LCD3 Measurements

[arnaud]

As Purrin said essentially but my take is:
> One thing I enjoy about headphones is ability to hear the recording without the coloration from the room
> Listening to speakers from a dead sounding room isn't optimal because they've typically been designed accounting for some wall reflection (in particular in regards to tonal balance at the listening position)
> The analogy with headphones earcup reflections would be more like not controlling the enclosure structural and acoustic resonances to add color. In either case it's not a strategy I would favor for the reasons marv mentionned (shows up as nasty peaks in the response). It's worse for heaphones because these resonances are at mid high frequencies and there are few of them so clearly audible
> I could imagine somehow that adding some homogenous reverberation over the whole frequency range could help with perceived imaging from headphones but you wouldn't be able to realize that acoustically (need a very large room with zillions of modes/reflections to achieve an homogenous RT over the f range)

 

[shamu144]

 
[size=10pt][size=10pt]Thank you both for sharing your point of view.[/size][/size]
 
[size=10pt][size=10pt]Somehow, I would rather think that those ear cups reflections do indeed play a role in the spatial presentation of the headphone. [/size][/size]
 
[size=10pt][size=10pt]Sure, we know that our human auditory sense do not allow us to distinguish early reflected sounds that arrive at the ear with less than 3 to 5 ms delay (which would make those ringing phenomenon observed in the high frequencies quite difficult to perceive when the decay is fast), but apparently, the brain is able to differentiate a reflected sound if it is separated by as little as 30ms (or so I read). [/size][/size]
[size=10pt] [/size]
[size=10pt][size=10pt]It would be interesting to see a CSD plots but down at -60dB (which is what we could consider sufficient attenuation for achieving silence from a 90dB level) without trying to minimize those early reflections as per your current measurement procedure (absorbing baffle). We might see quite enough energy from those ear cup reflections after 30ms especially below 1000Hz. [/size][/size]
 
[size=10pt][size=10pt]I have tried to estimate the T60 reverberation time for a headphone based on the Sabine equation (found on the web, I am no expert at all) as follow: [/size][/size]
[size=10pt][size=10pt]T60 = 55,25 * V / c * S * a where [/size][/size]
[size=10pt][size=10pt]V is the volume (I use dimension 12cm x 9cm x 3cm),[/size][/size]
[size=10pt][size=10pt]c the speed of sound in the air[/size][/size]
[size=10pt][size=10pt]S the surface (I use 12cm x 9 cm) and [/size][/size]
[size=10pt][size=10pt]a is the absorption coefficient (here the material used can play a significant impact, because absorption values can range from 0,1 to more than 1)[/size][/size]
 
[size=10pt][size=10pt]Well, If I am not wrong with the calculation, that gives us a range of 5ms (absorption coefficient 1) to 50 ms (absorption coefficient 0,1)!!!  So the latter value could indeed start to generate audible reflected sounds ![/size][/size]
 
[size=10pt][size=10pt]Here is a CSD plot for the K701, known for its cavernous soundstage:[/size][/size]
 
[size=10pt][size=10pt][/size][/size]
 
[size=10pt][size=10pt]And here is a Grado SR225, known for their lack of soundstage:[/size][/size]
 
[size=10pt][size=10pt][/size][/size]
 
[size=10pt][size=10pt]I don't know, but there seem to be a pattern indeed to me.[/size][/size]

 

[khaos974]

Quote:

 
[size=10pt][size=10pt]Thank you both for sharing your point of view.[/size][/size]
 
[size=10pt][size=10pt]Somehow, I would rather think that those ear cups reflections do indeed play a role in the spatial presentation of the headphone. [/size][/size]
 
[size=10pt][size=10pt]Sure, we know that our human auditory sense do not allow us to distinguish early reflected sounds that arrive at the ear with less than 3 to 5 ms delay (which would make those ringing phenomenon observed in the high frequencies quite difficult to perceive when the decay is fast), but apparently, the brain is able to differentiate a reflected sound if it is separated by as little as 30ms (or so I read). [/size][/size]
[size=10pt] [/size]
[size=10pt][size=10pt]It would be interesting to see a CSD plots but down at -60dB (which is what we could consider sufficient attenuation for achieving silence from a 90dB level) without trying to minimize those early reflections as per your current measurement procedure (absorbing baffle). We might see quite enough energy from those ear cup reflections after 30ms especially below 1000Hz. [/size][/size]
 
[size=10pt][size=10pt]I have tried to estimate the T60 reverberation time for a headphone based on the Sabine equation (found on the web, I am no expert at all) as follow: [/size][/size]
[size=10pt][size=10pt]T60 = 55,25 * V / c * S * a where [/size][/size]
[size=10pt][size=10pt]V is the volume (I use dimension 12cm x 9cm x 3cm),[/size][/size]
[size=10pt][size=10pt]c the speed of sound in the air[/size][/size]
[size=10pt][size=10pt]S the surface (I use 12cm x 9 cm) and [/size][/size]
[size=10pt][size=10pt]a is the absorption coefficient (here the material used can play a significant impact, because absorption values can range from 0,1 to more than 1)[/size][/size]
 
[size=10pt][size=10pt]Well, If I am not wrong with the calculation, that gives us a range of 5ms (absorption coefficient 1) to 50 ms (absorption coefficient 0,1)!!!  So the latter value could indeed start to generate audible reflected sounds ![/size][/size]
 
[size=10pt][size=10pt]Here is a CSD plot for the K701, known for its cavernous soundstage:[/size][/size]
 
[size=10pt][size=10pt][/size][/size]
 
[size=10pt][size=10pt]And here is a Grado SR225, known for their lack of soundstage:[/size][/size]
 
[size=10pt][size=10pt][/size][/size]
 
[size=10pt][size=10pt]I don't know, but there seem to be a pattern indeed to me.[/size][/size]

You can't really compare these two CSD, the Grado one is some 10 dB softer that the AKG one, all the jaggies that exist between 200 and 2000 Hz on the AKG CSD are basically hidden by the graph's floor on the Grado CSD.
Cut the lower 20 dB on the AKG CSD, and both graphs will be closer in scale.
 
 

 

[shamu144]

Good observation.
 
This is anyway just speculation and hypothesis, but seem fun to play with. I would be curious though to see those CSD adjusted for -60dB.

 

[arnaud]

shamu: the sabine equation is strictly only valid for "large" rooms where you have a large number of reflections (or modes, whichever way you want to see it). In the case of headphones, we're talking about 10 acoustic resonances or so, maximum, in the 2kHz - 20kHz frequency range, far and few in between. Even then, I guess you could use the expression to approximate the reverberation time for such acoustic resonances, but somehow I see the 50ms as on the unreasonable side. The worst I've seen was grado's (rightfully see because of their intentional use of resonating open chambers with no damping) and it's maybe more like 5-10ms at worst for 30dB decay or so.
 
In any case, I fail to link this to some better appreciation of soundstage. As I mentioned before, only way to make such reverberation potentially useful would be to have a very large number of reflections / large overlap of acoustic resonances with all some target reverberation time. This is never gonna happen in a headphone where the first acoustic resonance comes above 2kHz, and the next one is maybe 2kHz away...
 
Another point is that an open headphones with strong emphasis toward providing essentially a direct sound field to the ear like the HD800 is notorious for providing a deep / wide soundstage (I found it so no precise compared to Omega 2 / sr009 though). So where does your theory fit then exactly then? My (little) understanding of soundstage experience through headphone is that it's largely related to the tonal balance.
 
Finally, while a typical listener position, in a standard room is dominated by reverberant field or at least a clear mix of direct and reverberant field, the headphone response at ear location is a bit different physics. You're in the acoustic nearfield of the driver and there's no clear separation between direct and reverberated field, except canceling sound waves show up as peaks and notches in the response curve at a given measurement location.
 
 

 

[shamu144]

Yes, a wide and deep soundstage, but to my ears, at the expense of harmonic tones presence and continuity (unnatural).
 
Decay time doesn't have to be so long, as 50ms delay is enough to generate a distinct echo I believe. My "theory" (let's call this an hipothetical conjecture please) is that adding some controlled reverberation through the earcup acustical properties to the direct sound of a headphone driver could recreate the illusion of a bigger soundfield (cathedral instead of church) or modulate the perception of speech and tones, making it sound less dry (HD800 again) than when heard without any reverberation.
 
But I am definitely no acoustician engineer of course and haven't got a clue (only guesses) at this art of designing headphones !
 
 
Quote:

 
Another point is that an open headphones with strong emphasis toward providing essentially a direct sound field to the ear like the HD800 is notorious for providing a deep / wide soundstage (I found it so no precise compared to Omega 2 / sr009 though).
 

 

 

[Argo Duck]

Interesting speculations none the less. You sound like a scientist! You have to have about 9 wrong ideas to get one good one 

 

[purrin]

Quote:

 
Decay time doesn't have to be so long, as 50ms delay is enough to generate a distinct echo I believe. My "theory" (let's call this an hipothetical conjecture please) is that adding some controlled reverberation through the earcup acustical properties to the direct sound of a headphone driver could recreate the illusion of a bigger soundfield (cathedral instead of church) or modulate the perception of speech and tones, making it sound less dry (HD800 again) than when heard without any reverberation.

 
IMO, that's the real secret of the Sony R10 - the cups.

 

[shamu144]

There are probably many auditory cues that our brain use for soundstage and distance appreciation. The tonal balance you mentionned could be one of those (with the distance, high frequencies are attenuated more than lower frequencies resulting in a rolled off tonal balance for larger rooms).
 
Our brain matches our own psychoacoustic existing maps (that we learn through the course of our lives) to the incoming sounds. I have the feeling that an appropriate decay time and proper balance across the different frequencies between those early reflections from the earcups (less than 30ms), late reflections (reverberation >30ms) and direct sound could also "trick" our brain into believing that this little cavity between our ears and the HP drivers is a lot bigger than it appears.
 
Quote:

In any case, I fail to link this to some better appreciation of soundstage. As I mentioned before, only way to make such reverberation potentially useful would be to have a very large number of reflections / large overlap of acoustic resonances with all some target reverberation time. This is never gonna happen in a headphone where the first acoustic resonance comes above 2kHz, and the next one is maybe 2kHz away...

 

 

[purrin]

Quote:

There are probably many auditory cues that our brain use for soundstage and distance appreciation. The tonal balance you mentionned could be one of those (with the distance, high frequencies are attenuated more than lower frequencies resulting in a rolled off tonal balance for larger rooms).
 
Our brain matches our own psychoacoustic existing maps (that we learn through the course of our lives) to the incoming sounds. I have the feeling that an appropriate decay time and proper balance across the different frequencies between those early reflections from the earcups (less than 30ms), late reflections (reverberation >30ms) and direct sound could also "trick" our brain into believing that this little cavity between our ears and the HP drivers is a lot bigger than it appears.

 
That's why I mentioned R10. I have the W5000ANV which is kind of like an R10 lite. Both these headphones utilize the space in the cups to provide a "reverb" effect.

 

[purrin]

LCD3 Measurements - Then and Now (2+ Months Later):

 
Left Channel

 
Right Channel

 

• Black line = taken on 1/24/2012. Green/Red line = taken on November 2011.
• The 12k null is a measurement artifact. It should be ignored. This was confirmed by listening to sine sweeps covering that area (not fun.)
• The graphs above are raw uncompensated data. They should not be compared to anything else.
• The earpad material maintains a remarkably consistent shape. Differences in earpad placement and compression can result in significant measurement differences.

 

[Anaxilus]

Needs more burn-in.  Perhaps a bonfire?

 

[arnaud]

Two things:
> amazing consistency in your tests, looks like you're at the same quality level as professional lab!
> amazing consistency in the pad behaviors 
 
I guess then that not all lcd3s are created equal and better understand the need for audeze to ship 100dB scale smoothed curves.
 
Neat work as usual marv,
 
Arnaud
 

 

[purrin]

Quote:

Two things:
> amazing consistency in your tests, looks like you're at the same quality level as professional lab!
> amazing consistency in the pad behaviors 
 
I guess then that not all lcd3s are created equal and better understand the need for audeze to ship 100dB scale smoothed curves.
 
Neat work as usual marv,
 
Arnaud

 
Thx. It's not me, it's the CLIO measurement system. CLIO does have calibrate button which I do press on occasion. If figured it was worthwhile to post the LCD3's consistency before and after it went on the statewide tour being molested by at least a dozen Head-Fi'ers. Actually, when I got it back, I thought it was broken because it seemed to lack air. I had to measure it again to make sure it wasn't in my head.
 

 

[wink]

I'm deflated.........