Headphone speed and its effect

[b0dhi]

Quote:

Originally Posted by toxicsweet /img/forum/go_quote.gif Balanced armatures are much faster than dynamic headphones. All in all, dynamic headphones are the slowest type of headphones(AFAIK).

What makes you say that armatures are faster than dynamics? I would've thought that since a typical full range armature can't oscillate as fast as a typical dynamic, dynamics would be faster.

 

[PiccoloNamek]

I think what he meant is that because of their small size and low mass, armatures have a faster transient response.

 

[b0dhi]

A faster transient response implies a higher maximum relative frequency response. Since that isn't the case with armatures, it would mean they have a slower transient response. That is especially true considering the fact that armatures need to displace less total air to get the same dBSPL at the eardrum, due to the fact that they have a tighter and closer coupling to it.

 

[PiccoloNamek]

Quote:

A faster transient response implies a higher maximum relative frequency response. Since that isn't the case with armatures, it would mean they have a slower transient response.

But they clearly don't. Not only do they seem to have a faster attack and decay, but they also resolve more inner detail than dynamic phones as well, at least in my experience, although that could be due to the close coupling between driver and eardrum.

That being said, I don't believe a headphone's start/stop time is necessarily dependent on its frequency response. A headphone could have a very poor frequency response but still have the ability to initiate and cease movement quickly.

 

[b0dhi]

Quote:

Originally Posted by PiccoloNamek /img/forum/go_quote.gif But they clearly don't. Not only do they seem to have a faster attack and decay, but they also resolve more inner detail than dynamic phones as well, at least in my experience, although that could be due to the close coupling between driver and eardrum.

This might have something to do with the nature of the enclosure in a balanced armature. The reflections in the chamber are very short because it's so small, meaning that when the driver makes a sound, you don't get a flood of reflections from the enclosure a relatively long time later (relative to the sampling rate). With a typical dynamic, even with a well designed chamber, these reflections arrive at the ear much later.

Quote:

Originally Posted by PiccoloNamek /img/forum/go_quote.gif That being said, I don't believe a headphone's start/stop time is necessarily dependent on its frequency response. A headphone could have a very poor frequency response but still have the ability to initiate and cease movement quickly.

Start/stop time, transient response, and frequency response are all closely related and describe how the driver responds to input. It would be hard to envision a driver that has a very quick start/stop time while at the same time having a low maximum frequency response. Although I would be happy if someone would provide an example of such a driver if it exists.

 

[ericj]

Quote:

Originally Posted by b0dhi /img/forum/go_quote.gif Start/stop time, transient response, and frequency response are all closely related and describe how the driver responds to input. It would be hard to envision a driver that has a very quick start/stop time while at the same time having a low maximum frequency response. Although I would be happy if someone would provide an example of such a driver if it exists.

I'm amazed that you don't seem to understand the concept of resonance.

While it's true that on average, larger, looser drivers tend to resonate lower, it is in fact not hard to envision a driver that has both excellent transient response as a low end to the frequency response.

For decades, Stax has been making these big drivers that are almost three times as long as they are wide. lambdas, sigmas, and the newer 202/303/404/4070 models all have excellent transient response and low low bass.

And it only works properly because it's planar, and because of the thoroughly push/pull nature of the electrostatic driver.

 

[b0dhi]

Quote:

Originally Posted by ericj /img/forum/go_quote.gif I'm amazed that you don't seem to understand the concept of resonance. While it's true that on average, larger, looser drivers tend to resonate lower, it is in fact not hard to envision a driver that has both excellent transient response as a low end to the frequency response.

Maybe I wasn't clear. I didn't say low end, I said "low maximum frequency response", referring a good armature's ~16Khz rolloff relative to a good dynamic's ~25Khz or higher rolloff.

 

[catscratch]

Well, you said it yourself. Balanced armatures. Definitely pretty quick, not quite as fast as a 'stat but well above your average dynamic driver, and most have yet to demonstrate any meaningful output past 16khz.

I'm a speed freak in my headphones. I have speed issues with a balanced HD650 which is the 2d fastest dynamic I've ever heard (Qualia 010 being faster). I even have speed issues with the balanced-armature Westone ES2, which is plenty fast even by balanced armature standards. 'Stats, the Taket H2 (the fastest driver I've heard so far), the K340, they're all fast enough, but most have had other problems that I found hard to live with.

The easiest way to differentiate is to load up the headphone with some seriously fast, dense music. I use well-produced death metal and very fast, heavy symphonic pieces. You don't even have to listen for any kind of specifics - if the headphone can keep up and not lose sight of the tone and texture of each separate instrument and if it can still image convincingly, it's fast enough for me. Or rather, should be, but never is.

 

[b0dhi]

Can someone explain to me why it is that in an ABX test comparing -v2 MP3 and FLAC, I get 100% accuracy (p < 0.001) with the PK1, but much lower accuracy with the H2? Bear in mind the PK1 is unamplified while the H2 is well amplified, giving it every advantage. Also, the H2 has much more forward (and much more extended) treble than the PK1, so "fake" detail due to frequency response can't be the reason.

 

[Cool_Torpedo]

Quote:

Originally Posted by b0dhi /img/forum/go_quote.gif Can someone explain to me why it is that in an ABX test comparing -v2 MP3 and FLAC, I get 100% accuracy (p < 0.001) with the PK1, but much lower accuracy with the H2? Bear in mind the PK1 is unamplified while the H2 is well amplified, giving it every advantage. Also, the H2 has much more forward (and much more extended) treble than the PK1, so "fake" detail due to frequency response can't be the reason.

If you owning both can't say, it's hard for us to tell.
However I have two guesses. One is the amp used, which might be not as transparent to the signal passed through it as it should.
The other one could be related to sensitivity. Were the PK1 more sensitive, they'd be converting into sound any minute voltage variation in the signal, while the H2 could be "eating" those wee changes. The same reason why some phones hum a lot with a source despite using a low volume, while others don't hum at all even though you pump up the volume.

Rgrds

 

[b0dhi]

The point is that it still sounds like an extremely fast driver (and this is other peoples' impression too, with different amps and sources driving it). And if it sounds so fast, and yet as far as personal ABX testing shows, can't match the resolution of the much "slower" PK1, then of what value is the subjective impression of a headphone being "fast"? While no doubt the H2 (and others) have very fast drivers, I question whether that in itself says anything useful about the quality of sound it will produce. At least in this case, it doesn't translate into better definition. (I very much doubt my amp is the cause, because it's leagues above the source itself.)

 

[Cool_Torpedo]

I'm not saying they're not fast at all, but if they need a higher voltage signal to show it and don't move at all when small signals are present, they're somehow compressing the whole, which is one of my findings when listening to planar drivers in general. I don't know the H2, never listened to them.

There's also a thing associated to "speed" which is the lack of stored signal in the driver after the "sound event" ends. The driver can be not very speedy but the cleanliness associated to the lack of stored energy, could be very noticeable, hence the perceived effect of speed. Also some excess of energy in the treble and some lack of it in the bass can make for a "speedy" sound perceptually, which not necessarily is due to a driver being trully fast.

Who knows, but my guess is that if they can't show minute details and don't allow to distinguish a compressed file from a lossless, they're probably not very resolving.

Rgrds

 

[Catharsis]

Quote:

Originally Posted by b0dhi /img/forum/go_quote.gif What you're describing is the frequency response. "Faster" by this definition is basically how high the (true - not manufacturer's) frequency response goes (comparatively). Assuming most drivers have a strong response at 20Khz, how would having a higher frequency response beyond the audible range make drum hits or voices sound any different?

Er yes....

If a driver can produce a frequency of 35,000hz, then effectively that driver has to go from "off" to "on" and back again 35,000 times in a single second. This would effectively mean that the time required by the driver to return to it's stable position of "0" would be 1 / 35,000th of a second or 0.00002 seconds.

I would say that this is far below the perceptions of humans in a technical sense. I'm beginning to think that speed is a relative concept of how we perceive sound.

Let me put it to you this way:

Most loudspeakers don't have very high maximum audible frequencies; a result of high mass drivers etc. Most headphones have incredibly small and lightweight drivers and are thus able to easily produce frequencies well above the human threshold of perception. Might this simple fact of measuring driver speed boil down to frequency response as b0dhi had stated?

After all, we don't have any objective evidence with which to discuss speed other than through measurements such as Fq response, slew rate etc. Most of this discussion is assumption.

 

[Catharsis]

Quote:

Originally Posted by billinkansas /img/forum/go_quote.gif The drivers ability to stop responding to a frequency very quickly is at least as important as its ability to start. Slow phones can leave a lot of junk noise following your music that clouds things up a bit.

So does this mean that a 10,000 hz input signal would become a 9,800hz output signal through "slow" headphones because it any given frame of "real" time, there are a few milliseconds where the driver is warming up and a few milliseonds of cooling down (attack and decay)?

These concepts don't make sense to me. I can see how we might perceive sound differently (for a great number of reasons), but you technically can't have a delayed attack and response, otherwise you would have an incorrect number of vibrations per second, which in essence would alter the pitch of the sound.

 

[Catharsis]

Quote:

Originally Posted by billinkansas /img/forum/go_quote.gif Here’s an example of speed at the bottom end of the spectrum: This evening I was listening to a CD of the London Symphony playing Edward Elgar’s “Cockainge” – great recording BTW. About 5 minutes into the piece there is a soft section that as 4 very light deep tympani taps a few seconds apart - so soft that it would be easy to not notice tham. I listened carefully to this over and over for several minutes. Here’s what some of my different headphones thought about that passage: Stax Lambda – These don’t go quite deep enough to capture the whole sound, but most of it was there crystal clear – the drum had a soft but fast and precise attack with a natural decay – just like real life. Stax SRX-MKIII - Similar to Lambda, less deep response but even clearer. I don't have my Stax SR303's set up right now, but I'll try them sometime as well. Yamaha YH-1 Orthodynamic - dampened with thick felt – There it was, the entire sound of the drum tone, and just about as clear and quick as the Lambdas. Best of the bunch. Yamaha YH-2 Orthodynamic stock – Similar to the YH-1, almost as deep as well. 2nd best of the bunch. Denon D2000 – the depth of the tone was there ok as one would expect from these phones, but it the effect was neither clear nor precise. Sennheiser HD600 – the bass drum sounded like someone down the street kicking a barrel. The attack was nebulous, and the whole thing was kind of just a thud. I really wouldn’t have been able to tell what it was with any certainty if I didn’t already know. Not good.

If you're now referring to detail retreival as being a direct product of driver speed, then I guess we can say that the SA5000, DT880, DT48, CD3000, and Ety's are all fast.