[Bizzlebin]

I've got a few more questions now:
 
Quote:

Additionally, they used all-pass filters that cause huge phase shifts. Not something that would happen in a single driver headphone.

 
Do you mean that all all-pass filters cause large phase shifts, but only in multi-driver headphones? What about an L-R filter, which I've read behaves like an all-pass?
 
Quote:

The only headphone (with measurements available) that I think could have problems with phase is the DT 48 E, but the magnitude response is a much, much greater problem and there's no question about the audibility.

 
I looked at the phase response vs impedance on InnerFidelity for the 120 Ohm DT 48 E (http://www.innerfidelity.com/images/BeyerdynamicDT48E120Ohm.pdf) and it looks about as flat as I've ever seen on a headphone graph. Am I understanding the graph incorrectly?
 
Quote:

Isn't IMD related to THD? I mean, isn't there a way to estimate (if not directly measure) the IMD based on a THD graph? If not, how does one measure it and what are the sonic clues that we can use to pinpoint it audibly?

 
Finally, I don't think my question a while back was addressed; it got lost at the end of page 3!
 
-Bizzlebin

 

[ultrabike]

My answers in bold:
 
Quote:

I've got a few more questions now:
 
 
Do you mean that all all-pass filters cause large phase shifts, but only in multi-driver headphones? What about an L-R filter, which I've read behaves like an all-pass?
 
A headphone can be visualized as a filter, and it is possible that single driver headphones will not induce large phase shifts to their input signal. In general though, external all pass filters can cause large phase shifts on single and multi-driver headphones. The amount and details of the phase shift are part of the design of the filter.
 
I looked at the phase response vs impedance on InnerFidelity for the 120 Ohm DT 48 E (http://www.innerfidelity.com/images/BeyerdynamicDT48E120Ohm.pdf) and it looks about as flat as I've ever seen on a headphone graph. Am I understanding the graph incorrectly?
 
That (second row first column) plot is electrical impedance frequency response phase and magnitude, not acoustical. The acoustical phase response is missing and is likely not flat. Only the acoustical magnitude response plot is available (first row first column)
 
 
Quote:

Isn't IMD related to THD? I mean, isn't there a way to estimate (if not directly measure) the IMD based on a THD graph? If not, how does one measure it and what are the sonic clues that we can use to pinpoint it audibly?

 
Finally, I don't think my question a while back was addressed; it got lost at the end of page 3!
 
THD measures non-linear harmonic distortion due to a single tone. IMD measures non-linear harmonic distortion due to the inter-modulation of two tones. They don't look at the exact same harmonics (tones).
 
-Bizzlebin

 
Hope this helps 

 

[streetdragon]

Eq cannot remove the biting trebles of a bad or cheap headphone, or increase the bass to match Sony's XB series without distorting at high volumes, but they certainly can change the overall freq graph to match more expensive sounding headphones.

 

[Vkamicht]

Quote:

Eq cannot remove the biting trebles of a bad or cheap headphone,

 
Why not? What is "biting treble"? EQ can always -remove- something, it's the attempt to add to a driver's response that isn't there to begin with, where you encounter problems. Bass is the typical one there. Anyway, when you say biting treble I'm imagining a spike in the response that is difficult if not impossible to remove using graphical EQ with predefined cycles and bandwidth, but a parametric can take care of it no problem.
 
For me EQ is the difference between "wow, that sounds great", and "wow, that sounds real."

 

[streetdragon]

well... yeah, you're right. my mistake. guess i was refering noisy response from the driver (those ringing waveforms from the 500hz test) that could cause indistinct and harsh treble when tried to make brighter and sharper
how is the parematric eq better than the graphical eq too? is it better in terms of flexability or output sound quality?
i only have the 10 band graphic eq to play with on my phone and com

 

[firev1]

Parametric eqs can be made to really tailor fit the response curves you want to modify. For example, in one filter, I can have a 1khz wide reduction in frequency or a very surgical cut. Tl;dr it is more precise and flexible.

 

[ultrabike]

Quote:

 
Why not? What is "biting treble"? EQ can always -remove- something, it's the attempt to add to a driver's response that isn't there to begin with, where you encounter problems. Bass is the typical one there. Anyway, when you say biting treble I'm imagining a spike in the response that is difficult if not impossible to remove using graphical EQ with predefined cycles and bandwidth, but a parametric can take care of it no problem.
 
For me EQ is the difference between "wow, that sounds great", and "wow, that sounds real."

 
I used Audacity's parametric and graphic equalizers on some songs to compensate for my DT990-250 (before I sold it) and was pleased with the results. Wished good portable equalizers like those ones existed to equalize songs real time (no agenda here, just really looking forward for one on a portable DAC/Amp). Problems with many graphic equalizers might be rooted to some of them being poorly implemented. I also agree that narrow spikes might be difficult on a well designed graphical EQ depending on the details (my best guess is that a 5-band will have a difficult of a time). 

 

[ultrabike]

Quote:

well... yeah, you're right. my mistake. guess i was refering noisy response from the driver (those ringing waveforms from the 500hz test) that could cause indistinct and harsh treble when tried to make brighter and sharper
how is the parematric eq better than the graphical eq too? is it better in terms of flexability or output sound quality?
i only have the 10 band graphic eq to play with on my phone and com

 
I'm inclined to believe that an appropriate, well designed and well used EQs can counter some driver ringing issues if those issues linear (described by the impulse response and displayed in the CSD plots). Also, some people with more experience on phone and portable players might chime in, but I can say that in my experience, my Rockboxed Sansa Zip 5-band graphical EQ, and the 10-band EQ that comes with VLC are absolutely no match to Audacity set of EQ.
 
Quote:

Parametric eqs can be made to really tailor fit the response curves you want to modify. For example, in one filter, I can have a 1khz wide reduction in frequency or a very surgical cut. Tl;dr it is more precise and flexible.

 
True. I like those a lot personally. A 31-band graphical might still do a very decent job though.

 

[bigshot]

well... yeah, you're right. my mistake. guess i was refering noisy response from the driver (those ringing waveforms from the 500hz test) that could cause indistinct and harsh treble when tried to make brighter and sharper

That is likely due to clipping caused by boosting aspecific frequency instead of pulling back all the other frequencies.

Parametric equalizers are better for overall curves. Graphic equalizers are better for fixing a lot of little bumps and dips.

 

[streetdragon]

Quote:

 
I'm inclined to believe that an appropriate, well designed and well used EQs can counter some driver ringing issues if those issues linear (described by the impulse response and displayed in the CSD plots). Also, some people with more experience on phone and portable players might chime in, but I can say that in my experience, my Rockboxed Sansa Zip 5-band graphical EQ, and the 10-band EQ that comes with VLC are absolutely no match to Audacity set of EQ.
 

guess for me that's why soundstage appears to improve significantly after tuning the headphones? (tuned to mostly electronic music like trance and house and DnB , but seemed to improve with other genres too like rock and pop) or is it something else?

 

[ultrabike]

Probably. The type of EQs discussed so far (such as parametric and graphical) enable us to compensate for rig issues, or to custom fine tune our setups to our liking. However, other types of effects/EQs may affect soundstage perception quite a bit, like crossfeed (such as the passive one in my Headroom Total Bithead) and Dolby Headphone (among others.) I use crossfeed often and like it (doesn't mean everyone has to like it though.) BTW, the Sansa Zip Rockbox FW offers some nice control over the crossfeed effect.
 
Furthermore, if crossfeed (and other) effects are properly captured during the recording such as in the case of binaural recordings... man talk about soundstage!

 

[streetdragon]

speaking of the binaural recordings.... that person in the left/right channel ID test demo, i almost took off my headphones because his voice right next to my ear is way too creepy >_<"

 

[Norman314]

Some of my earlier comments on square waves might be misguided -- check out this link http://www.innerfidelity.com/content/presto-change-o-measurement-transformations-page-2
Apparently, square waves response can be predicted pretty accurately just based on frequency response. I had thought it also incorporated other information. 
 
So maybe the question we should all be asking is, what are the best equalizers to be buying, and how do you go about equalizing effectively? Are there already good threads on those topics somewhere?

 

[bigshot]

That's a bingo.

I keep returning to this, and I can't find many people who are addressing the issue of equalization.

In my opinion, the ideal is to have both a parametric AND a graphic equalizer. A parametric is great at achieving the overall response curve, but it desn't generally have enough bands to cancel out all of the small bumps and dips. You need a 1/4 octave graphic equalizer for that. Digital EQs don't drift as much as analogue ones.

The best way to test for flatness is a signal generator that you can dial up and down quickly. That is the easiest way to hear bumps in the response. You take an octave at a time and balance it starting at the bottom and working your way up. When you get to the end, boost the volume a bit and do another refining pass. Repeat this process until it's as loud as you can stand.

 

[streetdragon]

IMO EQ can really improve listening experience, so much so that after tuning my pair of HD202 for about 1 year (small tweaks and improvements made over time) it placed it quite close to the regular untuned M-50 in listening experience. just the treble in HD202 was more harsh and less crisp. talk about free performance! (of course it applies differently to preferences too)