Transient response...

[JimChjones]

Something that comes up in the numerous speaker size threads is "10s respond better to transients". I've never seen it covered in detail, but it bemuses me. What I understand as transients in this context is short duration events like note attack and so on. But when I think about the actual wave form presented to the speaker, then it seems to me all these events are of very long duration compared to the actual waveform, so I don't understand how they can have any relevance, as the speaker has to move far faster from one extreme to the other just to reproduce the notes as it would have to for anything to do with these transients.
And yes, I accept that a big heavy cone may be accelerated by the motor slower than a small light one: I guess this is why little tiny tweeters have better HF response than big woofers, but it still seems to me that the event duration of this things is still so great that it is hardly going to be affected - and in any case this is just a matter of frequency response. So is there a phenomenon I haven't worked out, or is this just one of these semi myths, and "response to transients" is really just a secondary effect of frequency response?

 

[Downunderwonder]

The "transient" indeed comes from the initial attack of the note. All notes have them. I believe you are only referencing the high end where you find greater variance in response. A boffin will be along to tell the whole story soon enough.

 

[chadds]

There are several types of transient response. One is electrical another is the way a car’s suspension sets up in a curve. The third is, “get a job!”

 

[Rick James]

This tends to come up in the constantly recurring ad nauseam driver size threads. As is the case with almost every aspect of driver performance transient response is affected by a number of factors, but driver size isn't one of them. For every fifteen than has poor transient response you can find a ten that's worse. This is a pretty accurate assessment of what transient response is:
Barefaced Bass - Transient Response

Cutting to the chase this is what mainly results in good transient response:

1. High motor strength : moving mass ratio
2. Good control of non-pistonic cone output
3. Low system inductance
4. Minimum resistance in series crossover components
5. Gentle LF roll-off slope
6. Linear output at the required high SPLs

Driver size isn't on the list.

I accept that a big heavy cone may be accelerated by the motor slower than a small light one

Don't be so accepting, because like most of the opinions related to driver size this one isn't true.

 

[JRA]

A boffin will be along to tell the whole story soon enough.

 

[seamonkey]

Transients can be measured fairly simply, for example
Make a recording of a signal going into a speaker and another track with a mic on the speaker. Correct for the speed of sound in air which actually is very easy with any DAW these days. Then look, listen, and measure the envelop.

Transients usually improve as the piston area and motor strength is increased. Measure it. For good professional driver and cabinet design it's not going to be anything a DBT would differentiate.

 

[beans-on-toast]

The coil in a speaker has inductance. It acts like a buffer, resisting change. The mechanical parts of a speaker resist movement.

It’s pretty amazing that speakers, which are generally pretty inefficient, and filter the sound and associated transients in their own way, reproduce as well as they do. Our hearing steps in and makes it all work.


Activity: Measuring a Loudspeaker Impedance Profile [Analog Devices Wiki]

 

[Rick James]

The coil in a speaker has inductance. It acts like a buffer, resisting change.

Now that you mentioned it:
https://adireaudio.com/wp-content/.../Adire-Audio-Woofer-Speed-by-Dan-Wiggins.p...

 

[JimChjones]

Nice link on transient response
I wonder now though if there's another confusion going on. Transient response as the science has it is now reasonably clear to me, but when some at least talk about transients they seem to me to mean something different, to do with short duration signals. I'll have to mull over this.

 

[chadds]

He said professional driver. Must be about race cars again.

 

[JimChjones]

Reading that link on transient response again I'm having trouble with the logic. My serious maths was always poor, and it seems to be worse now, but what I'm struggling with is the transformation of equation 2. It says that you can treat mass as a constant because the moving parts of a driver don't change. And therefore you end up with acceleration proportional to current. Now clearly that's correct for any given driver. But I thought our starting point - and the experiment too - was what happens if BL is constant and M varies? So I don't understand why his transformation is valid.

He then does the experiment which appears to me to demonstrate that varying the mass of the cone has a large effect on bass frequencies, he says because of FS, and a minimal effect on high frequencies. It also demonstrates that the driver inductance has a big effect. Fair enough. I'm sure the conclusion of the paper - that the inductance of the driver is far more important than the mass - is correct.

The other thing that worries me about the experiment is displacement. I've always understood that all else being equal, bass frequencies require a larger cone movement than high ones. If at 60 hz the cone moves through 6mm, and at 6khz the cone moves through 0.6mm then there's an effect on the required acceleration of the cone. There's also an opposite effect because that movement is required at the higher frequency. I wonder if they balance out, or whether one is dominant. Is the speaker velocity (which presumably you'd need to establish an RMS value for) greater at low frequencies or high ones?

[Added] thinks, if RMS current is constant at low or high frequencies, then if the cone is accelerated at 60 hz or 60khz, does that mean intrinsically the cone movement is reduced in proportion to the increasing frequency? Not sure I've expressed that very well.

 

[agedhorse]

The simple explanation is that transient response is directly related to high frequency bandwidth. The greater the bandwidth, the "faster" the transient response.

Everything that affects bandwidth also affects transient response.

 

[JimChjones]

Thanks, that's kinda the practical conclusion I was coming to.

 

[micguy]

Driver size isn't on the list.

It isn't specifically, but the moving mass is high on the list, and that depends on the driver size. A 15 will have about 2.25 times the moving cone mass of a ten, assuming the same cone material and thickness. If you look at the mass of air coupled to the cone, the situation is even more in favor of the 10 - the mass of that slug of air is proportional to the diameter cubed, giving you 3.375 times the mass of the slug coupled to a 10.

Driver size is definitely correlated to transient response. Yes, you can always make a 10 bad if you want to, but if you're trying to make good drivers, it's easier to make a good 10 in that respect than a good 15. You can't make a 15 as good as a great 10 - if the 10 is optimized for transient response, a similarly constructed 15 just won't keep up.

 

[micguy]

The simple explanation is that transient response is directly related to high frequency bandwidth. The greater the bandwidth, the "faster" the transient response.

Everything that affects bandwidth also affects transient response.

Mr. Fourier agrees.

 

[Rick James]

A 15 will have about 2.25 times the moving cone mass of a ten, assuming the same cone material and thickness.

And how likely is that? I never assume anything. It's assumptions based on pure conjecture instead of engineering facts that result in threads that waste everyone's time, like speaker size threads, mixing cabs threads, underpowering threads, magic cable threads and tube versus SS threads.

 

[ThisBass]

Something that comes up in the numerous speaker size threads is "10s respond better to transients". I've never seen it covered in detail, but it bemuses me.

If the tens lack some low end (versus different sized drivers) then the "assumption" is right that the drivers respond "better" to transients. In real practice a lack of low end response mutes transient noise at the lowish side of the spectrum which in total may soundwise be "experienced" as "improved" transient response.

What I understand as transients in this context is short duration events like note attack and so on.

Good point. In real practice any note attack generates also transient noise due to the sharp envelope shape raise and decay.

But when I think about the actual wave form presented to the speaker,

In first place any note attack generates noise
The envelope decay can be considered as an arrangement of transient noise and harmonics.
The settled oscillation of the string can be considered as an arrangement of harmonics.

Here you are with a couple of plots which might help to comprehent "transient" things.
Interestingly transient things are the main reason WHY its so very beneficial to protect drivers with a HPF.

Transient noise and what it means to woofer excursion

 

[-Asdfgh-]

The "transient" indeed comes from the initial attack of the note. All notes have them. I believe you are only referencing the high end where you find greater variance in response. A boffin will be along to tell the whole story soon enough.

You mean @agedhorse?

 

[Downunderwonder]

You mean @agedhorse?

One of his European cousins would generally be out of bed first.

 

[micguy]

And how likely is that? I never assume anything. It's assumptions based on pure conjecture instead of engineering facts that result in threads that waste everyone's time, like speaker size threads, mixing cabs threads, underpowering threads, magic cable threads and tube versus SS threads.

The likelihood (given my Lifetime of Engineering audio gear, with a lot of it being transducers, I can speak to this), is that a larger speaker would need a thicker cone, so the 15 would be even more hampered. My point was, the 15 is hampered by its size (size does matter in this case), and if you remove my simplification assumption of constant thickness between the two cones, it doesn't change that conclusion - the conclusion gets stronger.