The No Headphone Amp!

[AudioLapDance]

Good old NwAvGuy:

IS THERE A STANDARD FOR OUTPUT IMPEDANCE? The only standard I’m aware of is IEC 61938 from 1996. It specifies an output impedance of 120 ohms. There are numerous reasons why this is standard is way out of data and a really bad idea. In a Stereophile article about headphones, they said of the 120 ohm standard:

“Whoever wrote that must live in a fantasy world.”

I have to agree with Stereophile. The 120 ohm standard might have been (barely!) tolerable before the iPod and other portable music sources became immensely popular, but it’s not any more. Most headphones are designed very differently today.

PSUEDO STANDARDS: A lot of professional gear has a 20 – 50 ohm headphone output impedance. I’m not aware of any that follows the 120 ohm IEC standard. Consumer gear tends to be in the range of 0 – 20 ohms and, with the exception of tube and certain other esoteric designs, most high-end audiophile headphone sources are well under 2 ohms.

120 ohm produces a 3db bass boost and 1.5dB lift @20kHz on my K301s
120 ohm Rs.png

[counter culture]

It hasn't been unknown for me to find it necessary to explicitly advise the occasional individual as to exactly how they should manipulate their own anatomy. You, however, do appear to betray at least a capacity to read, and a moderate propensity to learn, which is more promising than many.

[burgunder]

Take a look at these thoughts ecp.cc

[xnor]

$200 to 350 for an (active) attenuator makes sense if you can build a nice headphone amp with configurable gain, output impedance for less than 1/4th to 1/2 the price... NOT!

[xnor]

I might add: an attenuator with questionable performance

[PlasticIsGood]

Originally Posted by AudioLapDance

Good old NwAvGuy:

IS THERE A STANDARD FOR OUTPUT IMPEDANCE? The only standard I’m aware of is IEC 61938 from 1996. It specifies an output impedance of 120 ohms. There are numerous reasons why this is standard is way out of data and a really bad idea. In a Stereophile article about headphones, they said of the 120 ohm standard:

“Whoever wrote that must live in a fantasy world.”

I have to agree with Stereophile. The 120 ohm standard might have been (barely!) tolerable before the iPod and other portable music sources became immensely popular, but it’s not any more. Most headphones are designed very differently today.

PSUEDO STANDARDS: A lot of professional gear has a 20 – 50 ohm headphone output impedance. I’m not aware of any that follows the 120 ohm IEC standard. Consumer gear tends to be in the range of 0 – 20 ohms and, with the exception of tube and certain other esoteric designs, most high-end audiophile headphone sources are well under 2 ohms.

120 ohm produces a 3db bass boost and 1.5dB lift @20kHz on my K301s
Attachment 329861

Measured how, please? Since you say there is no common standard to compare with, there is no definition of 0dB, and I could equally say that your low source impedance leads to a midrange lift, or a loss of bass and treble.

Measuring the impedance of my k702 phones, I get a similar curve and, if driven by a source of zero impedance, this implies a power dissipation of a similar shape. Power dissipation is not the same as sound level, however, because efficiency is not constant.

If you actually measured sound pressure level to arrive at the graph, then again there is the problem of no defined 0dB, because unlike speakers, flat sound output is not generally the aim of headphones.

Note also how your "fantasy world" source impedance leads to a curve remarkably similar in shape to an equal loudness contour:
Equal-loudness contour - Wikipedia, the free encyclopedia

Stereophile is written to sell "high end", "audiophile" products to people foolish enough to take magazines seriously.

Incidentally, I note that the k702 has a little blip in its impedance curve at around 2k. This appears, by accident or design, to match the mid-range blip of the equal-loudness contour.

Running from a sample of 80 "consumer gear" headphone outputs which, contrary to what you say, all have around 100 ohm source impedance, the k702 sounds well balanced. From my own headphone amp, with a source impedance of around 0.1 ohms, the sound is much the same except a bit clearer because my amp has less distortion. The commonly-perceived weakness in bass sounds slightly emphasised with the low-impedance source, but hearing seems to adapt quite quickly.

I guess so much consumer gear has that source impedance because it guarantees that the opamp driver is not overloaded by low-impedance phones, and provides a measure of protection for phones and ears when transferring phones from one source to another. Headphones are portable in a way that speakers aren't.

[xnor]

Originally Posted by PlasticIsGood

Measured how, please? Since you say there is no common standard to compare with, there is no definition of 0dB, and I could equally say that your low source impedance leads to a midrange lift, or a loss of bass and treble.

First of all, nwavguy (rocketscientist on this forum) wrote this.

Measured relative to a voltage source (0 ohm output impedance).

No, you couldn't say a low source impedance leads to a midrange lift because with 0 output impedance output = input * gain, and not some fraction of the output that is dependent on the headphone's impedance (see voltage divider).

Measuring the impedance of my k702 phones, I get a similar curve and, if driven by a source of zero impedance, this implies a power dissipation of a similar shape. Power dissipation is not the same as sound level, however, because efficiency is not constant.

No, it's not of a similar shape.
Why is power of any interest? What matters is voltage output.

If you actually measured sound pressure level to arrive at the graph, then again there is the problem of no defined 0dB, because unlike speakers, flat sound output is not generally the aim of headphones.

Note also how your "fantasy world" source impedance leads to a curve remarkably similar in shape to an equal loudness contour:
Equal-loudness contour - Wikipedia, the free encyclopedia

It doesn't matter (see above), but the reference point is usually 1 kHz.
If in the low output impedance case 100 Hz is at +3 dB and in the high output impedance case it is at +6 dB the relative change is +3 dB.

You really don't want to mention the equal loudness contours in conjunction with flat frequency response.

Incidentally, I note that the k702 has a little blip in its impedance curve at around 2k. This appears, by accident or design, to match the mid-range blip of the equal-loudness contour.

That's just painful to read.

Running from a sample of 80 "consumer gear" headphone outputs which, contrary to what you say, all have around 100 ohm source impedance, the k702 sounds well balanced. From my own headphone amp, with a source impedance of around 0.1 ohms, the sound is much the same except a bit clearer because my amp has less distortion. The commonly-perceived weakness in bass sounds slightly emphasised with the low-impedance source, but hearing seems to adapt quite quickly.

What consumer gear are you talking about? Most portable players have < 10 ohm.

[PlasticIsGood]

xnor:
My headphones don't have a voltage output, they have a pressure output that has an apparent sound level related to power. The decibel, including dBV, is always related to power. The only reason for the log is that power is a square. An input voltage graph expressed in dB is silly if it's not related to power output. You should read up on this if you don't know.

It is obvious that source impedance has an effect on voltage input, but that is not directly related to sound level output. If phones were designed to have a flat sound level output with a 120ohm source impedance, then they would only be flat with a lower source impedance if their own impedance were constant, which they never are. In fact there is no standard. A 3dB lift in bass according to one arbitrary notion of correctness can easily be a 3dB drop in midrange and treble according to another equally arbitrary standard.

I'm not going to list the 80 "consumer gear" (nasty elitist term) items I have tested with my phones, obviously. They were mostly CD players from the first to some of the most recent. Also several amps with phone outputs.

It is true that recent ipods have an output impedance of less than 10 ohms. I don't know about other portable devices. Mobile telephones are possibly the most common source these days. It's not easy to find out what their headphone output impedance is without testing them, as it doesn't generally appear in the published specs that I can find. "Most" is a big word that I doubt you can justify. How do you know?

There is no standard for sound pressure output of headphones. That's the main reason why they sound so different from each other. They are not designed to be flat. If speaker output is ideally flat, then headphone output is certainly not ideally flat because the ear perceives headphones differently from speakers. Since it is known flat is not ideal, this begs the question of what is correct.

Essentially, this question has no straight answer. The equal loudness curves clearly have some bearing on the matter, but there appears to be no general agreement about exactly what.

Painful? No likey, no readey. I am happy for you to ignore my posts

[xnor]

Originally Posted by PlasticIsGood

My headphones don't have a voltage output, they have a pressure output that has an apparent sound level related to power.

And power is a function of voltage and resistance (Joule's and Ohm's law, P = V^2/R).
So we can say that the output voltage of a device is related to the SPL of the connected headphones.

The decibel, including dBV, is always related to power. The only reason for the log is that power is a square. An input voltage graph expressed in dB is silly if it's not related to power output. You should read up on this if you don't know.

dBV is defined as voltage re 1 V regardless of impedance. So it has nothing to do with power.
I don't know what the other stuff you're talking about has to do with what I wrote about the power graph not looking like the graph posted before, especially not with low output impedance.

It is obvious that source impedance has an effect on voltage input, but that is not directly related to sound level output.

Yes it is. Double the voltage at frequency X and with an ideal headphone you'll get +6.02 dB (sound pressure) at frequency X.
That is, of course, on top of the headphones' frequency response.

If phones were designed to have a flat sound level output with a 120ohm source impedance, then they would only be flat with a lower source impedance if their own impedance were constant, which they never are. In fact there is no standard. A 3dB lift in bass according to one arbitrary notion of correctness can easily be a 3dB drop in midrange and treble according to another equally arbitrary standard.

Some headphones do have a pretty flat impedance, but let's ignore exceptions.
Yeah, a headphone designed for 120 ohm Zs will most likely have a less peaky bass response and probably less strong upper treble with a ~0 ohm device.
Anyway, I'd argue that such headphones are also exceptions. And even if the "flatter" response bothers you, you can always add a peaking EQ at around 100 Hz and maybe a highshelf to boost treble a bit.

I'm not going to list the 80 "consumer gear" (nasty elitist term) items I have tested with my phones, obviously. They were mostly CD players from the first to some of the most recent. Also several amps with phone outputs.

If you're going to list amps I have to tell you that on such amps with high headphone Zout the headphone jack is usually connected internally using a couple of resistors. Some receivers have as high as 470 ohm output impedance. Those outputs sound pretty bad with most dynamic heapdhones..

As for portable players, (smart)phones etc. all recent and even older devices I've seen have low output impedance.
Many Apple, Samsung, LG, Sony Ericsson, ... device have <10 ohms (some even <1 ohm), with the occasional portable device having ~15 to ~20 ohms.

"Most" is a big word that I doubt you can justify. How do you know?

Look up measurements. There are many on the internets.

There is no standard for sound pressure output of headphones. That's the main reason why they sound so different from each other. They are not designed to be flat. If speaker output is ideally flat, then headphone output is certainly not ideally flat because the ear perceives headphones differently from speakers. Since it is known flat is not ideal, this begs the question of what is correct.

We've had diffuse field equalization for a few centuries now. The reasons headphones sound different are:
- limits to mechanical and electrical equalization
- manufacturer doesn't care
- different target markets (eXtra Bass headphones ...)

Essentially, this question has no straight answer. The equal loudness curves clearly have some bearing on the matter, but there appears to be no general agreement about exactly what.

Equal loudness curves seem completely irrelevant to me, both for speakers or headphones.

[PlasticIsGood]

And power is a function of voltage and resistance (Joule's and Ohm's law, P = V^2/R).
So we can say that the output voltage of a device is related to the SPL of the connected headphones.

Not directly, because efficiency is not constant, so the proportion of power dissipation output as sound is not constant with respect to frequency. If that were not the case, then all headphones would have a flat frequency response with respect to the input, and we would have nothing to discuss.

dBV is defined as voltage re 1 V regardless of impedance. So it has nothing to do with power.

Just before that sentence in Wikipedia's entry on the dB is this one:

"Since the decibel is defined with respect to power, not amplitude, conversions of voltage ratios to decibels must square the amplitude"

Which is where the log comes from in the conversion from V to dBV. The dB is all about power. It is not at all about anything else but power.

I don't know what the other stuff you're talking about has to do with what I wrote about the power graph not looking like the graph posted before, especially not with low output impedance.

What other stuff? Actually don't bother, it'll get too convoluted for this format to handle. Let's drop that bit. When I said it would be a similar shape I just had in mind that they go up and down together. The difference is that one would have a log vertical scale which, for low variance appears similar, just as sine squared looks similar to sine.

Yes it is. Double the voltage at frequency X and with an ideal headphone you'll get +6.02 dB (sound pressure) at frequency X.
That is, of course, on top of the headphones' frequency response.

If headphones were ideal we would have nothing to discuss. I'm talking about real headphones. The discussion is about frequency response not amplitude response. Maintain the same voltage input over a range of frequencies and the output will vary. That's what I meant when I said that voltage input is not directly related to sound output.

If you examine your maths you will also see how the dBV relates to power. Double the voltage doesn't result in double the dBV, notice? Why not? Where does the 6 come from, don't you wonder? Could it be because double the voltage results in four times the power, which is a rise in power of 6dB? How can you say the dBV is not about power when the relationship is so clear? dBV may be regardless of resistance, and therefore regardless of absolute power, but it assumes constant resistance, and is therefore directly proportional to relative power. The dB is essentially about power and nothing else but power.

Some headphones do have a pretty flat impedance, but let's ignore exceptions.

Compared to most speakers, most headphones have a pretty flat impedance response.

Yeah, a headphone designed for 120 ohm Zs will most likely have a less peaky bass response and probably less strong upper treble with a ~0 ohm device.

That sneaks in a judgement that makes your argument circular. Headphones designed with 120 ohm source in mind may not be peaky in the bass output. If they were flat in the bass, then a lower source impedance would render them lacking at whatever frequencies the headphone impedance were higher, or peaky wherever it was lower.

... even if the "flatter" response bothers you, you can always add a peaking EQ at around 100 Hz and maybe a highshelf to boost treble a bit

Rhetorical argument. Flat doesn't bother me. As I have said, my perception soon adjusts to quite a wide variation in response. Headphones isolate from the real world and create their own. We are adaptable animals. If I were to test lots of headphones in succession with the same source, I may be fooled into thinking differences are more significant than they actually are under normal circumstances. Anyone who chooses headphones by such a listening test is under an illusion. Much better to decide on clarity and reputation...but I digress.

For similar reasons I have no need for equalisation, and prefer simplicity.

If you're going to list amps I have to tell you that on such amps with high headphone Zout the headphone jack is usually connected internally using a couple of resistors.

Of course. As I said, added series resistance allows a normal opamp to drive any headphones without fear of overload. For headphones with 120 ohms source in mind, this gives cheap, bomb-proof high fidelity. The recent fad for low source impedance headphone amps rather fouls the plot.

We've had diffuse field equalization for a few centuries now. The reasons headphones sound different are:
- limits to mechanical and electrical equalization
- manufacturer doesn't care
- different target markets (eXtra Bass headphones ...)

This is an overly cynical view of how markets work I think. Headphones are not flat because flat doesn't sound flat. Since nobody really knows what flat is with respect to headphones, the market chooses whatever sounds best. One thing clear in my mind is that if you are hearing different from what most listeners hear, it ain't hi-fi. Elitism destroys fidelity like nothing else. Consequently Little Richard sounds daft on anything other than a Dansette. You may not agree. My handle isn't a joke, I mean it.

Equal loudness curves seem completely irrelevant to me, both for speakers or headphones.

Certainly irrelevant for speakers, but not necessarily for headphones. I'll try to explain the commonly accepted view even though I don't necessarily agree with the common conception of fidelity:

The aim of speakers is to produce sound in the free air around you that is the same as the sound in the free air wherever it was recorded. It is reasonable to assume that ears in that free air around you would then hear the same as they would in the original free air, give or take all the stuff we know about the failings of speakers and rooms. It's generally and quite fairly accepted that that's the best that can be done, and therefore speakers should have a flat frequency response. No place for the "equal loudness" curves in that argument.

With headphones, you are no longer in free air, so much of what happens when your ears retrieve sound from free air in a room is bypassed. Now, a flat sound in free air is not flat in your ears, and it is that not-flat that headphones need to recreate if they are to approximate to that sound you would have heard in the original space where the sound was recorded. So we know flat is definitely not what we want for the generally-accepted view of hi-fi. What not-flatness do we need, exactly? Nobody knows, exactly, but we do know about the equal-loudness curves, so perhaps we should adopt them instead of flat. But external headphones do have a bit of outside-ear space, so maybe we should aim for something in between flat and equal loudness.

That appears to be the gist of the argument. A similar argument applies to the frequency-dependent cross-talk that happens in free air but not in headphones, but that's even more difficult to deal with so most headphones and amps ignore it, leading to a sound that, at best, is mostly pasted to the inside front of your head.

Considering the result is a confection, you may as well just forget hi-fi and go with whatever sounds best.

Better, in my view, is to follow the market...not the advertising hype, but rather what people actually buy and listen with. Considering music is essentially a social enterprise rather than a private phenomenon, I believe high fidelity must be socially defined. That's why snooty elitism can't work...you can't have higher fidelity than most people because most people define what high fidelity is.

There may be more chance of my argument finding favour in China, but we don't do politics here.