. Once you hear it isolated, you start hearing it easily and understand that it was exactly what bothered you. I consulted my theoretical physicist friends, and they confirmed I was not daydreaming: that's why experimental science uses current to create a constant magnetic field, never a permanent magnet. Yet, it is absolutely normal for a novel theory to be rejected. That's how it was, is, and likely always will be. But please, argue with me based on hard data.
BTW, a non-scientific theory aka belief does not need experimentations or proofs and can not (and shall not) be argued with.
The use of a sensing resistor indeed seems promising. Come to think of it, the acoustic ‘approach’ on measuring this noise isn’t the most logical.
Maybe Barkhausen noise is not the most important kind of error made by a driver. But according to John Watkinson it limits the achievable SNR of drivers with ferrite magnets to less than what the 16 bits from the Redbook CD standard offer.
Neo magnets are less prone to it and everything else that is stiffening the permanent field, like shorting rings etc, helps to reduce it.
In metallurgy it is used to acoustically check alloys BTW.
Regards
Charles
Neo magnets are less prone to it and everything else that is stiffening the permanent field, like shorting rings etc, helps to reduce it.
In metallurgy it is used to acoustically check alloys BTW.
Regards
Charles
@DualTriode
Please don't alter quotes in any way. This shifts the emphasis, which changes the meaning in a way the original poster did not intend.
Please don't alter quotes in any way. This shifts the emphasis, which changes the meaning in a way the original poster did not intend.
Also consider the s/n ratio of the acoustic environment.
One danger in this hobby is chasing things that are far beyond mattering.
One danger in this hobby is chasing things that are far beyond mattering.
True. But don't forget that sometimes the sum of all the little things, of which each one alone doesn't seem to matter, may get quite important if they are all being disregarded.
I would definitley try current drive (or output impedance that is rising with frequency) if I had the time.
Regards
Charles
I would definitley try current drive (or output impedance that is rising with frequency) if I had the time.
Regards
Charles
Hi Charles,
Well, exactly. However you still have to consider what the limits of perception are. Humans don't have built-in lock-in amplifiers synced to the signal source, so the noise floor really does put a limit on what could possibly ever be heard. That is simply physics. If your listening environment has may 80 or so dB dynamic range between background noise and hearing damage (not unreasonable), the dynamic range from a 16 bit source is more than enough. Now we have oversampling and other methods to create an even wider dynamic range. I'm all for the best quality, but loudspeakers suffer huge distortion at higher levels - as do our ears. So we have to be reasonable in our expectations as well.
Current drive? If it was the way to go, we would be there. You probably want to design a different driver and then go for it. Early engineers weren't stupid by any means. They followed the correct technical path for the technology we have. If there is a breakthrough, fantastic. Maybe the future is direct manipulation of air molecules. That sinks all diaphragm based systems no matter how they are driven.
Well, exactly. However you still have to consider what the limits of perception are. Humans don't have built-in lock-in amplifiers synced to the signal source, so the noise floor really does put a limit on what could possibly ever be heard. That is simply physics. If your listening environment has may 80 or so dB dynamic range between background noise and hearing damage (not unreasonable), the dynamic range from a 16 bit source is more than enough. Now we have oversampling and other methods to create an even wider dynamic range. I'm all for the best quality, but loudspeakers suffer huge distortion at higher levels - as do our ears. So we have to be reasonable in our expectations as well.
Current drive? If it was the way to go, we would be there. You probably want to design a different driver and then go for it. Early engineers weren't stupid by any means. They followed the correct technical path for the technology we have. If there is a breakthrough, fantastic. Maybe the future is direct manipulation of air molecules. That sinks all diaphragm based systems no matter how they are driven.
D
Deleted member 375592
@anatech: I agree - Early engineers weren't stupid by any means. They followed the golden KISS principle, which was and still is the right thing to do. Unfortunately, the current drive is more complex, requires an amp tightly bound to a specific driver, and requires a good understanding of the control theory, which was not taught at the time to EE engineers.... nor is it now.
Douglas Self, "Power Amplifier", Chapter 13, page 350 "One implementation of nested feedback loops is the concept of Nested Differentiating Feedback Loops (NDFL), introduced by Edward Cherry in 1982. The original JAES paper [36] is tough going mathematically. A somewhat more readable account was published in Electronics Today International in 1983 [37] including a practical design for a 60W NDFL amplifier, though I cannot help thinking that Cherry lost 99% of his audience when he launched suddenly into complex algebra and Laplace variables."
I suspect that economics was also a factor. The cost of electronics used to be quite high but it dropped a bit in recent years. Now, a DSP+3 amps may cost less than a good passive crossover but a steep long learning curve can not be jumped over.
Douglas Self, "Power Amplifier", Chapter 13, page 350 "One implementation of nested feedback loops is the concept of Nested Differentiating Feedback Loops (NDFL), introduced by Edward Cherry in 1982. The original JAES paper [36] is tough going mathematically. A somewhat more readable account was published in Electronics Today International in 1983 [37] including a practical design for a 60W NDFL amplifier, though I cannot help thinking that Cherry lost 99% of his audience when he launched suddenly into complex algebra and Laplace variables."
I suspect that economics was also a factor. The cost of electronics used to be quite high but it dropped a bit in recent years. Now, a DSP+3 amps may cost less than a good passive crossover but a steep long learning curve can not be jumped over.
Not to mention that, at this time, current drive amplifiers are vaporware. They exist in theory, but there are no commercially available options, and no kits available for DIY (as far as I know...)
j.
j.
Hi stv,
Plasma is messy and expensive. Our PM would tax it to death as well (carbon tax). lol! Magnat had some plasma tweeters for example. Can you imagine a full range plasma driver (and BBQ)?
Hi mikets42,
Cost is a huge factor, then reliability on top of that. Then, people would not accept having to bind an amplifier to a specific set of speakers. That doesn't fly. Powered speakers are poorly supported and when they die, you're on your way to buy a new pair. Not sustainable. Imagine trying to explain why your separate amplifiers have to be used with one specific model of speaker. There's an uphill battle. People would simply opt for the way we have always done it. I know I would.
Hi hifijim,
There isn't a compelling need for current drive amplifiers on a consumer level. No reason to develop a kit or product. Even if you did, the public would mostly be confused over which amplifier type went with what speaker. Just imagine the confusion in audio stores, havoc on-line. Then there is the used market - disaster.
At the moment, current drive is a non-starter.
Plasma is messy and expensive. Our PM would tax it to death as well (carbon tax). lol! Magnat had some plasma tweeters for example. Can you imagine a full range plasma driver (and BBQ)?
Hi mikets42,
Cost is a huge factor, then reliability on top of that. Then, people would not accept having to bind an amplifier to a specific set of speakers. That doesn't fly. Powered speakers are poorly supported and when they die, you're on your way to buy a new pair. Not sustainable. Imagine trying to explain why your separate amplifiers have to be used with one specific model of speaker. There's an uphill battle. People would simply opt for the way we have always done it. I know I would.
Hi hifijim,
There isn't a compelling need for current drive amplifiers on a consumer level. No reason to develop a kit or product. Even if you did, the public would mostly be confused over which amplifier type went with what speaker. Just imagine the confusion in audio stores, havoc on-line. Then there is the used market - disaster.
At the moment, current drive is a non-starter.
Commercially that is! Luckily a DIY person isn't bound by profit margins and customer segments, as they are the same entity, and thus free to optimize system not for profitability but for what ever they feel important.
In my opinion this stuff is just another design freedom for DIYer. I know one dude in Finnish DIY forum who has built his own ironless drivers for current drive, I think Re is in tens of ohms for example. Highly optimized system it seems, big subs and cardioid and all, tailor fitted for himself. Never saw or heard them though.
In my opinion this stuff is just another design freedom for DIYer. I know one dude in Finnish DIY forum who has built his own ironless drivers for current drive, I think Re is in tens of ohms for example. Highly optimized system it seems, big subs and cardioid and all, tailor fitted for himself. Never saw or heard them though.
Last edited:
Absolutely agree with you tmuikku.
The problem is also the amount of research and time that someone really knowledgeable will spend on a technology that doesn't have huge interest. Some will, but the big guns probably won't. Creating a totally new driver using material and technology not commonly used costs a massive amount of money. Everything is custom.
You can expect some successes (ie: it makes sound), but to get the performance up to a comparable level is going to require real effort. Remember, you're playing catch-up to over a century worth of development with Dynamic drivers. Current drive has been investigated on and off in the past.
The problem is also the amount of research and time that someone really knowledgeable will spend on a technology that doesn't have huge interest. Some will, but the big guns probably won't. Creating a totally new driver using material and technology not commonly used costs a massive amount of money. Everything is custom.
You can expect some successes (ie: it makes sound), but to get the performance up to a comparable level is going to require real effort. Remember, you're playing catch-up to over a century worth of development with Dynamic drivers. Current drive has been investigated on and off in the past.
I hate to say this and it may rub a few people of higher academic qualifications the wrong way, but trying to use absolutely scientific analysis to rank the overall objective sound quality of a midrange driver is impossible. You can isolate some of the basic specs, which is essentially what has been done the same way for decades now despite advances in measurment capabilities.
The fact you're dealing with human beings as the final judging link in the chain makes the exclusive use of scientific testing and analysis inconclusive when trying to quantify and / or qualify any type of driver being ranked among its peers.
A major issue is always going to be the emotional factor of how a given piece of music is conveyed over a given speaker, depending on a whole variable set of psychological factors and basic human emotions, such as mood, bias, taste in music, tonal balance, hearing ability, etc. All of this influences or clouds our sense of judgment and ability to place absolute ratings on some things which should be quantifiable through scientific analysis and methodology. You'd essentially need to take the human factor out of the equation which you'll never be able to do.
In theory, evenly low THD, smooth, wide banded FR, clean CSD and transient response should be able to show the technical abilities of a well designed speaker. The practical truth is however pretty far from it.
The fact you're dealing with human beings as the final judging link in the chain makes the exclusive use of scientific testing and analysis inconclusive when trying to quantify and / or qualify any type of driver being ranked among its peers.
A major issue is always going to be the emotional factor of how a given piece of music is conveyed over a given speaker, depending on a whole variable set of psychological factors and basic human emotions, such as mood, bias, taste in music, tonal balance, hearing ability, etc. All of this influences or clouds our sense of judgment and ability to place absolute ratings on some things which should be quantifiable through scientific analysis and methodology. You'd essentially need to take the human factor out of the equation which you'll never be able to do.
In theory, evenly low THD, smooth, wide banded FR, clean CSD and transient response should be able to show the technical abilities of a well designed speaker. The practical truth is however pretty far from it.
From a technology POV, getting rid of dead moving mass, making the VC the actual moving diaphragm (like a planar driver) is a good way to make the electromagnetic type of driver more efficient and wider banded. It requires very precise mechanics to realize and therefor expensive manufacturing to produce.
Another area not much explored is piezoelectrics. These can be very accurate and linear based on the type of materials employed and would eliminate heavy, expensive magnets and copper. Problem is simply lack of interest. It was looked at closer when people were experimenting with HF carrier modulated audio but ended up as an "LRAD" sonic weapon for the military.
Electrostatic drivers are still very unexplored IMO. They also possess the benefit of being inexpensive to assemble based on the ultimate propulsion method and membrane materials chosen. Now that we have cheap, decent sounding class D, it may be an option.
Another area not much explored is piezoelectrics. These can be very accurate and linear based on the type of materials employed and would eliminate heavy, expensive magnets and copper. Problem is simply lack of interest. It was looked at closer when people were experimenting with HF carrier modulated audio but ended up as an "LRAD" sonic weapon for the military.
Electrostatic drivers are still very unexplored IMO. They also possess the benefit of being inexpensive to assemble based on the ultimate propulsion method and membrane materials chosen. Now that we have cheap, decent sounding class D, it may be an option.
Hi profiguy,
As soon as you try to involve the observers brain in any of this, you're off into the weeds. Sorry.
If you recreate the same sound pressure variations in the air near the ear, you have successfully recreated the sound accurately a with high fidelity. Period. How a specific person hears falls out of the equation. They will, by definition hear the same thing as they did live or the first time. Now if you say mood comes into it, you've thrown everything out the window and nothing matters. if they hear the same live performance again, they "hear" it differently. So no standard, just a feeling of niceness (or not).
The practical truth is that the better and more accurate the speaker - room is, the better (more close) to the original event it will sound. No emotion in it, no mood. That is entirely the person's problem that is listening. No way to cater to that, maybe drugs I guess. So what you are saying is merely dust and smoke in the air to confuse the issue.
As soon as you try to involve the observers brain in any of this, you're off into the weeds. Sorry.
If you recreate the same sound pressure variations in the air near the ear, you have successfully recreated the sound accurately a with high fidelity. Period. How a specific person hears falls out of the equation. They will, by definition hear the same thing as they did live or the first time. Now if you say mood comes into it, you've thrown everything out the window and nothing matters. if they hear the same live performance again, they "hear" it differently. So no standard, just a feeling of niceness (or not).
The practical truth is that the better and more accurate the speaker - room is, the better (more close) to the original event it will sound. No emotion in it, no mood. That is entirely the person's problem that is listening. No way to cater to that, maybe drugs I guess. So what you are saying is merely dust and smoke in the air to confuse the issue.