** Confessions of a Failing Audiophile ** - I hated the LCD2 - Page 37  

Quote:

Originally Posted by Head Injury 

Square wave response is very closely related to the frequency response. A little bit of EQ down in the HD800's lower treble, maybe a boost in the upper mid-range, and its square wave response will improve as well. And no ringing.

The LCD2 outperforms the HD800 in my subjective pipe organ test.  I doubt there is any other headphone, other than the LCD3, that can match the LCD2 in pipe organ awesomeness.

It will be interesting to see Purrin's measurements of a LCD-2 rev 2 when he gets a chance to measure a rev 2.  Ditto with the LCD-3.  I wonder if any ringing will still be there?  And is driver matching better?

Quote:

Originally Posted by MacedonianHero 

Really? Sorry, I'm an engineer and I want to see the measurements. FWIW...the 50Hz square wave is the output of the headphones to an input 50Hz square wave.....only 50Hz. EQ will do nothing as the input is again...only 50Hz.

the frequency response, impulse response and step response are only different way to display for the *same* information, play a little with the Laplace transform and to go from one to another.

Of course that's for linear systems, but low distortion headphones are fairly linear. And I remember someone in the sound science sub forum playing doing the transforms and recreating the square waves response and impulse response purely based on the frequency response graphs. He used Tyll's FR graphs and the the square waves and impulse response he calculated matched nearly perfectly the graphs Tyll measured.

Or for a more intuitive approach, you could consider the Fourier decomposition of a square wave , so for a 50 Hz square wave, the signal is a sum of sin(50t) + 1/3 sin(150t) + 1/5 sin(250t) + 1/7  sin(350t) ..., the squareness of the 50 Hz square wave is only the direct consequence of the flatness of the FR of the LCD-2 from 10 to 1000 Hz: the 50, 150, 250 Hz sinwave are reproduced in the right proportions, leading to a sum that looks like a square, skewer those proportions a bit, and the square looks a little less squarish.

Quote:

Originally Posted by khaos974 

the frequency response, impulse response and step response are only different way to display for the *same* information, play a little with the Laplace transform and to go from one to another.

Of course that's for linear systems, but low distortion headphones are fairly linear. And I remember someone in the sound science sub forum playing doing the transforms and recreating the square waves response and impulse response purely based on the frequency response graphs. He used Tyll's FR graphs and the the square waves and impulse response he calculated matched nearly perfectly the graphs Tyll measured.

 

Or for a more intuitive approach, you could consider the Fourier decomposition of a square wave , so for a 50 Hz square wave, the signal is a sum of sin(50t) + 1/3 sin(150t) + 1/5 sin(250t) + 1/7  sin(350t) ..., the squareness of the 50 Hz square wave is only the direct consequence of the flatness of the FR of the LCD-2 from 10 to 1000 Hz: the 50, 150, 250 Hz sinwave are reproduced in the right proportions, leading to a sum that looks like a square, skewer those proportions a bit, and the square looks a little less squarish.

Very easy to test.  C/O measurements with a high quality EQ and see if we can generate perfect square waves from a headphone with an imperfect one.  From my understanding, the closer we get to the square wave, the closer we are to neutrality.

Are you sure that the good 30 hz square response is purely because of flatness from 10hz to 1 khz?  Even within a perfect 30 hz square wave, there are countless frequency contents that contribute to this response - closer to infinity.

Anyway this was posted in the science forum.

http://www.physics.mcgill.ca/~moore/P101/Lectures/Lecture-10.pdf

Quote:

Originally Posted by SP Wild 

Very easy to test.  C/O measurements with a high quality EQ and see if we can generate perfect square waves from a headphone with an imperfect one.  From my understanding, the closer we get to the square wave, the closer we are to neutrality.

Yes with speaker, with headphones, it's another story due the the HRTF. It's relatively true for bass and low mids, where all HRTF assume a unitary gain, but with high mids and treble, it's another story, and the corrections to apply are subject to more debate, hence is what is a neutral frequency response is subject to debate.

Quote:

Originally Posted by MacedonianHero 

 

To make assertions that one's opinions are purely objective is pure folly.

This, QFT. At least we still have a completely ridiculous LCD-2 thread for our online entertainment 

My subjective opinion is that the LCD-2 is the best all-around headphone in that it performs well with the widest variety of musical genres, recording qualities, and variety of amplification factors. Some headphones may be (subjectively) better for certain genres or perfect recordings within those genres but the LCD-2 is the most versatile that I have tried in these respects and works pretty well with a wide variety of amps.

Quote:

Originally Posted by SP Wild 

Unfortunately a few batches of LCD2 driver failures have been recorded.

I'm pretty sure that it was just a single, albeit large, batch of driver failures due to a single bad batch of driver material delivered to Audeze by a vendor. When they could no longer trust that material's quality they found a new material and this is why we have the LCD-2 rev.2.

Quote:

Originally Posted by khaos974 

Yes with speaker, with headphones, it's another story due the the HRTF. It's relatively true for bass and low mids, where all HRTF assume a unitary gain, but with high mids and treble, it's another story, and the corrections to apply are subject to more debate, hence is what is a neutral frequency response is subject to debate.

 

 

I don't believe it is as complex as that.  HRTF doesn't add up in my experience - recording variations and listening levels do seem more plausible than HRTF.  I am subscribing to the pure physics of air pressure variation being detected by the sensor that we have only one of on each side of our head.  Within that single factor of air pressure contains our frequency response and neutrality, staging, imaging and attack and decay.  This is why bright cans cannot produce a good square wave IMO.

Quote:

Originally Posted by SP Wild 

I don't believe it is as complex as that.  HRTF doesn't add up in my experience - recording variations and listening levels do seem more plausible than HRTF.  I am subscribing to the pure physics of air pressure variation being detected by the sensor that we have only one of on each side of our head.  Within that single factor of air pressure contains our frequency response and neutrality, staging, imaging and attack and decay.  This is why bright cans cannot produce a good square wave IMO.

Here's the relevant link http://www.head-fi.org/t/566163/headphones-are-iir-filters-graphs

Anything with a flat and straight FR from 20 to 1000 Hz will produce a nice response for a 30 Hz square signal

EDIT: never mind, I see you already posted in that thread.

Quote:

Originally Posted by SP Wild 

I can't agree with this.  The HD800 was more impressive in one case, in one occasion, by one member, I don not dispute the findings, but of all people - yourself, should be aware of the problem with sample size.  Unfortunately a few batches of LCD2 driver failures have been recorded.

 

Tyll offers a more comprehensive scope of measurements tailored specifically to explain what we hear, a much more complete data set.  When we look at the data as a whole, we are more able to define what we hear as opposed to a single factor.  It is the more comprehensive data set by Tyll which suggests, and mirrors my experience, of how a headphone might sound IMO.

 

Tyll makes comprehensive measurements, but he doesn't include CSD graphs which are incredibly interesting.

It's true purrin may have a bad LCD-2 and good HD800, and he does mention that the LCD-2s he's heard had a wide variety of sounds. However, given the great channel balancing in the HD800, and the fact that it's a big company, I'd expect tolerances in product variation to be quite tight. Good or bad LCD-2, the fact that differences could be that large between them doesn't speak much for quality either. Likewise, Tyll is only one person as well. He's only measured one HD800. He's got three LCD-2 measurements, which is great, but the first one is quite different (and much less impressive) from the second despite them both seeming to be rev. 1. Product variation, or pad changes? I hope the latter.

All in all, though, there is evidence to suggest that the LCD-2 is not as perfect as it's widely believed. There isn't enough to make any authoritative claims, but there was never enough evidence to support "perfection" in the first place. You have to take all the evidence you can get. I don't think we can use Tyll's measurements to suggest that the LCD-2 is technically superior to anything when there's purrin's measurements making the claim that it's not. If purrin can measure more LCD-2, and he can confirm or deny that this is a problem, fantastic. Until then, use any measurements carefully, and use them all. To do anything else is to be biased, which I try not to be. It's not like I want to attack the LCD-2, I love it.

Quote:

Originally Posted by Ham Sandwich 

It will be interesting to see Purrin's measurements of a LCD-2 rev 2 when he gets a chance to measure a rev 2.  Ditto with the LCD-3.  I wonder if any ringing will still be there?  And is driver matching better?

I'm interested in this as well, because I don't hear excessive ringing. I don't really know what to listen for, though. 

Send him your LCD-2rev2!  Someone needs to send them to him to get measurements....

Quote:

Originally Posted by rhythmdevils 

Send him your LCD-2rev2!  Someone needs to send them to him to get measurements....

Is he bringing his testing rig to the next local CA meet in a week or so?  He could almost certainly get a rev 2 for testing then.

I don't want to give up my rev 2 for even a week.  Nothing else in my headphone collection will make do in their absence. :(

Maybe if I had a Thunderpants to tide me over...

Quote:

Originally Posted by Head Injury 

Tyll makes comprehensive measurements, but he doesn't include CSD graphs which are incredibly interesting.

 

It's true purrin may have a bad LCD-2 and good HD800, and he does mention that the LCD-2s he's heard had a wide variety of sounds. However, given the great channel balancing in the HD800, and the fact that it's a big company, I'd expect tolerances in product variation to be quite tight. Good or bad LCD-2, the fact that differences could be that large between them doesn't speak much for quality either. Likewise, Tyll is only one person as well. He's only measured one HD800. He's got three LCD-2 measurements, which is great, but the first one is quite different (and much less impressive) from the second despite them both seeming to be rev. 1. Product variation, or pad changes? I hope the latter.

 

All in all, though, there is evidence to suggest that the LCD-2 is not as perfect as it's widely believed. There isn't enough to make any authoritative claims, but there was never enough evidence to support "perfection" in the first place. You have to take all the evidence you can get. I don't think we can use Tyll's measurements to suggest that the LCD-2 is technically superior to anything when there's purrin's measurements making the claim that it's not. If purrin can measure more LCD-2, and he can confirm or deny that this is a problem, fantastic. Until then, use any measurements carefully, and use them all. To do anything else is to be biased, which I try not to be. It's not like I want to attack the LCD-2, I love it.

 

I'm interested in this as well, because I don't hear excessive ringing. I don't really know what to listen for, though. 

The great thing about headphone measurements is that it is measuring a tangible phenomena in a tangible world - as opposed to measuring amps and dacs.  It is mechanical in nature and is measuring a physical existence - ie, liquid, solid or gas.  I think we have the full capability in terms of newtonian physics.  I think we have the capability to measure most parameters of a tangible world. 

But "perfection", yes the LCD2s were implied by many that they were, this is very subjective when speaking experiences.  On many recordings, they do sound kinda "perfect" to me and I have old pads R1 (Tylls were an old pad R1 and a new pad R1).  Aside from that, they still come closer to "perfection" in many objective parameters than most other headphones.

Purrins measurement is designed to measure one parameter - and it does not contradict all the parameters measured by Tyll (however there is a post in Purrins thread where a CSD was posted of a R2 measured by Tyll that does).  I am simply saying that multiple parameters are involved in audio perception and Tyll is simply covering more.

To allude that the LCD2s measure poorly cannot be sustained.  Excelled at all parameters except one in one sample - is more accurate. 

Quote:

Originally Posted by SP Wild 

The great thing about headphone measurements is that it is measuring a tangible phenomena in a tangible world - as opposed to measuring amps and dacs.  It is mechanical in nature and is measuring a physical existence - ie, liquid, solid or gas.  I think we have the full capability in terms of newtonian physics.  I think we have the capability to measure most parameters of a tangible world.

 

Purrins measurement is designed to measure one parameter - and it does not contradict all the parameters measured by Tyll (however there is a post in Purrins thread where a CSD was posted of a R2 measured by Tyll that does).  I am simply saying that multiple parameters are involved in audio perception and Tyll is simply covering more.

 

A statement declaring that the LCD2s measure poorly cannot be sustained.  Excelled at all parameters except one is more accurate. 

A CSD graph measures both frequency response and ringing, at the same time. The only thing the LCD-2 appears to best the HD800 in according to Tyll's graphs are frequency response linearity (contradicted by purrin's graphs) and square wave response. The difference in THD is mostly negligible, as is isolation. Square wave response is mostly just frequency response. The CSD graphs are frequency response combined with impulse response, measured at all frequencies at once. That's more comprehensive than any one graph by Tyll.

I'm not saying the LCD-2 measured poorly. I'm saying it measures no better than other headphones overall, better in some ways and worse in others. There's no evidence that it's really technically superior, because it has flaws all the same, and ones that can't be fixed by EQ. MacedonianHero actually said that there is "no comparison" between the LCD-2's measurements and others, the LCD-2 is that much better. It's not.

I'd send my LCD-2 in, but I usually have them on my head 

Somehow the first paragraph reads like you think amps and DACs can't be measured. They're much, much easier to measure. Signal in vs. signal out. Headphones are much more complex, in how they're measured, what is measured, and problems associated with both.

Quote:

Originally Posted by Head Injury 

I'm interested in this as well, because I don't hear excessive ringing. I don't really know what to listen for, though. 

I haven't noticed a ringing in my LCD-2 rev 2.  But I'm also not entirely sure what to listen for and I'm also not sure if I'm sensitive to whatever it is the LCD-2 ringing does.

I do hear the ringing in Grados.  That's easy.  The 2 kHz ringing is dead easy (shouty).  The higher frequency ringing peaks is also noticeable.  I previously attributed the higher frequency ringing to "weird spacial effects".  Now that I've seen the CSD plots I know it's actually due to ringing.

I hear the ringing peak in my D2000 if I use my Asgard amp.  I don't really notice it much if I use my AV123 x-head amp.  With the Asgard that treble peak (ringing peak) is not pleasant.

But I don't hear anything obnoxious or odd in my LCD-2 rev 2.  I still need to play around and experiment with them.  They're only a few weeks old.

Something neat as an artifact...  I was experimenting with a virtual subwoofer DSP effect in J River Media Center and using my Grado SR325is.  I was getting a weird phasing like effect in the "spacial treble".  I assume that artifact was due to the upper treble Grado ringing.  I didn't hear that effect with other headphones.  So perhaps some various DSP effects can help accentuate ringing artifacts.

Quote:

Originally Posted by rhythmdevils 

Send him your LCD-2rev2!  Someone needs to send them to him to get measurements....

But than I'll find out that the LCD-2 Rev2 would have terrible measurements and thus I would view the headphones I spent months working to get as a worse headphone compared to everyone else :-(.

Quote:

Originally Posted by Head Injury 

A CSD graph measures both frequency response and ringing, at the same time. The only thing the LCD-2 appears to best the HD800 in according to Tyll's graphs are frequency response linearity (contradicted by purrin's graphs) and square wave response. The difference in THD is mostly negligible, as is isolation. Square wave response is mostly just frequency response. The CSD graphs are frequency response combined with impulse response, measured at all frequencies at once. That's more comprehensive than any one graph by Tyll.

 

I'm not saying the LCD-2 measured poorly. I'm saying it measures no better than other headphones overall, better in some ways and worse in others. There's no evidence that it's really technically superior, because it has flaws all the same, and ones that can't be fixed by EQ. MacedonianHero actually said that there is "no comparison" between the LCD-2's measurements and others, the LCD-2 is that much better. It's not.

 

I'd send my LCD-2 in, but I usually have them on my head 

 

Somehow the first paragraph reads like you think amps and DACs can't be measured. They're much, much easier to measure. Signal in vs. signal out. Headphones are much more complex, in how they're measured, what is measured, and problems associated with both.

Lets just say I cannot correlate how a headphone's overall tonal balance is by a CSD or impulse response graph.  The frequency response graph is generated by a sineweep, one frequency at a time by Tyll.  This graph actually measures the entire spectrum - I will purchase or not purchase a head phone according to this graph.

Ignoring frequecies  below 1 khz is not wise, this is where most voice and instruments lay.  All amps and dacs measure "perfectly" flat easily and we still argue which is "neutral".  How difficult it is get a headphone to measure flat, and the LCD2s was the first to nail it for what is known and objective for dynamic headphones - all the way up to 1khz, after that we can question "HRTF" - a CSD plot showing only signal fall times after 1 khz does not overide neutrality below 1 khz.

So going by your evidence, I should have just kept my HD650s and call it a day, because it excelled at CSD, as good as the best.  I have them both here right now - there is no contest.  Tylls other measurements explain the differences I hear between these two.