Yes, I think 5% 2nd is excessive but I suspect it is system dependent. I wonder about the quality of the headphones?
..but the broad conclusion of the test is good for me.
Member
Joined 2009
Paid Member
I didn't say he was wrong, Earl knows more about speakers than I'll ever know. I was laughing at the implication that because he has published it, that somehow means we should accept it as true.
We are lucky - or it is remarkable - that most of us find speaker distortions so acceptable.
Last edited:
I do get that no one is really saying they all sound the same, HD rarely is the number one character when defining sound quality.
I do have an example where cleaner woofers make an audible difference that shows up in the frequency domain due to reduced harmonic distortion of the drivers.
I have a pair of JBL 4430’s. They have a single 2235 woofer and they sound pretty good. I also have 4435’s with the dual 2234’s in them and they sound a lot better, on the bottom, much cleaner and tighter.
The one inch drivers on the 4435’s are the limiting factor for listening loud, they start to get hard when you lean on them. The 4430’s are pretty ballanced as both drivers approach their limits about the same time.
Here’s the main difference, the dual 15’s have about 1/10 the distortion at the same volume as the single, according to the JBL owners manual for the 4430-4435.
So here comes a pair of 4350’s with 2235’s. They have the same sound character on the bottom (250Hz and down) as the 4435’s but will play way louder than the 4435’s overall, as expected.
Where is this all going? The 4430 has midbass growth at high volume that I originally attributed to something I have heard explained as “overdriving the room” or stored energy in the modal region. Pretty much all my systems have displayed this to some extent. The midbass gets muddy with volume BUT,, the dual 15” 4435 and 4350 exhibit this less.
Here’s my big argument to this whole distortion in woofers is meaningless:
I bought four JBL 1500Al woofers as used in the DD6600. These are what JBL calls ultra low distortion woofers, like 0.1% HD at 100 Watts input power as compared to the several whole number percent HD 2235’s and guess what?
The mid bass growth phenomenon is totally absent in the 4350’s with these ultra low distortion drivers.
I attribute this to the absence of increasing false additional mid-bass generated by the driver when operating at high levels.
Increasing HD as drive level increases is a nonlinear distortion that I can easily discern in woofers in this case.
Barry.
I do have an example where cleaner woofers make an audible difference that shows up in the frequency domain due to reduced harmonic distortion of the drivers.
I have a pair of JBL 4430’s. They have a single 2235 woofer and they sound pretty good. I also have 4435’s with the dual 2234’s in them and they sound a lot better, on the bottom, much cleaner and tighter.
The one inch drivers on the 4435’s are the limiting factor for listening loud, they start to get hard when you lean on them. The 4430’s are pretty ballanced as both drivers approach their limits about the same time.
Here’s the main difference, the dual 15’s have about 1/10 the distortion at the same volume as the single, according to the JBL owners manual for the 4430-4435.
So here comes a pair of 4350’s with 2235’s. They have the same sound character on the bottom (250Hz and down) as the 4435’s but will play way louder than the 4435’s overall, as expected.
Where is this all going? The 4430 has midbass growth at high volume that I originally attributed to something I have heard explained as “overdriving the room” or stored energy in the modal region. Pretty much all my systems have displayed this to some extent. The midbass gets muddy with volume BUT,, the dual 15” 4435 and 4350 exhibit this less.
Here’s my big argument to this whole distortion in woofers is meaningless:
I bought four JBL 1500Al woofers as used in the DD6600. These are what JBL calls ultra low distortion woofers, like 0.1% HD at 100 Watts input power as compared to the several whole number percent HD 2235’s and guess what?
The mid bass growth phenomenon is totally absent in the 4350’s with these ultra low distortion drivers.
I attribute this to the absence of increasing false additional mid-bass generated by the driver when operating at high levels.
Increasing HD as drive level increases is a nonlinear distortion that I can easily discern in woofers in this case.
Barry.
I do not see how my statements implied that my work was right, only that I have done substantial work in this area. Surely one who has studied a subject is more likely to be "right" than one who hasn't.
And yes, we are lucking that most loudspeaker NONLINEAR distortions are so benign. Of course their linear distortions totally dominate our perceptions of audio.
Back to the question at hand: Why is THD wrong?
Its all about masking. Masking is the major reason that we have MP3 and all the major "perceptual encoders". Masking hides vast portions of an audio signal that we can measure but the ear is completely impervious to. A good codec can eliminate 80-90% of the information content in a digital signal and yet the ear is unable to detect this. That's how dominate masking can be.
THD (and IMD) do not take masking into account in any way. This makes them totally unreliable as measures of "what we hear". A loudspeaker will typically have large amounts of 2nd harmonic, but this harmonic is almost always completely masked. Hence we do not detect a problem with a loudspeaker that has 15% 2nd order nonlinearity. On the other hand an amplifier with crossover distortion generates orders as high as ten or twenty in the spectrum. These very high harmonics are not masked at all, which is why we can hear as low as .1% THD (or less, depending) of crossover distortion in an amp. How can one thus use a measure like this that has no correlation to what we hear to "rate" the audible nonlinearity of any audio component? We can't!!
You are talking about how audible a loudspeaker is once it reaches into SPLs beyond what it was designed for. Nothing that I claim has any relevance in that case. You simply need to stay below the level at which this occurs (or buy better speakers.) That's an issue with "dynamics" not THD.
This is not scientific since the woofers are identical and using two should only half the nonlinearity not 1/10. The additional 6 dB of headroom would be a significant factor, yes.
I owned 4430's for 15 years and the bass was very good. The rest was pretty disappointing however.
To limit the effect of overloading a woofer, I have always recommended multiple subs. In every case of systems that I have designed this has completely eliminated problems associated with overloaded bass. With singular woofers, especially smaller one, overloading the bass is a serious issue that is extremely common, but again, this is an overloading issue and not one relating to the perception of THD at normal listening levels.
We used balanced armature insert earphones (Etymotic 4a) at fairly low levels (OSHA requirement.) This kind of transducer does not have the usual nonlinearity that a loudspeaker has because its amplitude of motion is negligible - it is playing into a very small closed cavity. The nonlinear distortion of these types of headphones is always negligible when compared to a loudspeaker playing into a room.
Over-the-ear headphones are different. They are moving coil and do act more like loudspeakers.
hi Dr. Geddes- consider the case of a direct radiator vented box tuned by an array of narrow slits or relatively small holes bored in the front panel as popular for a while in the 1950's. That type even at low power, when driven at and in the vicinity of tuning with sine, generally exhibits very high vent velocity and many strong overtones. I'm guessing with a lot of music (assuming no sine bass drops or certain organ pedal stops) a listener might perceive "distortion" and neither might one notice much if playing an electric bass guitar through such a system.
Is this mainly due to the transient nature of (traditional instruments) music vs a sine wave? Is this part of masking ?
Is this mainly due to the transient nature of (traditional instruments) music vs a sine wave? Is this part of masking ?
How could you qualify the bass and the sub-bass rendering of the Etymotic 4a at low levels please?
Sorry for my evasive question...
I find them extremely "thin sounding" (overdamped) in the bass departement and there is no sub-bass rendering at all.
Perhaps i'm wrong, but this example, for me seems to eliminate the problem of non-linearities by an excess of control.
The linear response of the devices would have no effect on our study since it was done as a comparison of different clips all through the same device. One may or may not actually like the sound quality of the devices, but that would be irrelevant. They were asked to respond to the various levels of simulated nonlinearity is the samples. High nonlinearity in the headphones would have been an issue because of masking, but the linear responses would have always been the same so that would not have been a factor.
Member
Joined 2009
Paid Member
Agreed and respected.
Yes, to a point, but conventional wisdom at the time, or whatever it should be called, can influence the conclusions we make from what we study.
Anyhow, I'd be interested in your opinion on a distortion mode that others have reported and which bothers me somewhat as I have never been comfortable with the descriptions. Doppler distortion.
As I understand, it is claimed that this distortion is produced when a driver plays both low and high frequencies at the same time. However, providing the speaker is operating linearly, this should not be possible. If it is operating in a non-linear mode with cone break up and local resonances then fair enough. However, I've not seen the concern of Doppler distortion linked with non-linear cone behaviour.
In a linear mode, the motion of the cone at all frequencies is relative to the magnet and therefore high and low frequencies do not mix.
Perhaps not in the speaker forum, but many of the 'amplifier guys' know all about it and exploit it at times.
Last edited:
Doppler distortion happens to be an entirely LINEAR effect. It's predicted in astronomy in Red-Shift.
When the Cops approach you "criminals" with their awful high-pitched sirens, doubtless designed to terrify and distract you, it sounds higher pitched than when they recede into the distance.
Don't waste time on it. Non-linear effects are the main event for people who love good sound. It's actually pure mathematics. A three-way is far more linear than a two way. Hence sounds less distorted. OK, we know that in this real Universe, crossover equates to a power hole. Unless you work at odd-order filters.
You single driver folks just must take your chances.
When the Cops approach you "criminals" with their awful high-pitched sirens, doubtless designed to terrify and distract you, it sounds higher pitched than when they recede into the distance.
Don't waste time on it. Non-linear effects are the main event for people who love good sound. It's actually pure mathematics. A three-way is far more linear than a two way. Hence sounds less distorted. OK, we know that in this real Universe, crossover equates to a power hole. Unless you work at odd-order filters.
You single driver folks just must take your chances.
Member
Joined 2009
Paid Member