You are here

Log in or register to post comments
bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
Flaw with Tyll's headphone measurements

I just listened to two headphones, one on the left ear and the other on the right. Both were supposed to be outputting 4,000 hz, but one headphone sounded brighter than the other one. I can see how two headphones would have different degrees of brightness when playing a song, but that should not be the case when they are playing the exact same frequency. You can try this test with any tone generator you like, even a crude one like this http://plasticity.szynalski.com/tone-generator.htm

Tyll's headphone measurements assume that a headphone that claims to be putting out x hz is actually creating that sound, and not some other frequency. In reality, they're not. They're creating a frequency other than the one they're supposed to be putting out. Some headphones move it up, some move it down. What Tyll needs is a new chart that shows how much the sound the headphone actually puts out differs from the sound it claims to be putting out.

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
This is to some extent already done

Usually a headphone would put out a 4000 Hz tone when supposed to. However, it could produce other frequencies along with the 4000 Hz tone. This is called non-linear distortion and it is shown in Tyll's to some extent in the "%THD+noise @ 90dB and 100dB" plot.

If non-linear distortion dominates the fundamental (in this case 4000 Hz), then I would say that headphone (or headphone design) is somewhat flawed.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
I think it's worse than the

I think it's much worse than the tiny numbers shown in the distortion plots. I did the aforementioned test with two very high end IEMs.

On a basic level, does Tyll even have an instrument that can measure the actual hz of the sound being put out by a headphone? I'm not talking about the hz the headphone should be putting out, I mean the hz it is actually putting out.

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
Yes

The frequency response plots are actual AFAIK.

Here is Tyll's Frequency Response Headphone Measurement Procedure:
http://www.innerfidelity.com/content/headphone-measurement-proceedures-f...

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
No I think they measure the

No I think those charts measure the db being put out by the headphone when a certain frequency is input into the headphone. They don't measure whether the headphone is actually outputting that frequency, or if it's outputting another frequency. They just measure the db being put out by the headphone.

I'm not sure if Tyll even has equipment to measure the frequency actually being output by a headphone.

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
Frequency and Sound Pressure

What you are seeing in those plots is sound pressure level (SPL) in dB (power) at particular frequencies. In other words, you are seeing how much power there is for all audible frequencies.

Frequency content is the x-axis (Hz), how much there is of it is in the y-axis (dB).

I believe the equipment that Tyll is using is in Audio Spectrum Analyzer mode. Spectrum Analyzers are pretty standard in multiple applications.

What you are describing as missing in Tyll's plots is IMO pretty clearly there in his characterization sheets. I think you should read a little more about Spectrum Analyzers.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
I'll try one last time, as

I'll try one last time, as simply as possible.

Assume Headphone 1 is such that when given a 4,000 hz source. It outputs 3,500 hz frequency sound.

Assume Headphone 2 is such that when given a 4,000 hz source. It outputs 4,500 hz frequency sound.

Tyll inputs a 4,000 hz source into Headphone 1 and his equipment says 90 db are being produced (his equipment doesn't know that it's 90 db of 3,500 hz sound. All it knows is that it's 90 db).

Tyll inputs a 4,000 hz source into Headphone 2 and his equipment also says that 90db are being produced.

Thus Tyll's equipment falsely concludes that they are identical at reproducing 4,000 hz. They're not, because Tyll's equipment is only measuring the frequency going into the headphone, and the db of pressure being output by the headphone. His equipment does not measure the frequency of sound actually being put out by the headphone.

Again, some headphones are brighter than others when reproducing the exact same frequency, from the exact same source. Tyll's measurements are flawed because they don't capture this.

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
I think I understand.

I think I understand what your are saying, and indeed Tyll's plots don't provide non-linear behavior per excitation frequency. They produce a lumped sum of all frequency powers that do not correspond to the fundamental excitation in the form of THD%+noise plots.

However, what you are describing in your particular example I have never seen. Particularly at 4,000 hz. Not even earbuds.

If you are hearing 3,500 hz or 4,500 hz when providing 4,000 hz to your headphones, then something is very wrong. Headphones usually produce a dominant fundamental which would be 4,000 hz in your example and other much lower power harmonics at multiples of the fundamental, none of which correspond to 3,500 or 4,500 hz.

This is leading me to believe that you would benefit from learning what a Frequency Response plot is and what a THD plot is. Also, you are likely going to benefit from understanding what non-linear behavior is and how it applies to headphones. Otherwise you are not going to understand what Tyll's measurements provide and call the flawed erroneously.

I'll give you an example as to what to look for. Consider the HE-400:

http://www.innerfidelity.com/images/HiFiMANHE400.pdf

"%THD+noise @ 90dB and 100dB" shows that this headphone exhibits quite a bit of power (relative to other headphones) in frequencies that do not correspond to the excitation frequencies. In particular, this headphone will generate over 1% of power not corresponding to the excitation in frequencies in the 200 hz to 1,000 hz ranges. It will not tell you what frequencies generated the wrong frequencies in the 200 to 1,000 hz region, but more than likely these offending frequencies are south of 500 hz me thinks.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
I don't know why you keep

I don't know why you keep responding to this very simple and fixable flaw with distractions and irrelevant comments.

The point is that when tyll tells you a certain headphone puts out a certain decibel at a certain hz, that's not true because his system can't measure what hz the headphone is actually putting out. His system only knows what hz he fed into it. but your ears can do what his equipment can't -- tell you the hz coming out of the headphone.

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
There is one flaw...

... but that flaw is not in Tyll's equipment and work. It is your lack of understanding of what you are looking at. You can't really judge something that you obviously don't understand.

I would say to stop looking at the characterization sheets, since apparently you can't make heads or tails about them, and use your ears. But then you are saying that you keep hearing 3,500 and 4,500 hz when you apply 4,000 hz. If your cans are not broken I suggest get yourself checked. What you are hearing is not normal.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
No, Tyll's equipment does not

No, Tyll's equipment does not "work."

Let me give you a paint by numbers example. I can't make it any simpler that this so if you still do not understand please move along to another thread.

Imagine three sounds, one playing at 3,500 hz, one playing at 4,000 hz, and one playing at 4,500 hz. They sound different, right? In addition, it doesn't have to be those three frequencies. It can be 3,900 hz, 4,000 hz, and 4,100 hz, whatever. Sounds of different frequencies sound different, right? Sounds of different volumes sound different, and sounds of different frequencies sound different. They're two separate dimensions. In other words, if I play 3,900 at 80 volume, 90 volume, and 100 volume, those three sounds will be distinguishable. But if I play 3,900 hz, 4,000 hz and 4,100 hz at 80 volume each, those three will also be distinguishable. Again, the frequency of a sound is not the same as it's volume. They're two separate dimensions.

Now imagine two headphones. Imagine that using the same dac/amp and tone generator (and assume this source works properly) you feed them a 4,000 hz sound. In other words, the dac/amp has a headphone out, and you plug the headphone into that headphone out, and you tell the source to play a 4,000 hz sound. Perhaps you told the source to play a wav file. Perhaps you used some other technique. But the source is playing 4,000 hz.

Is it not correct that when each headphone plays this sound, they're only outputting one frequency? Either they're putting out 4,000 hz, 4,001 hz, 3,900 hz etc. whatever. But it's only one frequency. A trained ear or system can figure out which frequency the headphone is putting out, but it's just one frequency.

So the question is -- is the headphone putting out the frequency it's supposed to be putting out?! The answer is no, even the best headphones vary dramatically on this point and Tyll's system isn't catching this error.

So if a headphone set up as above puts out 3,500 hz, Tyll's system won't notice that it's making a 3,500 hz sound when it's supposed to be making a 4,000 hz sound. Tyll's system has no way of catching this error. His system doesn't even attempt to measure the frequency actually being put out by the headphone. All Tyll's system does is measure the db being put out by the headphone when a particular frequency is fed into it.

Tyll should update his system so that he measures the difference between the frequency the headphone is supposed to be putting out, and the frequency it actually is putting out.

This is very easy to do. You don't even need to do it for the entire frequency range. just pick say 10 key frequencies, 50 hz, 100 hz, 250 hz, 500 hz, 1000 hz, 2,000 hz, 4,000 hz, 6,000 hz, 8,000 hz, 10,000 hz and 12,000 hz.

Measure the frequency being put out by the headphone when each of these ten frequencies is fed into the headphone, and plot the errors. You'll see they're quite bad for most headphones, even those that claim to be "neutral."

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
Incorrect in many ways.

Dude, you are very confused. If a headphone is told to put 4,000 hz, it will damn well put 4,000 hz along with some other frequencies at the same time but at a much lower level. In most cases less than 1% or 0.1% of the power will not be 4,000 hz.

Headphones don't put out some random frequency when given 4,000 hz as you suppose. They do mainly put 4,000 hz.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
By your logic we shouldn't

By your logic we shouldn't test to see if a headphone is actually putting out 4,000 hz -- the frequency it claims to be doing so -- because in reality it's putting out millions of frequencies all blended together into some unknown unmeasurable frequency.

So I guess we also shouldn't bother to see if an opera singer can actually sing a particular note before hiring them. Rather, we should just hire whoever claims to be singing a particular note without using our ears to confirm it, because the voice contains millions of frequencies and it's impossible to see if they can actually hit the note they claim to be hitting . . . what the **** are you talking about?

4,000 hz has a sound, and it sounds different from 4,001 hz, and from 4,500 hz. Either the headphone is creating a 4,000 hz sound when it's supposed to, or it's not. Tyll's equipment and his measurements are flawed because he doesn't test that. His equipment assumes that a headphone that is supposed to be putting out 4,000 hz is actually putting out that frequency, and he only measures the db of the sound being put out. He should be measuring both the db being put out, and whether it's actually putting out that frequency or some other frequency.

Thank you for interrupting the thread with your inane spam, which I assume will be followed by another unrequested (in other words go away) post.

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
well...
I'm obviously not helping you out here. Hopefully you'll get your answers eventually.
veggieboy2001
veggieboy2001's picture
Offline
Last seen: 1 year 7 months ago
Joined: May 1 2013 - 3:34pm
Ultra...

I give you a lot of credit...you really, REALLY tried to help.
But either way, thanks...I think I understand THD much better myself, now.

Tyll Hertsens
Tyll Hertsens's picture
Offline
Last seen: 1 year 1 month ago
Joined: Feb 16 2011 - 12:52pm
Thanks for trying Ultrabike
Bluecap: The AP could absolutely measure the output frequency if I so chose to do so, but during FR measurements it is, in fact, measuring output power at given input frequencies step by step.

The thing you're having a hard time with is the idea that if you put in one frequency it's possible to get a different frequency output. No way. Put in 4kHz, you WILL get out 4kHz. There's simply no way the voice coil could respond in any other way than to follow the input signal fairly closely. The thing that does happen is that other frequencies might come out. Yes, you can put in one tone and get out more than one tone. Usually when this happens you get a series of harmonic overtones. Put in 1kHz, get out a little 2kHz, 3kHz, 4kHz, etc.

Bottom line: You assumption that you heard two different tones coming out of your headphones with the same tone going in is erroneous in some way. But you might have heard more or less harmonic over-tones with the two different headphones.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
Tyll Hertsens wrote:Bluecap:
Tyll Hertsens wrote:

Bluecap: The AP could absolutely measure the output frequency if I so chose to do so, but during FR measurements it is, in fact, measuring output power at given input frequencies step by step.

The thing you're having a hard time with is the idea that if you put in one frequency it's possible to get a different frequency output. No way. Put in 4kHz, you WILL get out 4kHz. There's simply no way the voice coil could respond in any other way than to follow the input signal fairly closely. The thing that does happen is that other frequencies might come out. Yes, you can put in one tone and get out more than one tone. Usually when this happens you get a series of harmonic overtones. Put in 1kHz, get out a little 2kHz, 3kHz, 4kHz, etc.

Bottom line: You assumption that you heard two different tones coming out of your headphones with the same tone going in is erroneous in some way. But you might have heard more or less harmonic over-tones with the two different headphones.

I'm glad we can finally agree that you are making an assumption (an incorrect one). You assume that the headphone is putting out the frequency it claims to be putting out.

You also claim that your "AP could absolutely measure the output frequency if I so chose to do so" -- so instead of arguing about this assumption, why don't you just measure it and settle this?

Pick five random headphones from your list, and see what frequency they put out when you feed them 50hz, 100 hz, 250 hz, 500 hz, 1,000 hz, 1,500 hz, 2,500 hz, 3,500 hz, 5,000 hz, 7,500 hz, 10,000 hz, 12,500 hz. (Actually, pick five random IEMs; that's what I used; IEMs are probably less reliable than headphones in this regard)

And please don't tell me that it's impossible to see what frequency the headphone is putting out because they put out a mix of frequencies. Whatever mix of frequencies the headphone puts out coalesces into the sound created by one frequency. For example, if a headphone fed 4,000 hz puts out a little 3,900 hz, a little 4,050 hz and a ton of 4,000 hz, at the end of the day that mix sounds like one frequency (say, 3,975 hz). In other words, if you play 3,975 hz and you play this headphone fed 4,000 hz they will sound identical. So for all intents and purposes, when fed 4,000 hz, this headphone puts out 3,975 hz. Look at that one coalesced frequency being put out and compare it to what the headphone should be putting out, which in this case would be an error of 25/4,000 hz.

If you have the equipment this should be super easy to do.

And when you do it, you'll be shocked by the disparities.

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
Alright...

What you are looking for is not new or out of this world. It is usually called a THD measurement at a particular frequency. It is tedious and paints a very incomplete picture relative to what you get from a linear frequency response measurement and a complete THD%+N measurement, but here you go:

Response of an HD600 to a 4,000 hz tone:

http://www.head-fi.org/image/id/7136454/width/900/height/900/flags/LL

Note the cursor peaks at exactly 4,000 hz. There is an obvious harmonic contribution at 8,000 hz and other higher harmonics which are below the noise floor. The visible 8,000 hz is 0.078% of the 4,000 hz power.

So no 3,500 or 4,500 hz.

Even if you were to stretch your definition by using your "coalescing" frequencies concept we are dealing with 0.078% of distortion, and you are not taking into consideration frequency masking and perhaps assuming too much about how the human ear works by using the "coalescing" concept.

Am I surprised by these results? Nope.

Can Tyll's equipment do this? Yup. He can run custom THD measurements on headphones just as he has done with amplifiers in the pass using the AP.

Want more data points?

HD600 10,000 hz tone: http://www.head-fi.org/image/id/7136549/width/900/height/900/flags/LL

As you can see, a 10,000 hz source puts out exactly 10,000 hz plus harmonics at 20,000 hz and 30,000 hz all of which are 0.08% of the 10,000 hz power.

We could go on and on with other headphones and frequencies, and you shall find similar behavior all of which has previously described to you.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
First of all those charts

First of all those charts show a ton of noise coming from other frequencies which suggests an error rate higher than 0.07%. But now that you understand what it is, why don't you do it for a few IEMs and see what you get.

ultrabike
ultrabike's picture
Offline
Last seen: 1 year 4 months ago
Joined: Feb 15 2012 - 3:56pm
LOL!

LOL! Now that I understand what it is?

The noise at the lower frequencies is likely environment noise and perhaps the mic pre-amp, not the headphone. I've done some of this with quite a few IEMs in the past and many other headphones and obviously so has Tyll. On the other hand you probably barely know how to put headphones on your head.

Like I've said before, it is obvious you have no idea about what you are talking about and very quick to talk out of your ass. Pick up a book. At the very least read a few wikipedia articles. Work on your approach to request information. In your posts you come across as an uneducated arrogant prick.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
ultrabike wrote:LOL! Now that
ultrabike wrote:

LOL! Now that I understand what it is?

Like I've said before, it is obvious you have no idea about what you are talking about and very quick to talk out of your ass. Pick up a book. At the very least read a few wikipedia articles. Work on your approach to request information. In your posts you come across as an uneducated arrogant prick.

Didn't I ask you above to stop spamming this thread?

Bottom line is that the HD600 in the graph above is not going to sound like true 4,000 hz. And that's a good headphone. For an IEM the discrepancy is far far worse.

Tyll needs to update his measurements to better distinguish between headphones that actually produce the hz they're claiming to produce and those that don't. If he added that dimension, you would quickly see a bunch of high-end IEMs exposed as deficient. In addition, I would bet that the results would be better for full-sized headphones, and best for speakers, showing you that IEMs are fundamentally flawed as a listening tool.

Bob Katz
Bob Katz's picture
Offline
Last seen: 1 year 3 months ago
Joined: Feb 18 2015 - 10:08am
Flaw?

Every measurement method is a small "window" on the complex device we call a headphone. Consider the frequency response plot to be a measurement of the FUNDAMENTAL amplitude at any frequency of interest. A separate plot of harmonic distortion vs. frequency can tell you what OTHER frequencies are being generated when a given source frequency is being generated.

Jay_WJ
Jay_WJ's picture
Offline
Last seen: 4 years 2 months ago
Joined: Dec 13 2013 - 6:50pm
ultrabike,

ultrabike,

You have a lot of patience. I admire you! If I were you, I wouldn't even reply twice. ^^ I met too many people who are trying to reinvent the wheel. Audio measurement is scientifically mature. Apparently, there are many hobbyists who simply do not appreciate this rightly. I am a scientist with a PhD. You would be surprised if you knew that even in scientific communities, there are such people. Then, in the hobbyist world? No further comment ^^

bluecap,

It is not ultrabike who spams this thread. It is you who spam this forum.

veggieboy2001
veggieboy2001's picture
Offline
Last seen: 1 year 7 months ago
Joined: May 1 2013 - 3:34pm
Thanks Jay_WJ

It's times like these I actually wish there was a "like" button on Innerfidelity.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
Jay_WJ wrote:ultrabike,
Jay_WJ wrote:

ultrabike,

You have a lot of patience. I admire you! If I were you, I wouldn't even replied twice. ^^ I met too many people who are trying to reinvent the wheel. Audio measurement is scientifically mature. Apparently, there are many hobbyists who simply do not appreciate this rightly. I am a scientist with PhD. You would be surprised if you knew that even in scientific communities, there are such people. Then, in the hobbyist world? No further comment ^^

bluecap,

It is not ultrabike who spams this thread. It is you who spam this forum.

Well that was a helpful and substantive post and certainly not non-contributing spam.

By the way, my cracker jack box didn't contain a phd but it had a nice toy tattoo.

xnor
xnor's picture
Offline
Last seen: 5 months 1 week ago
Joined: Apr 11 2011 - 1:10pm
Bluecap, it is you who is

Bluecap, it is you who is making false assumptions.

bluecap wrote:

I just listened to two headphones, one on the left ear and the other on the right. Both were supposed to be outputting 4,000 hz, but one headphone sounded brighter than the other one. I can see how two headphones would have different degrees of brightness when playing a song, but that should not be the case when they are playing the exact same frequency.

As has been pointed out to you, if your headphones are crap then they will produce large amounts of nonlinear distortion (harmonics) that - combined - of course sound differently than a pure tone.

The fundamental is at 4k. Period.
You falsely assume that hearing differences means that the fundamental shifted. That's nonsense.
This only could be the case in active headphones with electric circuitry in them that changes the input. And if that were the case, you'd see something like 100% THD+N.

Quote:

Tyll's headphone measurements assume that a headphone that claims to be putting out x hz is actually creating that sound, and not some other frequency. In reality, they're not. They're creating a frequency other than the one they're supposed to be putting out. Some headphones move it up, some move it down. What Tyll needs is a new chart that shows how much the sound the headphone actually puts out differs from the sound it claims to be putting out.

Nope.
The fundamental stays the same. The additional frequencies are called harmonics. Tyll does measure and display this.

RazrLeaf
RazrLeaf's picture
Offline
Last seen: 2 years 7 months ago
Joined: Oct 24 2014 - 12:20pm
You should really read a physics book before posting...

Bluecap, if you understood the physics behind how a headphone converts electrical signal into the pressure wave that human ears perceive as sound, you would realize just how stupid you are. Unless your headphones are straight up broken, it'd not possible to put a given frequency in and not get that frequency out. It could be distorted, but that's the result of the summation of other frequencies present in addition to the fundamental, not the absence of the fundamental.

Go read a physics book about electromagnetics before you come back.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
If there is a more useless

If there is a more useless response, to a substantive question with detailed procedure by which you can test what I'm saying, I don't know what it is.

Why don't you go read an organic chemistry book, an anatomy book, and Planet of the Apes before posting again? It'll answer all of your questions.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
xnor wrote:Bluecap, it is you
xnor wrote:

Bluecap, it is you who is making false assumptions.

You falsely assume that hearing differences means that the fundamental shifted. That's nonsense.

lol. I need to save this quote as it perfectly summarizes the problem here. It proves that you're not concerned about how something actually sounds. That's the flaw I'm talking about. Tyll's headphone measurements aren't concerned with the coalesced frequency actually being produced when you feed a headphone N hz. Thus they fail to distinguish between headphones that actually produce the sounds they're supposed to produce and those that don't. That's why the charts don't represent what you actually hear, especially with IEMs.

xnor
xnor's picture
Offline
Last seen: 5 months 1 week ago
Joined: Apr 11 2011 - 1:10pm
bluecap wrote:xnor wrote
bluecap wrote:
xnor wrote:

Bluecap, it is you who is making false assumptions.

You falsely assume that hearing differences means that the fundamental shifted. That's nonsense.

lol. I need to save this quote as it perfectly summarizes the problem here. It proves that you're not concerned about how something actually sounds.

No, it proves that you are clueless.

Let's assume for one moment that your brainfart - the idea that the fundamental shifted - is actually true.

We feed a 4 kHz tone, get a fundamental at 4.5 kHz plus harmonics above that. During the analysis the notch filter at 4 kHz will not significantly change the shifted fundamental at 4.5 kHz, so the analyzer will see huge amounts of energy that is not harmonically related to the fundamental.

That is noise. N. As in THD+N.

As such you would see 100% THD+N. We never see that, because your idea stems from absolute cluelessness.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
You're just playing inane

You're just playing semantic games by using terms like "fundamental" tone.

When fed a signal, a headphone produces ONE tone. One single tone. You can determine that tone by playing other tones and asking yourself, is this tone higher or lower in frequency than the tone being produced by the headphone? Eventually you'll zero in on the tone the headphone is actually producing, call this the coalesced tone using my terminology above, and it won't be the tone it's supposed to produce.

You want to say that it's producing the true "fundamental" tone plus distortion.

I could give you countless examples of why your way of analyzing the issue is inept, but this one should suffice. A Porsche 911 Turbo S goes from 0-60 in 2.9 seconds. A Toyota Prius does it in 9 seconds.

Your way of viewing the world would say the Toyota Prius and Porsche 911 have the same 0-60 time because the Prius has a "fundamental" 0-60 time of 2.9 seconds with an insignificant distortion of -6.1 seconds. lol. No. One has a 0-60 time of 2.9 seconds, the other a 0-60 of 9 seconds. End of story.

And even if you wanted to use that fundamental tone + distortion framework, then at least measure the distortion properly, which is not being done as described above.

The fact that none of you are actually concerned about the true tone being heard when a headphone is fed a signal proves that none of you are concerned about the most important thing about a headphone.

If anyone else would like, please feel free to step up and get your verbal thrashing, which I have dispensed many times in this thread. Or you can devolve into more incoherent trolling arguments and name-calling which is what often happens when you have nothing substantive to say. But rather than do that, simply acknowledge that I'm right because my premise is right -- it's the sound that matters, not (incorrectly) imagined physics, graphs or other irrelevant nonsense. Tyll's measurements don't capture the tone (the hz) actually being heard when a headphone is fed a particular signal, and it's why his measurements don't tell you how a headphone (an IEM specifically) will actually sound.

xnor
xnor's picture
Offline
Last seen: 5 months 1 week ago
Joined: Apr 11 2011 - 1:10pm
bluecap wrote:You're just
bluecap wrote:

You're just playing semantic games by using terms like "fundamental" tone.

When fed a signal, a headphone produces ONE tone. One single tone.

No, it doesn't. Your damaged hearing might perceive it as a single tone, but it isn't.

You are a troll. Q.E.D.

bluecap
bluecap's picture
Offline
Last seen: 4 years 6 months ago
Joined: May 8 2014 - 5:10pm
xnor wrote:bluecap wrote:You
xnor wrote:
bluecap wrote:

You're just playing semantic games by using terms like "fundamental" tone.

When fed a signal, a headphone produces ONE tone. One single tone.

No, it doesn't. Your damaged hearing might perceive it as a single tone, but it isn't.

You are a troll. Q.E.D.

I just explained how you can determine that one tone. Say a headphone is putting out a bunch of tones that together sound like 3,700 hz (when it's supposed to be putting out 4,000 hz). Just play a 3,800 hz tone, and ask yourself if the headphone's sound is lower. Then play a 3,500 hz tone and ask yourself if the headphone's sound is higher. Keep doing this, assuming your hearing is sufficiently precise, and you'll zero in on the actual coalesced tone being produced. Sound is linear. Either a tone is higher in frequency than another tone, or it's lower in frequency than the other tone. You can use this linearity to determine the coalesced frequency of any tone.

I don't know how many times I have to provide you people with detailed, step by step techniques that prove what I'm saying, or which you can use to measure the actual distortion.

Tyll Hertsens
Tyll Hertsens's picture
Offline
Last seen: 1 year 1 month ago
Joined: Feb 16 2011 - 12:52pm
Oh well....
The fundamentally important thing about InnerFideity is getting down to the truth of things and helping people understand those truths.

Bluecap here seems bound and determined to speak what he considers truth, but, unfortunately, he is patently wrong and is unwilling to re-evaluate his position. The big problem is that folks who come here looking for an education might stumble across this thread and may come to think Bluecap has a point. He doesn't, but some folks won't know that. He is, in essence, sowing seeds of doubt and confusion for InnerFidelity readers. I don't know if he's doing this on purpose, or if he is sincerely as stubbornly ignorant as he portrays himself. Regardless, he is doing a significant disservice to InnerFidelity readers, and wasting the time of respected and learned enthusiasts who have tried to help him see the error of his thoughts here in this thread.

I'm not going to let him lead folks astray on InnerFidelity any longer. Off to banned camp for you.

Harry Manback
Harry Manback's picture
Offline
Last seen: 1 year 2 months ago
Joined: Jan 12 2015 - 12:22pm
I can see both sides of this

I agree and understand that you measure THD as the amount of "junk" that is coming out when you expect to hear frequency X. I also understand that when you send a signal of frequency X to a voice coil, it moves up and down at that frequency, ignoring any mechanical losses due to friction and resistance of the driver. I think that the OP was trying to state the need for a measurement that could graphically display the sum total result of the THD. It does, effectively, shift the frequency output when summed. This would include those losses as well as any effects of the design of the enclosure. If this tonal summation didn't happen, a dynamic headphone driver could not reproduce music, as the music itself is the summation of all of the sounds made by the instruments together. I would call such a measurment "Frequency Accuracy". THD is the standard and a graph shown previously shows the THD as the ratio of the primary frequency in comparison to the THD as lines of relative height. If you could show this as a single frequency that is averaged, would it not shift the fundamental frequency actually produced to the left or right to some extent? Not the frequency of the voice coil, but the frequency of the produced tone. I can envision a heat map of frequency vs accuracy at say across a volume level sweep of -10db to 0db.

Please do correct me if I've made some mistakes, but bear in mind that I don't accept things I'm told because someone is an "expert". I don't put a lot of faith in credentials or social standing. I'm only interested in the facts and the truth.

FYI - I'm not trolling. I'm not expert in audio science, but I do have a degree in Engineering, and have mathematics classes from Calculus 1 through Differential Equations and Numerical methods, so I can usually understand what is explained to me, and I believe I understand what the OP was getting at. He just reacted badly when Ultrabike began to show frustration. In my opinion, only Tyll and Bob Katz kept their cool in this thread. Things went south around post 8. Let's not forget that it's good to question things. Perhaps having a new kind of graph would be a good thing for the layperson to understand.

Tyll Hertsens
Tyll Hertsens's picture
Offline
Last seen: 1 year 1 month ago
Joined: Feb 16 2011 - 12:52pm
Problem is the OP was

Problem is the OP was focussed on a perception that one could input one tone into the headphones and get a different tone out. Doesn't happen, headphones are a minimum phase device so the driver is always within 360 degrees of the input signal (probably far less, but I don't know the number.

Your thought that the average frequency may shift up and away from the original tone due to harmonic distortions is absolutely correct, but I tend to think the result is more like the difference between a tone and a chord as opposed to a tone and a tone that's higher in frequency.

That's best represented by a spectral plot from a single tone, which can be done, but I don't do it.

I can point you to HeadRoom's "Build a Graph" which does display it. Here's an example.

http://graphs.headphone.com/graphCompare.php?graphType=1&graphID[]=3791&...

castleofargh
castleofargh's picture
Offline
Last seen: 6 months 1 week ago
Joined: Feb 22 2013 - 11:30pm
wow I just realized there was

wow I just realized there was a forum(or maybe I forgot?).
that was an entertaning topic. and kudos to ultrabike for not turning into the hulk.

eheh it would be fun if bluecap dream was real, I can't imagine what it would sound like to have a driver able to part ways with the speed of the signal wave. it would probably mean a good 100% distortion value almost everywhere, that must sound nice. ^_^

Omniryx
Omniryx's picture
Offline
Last seen: 4 years 5 months ago
Joined: Apr 15 2015 - 9:00pm
Ain't no fix for stupid

There is no cure for invincible ignorance. In bluecap's world, you could cue up Diana Krall and Willie Nelson would come out of the phones.

Seriously, I thought you folks were wonderfully tolerant. On some audio forums, bluecap would have been flamed to a smoking cinder.

  • X