[cegras]

I don't think decay is all about how quickly the sound fades to nothingness. Just from thinking about it a little, a very fast attack could come from something that fades very, very fast to a steady volume level and holds it, instead of having it fade quickly to nothingness overtime. This is what the STAX graph is showing.

 

[Shike]

Quote:

Originally Posted by kool bubba ice /img/forum/go_quote.gif At some point, does speed become about bragging rights.. If a headphone is a bit faster then a already fast headphone, will it even matter in the long run, in terms of musical enjoyment..

I've presented this argument a few times already. How fast is fast enough? I'm of the belief that most good quality headphones are already fast enough, so they should be judged on other merits.

That's just me though.

 

[kool bubba ice]

Quote:

Originally Posted by Shike /img/forum/go_quote.gif I've presented this argument a few times already. How fast is fast enough? I'm of the belief that most good quality headphones are already fast enough, so they should be judged on other merits. That's just me though.

So it comes down to, 'mine is bigger then yours' sort of thing..

 

[Pyriel0]

When someone can tell me the picosecond(or less) a guitar string stops vibrating, I will isten to their time theories.

 

[rds]

Quote:

Originally Posted by kool bubba ice /img/forum/go_quote.gif At some point, does speed become about bragging rights.. If a headphone is a bit faster then a already fast headphone, will it even matter in the long run, in terms of musical enjoyment..

Like everything at some point it won't be limiting factor. In other words it will be so fast that we cannot percieve a difference. Are these speeds at that level already? I actually think there is a way to sort of answer that question.

We look at the resolution of our source material. Let's say a redbook CD:

sample rate: 44 100 samples/second, or 1 sample 22us (micro-second).
sample resolution: 16 bits, or a range 65536 steps. If my calculation is right (20log(65536)) this means 96dB of dynamic range.

This would mean that in theory we can jump 96dB in 22us or 4.4dB/us.

That would be the limit of CD resolution assuming my math is reasonable. So looking at our headphone graphs we are nowhere close to this limit.

Now when it comes to what we can perceive I have no idea.

 

[grokit]

So the opposite of a fast headphone might be the double-dampening W5000 "Raffinatos", which I have read actually resonate for a second or two with no source present?

They are very highly regarded by most, dissed by a few others from what I have seen. I have never heard them, but I will someday as have tried the W1000s and liked them enough to want MORE.

So they are on my "list", for better or worse.

It doesn't take much to drive them (edit: in comparison to some others, and without regard to proper "synergy), and I have noticed that the higher-rated dynamics in this thread do, so that's another opposite.

But "double-dampening" the bass just sounds slow!

And I really don't understand these graphs

 

[MrGreen]

Quote:

Originally Posted by rds /img/forum/go_quote.gif Keep in mind that the graphs above show decay which is related to attack, but is not the same thing. Interesting that the stax have the slowest decay of the group by a good margin

Didn't realise that I suggested that. Edited the post to make it more clear.


Out of curiosity, what graphs measure speed of attack?

 

[Shike]

Quote:

Originally Posted by kool bubba ice /img/forum/go_quote.gif So it comes down to, 'mine is bigger then yours' sort of thing..

That's my thinking, but I'm a Stax cult reject apparently.

 

[cegras]

STAX probably seems fast because it decays very fast to a fixed volume level below. Grado headphones are, at this point, still 'fading.'

 

[b0dhi]

Quote:

Originally Posted by rds /img/forum/go_quote.gif Interesting that the stax have the slowest decay of the group by a good margin

I'd venture a guess that this is done to give them more deep bass (<40hz).

 

[FourierMakesFunk]

Quote:

Originally Posted by rds /img/forum/go_quote.gif That would be the limit of CD resolution assuming my math is reasonable. So looking at our headphone graphs we are nowhere close to this limit. Now when it comes to what we can perceive I have no idea.

CD resolution don't really correspond to this like that (a driver being slower than the bit rate actually means it is rounding off the hard edges between steps a bit), but if memory serves me right we can hear a little bit worse than that, and it all gets complicated further by the fact that were are never listening to straight impulses. During that decay time other sounds are being made and the driver continues to get forced around.

Ignoring some psychoacoustic filtering that is probably going on, these headphones are probably pretty darn near the limit of what we can perceive if you listen really hard to actual audio. In terms of getting a good sound, other factors are going to be far more important and limiting to a headphone.

That isn't to say that speed isn't unimportant, just designs have gotten good enough for now.

BUT we are Head-Fi-ers! Good enough is never good enough! Quick, fetch me my 192k / 24 bit audio and anechoic listening room! I want to hear Leslie Gore's "Sunshine Lollypops and Rainbows" with fidelity that the gods themselves could envy!

.

 

[paranoize]

How do you interpret these strange graphs?

 

[rds]

Quote:

Originally Posted by FourierMakesFunk /img/forum/go_quote.gif CD resolution don't really correspond to this like that (a driver being slower than the bit rate actually means it is rounding off the hard edges between steps a bit), but if memory serves me right we can hear a little bit worse than that, and it all gets complicated further by the fact that were are never listening to straight impulses. During that decay time other sounds are being made and the driver continues to get forced around. Ignoring some psychoacoustic filtering that is probably going on, these headphones are probably pretty darn near the limit of what we can perceive if you listen really hard to actual audio. In terms of getting a good sound, other factors are going to be far more important and limiting to a headphone. That isn't to say that speed isn't unimportant, just designs have gotten good enough for now. BUT we are Head-Fi-ers! Good enough is never good enough! Quick, fetch me my 192k / 24 bit audio and anechoic listening room! I want to hear Leslie Gore's "Sunshine Lollypops and Rainbows" with fidelity that the gods themselves could envy! .

But what I'm trying to show with my calculations is that the state of the art headphones (that we've seen graphs for) are about 1000x slower in terms of slew rate than CD audio.

Of course the full dynamic range is never used and very rarely (if ever) would a signal rise anywhere close 96dB between samples.
However, if the driver still has significant inertia from the last sample when it goes to play the next sample that inertia will cause distortion. It seems me that resolution is a direct function of attack/decay time, and I think we all know that some phones resolve details better than others. Some resolve better in the bass and others better in the midrange or treble. Looking at the graphs we have here that's exactly what I'd expect.

PS I know my "model" is simplified and that DACs will interpolate and amps have a slew rate etc etc. But by talking about impulses we avoid a lot of complex math without losing much in the way of a general description.

 

[cegras]

Quote:

Originally Posted by paranoize /img/forum/go_quote.gif How do you interpret these strange graphs?

It's a waterfall graph. It's my first time looking at it too, but look at the axis. You have time on the y axis, decibels on the z, and frequency on x.

You're looking at how fast a signal at a certain frequency (x) decays in loudness (z) over time (y).

 

[unclepauly]

I want the slowest phones. Where can I find them?