The Impedance Ballad (image heavy)

[creatip123]

There have been lots of people with the mindset of 'the higher the impedance of a headphone, the harder it is to drive them, or the harder it is to push it louder'. Consequently, this leads to the mindset of 'the higher the impedance of a headphone, the more it needs an amp' 

 

It's been said a lot of times in a lot of topics here. The regulars here always go against this mindset, but of course with words alone, it's kinda like a 3rd grader's argument. Something like, 'afraid not....' 'afraid so....'

 

Well, I thought of a simple and rough experiment, for this matter. For the base theory, let's just go with that mindset for a while, 

 

 the higher the impedance of a headphone, the harder it is to drive it louder, the more it needs an amp

 

 

 

That's the base theory, so let's see if it applies to the reality.

 

For this experiment, I'd need some sort of loudness measuring device. This is what I'm using:

 

 

It's just a regular sound level meter. It will 'listen' to ambient sound, and calculate the average (A weighted) loudness.

 

These are the specs:

 

 

 

Accuracy up to ±1.5dB.
Maximum value holding function to lock up the maximum reading.
Measuring range: 30-130dBA
Accuracy: ±1.5dB
Frequency range: 31.5Hz to 8.5KHz
Frequency weighting:A
Digital display:4digits
Resolution:0.1dB
Sample rate:2 times/second
Microphone: 1/2 inch electret condenser microphone

 

It got 'MAX' function, which is it will hold a max value of measurement.

 

 

It will 'listen' through this microphone. 

 

Now this is an ambient sound level, right? That means, it's quite sensitive, and it will pick up lots of ambient noises. For this, I need a way to give it some kind of sound insulation. 

 

 

This is the insulation. It's just a plain handycraft foam sheets, 12x12cm, 3 layers stacked. Cut a hole in the middle for the microphone to sit in.

 

 

This is how it looks.

 

 

This is how it's implemented. So I covered the pads of the headphones for testing with the foam insulation, and take the measurements. The foam minimizes sounds leaking in and out of the headphone. 

 

 

Yeah, so I got the measurement device ready. Next I'd need the test subjects, right?

 

 

These are all the headphones I got with me, sorted from the smallest impedance on the left, to highest impedance on the right.

 

1. ATH-AD700: 32 ohms (http://eu.audio-technica.com/en/products/product.asp?catID=5&subID=37&prodID=156)

2. Hifiman HE400: 35 ohms (http://www.head-direct.com/Products/?act=detail&id=115)

3. AKG Q701: 62 ohms (http://eu.akg.com/akg-product-detail_eu/q-701.html => click the 'specsheet' link)

4. Krezt DJ-9200 (local brand, China generic): 64 ohms (http://krezt.co.id/v3/?product=krezt-dj-9200)

 

So....it's already sorted out from smallest (32 ohms) to biggest/highest (64 ohms) impedance. According to the base theory, with the same source device, and same volume setting, it will sound the loudest on the 32 ohms one, and least loud on the 64 ohms, right? Well, let's see....

 

Measurement device ready, test subjects ready, next I'd need a source device. I'm not gonna use my PC, ipad, or some random DAP or amps. Instead, I'm gonna use this:

 

 

Good ol' Nokia 2730 classic. It's discontinued. It's just a plain basic phone. No touch screens, no iOS or android, hell it doesn't even have wi-fi capability, and can't play any videos. I bought it for $35 or so. The latest price I found on google is $50 or something. At least it got an MP3 player function, and a 3.5mm audio jack. I'd be plugging the test headphones directly to this phone, set the volume to max, with no amp in-between. 

 

This is the test track I'm using:

 

 

Obviously, the MP3 version, not the video, because of the phone's limitation. 

 

Why do I use a trance song? Why not a sine tone, something like 1kHz sine tone or something? Well, it's because the headphones got different frequency responses, so I thought it'd be better to get an average reading of various frequencies, rather than 1 particular tone. 

 

So it's all set, played the song for 30 seconds from the beginning. With the sound meter's 'MAX' function, it will lock and display the maximum loudness it measured.

 

 

Moment of truth, the results, sorted from number 1 to 4. Remember, same source, same volume, same test track, same sound meter, same method of testing:

 

 

AD700: 91.8 dB

 

 

HE400: 83.1 dB. So far so good, it went according to the base theory. Because HE400 is 35 ohms, so it's less loud than the 32 ohms AD700, right?

 

 

Q701: 83.2 dB. Hmmm that's strange. 35 ohms to 62 ohms impedance is quite a leap. So the Q701 should be considerably less loud than the HE400, right? Then why the hell did it measure basically the same as HE400 (let's waive the 0.1dB, chalk it to difference in frequency response)??

 

 

Krezt: 103.4 dB. Hot damn, I've broken a law of physics here. The highest impedance of all 4 test subjects is actually the loudest!! Yay, the 64 ohms headphone is 11dB louder than the headphone with half its impedance, 32 ohms.....

 

What gives, dude?

 

Seems like the base theory of 'the higher the impedance of a headphone, the harder it is to drive it louder, the more it needs an amp' is on the brink of being debunked. Oh man, the world's coming to an end....

 

Sooooo....if impedance is not the determining factor, then there have to be another determining factor, right? After all, the universe is bound to orderly laws of physics. 

 

So let's look at one other rating of the headphones, which is the sensitivity/efficiency. Something that looks like 'X dB', or 'Y dB/mW', or 'Z dB/V'.

 

1. AD700: 98 dB/mW

2. HE400: 93.5 dB/mW

3. Q701: 105 dB/V => this is not the same as dB/mW. To be able to measure it in the same context, this have to be converted to dB/mW, which results in ~93 dB/mW

4. Krezt DJ-9200: 107 dB/mW

 

Oh wow, now it all make sense!!. Using the order of the sensitivity ratings, the order of the least loud to loudest is: 2 or 3 (about the same ratings and measurements) - 1 - 4. This corresponds with the test results in the pictures. 

 

So what is this efficiency/sensitivity? In a nutshell:

 

When you 'inject' 1 mW or 1 V RMS, the headphone will respond with X dB of loudness. 

 

 

Just by the definition of the efficiency/sensitivity, it's already quite obvious that when we're talking about loudness, or how loud it will go, efficiency/sensitivity is the more important, and most determining factor. 

 

So at this point, I think we could all agree now that 'impedance rating' is NOT the determining factor of how easy or hard it is to drive a headphone to the desired loudness. Here's a fun trivia fact: the hardest to drive headphone in the market, the Hifiman HE-6, is only 50 ohms. Even less impedance than the Krezt DJ-9200. 

 

So the next time you see someone posting something like, 'My headphone is 250 ohms, so it DEFINITELY needs an amp. If only my headphone is 32 ohms, I won't need an amp at all', hope you'd remember this topic and this simple experiment...

[Guest]

omg you are so wrong. Higher impedance is harder to run but there are MUCH MORE factors that affect loudness. You can NOT compare different models and expect good results. Its like you take 2 different models with same impedance and say that they must sound same.

 

Post that does not have much sense.

[Nup]

Very insightful, thanks man!

[Lays]

omg you are so wrong. Higher impedance is harder to run but there are MUCH MORE factors that affect loudness. You can NOT compare different models and expect good results. Its like you take 2 different models with same impedance and say that they must sound same.

 

 

Sensitivity is like 90% of the equation. Honestly impedance doesn't really matter very much, the difference between a 250 ohm DT990 & 600 ohm is around 3 DB of loudness when both headphones receive the same amount of power.

 

 

People like @Dark_wizzie / @ShearMe / @SSL / @h264 / @EmoRarity can come in here and go into scientifics, they're much better at it than I am.

[mr moose]

omg you are so wrong. Higher impedance is harder to run but there are MUCH MORE factors that affect loudness. You can NOT compare different models and expect good results. Its like you take 2 different models with same impedance and say that they must sound same.

 

I think you missed the point of his post.  Give it another read.

[Lutkeveld]

+1000 internetz to you Creatip. Finally a post we can link to when the impedance war starts again.

Maybe we can even get this pinned.

 

omg you are so wrong. Higher impedance is harder to run but there are MUCH MORE factors that affect loudness. You can NOT compare different models and expect good results. Its like you take 2 different models with same impedance and say that they must sound same.

 

I can't really see if you're trolling or not. The whole point was that there are indeed MUCH MORE factors that affect loudness, but people tend to ignore them and and only look at the impedance. As creatip proofed, there is much more to loudness than only impedance.

[SSL]

Great post, thanks for doing this. Nice photography, too.

 

Might benefit from a definition of sensitivity in a large font.

 

omg you are so wrong. Higher impedance is harder to run but there are MUCH MORE factors that affect loudness. You can NOT compare different models and expect good results. Its like you take 2 different models with same impedance and say that they must sound same.

 

Do try explain for all of us, then. I'm dying to see what you come up with.

[creatip123]

Great post, thanks for doing this. Nice photography, too.

 

Might benefit from a definition of sensitivity in a large font.

 

 

Do try explain for all of us, then. I'm dying to see what you come up with.

 

Noted and edited....

[Guest]

@Lays @mr moose @Lutkeveld @SSL @creatip123

 

I'm sorry.

I am not good at english so I wrongly expressed myself. I didn't meant to start any "war".

What I wanted to say is that lower impedance will make it easier to "drive" speaker but that there are other things in final equation.

So it is wrong when people say that impedance does not affect loudness. It does affect a lot but its other things that make same impedance speaker to sound different.

I hope that I made myself clear this time

[creatip123]

+1000 internetz to you Creatip. Finally a post we can link to when the impedance war starts again.

Maybe we can even get this pinned.

 

 

I can't really see if you're trolling or not. The whole point was that there are indeed MUCH MORE factors that affect loudness, but people tend to ignore them and and only look at the impedance. As creatip proofed, there is much more to loudness than only impedance.

 

Too bad can't get myself a square meal with that currency....

 

 

@Lays @mr moose @Lutkeveld @SSL @creatip123

 

I'm sorry.

I am not good at english so I wrongly expressed myself. I didn't meant to start any "war".

What I wanted to say is that lower impedance will make it easier to "drive" speaker but that there are other things in final equation.

So it is wrong when people say that impedance does not affect loudness. It does affect a lot but its other things that make same impedance speaker to sound different.

I hope that I made myself clear this time

 

Impedance determine how high the voltage of an amp has to be to overcome the resistance. The higher the impedance, the higher the voltage has to be. Maybe 10+ years past, there aren't much small/portable devices that have sufficient voltage to overcome high impedance like 300 ohms, but that's not the case nowadays. Thing is, even with impedance value taken into equation, the effect is still rather small.

 

Take HD650 for example. The rating for HD650 is:

 

- Efficiency: 103dB/mW

- Impedance: 300 ohms

 

Using the headphone power calculator, we got:

 

 

You can see the required power, voltage RMS, and current, to drive it to 110dB SPL. Mind you, 110dB is roughly the equivalent of a chainsaw's sound 1 meter from your ears.

 

Now, if I double the impedance value, to 600 ohms (that's 300 ohms difference), it becomes:

 

 

You can see the required power stays the same, because the efficiency value stays the same (103dB/mW). Required voltage got raised to 1.73V (0.5V difference), and consequently required current dropped.

 

Now what happens if I lower the efficiency, just say by 10 dB/mW?

 

 

You can see significant differences/increases in the required factors. Also keep in mind, voltage gain circuits is much simpler and easier to implement, compared to current gain circuits.

 

That's why, the hardest to drive headphone I mentioned earlier, the HE-6 (http://www.head-direct.com/Products/?act=detail&id=92), although it's only 50 ohms, but it got a very low efficiency rating that's a nightmare to 90% of headphone amps out there. Plugged to the calculator:

 

 

You can see, although it's only 50 ohms, only 1/6 of HD650's impedance (first graph), it requires 89.4x more power, and 7.3x more current than HD650 to get the same 110dB loudness. 

[Guest]

Too bad can't get myself a square meal with that currency....

 

 

 

Impedance determine how high the voltage of an amp has to be to overcome the resistance. The higher the impedance, the higher the voltage has to be. Maybe 10+ years past, there aren't much small/portable devices that have sufficient voltage to overcome high impedance like 300 ohms, but that's not the case nowadays. Thing is, even with impedance value taken into equation, the effect is still rather small.

 

Take HD650 for example. The rating for HD650 is:

 

- Efficiency: 103dB/mW

- Impedance: 300 ohms

 

Using the headphone power calculator, we got:

 

 

You can see the required power, voltage RMS, and current, to drive it to 110dB SPL. Mind you, 110dB is roughly the equivalent of a chainsaw's sound 1 meter from your ears.

 

Now, if I double the impedance value, to 600 ohms (that's 300 ohms difference), it becomes:

 

 

You can see the required power stays the same, because the efficiency value stays the same (103dB/mW). Required voltage got raised to 1.73V (0.5V difference), and consequently required current dropped.

 

Now what happens if I lower the efficiency, just say by 10 dB/mW?

 

 

You can see significant differences/increases in the required factors. Also keep in mind, voltage gain circuits is much simpler and easier to implement, compared to current gain circuits.

 

That's why, the hardest to drive headphone I mentioned earlier, the HE-6 (http://www.head-direct.com/Products/?act=detail&id=92), although it's only 50 ohms, but it got a very low efficiency rating that's a nightmare to 90% of headphone amps out there. Plugged to the calculator:

 

 

You can see, although it's only 50 ohms, only 1/6 of HD650's impedance (first graph), it requires 89.4x more power, and 7.3x more current than HD650 to get the same 110dB loudness. 

So we both say same thing

[Lutkeveld]

Ceteris paribus a higher impedance headphone would indeed be more difficult to drive

[Valhallatier]

honestly the only things matter in headphones are, if its comfy or does it sounds good.
if its not loud enough but it sounds good, then im fine.
if its loud but it sounds weird/bad to you, then all of these equations are bullshiz~

[creatip123]

honestly the only things matter in headphones are, if its comfy or does it sounds good.

if its not loud enough but it sounds good, then im fine.

if its loud but it sounds weird/bad to you, then all of these equations are bullshiz~

 

Glad you understand the point of this topic then. All the 'OMG my headphone is 250 ohms, I DEFINITELY needs an external amp', or '50 ohms headphones is fine without an external amp' don't mean a thing. 

 

Get an amp if:

- it's deemed to be needed

- you want to, just because. 

[EndlessOyster]

Thats all fine for you headphones where you were using reasonably low impedance compared to some headphones out there.
Is it possible you could get information about the devices you used to power the headphones so the phone. It be nice to know what the manufacturer specs are.
Is it also possible you would do the test again but this time looking at frequency because if for say you plug a 300 ohm headphone into that nokia turned it up to produce the same SPL as the other headphones your high frequencies would just disappear. Also do be careful if you do do that since you may have some square wave fun.
Also it probably would of been better if you test audio you used was pink noise instead of a song which could be missing frequencies etc.

[LeeWonky]

Thats all fine for you headphones where you were using reasonably low impedance compared to some headphones out there.

Is it possible you could get information about the devices you used to power the headphones so the phone. It be nice to know what the manufacturer specs are.

Is it also possible you would do the test again but this time looking at frequency because if for say you plug a 300 ohm headphone into that nokia turned it up to produce the same SPL as the other headphones your high frequencies would just disappear. Also do be careful if you do do that since you may have some square wave fun.

Also it probably would of been better if you test audio you used was pink noise instead of a song which could be missing frequencies etc.

Please, read this:

 

 

Too bad can't get myself a square meal with that currency....

 

 

 

Impedance determine how high the voltage of an amp has to be to overcome the resistance. The higher the impedance, the higher the voltage has to be. Maybe 10+ years past, there aren't much small/portable devices that have sufficient voltage to overcome high impedance like 300 ohms, but that's not the case nowadays. Thing is, even with impedance value taken into equation, the effect is still rather small.

 

Take HD650 for example. The rating for HD650 is:

 

- Efficiency: 103dB/mW

- Impedance: 300 ohms

 

Using the headphone power calculator, we got:

 

 

You can see the required power, voltage RMS, and current, to drive it to 110dB SPL. Mind you, 110dB is roughly the equivalent of a chainsaw's sound 1 meter from your ears.

 

Now, if I double the impedance value, to 600 ohms (that's 300 ohms difference), it becomes:

 

 

You can see the required power stays the same, because the efficiency value stays the same (103dB/mW). Required voltage got raised to 1.73V (0.5V difference), and consequently required current dropped.

 

Now what happens if I lower the efficiency, just say by 10 dB/mW?

 

 

You can see significant differences/increases in the required factors. Also keep in mind, voltage gain circuits is much simpler and easier to implement, compared to current gain circuits.

 

That's why, the hardest to drive headphone I mentioned earlier, the HE-6 (http://www.head-direct.com/Products/?act=detail&id=92), although it's only 50 ohms, but it got a very low efficiency rating that's a nightmare to 90% of headphone amps out there. Plugged to the calculator:

 

 

You can see, although it's only 50 ohms, only 1/6 of HD650's impedance (first graph), it requires 89.4x more power, and 7.3x more current than HD650 to get the same 110dB loudness. 

[EndlessOyster]

Please, read this:

 

OO thank you that was a enjoyable read, I didnt scroll through the comments because just thought there be loads of pages because its a nice article and i though there be loads.

I still stand behind my comments that he should of used pink noise. I would like to the know the specs of the nokia (im sure its dissapeared to the land of brick phones ). I would also like if you did more about frequencies, because the amp may be able to make a certain headphone pump out 100dB but it may not be producing all frequencies.

[LeeWonky]

OO thank you that was a enjoyable read, I didnt scroll through the comments because just thought there be loads of pages because its a nice article and i though there be loads.

I still stand behind my comments that he should of used pink noise. I would like to the know the specs of the nokia (im sure its dissapeared to the land of brick phones ). I would also like if you did more about frequencies, because the amp may be able to make a certain headphone pump out 100dB but it may not be producing all frequencies.

I wonder if you actually use your headphones.

[EndlessOyster]

I wonder if you actually use your headphones.

I dont understand you point please elaborate, because my comments are perfectly acceptable and frequency response from the test would of made this a amazing article.

[SSL]

I would also like if you did more about frequencies, because the amp may be able to make a certain headphone pump out 100dB but it may not be producing all frequencies.

 

lol no. Again, I tested a 600ohm headphone from the original Fiio E10. Frequency sweeps, single sine tones, square waves, pink noise, actual music. There was NO loss of high frequencies.

[Lutkeveld]

Testing=measuring right?

[SSL]

Testing=measuring right?

 

Okay, I'll bite.

 

No, my "testing" was done via "listening". Maybe you're familiar with the activity?

[EndlessOyster]

lol no. Again, I tested a 600ohm headphone from the original Fiio E10. Frequency sweeps, single sine tones, square waves, pink noise, actual music. There was NO loss of high frequencies.

Surprising, still would love to see a frequency response graph.

[SSL]

Surprising, still would love to see a frequency response graph.

 

It isn't surprising at all. Most headphone amplifiers are voltage sources. That means that with a sufficient damping factor there will be no modification of frequency response under any circumstances, even if the gain on the amp is raised to the point of clipping. You claim that the high frequencies would roll off or be distorted (first) is erroneous and unfounded.

[EndlessOyster]

It isn't surprising at all. Most headphone amplifiers are voltage sources. That means that with a sufficient damping factor there will be no modification of frequency response under any circumstances, even if the gain on the amp is raised to the point of clipping. You claim that the high frequencies would roll off or be distorted (first) is erroneous and unfounded.

But if its clipping on the amp its clipping on the headphones, and the reason why I ask for such a thing is because then I have evidence instead of just your word.