Gedlee on distortion

[henry218]

Re: Gedlee on distortion

A quote from an earlier post....

"Instructions for trolling:
1. Stir the pot fairly regularly.
2. Fairly regularly claim you are not stirring the pot fairly regularly."

I think we now have two trolls on board

it is indicated in your post here that you are dismissing him

[henry218]

Re: Gedlee on distortion

he is citing what would be considered as authoritative sources, if you want to argue with him, then argue with the paper/sources presented in his posts.

[carlspeak]

Re: Gedlee on distortion

it is indicated in your post here that you are dismissing him

That's your opinion. I was stating mine. I'd love to see one of his projects. Who knows, I could learn something from him.

[henry218]

Re: Gedlee on distortion

Who knows, I could learn something from him.

i believe that is the ultimate goal of this forum

[hongrn]

Re: Gedlee on distortion

To be quite honest, I don't care either way. His ability or desire to design/build a loudspeaker doesn't refute or confirm his conclusion about distortion.

I find it a common theme:

Pallas makes a claim.

People disagree with the claim.

Rather than present evidence to dismiss his claim, we simply ask Pallas what he has designed and attempt to dismiss him by poisoning the well. Whether or not he has designed anything does not change the validity of his statements. I'll agree that sometimes Pallas can be condescending, but I'm not sure I blame him; when he claims something people disagree with they tend to get combative--often times before he does.

I think it would serve the discussion much better to ask Pallas for his evidence and to make an argument against his claim with counter evidence rather than dismiss him because he hasn't posted a build.

My 0.02.

Also, "PallasandhisToole" made me chuckle. :p

If you read this thread carefully, several people started by disagreeing with statements made by Gedlee, NOT by Pallas. Of course, Pallas had to jump in, and in his usual subtle ways, called into question people's ability to comprehend the written word, and therefore, the level of their intelligence. Now, I'm not a smart guy when it comes to distortion and harmonics discussion, but I find it completely disgusting for Pallas to conclude that if you can't read or comprehend a document, you're somehow are less intelligent than he is. My post is not about attacking Pallas, and call me "small minds" like some earlier posters did, but it's about having a decent and civil discourse without bringing out your high horse and your elitism. On Carl's argument that Pallas should share his designs, I agree with you Carl. This is a DIY forum, so back up your assertions, assumptions, theories, and postulations with actual designs. After all, if you don't use the scientific method to prove your point, do you really have a point?

[ksneote]

Re: Gedlee on distortion

Well, this thread has been a roller coaster, certainly not what I expected when starting it!

I was curious as to Gedlee's stance on distortion moreso in a practical sense, ie when looking at a driver for potential use, what parameters would help one determine if it is a good quality or not? In contrast to Gedlee, Zaph seems to put strong weight into distortion. ( http://www.zaphaudio.com/nonlinear.html )

I'm not trying to take any sides or make any grand claims, just trying to better understand how and when distortion becomes relevant, either when choosing drivers for a system, or after the speaker system is complete.

[dcibel]

Re: Gedlee on distortion

In contrast to Gedlee, Zaph seems to put strong weight into distortion. ( http://www.zaphaudio.com/nonlinear.html

First, the name is Earl Geddes. Gedlee is the name of his company. Second, I don't think I'm wrong in stating that both Zaph and Earl agree on the merits of THD. From the page you linked to:
THD (total harmonic distortion) is not useful because it's just a sum that is dominated by 2nd and 3rd harmonics. It really tells you nothing about how it sounds. I never bother to report a THD number.

I fear in this roller coaster you missed the first reply to your original post :p In any case, the article from Zaph Audio is a good read.

[ksneote]

Re: Gedlee on distortion

First, the name is Earl Geddes. Gedlee is the name of his company. Second, I don't think I'm wrong in stating that both Zaph and Earl agree on the merits of THD. From the page you linked to:
THD (total harmonic distortion) is not useful because it's just a sum that is dominated by 2nd and 3rd harmonics. It really tells you nothing about how it sounds. I never bother to report a THD number.

I fear in this roller coaster you missed the first reply to your original post :p In any case, the article from Zaph Audio is a good read.

Ah, excuse me, Geddes.

Ha, yes, seems like I did. So my next question would be, when (if ever) is THD actually relevant, or a useful number?

Also do you know of anyone who has done Zaph's "Real Test" on that link? A similar concept was mentioned earlier in this thread I believe.

[rickcraig]

Re: Gedlee on distortion

Ah, excuse me, Geddes.

Ha, yes, seems like I did. So my next question would be, when (if ever) is THD actually relevant, or a useful number?

Also do you know of anyone who has done Zaph's "Real Test" on that link? A similar concept was mentioned earlier in this thread I believe.

As someone else noted THD numbers can help when selecting crossover points / slopes. I think problems arise though when a builder fixates on one design aspect with possibly overlooking other issues. BTW, I used to live in Bloomington

[dcibel]

Re: Gedlee on distortion

When THD is abnormally high, it tells you something is likely broken. When THD is abnormally low, it tells you that all the harmonics are possibly completely inaudible and you can now happily focus your attention elsewhere ;)

[ksneote]

Re: Gedlee on distortion

As someone else noted THD numbers can help when selecting crossover points / slopes. I think problems arise though when a builder fixates on one design aspect with possibly overlooking other issues. BTW, I used to live in Bloomington

I see. So when THD is too high, you cross above or below that point? I doubt there is a steadfast rule, but when is it "too" much THD?

Oh really? Did you attend IU?

When THD is abnormally high, it tells you something is likely broken. When THD is abnormally low, it tells you that all the harmonics are possibly completely inaudible and you can now happily focus your attention elsewhere ;)

And when THD is "average", your driver is probably nothing special?

[Deward Hastings]

Re: Gedlee on distortion

When THD is abnormally high, it tells you something is likely broken. When THD is abnormally low, it tells you that all the harmonics are possibly completely inaudible and you can now happily focus your attention elsewhere ;)

That's about it . . .

I see. So when THD is too high, you cross above or below that point?

And when THD is "average", your driver is probably nothing special?

You really need to see the distortion spectrum. Mechanical systems (like drivers) rarely produce much in the way of higher harmonics, so you look at rising 2nd and 3rd at the low end as indicating that the driver is nearing its excursion limits (2nd is asymmetric, and might mean an off-center voice coil, 3rd suggests out of gap or suspension nearing its limit or both), while distortion peaks on the high end suggests cone (or other) resonances amplifying the distortion that is there. Taken together all the distortion curves suggest a "useable range" for the driver. Jeff has already discussed inductance effects . . . every motor should have a shorting ring. There are lots of decent drivers that do not produce objectionable, or even audible, distortion if used within a reasonable range and not pushed too hard. But you won't figure out which they are by a simple THD number. Within that "reasonable range" rising loudness masks rising distortion, so we luck out on that.

In general you won't hear 2nd and 3rd harmonic distortion that is more than 40dB down, but you may hear higher order harmonics that are less than 50-60dB down, and if the driver is wide range you may hear intermodulation products caused by the 2nd and 3rd (two and three way systems reduce that problem). While distortion curves are not the primary determining factor in choosing crossover frequency it is good to choose both drivers and crossover frequency that do not result in an abrupt discontinuity in the combined distortion curves.

Electronics are different. Not only are higher order (and objectionable) distortion components common, but in many amplifiers distortion rises as signal (and masking) falls. This is a Very Bad Thing (although not as well recognized as it should be). The most useless number in all amplifier specification is THD at high signal level . . . all it tells you is when the amplifier starts clipping (maximum output), which you can just as easily determine by simpler means . . . it tells nothing at all about how an amplifier will sound. If THD is at least 60dB down at 1 Watt (or better 100 milliWatt), though, you've probably got a pretty decent sounding amp.

[andy19191]

Re: Gedlee on distortion

So my next question would be, when (if ever) is THD actually relevant, or a useful number?

Distortion defines the operating range of a driver or speaker which is rather fundamental information. THD can be used something like this to define the operating range although I would prefer to see 1, 3 and 10% levels. The intrusiveness of the distortion is not completely irrelevant but is of little importance because of the steep rise. I would suggest the absence of information like this when it comes to considering the performance of speakers is rather telling.

[Sydney]

Re: Gedlee on distortion

...So my next question would be, when (if ever) is THD actually relevant, or a useful number?...

I find several things mentioned as poignant: Perceptual threshold, and fixation.
As an analogy: the difference in singing/playing in harmony vs being off-key - even at very low levels non-harmonics are disconcerting.
And fixation on aspects without a greater consideration of the entire system.

" I have not found a quantitative or qualitative relationship between the various
distortion types you can easily measure - and loudspeaker preference"
from

http://audioxpress.com/files/dappolito2960.pdf

[JRT]

Re: Gedlee on distortion

Electronics are different. Not only are higher order (and objectionable) distortion components common, but in many amplifiers distortion rises as signal (and masking) falls. This is a Very Bad Thing (although not as well recognized as it should be). The most useless number in all amplifier specification is THD at high signal level . . . all it tells you is when the amplifier starts clipping (maximum output), which you can just as easily determine by simpler means . . . it tells nothing at all about how an amplifier will sound. If THD is at least 60dB down at 1 Watt (or better 100 milliWatt), though, you've probably got a pretty decent sounding amp.

At low output level, noise often dominates the nonlinearities and THD drops below the noise threshold.

Not always but all too often for amplifiers intended for use within the full audio bandwidth the THD is specified at 1_kHz rather than across the full audio bandwidth. With a conventional class A/B amplifier using a large dose of negative feedback to reduce nonlinear distortion, that nonlinear distortion reduction is more effective at lower frequencies, and THD tends to rise at higher audio frequencies. So it is much more useful to see THD specified at full power over the range of 20_Hz to 20_kHz.

It is also useful to know useable headroom for worst case high level brief crests. And a good indicator useable headroom is the power level where THD passes thru 5% or 10%, providing a good coarse threshold for indicating a transition into significant clipping. And again, that too should be THD across the full audio bandwidth, not just at 1_kHz.

For a tube amplifier with an output transformer, full power bandwidth may be constrained at low frequencies where the output transformer transitions into saturation, where 3rd order harmonic distortion rises steeply with decreasing frequency, and pretty well defines the low frequency side of the useable bandwidth.