Which speaker type has a faster transient response?

I remember seeing some comparative waterfall diagrams, and BA was faster - however, I have no idea where to find the links :/ 

subscribed

I would guess BA, but unfortunately I haven't seen any waterfalls for them.

However, if you look at Ryumatsuba's website it's been noted that Stax stats are actually slower than certain "entry" and "midrange" level dynamic headphones in waterfall plots.

I think It would also depend on what is driving them. also most armature based have filters between the armature and your ear that shape the frequency curve but may have effects on the perceived speed of the sound. the fastest type of drivers should be plasma speakers it doesn't get much lower mass than ionized air

 (Quote:

Originally Posted by Shike 

I would guess BA, but unfortunately I haven't seen any waterfalls for them.

 

However, if you look at Ryumatsuba's website it's been noted that Stax stats are actually slower than certain "entry" and "midrange" level dynamic headphones in waterfall plots.

No, that's wrong.

The waterfall plots are cumulative spectral decay and not related to how fast the headphone is. The porta pro, for example, has a very good CSD. I'm not sure what's going on with the impulse response on Ryumatsuba's site. Could be a phase inversion, but I doubt it - I wouldn't trust it since it seems... odd. Simpy, a CSD is more related to how "open" a headphone or speaker is - i.e. how long the sound lingers after it is played

The less mass and higher the force applied, the faster the acceleration of the driver. In an impulse response measurement, you're getting a sudden +x voltage (x being any value) which drops off immediately and you're seeing how the speaker reacts in terms of damping and reflections. It doesn't have much to do with speed. In fact, I don't think I've seen any measurement that has much to do with speed, apart from perhaps the duration it takes the speaker to reach the initial voltage peak (which is always very quick). It also isn't that reliable when a constant flux is applied, as this will cause the speaker to move back and forth again. In most cases this is to measure the "room" of headphones (i.e. the cups). The secondary peaks are going to be virtually inaudible (if they are still present - which is debatable) during music playback.

I'm not sure why so many people here are incapable of reading graphs.

From experience, I'd say that stax in general are better at detail retrieval than say, an Etymotic balanced armature, which is better than most (but not all) dynamic headphones.

I'm not even sure what amp ryumatsuba uses for most of these measurements. Most of his stax measurements are done with energisers, however. He is in essence, testing the system, rather than the headphones, as energisers are less than ideal.

Quote:

Originally Posted by Nagasaki_Kid 

The fastest type of drivers should be plasma speakers it doesn't get much lower mass than ionized air

That's correct.

Quote:

Originally Posted by Tronz 

Quote:

Originally Posted by Nagasaki_Kid 

I think It would also depend on what is driving them. also most armature based have filters between the armature and your ear that shape the frequency curve but may have effects on the perceived speed of the sound. the fastest type of drivers should be plasma speakers it doesn't get much lower mass than ionized air

The way I test a speakers transient response is by how seperated and clear every sound is from one another. Also, if certain instruments have that snap. Can I still fall prey to frequency response by doing it this way as well?

 

The "frequency response shapes detail!11" argument is hugely flawed because it presumes that detail exists only in treble. Generally speaking, the people arguing this are also referring to phones that are very bright above 10khz, of which there is very little in any piece of music. At best you could argue that a huge bass response is going to drown out detail in other ranges - but frankly when you get to that point the bass response becomes more of a problem than the detail retrieval.

This argument is also frequently adopted by members who have headphones that, which they will not admit, have boosted bass response and recessed treble compared to what is generally considered ideal (a diffuse or free field equalisation).

Most headphones are more than fast enough to retrieve the detail in most pieces - after all they can extend to frequencies 20khz+. Stax, for example, are capable of >44.1khz (of course theres the infamous sony 110khz, but I've not actually seen measurements saying they are capable of this), which means they're capable of moving fast enough to play every sample in the simplest piece you can imagine (and most in the functional frequency range in most pieces)- which is an inaudibly small amount of time. The way headphones present the detail is vastly different. That said, don't forget that in most cases the frequency response measurement is essentially marketing in headphones - particularly in terms of treble extension.

The LCD-2, for example, is extremely subtle (too subtle IMO) about detail. The Stax lambda is very in-your-face about it. The ER4 lies somewhere in the middle. FWIW, the LCD-2 is very dark, the Lambda is less bright than the ER4 (as it is a free field EQ), and the ER4 is the brightest (as it is a diffuse field EQ).
 

Regardless, headphone measurements are notoriously difficult to perform accurately. On Ryumatsubas site, for example, you see what appears to be a poor fit on several models (possibly due to harder earpads against the harder dummy vs human skin).

But still, it's not nearly as simple as many make it out to be. For example, there's a slow reaction to voltage flux in several canalphones (SE530?) on headroom. It's probably also mentioning that many electrostats produce higher harmonic distortion than dynamics. Balanced armatures also have the issue of bandwidth availability - most actually can't produce above 16khz or so. I havent seen reliable enough measurements with enough resolution to make comments about the reaction of specific models of electrostatic.

EDIT: Even John Willet from Sennheiser agrees.

Quote:

Originally Posted by John Willett 

The speed of movement on a dynamic diaphragm depends on the power of the magnet, the stiffness of the surround and the weight of the diaphragm and voice-coil.

Powerful magnets with light diaphragms and voice-coils will be faster than heavy diaphragms / voice-coils and small magnets.

Having said all this; no dynamic can ever be as fast and as responsive as an electrostatic - in the same way that a dynamic microphone can never be as fast and responsive as a condenser microphone.

The above contains correct information by Newtons Law, not audiophile voodoo.

Quote:

Originally Posted by Tronz 

Thanks for the info MrGreen, but I've found a contradiction. You basically judge the speed of a diaphragm by how far it can extend in the treble? correct? At the end you stated that the info John posted was correct. He stated electrostatics can never be matched by dynamics in terms of speed and responsiveness. There are some dynamics that can extend further than electrostatics in the treble region. This means the diaphragms can vibrate faster, correct?
 

No, I've over simplified the example to provide a scope of how fast the average transducer is. Most speakers are capable of going well above 20khz or even higher, but there is no universally accepted tolerance. For example, a headphone that quotes up to 110k (you know the one) might only have up to 8khz within a 3dB tolerance. I merely provided the frequency response comment to point out that a CDs sampling rate will be totally presented by a lot of headphones.

In truth, headphones only need to produce 20khz to produce all frequencies present in CD audio.

Anyway, it's a gross oversimplification.

The stronger the magnet, and the lighter the diaphragm = The faster the response. Plasma speakers are the fastest as the diaphragm (the arc) has zero rest mass. I'm not sure why some balanced armatures appear to have slow rise in graphs, possibly due to magnet strength. I'm not a headphone engineer (obviously)

The ear isn't all that fast.

Dynamic drivers can be as fast as they want if they get as thin as they want. There's design problems when you start making them extra thin, however (especially since the drive force in a dynamic headphone is uneven).

Magneticplanar and electrostatic headphones don't have this problem (ortho is heavier than electrostat though in most cases). In fact there are electrostatic drivers taht are only one micron thick (lambda signature and sigma pro). Most stax these days use 1.35 micron, I think? (micron is a millionth of a meter by the way, or 1/1000th of a millimeter, just to give you an idea of how thin that is)