Hi
I’ve been carefully reading Zaph’s Design Mantras
http://zaphaudio.com/mantras.html
Amongts lots of good afvice, he says that distortion management is the most important issue in design . .
“When you start your design process armed with full distortion data, the chances you're going to get the design right on the first try or soon thereafter.
Without distortion data, you may never work out what's wrong with a design”
How is this easily done/ with a modest outlay?
Soundeasy? Clio??
I’ve been carefully reading Zaph’s Design Mantras
http://zaphaudio.com/mantras.html
Amongts lots of good afvice, he says that distortion management is the most important issue in design . .
“When you start your design process armed with full distortion data, the chances you're going to get the design right on the first try or soon thereafter.
Without distortion data, you may never work out what's wrong with a design”
How is this easily done/ with a modest outlay?
Soundeasy? Clio??
I have been surprised by distortion in a couple of recent designs and pay a bit more attention than I used to.
You need to make sure you're getting a true THD measurement and not a (much easier) THD+Noise figure. Check that your software measures the harmonic levels from an FFT spectrum and that it uses an appropriate stimulus. I'm not sure that you can get pure THD from pink noise or even MLS. I think it has to be swept sine or ESS/chirp. It's also useful to know how many harmonics were used in the calculation as sometimes higher orders pop up, plus, you can be bandwidth limited above 5KHz or so.
You need to make sure you're getting a true THD measurement and not a (much easier) THD+Noise figure. Check that your software measures the harmonic levels from an FFT spectrum and that it uses an appropriate stimulus. I'm not sure that you can get pure THD from pink noise or even MLS. I think it has to be swept sine or ESS/chirp. It's also useful to know how many harmonics were used in the calculation as sometimes higher orders pop up, plus, you can be bandwidth limited above 5KHz or so.
If you want to do this on the cheap, two options are ARTA and SpeakerWorkshop. I've never used ARTA so I can't comment.
SpeakerWorkshop can only perform harmonic distortion testing (not intermodular or other types AFAIK).
SW is also limited on my system to measuring F2 - F4. I don't know if this is a limitation of the sound card sampling rate (ie. SW might work out it can only go so high to obtain meaningful data).
SW uses sine waves. You specify the frequency range and number of points (frequencies) at which SW will sample, then it plays and records a sine at each frequency within the range. The more points, the longer it takes.
I'd be interested in how you go. Distortion measurement is something I want to understand more and its application, to produce accurate speakers.
Cheers,
David.
SpeakerWorkshop can only perform harmonic distortion testing (not intermodular or other types AFAIK).
SW is also limited on my system to measuring F2 - F4. I don't know if this is a limitation of the sound card sampling rate (ie. SW might work out it can only go so high to obtain meaningful data).
SW uses sine waves. You specify the frequency range and number of points (frequencies) at which SW will sample, then it plays and records a sine at each frequency within the range. The more points, the longer it takes.
I'd be interested in how you go. Distortion measurement is something I want to understand more and its application, to produce accurate speakers.
Cheers,
David.
It is an interesting question. I also believe that distortion is something best avoided.
There are some practical concerns however. Harmonic distortion, depending on its form and distribution, may not be very audible in some cases. However intermodulation distortion is certainly worth measuring. There are guidelines on what two components should be used (their relative level and freq). That would be a starting place, unfortunately the number of possible combinations creates a very large space to be searched.
Aside from the steady-state distortions (which are the most easily measured) there are also are transient forms. These would be tricky, since simple averaging will frequently "average out" what you are actually trying to measure. Certainly the nonlinearities that arise when a driver is being driven hard (into compression) would be of great interest. I am not at all sure there are conventional ways to measure these.
Anyhow, I have not been very helpful but I hope I have been encouraging.
There are some practical concerns however. Harmonic distortion, depending on its form and distribution, may not be very audible in some cases. However intermodulation distortion is certainly worth measuring. There are guidelines on what two components should be used (their relative level and freq). That would be a starting place, unfortunately the number of possible combinations creates a very large space to be searched.
Aside from the steady-state distortions (which are the most easily measured) there are also are transient forms. These would be tricky, since simple averaging will frequently "average out" what you are actually trying to measure. Certainly the nonlinearities that arise when a driver is being driven hard (into compression) would be of great interest. I am not at all sure there are conventional ways to measure these.
Anyhow, I have not been very helpful but I hope I have been encouraging.
Speakers' Distortion
You will find the gurus of speakers' distortion as the primary importance in determining sound qualities are,
Zalp, Mark K, Jay and many others.
On the other camp gurus like Gedlee, Magnetar and a few others don't care about this parameter that much.
What to decide? that is the burning question?
You will find the gurus of speakers' distortion as the primary importance in determining sound qualities are,
Zalp, Mark K, Jay and many others.
On the other camp gurus like Gedlee, Magnetar and a few others don't care about this parameter that much.
What to decide? that is the burning question?
Earl,
If you were selecting 3-7" midwoofers for 2 ways that had to fit on 5-9" wide baffles, would you be measuring distortion?
I see your speakers are bigger than most people's TVs, and you've already big high sensitivity drive units with faraday shielding in the motor systems, so distortion is kind of a moot point.
If you were selecting 3-7" midwoofers for 2 ways that had to fit on 5-9" wide baffles, would you be measuring distortion?
I see your speakers are bigger than most people's TVs, and you've already big high sensitivity drive units with faraday shielding in the motor systems, so distortion is kind of a moot point.
Earl,
I could try Zaph’s approach and see if it proves helpful.
I believe that ARTA may be easier to use than Speaker Workshop.
ARTA’s manual apparently shows it can measure both distortion by harmonic and intermodulation distortion.
WithTarragon,
I agree that IMD is more ‘serious’. However I don’t know to mitigate it’s ill effects . . any suggestions are welcome
(I need to get my shed completed & space sorted to do some testing . .)
“What to decide? that is the burning question?”
I don’t think a speaker’s distortion is the *primary consideration, but something to be factored in. If low distortion drivers are used, designing the XO to mitigate distortion isn’t necessary.
But for some projects, moderate distortion drivers may best meet other overall criteria, and/ or fit the budget. Hence the desirability of, if not making a silk purse from a sow's ear, trying to offset moderate distortion “deficiency”, for best overall balance.
I could try Zaph’s approach and see if it proves helpful.
I believe that ARTA may be easier to use than Speaker Workshop.
ARTA’s manual apparently shows it can measure both distortion by harmonic and intermodulation distortion.
WithTarragon,
I agree that IMD is more ‘serious’. However I don’t know to mitigate it’s ill effects . . any suggestions are welcome
(I need to get my shed completed & space sorted to do some testing . .)
“What to decide? that is the burning question?”
I don’t think a speaker’s distortion is the *primary consideration, but something to be factored in. If low distortion drivers are used, designing the XO to mitigate distortion isn’t necessary.
But for some projects, moderate distortion drivers may best meet other overall criteria, and/ or fit the budget. Hence the desirability of, if not making a silk purse from a sow's ear, trying to offset moderate distortion “deficiency”, for best overall balance.
tktran303 said:
If you are asking would I measure distortion is a very low end product? Maybe I would, but probably not. I could tell what I need to know about linearity just by looking at the driver design, I would not really need to measure it. So, in a sense, I do "measure the nonlinearity" by simply considering the design, but measuring something as meaningless as THD or IMD, no I wouldn't waste my time.
The point that I make over and over and over again, is not that one cannot create a bad design where distortion is audible, but that I CAN create a design where it is not the most significant audiblity problem, and the THD on any individual driver will have virtually no relavence to audibility. Thus, if distortion is audible in the design, then the system design is faulty.
Are there a lot of faulty system designs out there? - well most of them actually. They are designed not for audibility but for "spec sheet data" which tell virtually nothing about audibility, and yet they are what the consumer uses to judge their purchase. It's a ridiculous situation.
I make a speaker that is only 23" x 12" x 8" - thats not big. Would I ever bother to listen to Mini-speakers? No thanks, I like music too much to abuse my ears that way. There is a lot to be said for designing in a safety margin - headroom - in a loudspeaker by using larger drivers. It all comes down to SPL level. If very low listening levels is all that you listen to then the situation is quite different than wanting to hear some dynamics in the playback. The problem is that most people judge SPL by the speakers "power handling", but power handling only means power to failure NOT power to "bad sound". It will usually sound bad long before it blows up.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.