This reminds me of the "detail" discussions and how brightness/emphasized treble = fake detail, while faster transient response/attack/decay = real detail, even if the details aren't shoved in your face.
 
All I can say for sure is that consensus here suggests that dynamic/moving-coil designs are regarded as the slowest driver type, orthodynamic/planar-magnetic designs are significantly faster, and electrostatic designs are flat-out LUDICROUS SPEED! by comparison. Seems like planar driver types with their push-pull forces acting evenly over an entire diaphragm's surface help keep things controlled, while keeping that diaphragm as thin and massless as possible (easier when using electrostatic force instead of electromagnetic force like orthos) also helps significantly.

Quote:

yepimonfire said:

i don't think being bright IS fast. for example, i have a pair of etymotic HF5's that pretty much suck past 16khz and aren't anywhere close to bright, but they are very fast and bring out a lot of detail. I also have a pair of AD900's that are very fast and aren't bright. I had AD700's that were fast and ridiculously bright, i had a pair of Q701's which weren't even remotely bright that were fast. I don't think bloated bass will make a headphone sound slow either. I had a pair of IEM's once that were fast and had bloated bass, they sounded like crap however because of the bass, but they definitely handled transients well. i don't know what people here use as a reference for fast, but IMO a fast headphone is fast because no matter how complex the musical passages become, and no matter how heavy the bass is hitting, they continue to render the music flawlessly with no details smeared over. perhaps bright headphones do sound more detailed, but i dont see how that translates to fastness. i have a pair of radioshack bookshelf speakers that are bright and sound slow.

Click to expand...

Q701 not bright? I'm sure those treble peaks don't do anything at all. 

Quote:

lord voldemort said:

Q701 not bright? I'm sure those treble peaks don't do anything at all. 

Click to expand...

What treble peaks?

Quote:

namelesspfg said:

This reminds me of the "detail" discussions and how brightness/emphasized treble = fake detail, while faster transient response/attack/decay = real detail, even if the details aren't shoved in your face.
 
All I can say for sure is that consensus here suggests that dynamic/moving-coil designs are regarded as the slowest driver type, orthodynamic/planar-magnetic designs are significantly faster, and electrostatic designs are flat-out LUDICROUS SPEED! by comparison. Seems like planar driver types with their push-pull forces acting evenly over an entire diaphragm's surface help keep things controlled, while keeping that diaphragm as thin and massless as possible (easier when using electrostatic force instead of electromagnetic force like orthos) also helps significantly.

Click to expand...

 

 
This is a CSD plot of a dynamic driver (from the early '80s) with a convolution filter applied to the test signal before being fed into the phones. The decay measures as being instant over much of the frequency range, although the most stubborn spots of driver ringing persist. Not seen a single electrostat measure similarly by natural means.

vid said:

This is a CSD plot of a dynamic driver (from the early '80s) with a convolution filter applied to the test signal before being fed into the phones. The decay measures as being instant over much of the frequency range, although the most stubborn spots of driver ringing persist. Not seen a single electrostat measure similarly by natural means.

Click to expand...

Listen people.. This is all very informative and all. And we all appreciate the images but please stop. Stop, stop! ,STOP!, STOP!! You are making me want to own some electrostatics now!..... Just when I was thinking my collection was diverse enough... Please don't do this to me

No really though... thats useful and thanks!

Quote:

vid said:

 

 
This is a CSD plot of a dynamic driver (from the early '80s) with a convolution filter applied to the test signal before being fed into the phones. The decay measures as being instant over much of the frequency range, although the most stubborn spots of driver ringing persist. Not seen a single electrostat measure similarly by natural means.

Click to expand...

I think there is something else going on with electrostatics that has nothing to do with decay. The question is what?

I don't get the theory that attack and decay have nothing to do with speed. Its actually crazy if you think about it. Even my DT1350's go from faster using my portable amp to slow if I attempt to plug into my iTouch direct. The only difference is attack and decay.

If that ain't it, someone please post an actual definition that states what it is and why it has NOTHING to do with attack and decay. And please name a headphone with slow attack and decay in lets say bass and treble that is considered "fast" so I can learn for myself what fast really is.

Quote:

doc-holliday said:

I don't get the theory that attack and decay have nothing to do with speed. Its actually crazy if you think about it. Even my DT1350's go from faster using my portable amp to slow if I attempt to plug into my iTouch direct. The only difference is attack and decay.
If that ain't it, someone please post an actual definition that states what it is and why it has NOTHING to do with attack and decay. And please name a headphone with slow attack and decay in lets say bass and treble that is considered "fast" so I can learn for myself what fast really is.

Click to expand...

 
 
Quote:

doc-holliday said:

I don't get the theory that attack and decay have nothing to do with speed. Its actually crazy if you think about it. Even my DT1350's go from faster using my portable amp to slow if I attempt to plug into my iTouch direct. The only difference is attack and decay.
If that ain't it, someone please post an actual definition that states what it is and why it has NOTHING to do with attack and decay. And please name a headphone with slow attack and decay in lets say bass and treble that is considered "fast" so I can learn for myself what fast really is.

Click to expand...

the only difference i can think of between a dynamic headphone and a electrostatic/planar magnetic headphone is the fact that the drivers in the latter two are nearly flat, maybe it has something to do with phase? attack and decay definitely has something to do with speed, but it's just weird that dynamics seem to have better decay than electrostatics and everyone claims they sound faster/more detailed.
 
i am officially confused.

Here is one of the major problems with dynamic drivers, being that they are cone shaped, different frequency ranges leave the surface of the driver at different times, you also have problems (which you can easily observe by placing your loudspeakers below or above ear level) with higher frequencies "beaming" and being directional. With a flat driver (such as those used in electrostats and planar magnetic), this is mostly eliminated as frequencies radiate from all parts of the driver evenly. The HD800 also rectifies this problem with their ring driver. I don't think this equates to fastness, but it surely would influence the amount of detail resolution and clarity, which would be perceived as speed by most.
 
Notes about this image:
1. i suck at drawing
2. i particularly suck at MS paint
3. the bass is shown as radiating in a sphere because bass actually does radiate spherically, this is why you can place a subwoofer anywhere in the room and not tell its direction.

Lots of misinformation and speculation has been posted so far.
 
Impulse response is a measurement of headphone driver speed. Personally, I refer to "impulse response" when describing fast or slow headphones, as there's no ambiguity that way. Tyll of InnerFidelity has a PDF available somewhere of various headphones' measurements, it can probably be Google'd.
 
Impulse response partially depends on how well a headphone is being amped btw.
 
The fastest-sounding dynamic headphone I've heard is the Audio-Technica AD2000. Some of the slowest I've heard include the AKG K701, Beyerdynamic DT880, and Senn HD6xx. And electrostats are fast as already mentioned, by design.
 
Attack is not the same thing - look up "ADSR envelope" for more info on it. If we go back to the AD2K example, it happens to also have a fast attack, but its decay is very short - shorter than a lot of other dynamics.

Brightness has nothing to do with detail or speed, but I've noticed that "brightness" is a phrase that's started to morph into a catch-all to describe anything. Speed refers to the ability of a headphone (or speaker) to respond to the input signal and then *stop* responding to the input signal. Impulse response is a method of figuring distortion in a loudspeaker, related to phase alignment and resonance, and can get at showing "speed" just as a CSD can (CSD can be computed from IRF iirc). The convolving filter that vid demonstrated is upstream processing that attempts to change the resonance characteristics - you cannot beat any/every driver into perfect submission, but you can certainly change things up (similar acoustic simulation/recreation is used in modern DSP soundfield synthesis, although in reverse).

As far as 'stats being faster - they're just the evolution of ideas in speaker design: stiff and light is always the goal (it requires less energy to stop a very small mass from moving, than a very large mass from moving). But it isn't as simple as just how quickly the driver can start and stop, you also have to think about phase (which is where amplification can play into things to a minor extent), and delays that can be introduced there.

The easiest way to visualize this kind of stuff, imho, is via CSD waterfall.

Quote:

obobskivich said:

The convolving filter that vid demonstrated is upstream processing that attempts to change the resonance characteristics - you cannot beat any/every driver into perfect submission, but you can certainly change things up (similar acoustic simulation/recreation is used in modern DSP soundfield synthesis, although in reverse).

Click to expand...

 
$20 earbuds, original IR:

 
Same earbuds, convolution:
 

 
The impulse responses themselves aren't filtered, it's the sound as produced by the earbuds and as picked up by the microphone. Whether it's horribly distorted sound I can't say, since I can't measure the frequency response at my own eardrum and thus can't create a convolution filter for my own ears. Nonetheless, the impulse response claims that these cheap earbuds convolved are faster than any known electrostat - can the IR be trusted, then?
 
Also, if this is an 'upstream processing that attempts to change the resonance characteristics', so must physical phone modding (felt disc at the back of an ortho) be upstream processing, no?

Quote:

Tyll of InnerFidelity has a PDF available somewhere of various headphones' measurements, it can probably be Google'd.
 

Click to expand...

 
AllGraphs.pdf
 
Hi Steve!

vid said:

$20 earbuds, original IR:


Same earbuds, convolution:



The impulse responses themselves aren't filtered, it's the sound as produced by the earbuds and as picked up by the microphone. Whether it's horribly distorted sound I can't say, since I can't measure the frequency response at my own eardrum and thus can't create a convolution filter for my own ears. Nonetheless, the impulse response claims that these cheap earbuds convolved are faster than any known electrostat - can the IR be trusted, then?

Click to expand...

Kewlbeans. And regarding if it can be trusted, tough to say methinks. You'd need more points of data to paint a comprehensive picture. IR isn't telling you about "speed" in any sort of raw terms, it catches a number of variables (this isn't the same as shooting a car with a radar gun). So you're creating a filter that seeks to correct distortion that your IR measurement is showing (phase most likely). The driver isn't being physically modified though, so it starts and stops at the same rate as it did before - the signal is just being mangled in a way that when it hits the driver it becomes "unmangled." Same idea as any other sort of corrective EQ.

Distortion and FR could still be measured in your scenario, as it would show distortion through the same set-up you're shooting IR with. My guess is that electrostats will still be lower distortion and closer to full-range, and would benefit just as well from similar corrective filtering (and probably get you about as close to perfection as you could get; isodynamics could theoretically take you to the same place).

Also, if this is an 'upstream processing that attempts to change the resonance characteristics', so must physical phone modding (felt disc at the back of an ortho) be upstream processing, no?

Click to expand...

Nope. Two entirely unrelated concepts. At a high-level, you could consider it all "filtering" (since the output device is not passing the signal "straight up"), but physical modification of the output transducer or its enclosure is unrelated to EQ/DSP.

^ The spectrum of the convolution recording is flat between 40 Hz or so (lower limit of my recording capability) and about 15 kHz or whatever (where I set the lowpass), as is the original test signal, though without the lowpass. The convolution filter seeks to create a flat response, which it thus has done. The question is, is a flat response an automatic flat CSD (= perfect IR)? If so, why; and if not, why so?
 
If you would, please explain how physical modding and EQ are different in this case. Not that I doubt you, just that an explanation is nicer than a true/false statement.
 
The test signal was 'mangled' as you say to produce a desired output (flat response), but you say this can't change the rate at which the driver moves. Below is part of the waveform of (a) the original test signal, (b) the recording off the earbuds of the non-convolved signal, and (c) the recording of the convolved signal off the earbuds.
 
a
b
c
 
The rate at which the driver moves looks to have changed. But whether it did or not, it in any case seems that the driver is able to keep up with the test tone transients when fed with a convolved version of the signal, but having big trouble doing so otherwise.