[ultrabike]

Quote:

 
Sure, but just because someone believes FR is important, or highly indicative, doesn't make that so either.
 
You need evidence on both sides of the coin, not just the side you disagree with.

 
I believe FR is important because it shows SPL as a function of frequency. A properly measured headphone, should give an indication of the neutrality of a headphone, or at least an indication of what exactly you are buying besides the nice box + way cool build materials and design.
 
Limitations to what FR tells you is relative to the last part of the chain: your ears and brain. What I personally try to do, is see a FR graph (CDS, impulse response - and from as many web sites as I can) and try to match what I hear with what I see and decide what I want.
 
Another limitation that has been measured elsewhere is the manufacturing variations of the same headphone. If a manufacturer makes a HP with a certain FR that you like after matching it to your experience, and suddenly changes the build of the HP so that the FR of the previously measured models no longer match what you are about to buy, then yes FR no longer applies... But that is a flaw in the manufacturer's process. Such a change should call for a different naming of the HP so that consumer get what they pay for.

 

[ultrabike]

Quote:

I don't know why anyone thinks hardware EQ is superior to software.  A hardware EQ is an extra component in the audio chain that would degrade sound.  These days most hardware EQs are just a digital EQ with an ADC in front and a DAC behind--exactly what you'd get with software EQ, except with two potentially degrading conversions added.
 
Software EQ is transparent as long as you don't clip.

 
It really depends on the implementation. You might need to add more bits to the software EQ there as well... and just because its digital EQ, it doesn't mean is fully software, it can be digital HW (add a powerful audio processor to handle the arithmetic)

 

[RPGWiZaRD]

I'd prefer a hardware analog EQ. The digiZoid "amp" is purely analog with it's bass boost, probably partly why I love its sound so much hehe. 

 

[CrystalT]

In this thread: Kiteki proves that our perception of headphones is almost entirely cognitive bias.

 

[rhythmdevils]

Quote:

In this thread: Kiteki proves that our perception of headphones is almost entirely cognitive bias.

 
Really? 
 
In that case, I will now prove that human beings can fly because we also have 2 eyes and 2 legs. 
 

 
 
OR, maybe it's more complicated than that, and maybe I have no idea what I'm talking about and shouldn't try to draw complex conclusions from incorrect readings of graphics.

 

[CrystalT]

Or maybe people should actually go by what can be objectively measured, and keep their mouths shut when it comes to subjectivity?

 

[xnor]

Quote:

In this thread: Kiteki proves that our perception of headphones is almost entirely cognitive bias.

 
I doubt that's his intention and I wouldn't go that far (I guess you exaggerated intentionally). But it definitely is an issue which doesn't seem to be easy to eliminate with headphones. :/

 

[maverickronin]

Quote:

Lots of why's there...
 
1.  So you're saying I don't understand the graphs?
 
2.  Since goldenears calls above 10kHz and under 100Hz the "no guarantee zone", likewise Tyll says it isn't reliable, the raw data of headphones there is a bit all over the place.  That's not relevant though, I just got lucky with the MMM and LCD-3, the point I'm conveying is they look too similar to justify the differences.  I did look at all the data on the LCD-3 and MMM, like I said, can I 'golden FR' the MMM?  Is it just an LCD-3 with a tiny bit extra distortion and noise?  It's not a rhetorical question, it's open ended, i.e. can someone make an equalizer preset or convolver which turns the MMM into the LCD-3?  Looking at the responses, it seems like some in this thread do think so, so far!
 
3.  I don't.
 
4.  Yes CSD is useful in finding decay it seems (however it can't decipher if it's positive or negative sounding decay).  Afaik it's useful for room acousticians, when the goal is zero decay.

 
No kiteki you don't understand them.  You're just matching shapes together like its some sort of puzzle game or a slot machine.
 
Three squares!  Jackpot!
 
You don't what what the differences represent or why they're important.  You also don't understand how and why measurements at the extremes can be less reliable than the midrange.  Seal is important in bass response but you can look at Tyll's traces from different positions to determine how sensitive the headphone's seal is to proper placement.  In the high frequencies slight differences is position can easily change the exact frequencies of peaks and nulls in the FR.  It doesn't make ~25dB over the range of an entire octave appear or disappear.

 

[mikeaj]

You know, if you actually were to compare these two by listening, you might try a bandpass filter when comparing both at about 50 Hz to 5000 kHz or so, cutting out the rest.  Then you can see how similar they are in the range that people are actually claiming that they may be similar.

 

[purrin]

Two things:
 

• CSD to verify ringing (not just used for rooms, but also drivers for speakers, for years too, even before you were born I was using them for speaker drivers) See Zaph Audio: http://www.zaphaudio.com/tweetermishmash/compare.html (check CSD)
• FR in bass measures the same or similar, but more bass on one? Try non-linear distortion in bass. Try a 50Hz tone, full spectrum graph. The one with more bass is will probably have a doubling effect (2nd harmonic).
• SkullCandy looks to have serious issues with the last octave / high treble.

 
Moral of the story: gotta use the right tools. FR is only one of them. FR + CSD is better. FR + CSD + few scattered quality and meaningful non-linear distortion measurements gives us a much better picture.

 

[xnor]

Quote:

Two things:

But you posted three points.
 
Quote:

• CSD to verify ringing (not just used for rooms, but also drivers for speakers, for years too, even before you were born I was using them for speaker drivers) See Zaph Audio: http://www.zaphaudio.com/tweetermishmash/compare.html (check CSD)

Right, but what many don't seem to understand is that CSD is just a visualization technique. A visualization of the impulse response to show both frequency and time domain data on a single graph. That's all there is to it.
If you see ringing in the impulse response, you will see in the CSD too, of course.
 
Quote:

• FR in bass measures the same or similar, but more bass on one? Try non-linear distortion in bass. Try a 50Hz tone, full spectrum graph. The one with more bass is will probably have a doubling effect (2nd harmonic).

I don't get the first part. Regarding the second part, are you saying that distortion will add significant amounts of bass? Or the other way around, that a bassy headphone has inherently higher distortion?
(May I add that even 5% of distortion only adds about 0.01 dB to the FR at the frequency of the overtone.)
 
Quote:

• SkullCandy looks to have serious issues with the last octave / high treble.

True. The question is whether it's the drivers fault or if it's just acoustically non-transparent material in front of the driver.
 
 
Quote:

Moral of the story: gotta use the right tools. FR is only one of them. FR + CSD is better. FR + CSD + few scattered quality and meaningful non-linear distortion measurements gives us a much better picture.

I'd say impulse response first (--> FR, CSD ...). Distortion measurements second (THD, IMD).

 

[purrin]

Quote:

Right, but what many don't seem to understand is that CSD is just a visualization technique. A visualization of the impulse response to show both frequency and time domain data on a single graph. That's all there is to it.
If you see ringing in the impulse response, you will see in the CSD too, of course.

 
Ya ya, ya, it's all inter-related via math (Sweep|MLS->IR<->CSD|FR->Square Wave), tell me about it, but it's not as simple as you make it out to be.
 
Visualization techniques are key. DCT/FFT window sizes, window functions, rise times/boundary functions to avoid artifacts, and anchoring techniques, etc. all effect how the waterfall is going to look. (I wrote my own code and heavily experimented with different parameters to process the impulse response to waterfall plots for just headphones.)
 
Unless we can readily identify ringing at 8kHz or 20kHz just by looking at the impulse (I can, sort of, but would much prefer not to), crunching some math on the impulse to get a waterfall plot is much more effective means of visualization. For example: Ringing at 21kHz is probably harmless to anyone over 25, but it will look nasty in the impulse response. Short duration moderate magnitude ringing at 7kHz can sound nasty, but may be hidden or masked with other characteristics in the impulse response.
 
Unfortunately, most people (myself included) don't have Commander Data's processing abilities. If you do, that's awesome!
 

I don't get the first part. Regarding the second part, are you saying that distortion will add significant amounts of bass? Or the other way around, that a bassy headphone has inherently higher distortion?
(May I add that even 5% of distortion only adds about 0.01 dB to the FR at the frequency of the overtone.)

 
Let say two headphones have the same frequency response, but one has more 2nd order harmonic distortion. The one with more second order harmonic distortion (assuming it's sufficiently high) will sound more bassy because bass is being double an octave higher.
 
Check your math again. 5% distortion is ~ -26db. Remember to convert to decibels. 20*log(x/100)
 
If the second harmonic distortion is 5% for a 50Hz tone @85db. That means 100Hz will be -26db down at 59db. This amount of bass distortion is not uncommon with Best Buy or even audiophile headphones.

 

[xnor]

Quote:

Ya ya, it's all inter-related via math (Sweep|MLS<->IR<->CSD|FR->Square Wave), but it's not as simple as you make it out to be.
 
Visualization techniques are key. FFT window sizes, window functions, rise times, and anchoring techniques, etc. all effect how the waterfall is going to look.

 
Yes and yes. The emphasized part: that's why I stress the importance of impulse responses.
 
Sure, CSD is a very handy and nice means of visualization. I also agree that where our hearing is more sensitive, ringing is likely to be more audible too.
 
Quote:

Let say two headphones have the same frequency response, but one has more 2nd order harmonic distortion. The one with more second order harmonic distortion will sound more bassy because bass is being double an octave higher.
 
Check your math again. 5% distortion is ~ -26db. Remember to convert to decibels. 20*log(x/100)
 
If the second harmonic distortion is 5% for a 50Hz tone @85db. That means 100Hz will be -26db down at 59db. This amount of bass distortion is not uncommon with Best Buy or even audiophile headphones.

Yes -26 dB + 0 dB add up to 0.42 dB, just like 85 dB + 59 dB add up to 85.42 dB. You'd have to have a pretty significant dip in the frequency response at the frequency/ies of the overtone(s) for them to actually have a bigger effect (10 dB dip would cause +1.27 dB).
 
edit: corrected sums

 

[purrin]

Quote:

 
Yes and yes. The emphasized part: that's why I stress the importance of impulse responses.
 
Sure, CSD is a very handy and nice means of visualization. I also agree that where our hearing is more sensitive, ringing is likely to be more audible too.
 
Yes -26 dB + 0 dB add up to 0.01 dB, just like 85 dB + 59 dB add up to 85.01 dB. You'd have to have a pretty significant dip in the frequency response at the frequency/ies of the overtone(s) for them to actually have a bigger effect (10 dB dip would cause +0.1 dB, 20 dB dip about +1 dB).

 
What are you talking about? Do you actually understand what harmonic distortion is? You are confusing me as well as everyone else who may be trying to learn how is it that two headphones which measure the same (FR) in the bass, one can sound more bassy.
 
Assuming FR is the same, this ~5% second harmonic distortion at 50Hz (100Hz -26 down)

 
will sound more significantly more "bassy" than this: 
 
0.5% second harmonic distortion at 50Hz (100Hz @ -46db)

 

[Joe Bloggs]

Quote:

Yes -26 dB + 0 dB add up to 0.01 dB, just like 85 dB + 59 dB add up to 85.01 dB. You'd have to have a pretty significant dip in the frequency response at the frequency/ies of the overtone(s) for them to actually have a bigger effect (10 dB dip would cause +0.1 dB, 20 dB dip about +1 dB).

 
I don't understand the second part, but in the first part xnor is saying that a 85dB fundamental + a 59dB overtone adds up to a 85.01dB total volume.  which makes sense given the logarithmic nature of decibels.