Head-Fi.org › Forums › Equipment Forums › Sound Science › Would lack of amp power causes degradation to sound quality for high impedance headphones?
New Posts  All Forums:Forum Nav:

Would lack of amp power causes degradation to sound quality for high impedance headphones?

post #1 of 81
Thread Starter 

I'm opening this thread because recently, I had a debate with stv014 over at Full-Size Headphones and we both had quite a heated exchange.

 

I'll apologize again to stv014 for my attitude, and I hope we can have a more civil discussion this time.

 

I'll make the intention of this thread clear and simple:

 

Would lack of power for an amp causes lack of bass, lack of soundstage, congested midrange, among... other audiophile claims regarding sound quality degradation?

 

So... is there any technical truth to that? Or is it all just placebo?

 

From my experience, I'd say "yes" without a doubt, and the way I look at it, it's both technical and placebo.

 

Now, my backup for this statement is simply this:

 

For an amp that lacks power to drive high impedance headphones, a lack of volume would be notable. Attempting to increase the volume past a certain threshold will clip the amp and causes distortions. That's the technical part. Also technical is the fact that some amps have undue colorations or distortions even at low volume either intentional (bass boosting) or unintentional (ground loop issues, EMI issues, impedance mismatch, etc...), and I think that plays into this myth as well.

 

Now... solely lack of volume, in my opinions, can cause this lack of bass, or lack of soundstage, because different people react to volume differently, and I think subjectively that even a slight volume difference (approx. 1dB) can cause quite a noticeable difference in perception of the sound. Maybe it is due to this difference in perception that we read about claims of "improved bass", "improved soundstage", etc... And this is the placebo part of my backup. However, I'm not placing too much confidence in the placebo part since many reviewers volume match their different amps before reviewing and still find this "difference", so I'd say it's mostly technical.

post #2 of 81

What precisely do you mean by "lack of power for an amp"?

 

If the amp is clipping, that's definitely not good.  If the volume is not high enough, that's not good either (and affects sound quality; see Fletcher-Munson curves).  That said, for most amps people are talking about and most headphones, there are often no issues with clipping until unrealistically high volume levels.

post #3 of 81

I will only reply in detail tomorrow, but the claim from this post is not mentioned above, although it might be worth discussing, and was an important part of the "heated exchange".

 

Regarding subjective reviews, it is not enough to volume match by ear (that can easily be "off" by more than 1 dB, and tends to be biased towards equalizing the position of the volume knobs, rather than the actual loudness), the levels need to be matched with measurements (e.g. using a DMM). Additionally, the comparison should preferably be done double-blind, as it has been demonstrated that imaginary differences can be perceived between identical sounds if the listener expects - even subconsciously - to hear a difference. Finally, it should be ensured (preferably with measurements) that both amplifiers are operating within their limits, i.e. not clipping. Of course, if that happens, then the lower powered amplifier can indeed be considered "underpowered", at least for that particular listener and choice of music, but only a comparison at lower volume can prove that it is inherently worse at any listening level.


Edited by stv014 - 10/5/12 at 2:17pm
post #4 of 81
Thread Starter 
Quote:
Originally Posted by mikeaj View Post

What precisely do you mean by "lack of power for an amp"?

 

If the amp is clipping, that's definitely not good.  If the volume is not high enough, that's not good either (and affects sound quality; see Fletcher-Munson curves).  That said, for most amps people are talking about and most headphones, there are often no issues with clipping until unrealistically high volume levels.

 

The "lack of power" mentioned is just a general term. I think the technical term should be "lack of output voltage". But I wrote "lack of power" anyway since that's what most reviewers would say.

 

Quote:
Originally Posted by stv014 View Post

I will only reply in detail tomorrow, but the claim from this post is not mentioned above, although it might be worth discussing, and was an important part of the "heated exchange".

 

Regarding subjective reviews, it is not enough to volume match by ear (that can easily be "off" by more than 1 dB, and tends to be biased towards equalizing the position of the volume knobs, rather than the actual loudness), the levels need to be matched with measurements (e.g. using a DMM). Additionally, the comparison should preferably be done double-blind, as it has been demonstrated that imaginary differences can be perceived between identical sounds if the listener expects to hear a difference. Finally, it should be ensured (preferably with measurements) that both amplifiers are operating within their limits, i.e. not clipping. Of course, if that happens, then the lower powered amplifier can indeed be considered "underpowered", at least for that particular listener and choice of music, but only a comparison at lower volume can prove that it is inherently worse at any listening level.

 

I think that claim also falls into the same category (though not specifically), but I'll include it in this response since I think it's interesting as well.

 

For those who don't want to go back to the linked thread to look at the response, it's basically this:

 

Here's the impedance rating of the DT990:

 

 

Taken from HeadRoom.

 

Basically, I'm assuming that that graph tells us that the DT990 will lack bass if the amp driving it does not provide enough voltage even though all other frequencies will sound "even" or "as intended".

 

And some reviewers match volume with a dB meter or sound pressure leveler, so I'm not sure it's all human error or bias.


Edited by Bill-P - 10/5/12 at 2:25pm
post #5 of 81
Quote:
Originally Posted by Bill-P View Post

The "lack of power" mentioned is just a general term. I think the technical term should be "lack of output voltage". But I wrote "lack of power" anyway since that's what most reviewers would say.

 

Sometimes the limitation is in current, or voltage, or it could even be power too.  That's fine.  I was talking about the operating conditions, what "lack" means.  What's the criteria for declaring that something is lacking?

 

I would say that for a given listener using a given set of headphones, an amp lacks power if and only if it starts to distort significantly before the desired volume is reached.  There can be other problems relating to power; for example, the situation where the volume is not yet high enough even at maximum gain and volume setting—a result of a low gain setting and a low output level from the source—but that's a different issue.  There can also be problems not directly related to power, as already mentioned.

 

Anyway, are you talking about the same thing?

 

 

 


Quote:
Originally Posted by Bill-P View Post

Basically, I'm assuming that that graph tells us that the DT990 will lack bass if the amp driving it does not provide enough voltage even though all other frequencies will sound "even" or "as intended".

 

I don't follow this.  How do you determine whether or not an amp is providing enough voltage?  And what does this have to do with the impedance plot?  For an ideal voltage source (for most amps with lowish output impedance, close enough), if you send 1V rms at 20 Hz out of the amp, the headphones will get 1V rms at 20 Hz.  If you send 1V rms at 100 Hz, it will get 1V rms at 100 Hz.  If you send 1.5V rms at 3853 Hz, it will get 1.5V rms at 3853 Hz.  It doesn't have to do with the load impedance.

post #6 of 81

sorry to bump here out of nowhere (im quite noob in this part of soundscience) is there any chance that less expensive headphones will sound better than more expensive high impedance headphones just because both are unamped?

post #7 of 81

One other thing to take into account is the output impedance, since that will interact with the impedance curve of the headphone and affect the sound signature a bit (more than just the electrical damping, IIRC) - this is one reason higher output impedance OTL tube amplifiers *can* sound quite different with some high ohm cans (but not all). 

post #8 of 81
Quote:
Originally Posted by streetdragon View Post

sorry to bump here out of nowhere (im quite noob in this part of soundscience) is there any chance that less expensive headphones will sound better than more expensive high impedance headphones just because both are unamped?

 

There's plenty of chance for less expensive headphones to sound better than more expensive headphones, no matter the circumstances, but I don't think that's what you're asking.  Anyway, the answer is a "yes", but I don't think the usual rationale given in audiophile circles is the most accurate, and I think that's more likely to be the real question you're trying to ask.

 

First of all, it's lower-impedance and lower-sensitivity headphones that benefit more from a better amplifier.  Let's say you like the HiFiMAN HE-6 more than the Sennheiser HD 650 when given a great amp, but you're trying to assess the sound quality when running both out of a laptop.  The HE-6 may be way too quiet for some music and not enjoyable, never mind any other issues.  But even at proper volumes, some headphones may not work well with some sources.

 

That said, a lot of the audiophile obsession over amplifiers and general "common knowledge" is not particularly on the mark, from the perspective of sound quality.  You normally just want the better headphones, aside from some exceptions.

post #9 of 81
Thread Starter 
Quote:
Originally Posted by mikeaj View Post

 

Sometimes the limitation is in current, or voltage, or it could even be power too.  That's fine.  I was talking about the operating conditions, what "lack" means.  What's the criteria for declaring that something is lacking?

 

I would say that for a given listener using a given set of headphones, an amp lacks power if and only if it starts to distort significantly before the desired volume is reached.  There can be other problems relating to power; for example, the situation where the volume is not yet high enough even at maximum gain and volume setting—a result of a low gain setting and a low output level from the source—but that's a different issue.  There can also be problems not directly related to power, as already mentioned.

 

Anyway, are you talking about the same thing?

 

By "lack of output voltage", I'm referring to the fact that amps have a maximum output voltage that they cannot surpass. Any signal trying to gain past that output voltage simply clips... or distorts.

 

So yes, it's the same thing.

 

Quote:
I don't follow this.  How do you determine whether or not an amp is providing enough voltage?  And what does this have to do with the impedance plot?  For an ideal voltage source (for most amps with lowish output impedance, close enough), if you send 1V rms at 20 Hz out of the amp, the headphones will get 1V rms at 20 Hz.  If you send 1V rms at 100 Hz, it will get 1V rms at 100 Hz.  If you send 1.5V rms at 3853 Hz, it will get 1.5V rms at 3853 Hz.  It doesn't have to do with the load impedance.

 

Since volume depends on output power, which is then dependent on the output voltage and the load impedance (headphone impedance), the load impedance plays a role in determining how loud the output is to the listener. At clipping level, output voltage is limited, and so output power will drop if load impedance increases, which results in lower volume for certain frequencies.

 

So I'd say in those cases, the amp isn't providing enough voltage.

 

Edit: also, consequentially, the impedance plot shows a huge hump for mid bass and upper high frequencies, so I'd say that's basically saying that if the amp clips, the DT990 will "lack" bass volume and treble volume.


Edited by Bill-P - 10/5/12 at 4:56pm
post #10 of 81

I think it is important to first have a clear definition of exactly how an "ideal" amplifier should behave. Normally, it is meant to be a voltage source, that is, the output voltage (what is falling on the load) is expected to be equal to the input voltage multiplied by a constant gain value (that determines the volume) under any realistic conditions (i.e. not driving a short circuit, etc.). Do you agree with that definition ?

post #11 of 81
Quote:
Originally Posted by Bill-P View Post

Since volume depends on output power, which is then dependent on the output voltage and the load impedance (headphone impedance), the load impedance plays a role in determining how loud the output is to the listener.

 

It also depends on the efficiency of the transducer. Therefore, an impedance peak at a frequency where the drivers resonate does not necessarily mean that the SPL will be lower from the same voltage if the resonance (not unsurprisingly) also increases the efficiency. While the graph below may suggest that from the same voltage the power is lower at 100 Hz than it is at 1 kHz,

the following frequency response graph, which was created using a voltage source (= same voltage on the drivers at all frequencies from a constant input voltage, regardless of any impedance variations), does not show any drop at 100 Hz, quite the contrary in fact:

The graphs at both HeadRoom and InnerFidelity were created using Tyll Hertsens' equipment, who admittedly uses an amplifier with less than 2 Ω output impedance (see here) for measuring headphones. For a load impedance of 250+ Ω, that is fairly close to an ideal voltage source, and there should be no significant voltage variation at any audio frequency.

 

Of course, the important question is still whether you expect an ideal amplifier to behave as a voltage source, or something else ?


Edited by stv014 - 10/6/12 at 4:33am
post #12 of 81
Quote:
Originally Posted by Bill-P View Post

Since volume depends on output power, which is then dependent on the output voltage and the load impedance (headphone impedance), the load impedance plays a role in determining how loud the output is to the listener. At clipping level, output voltage is limited, and so output power will drop if load impedance increases, which results in lower volume for certain frequencies.

 

Clipping does not just limit voltage, but also significantly increases distortion. In fact, that is what usually becomes audible first, especially the presence of a large amount of high order distortion, and intermodulation between the various frequency components of the signal (e.g. the bass, which is usually present at the largest amplitude, periodically cutting out higher frequencies, resulting in the typical crackling sound of heavy clipping).

 

If we do assume that an ideal amplifier is a voltage source, then a frequency dependent increase in the load impedance is not an issue, unless it causes instability (it should not at low frequency with any reasonable design). Nor is it if the amplifier is expected to be a voltage source with some added output impedance (e.g. the old IEC 120 Ω standard) to provide some sort of partial "power equalization". In fact, it is frequency dependent impedance dips (as in multi-driver passive loudspeakers and IEMs) that tend to cause problems, because of the increased current draw. Basically, a voltage source tries to "force" a voltage on its output (usually regulating it using a strong negative feedback), and the only ways a load can interfere with that are:

- drawing too much current

- a reactive (particularly capacitive) load making the feedback loop unstable by adding excessive phase shift

- forming a voltage divider with the closed loop output impedance

All the above effects are worse with a lower impedance load.

 

With an amplifier that is designed to be a current source, i.e. one that uses feedback to actively regulate its output current, rather than voltage, clipping only at impedance peaks of the load would be a real problem. However, if there are any such amplifiers, they are at best rare, and I have yet to find one.


Edited by stv014 - 10/6/12 at 5:09am
post #13 of 81
Quote:
Originally Posted by Bill-P View Post


Would lack of power for an amp causes lack of bass, lack of soundstage, congested midrange, among... other audiophile claims regarding sound quality degradation?

Bass roll-off is usually

- a result of increasing output impedance with decreasing frequency. Some amps with <1 ohm output impedance at 1 kHz can easily have an output impedance >30 ohms at 20 Hz.

- too high corner frequency chosen for the AC coupling

 

Lack of soundstage, congested midrange: maybe due to high(er) distortion and lack of volume.

post #14 of 81
Thread Starter 

Quote:
Originally Posted by stv014 View Post

 

It also depends on the efficiency of the transducer. Therefore, an impedance peak at a frequency where the drivers resonate does not necessarily mean that the SPL will be lower from the same voltage if the resonance (not unsurprisingly) also increases the efficiency. While the graph below may suggest that from the same voltage the power is lower at 100 Hz than it is at 1 kHz,

the following frequency response graph, which was created using a voltage source (= same voltage on the drivers at all frequencies from a constant input voltage, regardless of any impedance variations), does not show any drop at 100 Hz, quite the contrary in fact:

The graphs at both HeadRoom and InnerFidelity were created using Tyll Hertsens' equipment, who admittedly uses an amplifier with less than 2 Ω output impedance (see here) for measuring headphones. For a load impedance of 250+ Ω, that is fairly close to an ideal voltage source, and there should be no significant voltage variation at any audio frequency.

 

Of course, the important question is still whether you expect an ideal amplifier to behave as a voltage source, or something else ?

 

You are right. I actually overlooked something. It wasn't voltage that I needed to look at. Voltage was only a factor in how easily the amplifier clips/distorts. The real culprit is actually the output impedance in this case:

 

 

 

We can see that an over-damped system will have a significant roll-off even if the amp has enough voltage to push the drivers at those resonant frequencies. Some balance between the amplifier and the headphone is necessary. Going back to the original discussion, I sincerely doubt the Fiio E11 has a dedicated feedback system to bump up output impedance in order to compensate for over-damping. As such, when the headphone impedance increases, damping factor increases, and the DT990 is further over-damped, causing some bass to roll off at any voltage level or listening level even if the amplifier can push the headphone just fine.

 

In which case, it's not "power" that's "lacking". It's more like an impedance matching issue.


Edited by Bill-P - 10/6/12 at 9:03am
post #15 of 81
Quote:
Originally Posted by Bill-P View Post

 

You are right. I actually overlooked something. It wasn't voltage that I needed to look at. Voltage was only a factor in how easily the amplifier clips/distorts. The real culprit is actually the output impedance in this case:

 

 

 

We can see that an over-damped system will have a significant roll-off even if the amp has enough voltage to push the drivers at those resonant frequencies. Some balance between the amplifier and the headphone is necessary. Going back to the original discussion, I sincerely doubt the Fiio E11 has a dedicated feedback system to bump up output impedance in order to compensate for over-damping. As such, when the headphone impedance increases, damping factor increases, and the DT990 is further over-damped, causing some bass to roll off at any voltage level or listening level even if the amplifier can push the headphone just fine.

 

In which case, it's not "power" that's "lacking". It's more like an impedance matching issue.

 

You're looking at a step response. The over-damped line shows high frequency roll-off.

The resonant frequency is usually a low frequency.

 

"Bumping up" the output impedance is the opposite of what feedback is there for. Also, there is no bass roll-off with the DT990, but tight control over the driver. 1V in equals 1V out regardless of frequency with an ideal amp with a gain of 1. With a higher output impedance 1V in might result in 0.5V at 1 kHz and 0.8V at 100 Hz, resulting in a bass "boost".


Edited by xnor - 10/6/12 at 10:18am
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Would lack of amp power causes degradation to sound quality for high impedance headphones?