Head-Fi.org › Forums › Equipment Forums › Sound Science › Help with impedance mismatching
New Posts  All Forums:Forum Nav:

Help with impedance mismatching

post #1 of 13
Thread Starter 

Specifically, I'm refering to a mismatch between amplifier output, and headphone resistance. My headphone amp is a SPL Auditor with a 9ohm output impedance, my soon to arrive new headphones are Denon AH-D7000s, which are 25ohm and considered to be sensitive (around 106-108db). I have read a great deal of information here on head-fi, and in other places and the information is mixed as it relates to just how much actual impact the mismatch in my equipment will actually have, meaning that I can and should hear fairly easily. For me if it is measurable, but not audible/perceptible then it isn't very important. I have heard from people who say they can't hear any negative impact from the type of impedance mismatching that I am going to be dealing with, while others suggest that the science guiding the impedance question means the sound will be impacted in such a way that I will hear the effects.

 

I think what is most commonly presented as the impact of impedance mismatching is that the bass response will be less well controlled, and that a higher overall distortion in the sound is likley. My question I hope can be answered outside of the realm of pure measurement, in other words will I for certain hear the negative impact of the mismatching or might it rear its ugly head in such a subtle fashion that I might be able to perceive it, but I would have to be trying hard. From what I have read the mismatch between a 9ohm output and a 25ohm resistance isn't very high all things being considered, and many people in similar situations to mine even suggest they love the combinations! From past experience I had a Valhalla and HF2 combination that I thought sounded quite good.

 

So I want to again ask, am I for certain going to be enduring significantly degraded audio quality, or is it actually a subtle, and perhaps not even perceptible effect? I have read the theory and it still seems somewhat ambiguous as it appears that the individual characteristics of the headphone matter, and that the actual impedance of the headphone varies throughout the frequency range, meaning the worst possible effects of mismatching aren't always going to be as pronounched. Any thoughts would be greatly appreciated. Thanks.

post #2 of 13

The impedance of the D7000 is fairly flat and the resonance low so change in FR will be quite small. The distortion that the driver produces is as far as I can tell a lot higher than what is being added through a low damping factor.

 

I wouldn't worry.

post #3 of 13
Thread Starter 

Thank you for your input, really appreciated.

post #4 of 13
Thread Starter 

I can concur from listening now that the Auditor/D7000 combination sounds fantastic, there isn't even any hiss during quiet passages in the songs, or at least not that I have heard. I have to acknowledge that without another similar quality amp with a low impedance output to test with, that I can't conclude that the 7000 would sound better with a better matched amp, but I doubt that. The combination is fantastic and vibrant, certainly the bass is less controlled than it was with my DT 880 600ohm, but you would expect that due to the open backed design of the 880. While I lose that ultimate control of the driver that the Auditor/880 combination had, I gain quite a bit in tonal richness and fluid musicality that the 7000 has, plus there is quite a bit more bass extension, even if it sounds less tight and focused in places.

 

Not that I couldn't live with the 7000 as is, but I am going to play around with different parametric EQ settings to see if the bass can be controlled a bit more. I'm thinking that adjusting the 80hz band with a -.5 decible cut and a Q notch that is narrow might help. Any thoughts on where the worst enclosure resonance comes from in the 7000?

post #5 of 13

The Dx000 series seems to have a lower resonant frequency than 80 Hz. More like in the 30 - 60 Hz range.

post #6 of 13
Thread Starter 

Thanks, I thought I had read somewhere that the resonance was more like what you are suggesting. I'm not sure if you have any experience with JRiver Media Center, I'm using version 18 and using the built in parametric EQ. My issue is that when using their Q adjustment, I have no guidance as to the scale of the increments, it seems I can put from 0 to quite a large number. I know that the wider the notch, the less the band you are adjusting effects the adjacent bands. I have for instance set the Q width to 12 for the low pass filter, but I have no idea if this is a narrow, or wide notch setting in JRiver? Any thoughts/suggestions are certainly welcome.

post #7 of 13

I'm assuming the filter is a peak used for boosts/cuts and that the Q is not based on -3 dB bandwidth but on "half of the gain" bandwidth.

 

Let's say f0 = 100 Hz, gain = +3 dB and you want +1.5 dB which is half the gain at 100/2=50 and 100*2=200 Hz: Q = 100 / (200-50) = 0.666

 

BW = 100 / 0.666 = 150 Hz = (200 - 50) Hz

post #8 of 13
Thread Starter 
Quote:
Originally Posted by xnor View Post

I'm assuming the filter is a peak used for boosts/cuts and that the Q is not based on -3 dB bandwidth but on "half of the gain" bandwidth.

 

Let's say f0 = 100 Hz, gain = +3 dB and you want +1.5 dB which is half the gain at 100/2=50 and 100*2=200 Hz: Q = 100 / (200-50) = 0.666

 

BW = 100 / 0.666 = 150 Hz = (200 - 50) Hz

 Okay, well I won't pretend that I completely understand what you are showing me here. If it is at all possible could you break down your example with a brief explanation for the pertinent operations? I believe that you are saying in your final answer that if I was attempting to adjust the 100hz frequency (and if the software does operate on the principle you are working from, which I'm sure you are correct about) then I would actually set the parametric EQ to center on the 150hz band? If that is the correct interpretation (hopefully) then I still remain confused about how to actually use the controls in JRiver.

 

So lets say that I want to adjust the 40hz band as it splits the 30-60hz that you suggest is a good target for the D7000 resonance, what would I input into the first control text field which is the target bandwidth field. The next control is the boost or cut amount in db you wish to make (I was entering -1). The last field is the Q where for instance I tried a setting of 12. I did think that I made a minor cleaning of the lower bass when I used 80hz as my target band in the first control field, but I suspect given the relationships you are trying to teach me about with your example that I was missing the mark by quite a wide margin. Thank you for your help. I hope that this might help others who happen into the thread.

post #9 of 13

The center frequency (f0) is what you should be able to configure, in the example it is 100 Hz and it is the frequency where the peak reaches its top.

 

 

The parametric EQ should also let you configure either Q or BW (bandwidth). I described how to calculate those above.

post #10 of 13
Quote:
Originally Posted by Sonic Defender View Post

 

So lets say that I want to adjust the 40hz band as it splits the 30-60hz that you suggest is a good target for the D7000 resonance, what would I input into the first control text field which is the target bandwidth field. The next control is the boost or cut amount in db you wish to make (I was entering -1). The last field is the Q where for instance I tried a setting of 12. I did think that I made a minor cleaning of the lower bass when I used 80hz as my target band in the first control field, but I suspect given the relationships you are trying to teach me about with your example that I was missing the mark by quite a wide margin. Thank you for your help. I hope that this might help others who happen into the thread.

 

I think the first parameter should be called "center frequency" instead of "target bandwidth".

 

40 Hz, -1.0 dB, Q=12 means that the bandwidth (distance between the lower and upper half-gain frequency) is only 3.3 Hz, so the filter is very narrow.

 

The half gain frequencies (-0.5 dB each) are about 38.4 Hz and 41.7 Hz.


Edited by xnor - 5/25/13 at 5:06pm
post #11 of 13
Thread Starter 

Thank you very much, the information is quite helpful. I study psychology and the notion of the half gain frequencies sounds very much like a confidence interval. May I ask how it is you derive the 3.3hz width of a Q =12? Cheers.


Edited by Sonic Defender - 5/25/13 at 6:57pm
post #12 of 13

I never knew exactly what the Q stood for until this thread, but I strongly presume it is just the center frequency divided by the bandwidth.

 

First example:  f0 = 100 Hz, bandwidth = 150 Hz.  Q = 100 / 150 = 0.667.

Second example:  f0 = 40 Hz, bandwidth = 3.333 Hz.  Q = 40 / 3.333 = 12.

 

The higher the Q, the narrower the band of frequencies adjusted, in terms of numbers of half steps or octaves or whatever. i.e. by ratios

 

Here, bandwidth is referring to the frequency range for half-gain (which appears to be some convention; in other contexts, bandwidth is counted at -3 dB or some other value).  You look at the two points where the gain is down to half of the max gain:  +1.5 dB points in the example where +3 dB is the desired max gain at the peak.


Edited by mikeaj - 5/25/13 at 7:30pm
post #13 of 13
Thread Starter 

Thank you both, again most helpful.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Help with impedance mismatching