[up late]

the fr doesn't support the claim made earlier in the thread that the bass extends further than the th900 (which i was sceptical of i must admit). it does show a mid-bass hump tho.

 

[Kyle 491]

This is a useful thread to separate out those who can hear and can communicate what they are hearing, and those who cannot. 

 

[Aimin Liao]

Hi. Guys
 
  You may try your Master 1 With Chord Hugo TT. I personally like the sound come out of TT

 

[npdang]

The measurements actually seem fairly good to me, with the exception of the driver breakup at ~19khz which is disappointing (you can see it in the FR plot as a +10db spike and the impedance plot as a tiny ripple).  There is also a ripple in the impedance curve at ~2.2khz, which might indicate some issues with loading the driver, or some type of surround edge resonance.  To my ears, a diaphragm breakup of at least 30khz or higher is necessary in order to fully avoid any kind of metallic/hardness coloration, although I can't explain why.  If you recall older titanium tweeters from the 90's typically show breakup at around this frequency, and you could definitely hear a hardness/metallic coloration to the sound, although that might have also been the result of high distortion motors.  I still think it would have been nice, especially at this price, to see a beryllium diaphragm used.  However, the impedance is otherwise impressively flat, probably a result of the copper shorting ring used in the motor.  Distortion is exceptionally low, especially above 2khz, and the rising distortion in the bass may just be the result of the rather steep drop off in the bass below 100hz, as opposed to poor motor design.  Something odd with the measurement below 100hz as well, as an increase of nearly 10db in output below 100hz does not show the expected increase in distortion.  I'm not familiar with square wave testing, but I believe it's generally used to test phase distortion, as opposed to energy storage as you suggest.  I do agree though, the impulse response does not look good, and indicates some issues with dampening the 19khz diaphragm breakup.  I'd be curious to see how it measures with a notch filter in place.  In any case, I'm sure many would argue you can't hear ringing at that high of a frequency, or at the least, it's not much of an audible issue.

 

[npdang]

It seems very odd, honestly, looking at the posted datasheets I'd have a hard time correlating any of those measurements with what people have been reporting here.  By any stretch though, almost all of the datasheets show exceptional performance at least compared to speakers

 

[takato14]

The measurements actually seem fairly good to me, with the exception of the driver breakup at ~19khz which is disappointing (you can see it in the FR plot as a +10db spike and the impedance plot as a tiny ripple).  There is also a ripple in the impedance curve at ~2.2khz, which might indicate some issues with loading the driver, or some type of surround edge resonance.  To my ears, a diaphragm breakup of at least 30khz or higher is necessary in order to fully avoid any kind of metallic/hardness coloration, although I can't explain why.  If you recall older titanium tweeters from the 90's typically show breakup at around this frequency, and you could definitely hear a hardness/metallic coloration to the sound, although that might have also been the result of high distortion motors.  I still think it would have been nice, especially at this price, to see a beryllium diaphragm used.  However, the impedance is otherwise impressively flat, probably a result of the copper shorting ring used in the motor.  Distortion is exceptionally low, especially above 2khz, and the rising distortion in the bass may just be the result of the rather steep drop off in the bass below 100hz, as opposed to poor motor design.  Something odd with the measurement below 100hz as well, as an increase of nearly 10db in output below 100hz does not show the expected increase in distortion.  I'm not familiar with square wave testing, but I believe it's generally used to test phase distortion, as opposed to energy storage as you suggest.  I do agree though, the impulse response does not look good, and indicates some issues with dampening the 19khz diaphragm breakup.  I'd be curious to see how it measures with a notch filter in place.  In any case, I'm sure many would argue you can't hear ringing at that high of a frequency, or at the least, it's not much of an audible issue.

"distortion is exceptionally low"

*rubs temples*

 

[npdang]

"distortion is exceptionally low"

*rubs temples*

 
Granted, I have never measured a headphone before, but I have measured hundreds of speakers and in comparison, these really are exceptional results.  That aside, several other observations:
 
1. Looking at the "raw data for 5 headphone positions" there is a significant deviation in the low frequency response; up to 10db.  I believe this deviation might result in an unreliable distortion measurement below 100hz.  Assuming a worst case scenario with the driver tested in the position yielding the lowest sensitivity, it would require significantly higher drive levels to reach the output test level, resulting in higher measured distortion. Even so, 10% THD at 20hz at the listening position for a speaker would be an amazing achievement and just at what most people would consider the threshold of audibility. 
 
2. Without knowing more about the test conditions and method, results from one headphone may not be directly comparable to another.  It appears (assuming here) that the driver sensitivity is referenced at 90db at 1khz, at 0.134 vrms.  Using a sweep with an output voltage referenced at 1khz for 90db, might result in relatively higher measured distortion for drivers with a more extended low end response.
 
3. The 100db test above 2khz show lower distortion levels than the 90db test, meaning that the actual measured distortion maybe buried in the noise floor.
 
4. Measured distortion does not change below 100hz between the 90db and 100db test.  This to me indicates that the measured distortion may not be accurate in absolute terms, or there is something else going on that I do not understand.  If the motor is close to being overdriven, a +10db increase in output should result in significantly higher measured distortion.  The fact that it remains the same, indicates that the driver is still operating well within its usable excursion limits.
 
5. Overall, I rarely pay much mind to THD below 100hz. It's just not that audible.  Non-linear distortion in the mid-high frequency range on the other hand, is far more offensive, and here the Pioneer does very well.  Also, the spectrum and character of the distortion products is perhaps just as important, if not more so, than the total amount.  That information, we do not have here. 

 

[takato14]

1: Absolutely wrong, distortion isn't affected by deviation any more than the averaged FR plot does.

2: Measurements from two different headphones are absolutely comparable because the exact same methodology is used for each of them. Look at measurements for Audeze's LCD-2 and you will see that bass extension has no correlation to bass distortion measurements. Regardless of that, the Pioneer falls off a cliff below 100Hz, indicating (regardless of distortion) that it has poor bass extension.

3: The 100dB distortion plots will demonstrate power handling, and are not affected by noise floor. If the 100dB plot is below the 90dB plot, this indicates a transducer with good power handling in that frequency range.

4: This is actually somewhat true, but even if the transducer is operating as intended here they have the bass tuned improperly so it becomes too thick.

5: I find distortion below 100Hz to be very audible. It tends to make anything in the bass sound flabby and poorly defined. As for the distortion orders, you can see some evidence of the individual distortion orders with the plot; as the line becomes flatter and less jagged you're approaching lower order distortions (D2) which make things sound thicker and more "full" at the cost of definition.

The Sennheiser HD800 measures far better than the Pioneer at $1500, so I still cannot see where you're getting the idea that these measure well.

 

[npdang]

1: Absolutely wrong, distortion isn't affected by deviation any more than the averaged FR plot does.

2: Measurements from two different headphones are absolutely comparable because the exact same methodology is used for each of them. Look at measurements for Audeze's LCD-2 and you will see that bass extension has no correlation to bass distortion measurements. Regardless of that, the Pioneer falls off a cliff below 100Hz, indicating (regardless of distortion) that it has poor bass extension.

3: The 100dB distortion plots will demonstrate power handling, and are not affected by noise floor. If the 100dB plot is below the 90dB plot, this indicates a transducer with good power handling in that frequency range.

4: This is actually somewhat true, but even if the transducer is operating as intended here they have the bass tuned improperly so it becomes too thick.

5: I find distortion below 100Hz to be very audible. It tends to make anything in the bass sound flabby and poorly defined. As for the distortion orders, you can see some evidence of the individual distortion orders with the plot; as the line becomes flatter and less jagged you're approaching lower order distortions (D2) which make things sound thicker and more "full" at the cost of definition.

The Sennheiser HD800 measures far better than the Pioneer at $1500, so I still cannot see where you're getting the idea that these measure well.

 
1. I think it does.  You may be measuring distortion at 90db for one position, and 100db for another with the exact same drive level.  That is usually a pretty significant difference, except here, where 90db and 100db seem to strangely produce the exact same level of distortion  
 
2. Depends on how you look at it. You could also say the response is flat to 20hz, with a 5db bump centered at ~100hz... and if you're willing to give up a little sensitivity, a wide Q notch filter could easily flatten the response.  Again, we just do not have enough information on how the sweep test is performed to conclusively say that driver sensitivity over the test frequency range does not impact measured distortion when expressed as a PERCENTAGE.  Take a look at this driver for example: http://accuton.com/drivers/detail.php?driver=57&matID=4&appID=2  ... there is a rather large difference in measured distortion where the driver's breakup is filtered, and where it is not.  Tyll also cautions not to place too much emphasis on his distortion measurements: "Maybe the most important thing to know about my THD+noise measurements is that they're the most likely plot to be erroneous."  
 
3. I think we are saying the same thing to a certain extent.  In the 90db plots, the measured distortion is buried in the noise floor.  In the 100db plots, it may not be.  Point being, we do not how how good the already exceptional performance at 100db might be improved upon if we were to lower the output level to 90db because the measured distortion at 90db is buried in the noise floor.  Now how could anyone say that is anything but exceptional performance is beyond me
 
4. Agreed, although it is still highly unusual and would make me question the accuracy of the results here.  
 
5. Not sure I can agree here, at least when looking at ALL of the measurements as a whole (and above 2khz where the ear is most sensitive), and especially given the threshold for audibility and in comparison to much more expensive loudspeaker systems.  Regardless of whether you might think these are good or bad measurements, I think that to draw any sort of definite conclusions about the performance of the Pioneers based on 1 simple distortion test which only expresses total harmonic distortion as a percentage, would be a huge mistake.  And while it is easy to see and knock the diaphragm break up in the impulse response with these measurements, it is not so easy to see the benefits a stiff diaphragm yields over a soft/plastic diaphragm as used in the HD800s... whereas, a laser vibrometer analysis would likely show a clear advantage to the Pioneer.  Same thing if you were to measure inductance related distortion, and so on.
 
I'd also suggest playing around with the Klippel distortion test.  Perhaps your ears are better than mine, but even with simple tones I have a hard time distinguishing past 10% with the 20hz-180hz test.  With more complex music, you can see that out of all the participants who took the test, ~10% distortion is just about the threshold for audibility for most people.  
 
http://www.klippel-listeningtest.de/lt/?page=instructions

 

[takato14]

 

1: Absolutely wrong, distortion isn't affected by deviation any more than the averaged FR plot does.

2: Measurements from two different headphones are absolutely comparable because the exact same methodology is used for each of them. Look at measurements for Audeze's LCD-2 and you will see that bass extension has no correlation to bass distortion measurements. Regardless of that, the Pioneer falls off a cliff below 100Hz, indicating (regardless of distortion) that it has poor bass extension.

3: The 100dB distortion plots will demonstrate power handling, and are not affected by noise floor. If the 100dB plot is below the 90dB plot, this indicates a transducer with good power handling in that frequency range.

4: This is actually somewhat true, but even if the transducer is operating as intended here they have the bass tuned improperly so it becomes too thick.

5: I find distortion below 100Hz to be very audible. It tends to make anything in the bass sound flabby and poorly defined. As for the distortion orders, you can see some evidence of the individual distortion orders with the plot; as the line becomes flatter and less jagged you're approaching lower order distortions (D2) which make things sound thicker and more "full" at the cost of definition.

The Sennheiser HD800 measures far better than the Pioneer at $1500, so I still cannot see where you're getting the idea that these measure well.

 
1. I think it does.  You may be measuring distortion at 90db for one position, and 100db for another with the exact same drive level.  That is usually a pretty significant difference, except here, where 90db and 100db seem to strangely produce the exact same level of distortion  
 
2. Depends on how you look at it. You could also say the response is flat to 20hz, with a 5db bump centered at ~100hz... and if you're willing to give up a little sensitivity, a wide Q notch filter could easily flatten the response.  Again, we just do not have enough information on how the sweep test is performed to conclusively say that driver sensitivity over the test frequency range does not impact measured distortion when expressed as a PERCENTAGE.  Take a look at this driver for example: http://accuton.com/drivers/detail.php?driver=57&matID=4&appID=2  ... there is a rather large difference in measured distortion where the driver's breakup is filtered, and where it is not.  Tyll also cautions not to place too much emphasis on his distortion measurements: "Maybe the most important thing to know about my THD+noise measurements is that they're the most likely plot to be erroneous."  
 
3. I think we are saying the same thing to a certain extent.  In the 90db plots, the measured distortion is buried in the noise floor.  In the 100db plots, it may not be.  Point being, we do not how how good the already exceptional performance at 100db might be improved upon if we were to lower the output level to 90db because the measured distortion at 90db is buried in the noise floor.  Now how could anyone say that is anything but exceptional performance is beyond me
 
4. Agreed, although it is still highly unusual and would make me question the accuracy of the results here.  
 
5. Not sure I can agree here, at least when looking at ALL of the measurements as a whole (and above 2khz where the ear is most sensitive), and especially given the threshold for audibility and in comparison to much more expensive loudspeaker systems.  Regardless of whether you might think these are good or bad measurements, I think that to draw any sort of definite conclusions about the performance of the Pioneers based on 1 simple distortion test which only expresses total harmonic distortion as a percentage, would be a huge mistake.  And while it is easy to see and knock the diaphragm break up in the impulse response with these measurements, it is not so easy to see the benefits a stiff diaphragm yields over a soft/plastic diaphragm as used in the HD800s... whereas, a laser vibrometer analysis would likely show a clear advantage to the Pioneer.  Same thing if you were to measure inductance related distortion, and so on.
 
I'd also suggest playing around with the Klippel distortion test.  Perhaps your ears are better than mine, but even with simple tones I have a hard time distinguishing past 10% with the 20hz-180hz test.  With more complex music, you can see that out of all the participants who took the test, ~10% distortion is just about the threshold for audibility for most people.  
 
http://www.klippel-listeningtest.de/lt/?page=instructions

1: No. 100dB is producing more distortion. If you raise the SPL, the distortion plot should move down by an increment equivalent to the amount by which you raised the SPL. Since Tyll gives his distortion in percentage, it'd be difficult to calculate the exact amount, but it is indeed rising. The fact that it lines up with the 90dB plot is mere coincidence and happens quite often. Alternately, the driver could be reaching the maximum amount of excursion it can create, causing it to stop actually pushing more volume -- which will also show up as more distortion on the plot.
 
2: Yes, you could say that, and it still means the same thing. It varies in the bass where other, cheaper headphones manage to maintain a flat response.
 
3: Noise floor has nothing to do with the distortion measurements. There is no noise floor, the headphones are all measured in an anechoic chamber. This is what I am saying, I have no clue what you're saying.
 
4: Highly unusual? Everyone tunes bass loosely to increase SPL as a method of compensating for poor upper range response of the transducer. The AKG K/Q701 is a prime example. So is the Shure SRH-1840.
 
5: When did I say I was going just off the harmonic distortion? Are you blind? The HD800 has a much smoother FR and better 30 and 300Hz square waves. The Pioneer has an enormous amount of variance and coloration in the upper regions and is overall far less flat than the HD800. Compare the impulse responses and the final nail is put into the coffin. The HD800 undeniably measures far better -- at $1000 less.
 
Another appropriate analog would be the Sennheiser HD700, which is measurably very similar to the Pioneer, yet still superior in all of the aforementioned categories. I think that one's down to $650 MSRP now? Really having a lot of trouble seeing where you're getting the "good measurements" claim from when there are other options that produce a similar or better result at a fraction of the price.
 
And as for the distortion test, what are you using to test it with? Headphones that don't distort in the bass are very rare, especially open headphones. You won't be able to discern bass distortion if you're used to hearing it all the time. My reference for bass is my Pioneer Monitor 10-II, which while not particularly perfect anywhere else, pushes a ridiculously clean, distortion-free bass with almost no variation from 300Hz to 10Hz.
 
EDIT: Can't seem to find what test you're talking about? I see no 80/120Hz or distortion percentages mentioned anywhere on the page you linked.

 

[npdang]

  1: No. 100dB is producing more distortion. If you raise the SPL, the distortion plot should move down by an increment equivalent to the amount by which you raised the SPL. Since Tyll gives his distortion in percentage, it'd be difficult to calculate the exact amount, but it is indeed rising. The fact that it lines up with the 90dB plot is mere coincidence and happens quite often. Alternately, the driver could be reaching the maximum amount of excursion it can create, causing it to stop actually pushing more volume -- which will also show up as more distortion on the plot.
 
2: Yes, you could say that, and it still means the same thing. It varies in the bass where other, cheaper headphones manage to maintain a flat response.
 
3: Noise floor has nothing to do with the distortion measurements. There is no noise floor, the headphones are all measured in an anechoic chamber. This is what I am saying, I have no clue what you're saying.
 
4: Highly unusual? Everyone tunes bass loosely to increase SPL as a method of compensating for poor upper range response of the transducer. The AKG K/Q701 is a prime example. So is the Shure SRH-1840.
 
5: When did I say I was going just off the harmonic distortion? Are you blind? The HD800 has a much smoother FR and better 30 and 300Hz square waves. The Pioneer has an enormous amount of variance and coloration in the upper regions and is overall far less flat than the HD800. Compare the impulse responses and the final nail is put into the coffin. The HD800 undeniably measures far better -- at $1000 less.
 
Another appropriate analog would be the Sennheiser HD700, which is measurably very similar to the Pioneer, yet still superior in all of the aforementioned categories. I think that one's down to $650 MSRP now? Really having a lot of trouble seeing where you're getting the "good measurements" claim from when there are other options that produce a similar or better result at a fraction of the price.
 
And as for the distortion test, what are you using to test it with? Headphones that don't distort in the bass are very rare, especially open headphones. You won't be able to discern bass distortion if you're used to hearing it all the time. My reference for bass is my Pioneer Monitor 10-II, which while not particularly perfect anywhere else, pushes a ridiculously clean, distortion-free bass with almost no variation from 300Hz to 10Hz.
 
EDIT: Can't seem to find what test you're talking about? I see no 80/120Hz or distortion percentages mentioned anywhere on the page you linked.

 
LOL, I've said all that I feel needs saying, so why don't we just agree to disagree  
 
On the Klippel page, in the section where it says "ready to go?" there's a blank box, you can select subwoofer for example, and then directly below that is another drop down box where you can select "two-tone 20hz 180hz." 

 

[npdang]

Sorry, try this link:  http://www.klippel-listeningtest.de/lt/

 

[plakat]

So distortion in the bass range seems somewhat high, still higher up it looks quite good to me. What does not correlate with what I hear is the impulse response (looking really bad actually) and the 300Hz square wave. I can live with that since I like their sound although it might explain why I never liked the pairing with the DAC2. Its measurements surely don't fit its price range -- but if they sound better to me than measurements might indicate thats better than the other way round.
 
I can also live with the fact that Tyll does not like them -- I value his opinion, but do not necessarily agree on taste.

 

[npdang]

  So distortion in the bass range seems somewhat high, still higher up it looks quite good to me. What does not correlate with what I hear is the impulse response (looking really bad actually) and the 300Hz square wave. I can live with that since I like their sound although it might explain why I never liked the pairing with the DAC2. Its measurements surely don't fit its price range -- but if they sound better to me than measurements might indicate thats better than the other way round.
 
I can also live with the fact that Tyll does not like them -- I value his opinion, but do not necessarily agree on taste.

 
The extended ringing in the time domain (ripples in the impulse response) is caused by the breakup of the diaphgram, which appears as a roughly +10db spike at 19khz in the frequency domain.  Whether or not that is audible, is debatable.  My experience is that it is audible, but as a slight "coolness" to the sound which some prefer, and others do not.  Using a high Q notch filter to suppress the breakup at 19khz would yield a much cleaner impulse response measurement.  If you have access to an EQ that allows you to easily dial in the Q and frequency, I would suggest trying a -10db cut at 19khz and see if you can hear any difference.  Also, the advantage of stiff diaphragms is that they tend to exhibit better distortion figures in the frequency regions where excursion is low (midrange/treble), and where the flexing of the cone plays a larger role in distortion performance, and I think we do see that here.  Where excursion requirements are higher (lower midrange/bass), the linearity of the motor dominates the distortion performance.  And in headphones, it also appears that the clamping force and cup design also plays a significant role.
 
With respect to the 300hz square wave, I can't hear phase distortion either (or at best its extremely subtle) so I'm not surprised you are unable to hear it as well.
 
Also, we do not have a complete distortion profile here, which is important because for example, many people find pleasant, or can tolerate, rather large amounts of 2nd and 3rd order (low order) distortion, (typical of single ended tube amplifiers), yet find objectionable extremely low levels of higher order products (5th order and above), that you might find in solid state amplifiers that otherwise measure overall very low levels of distortion but utilize significant negative feedback to achieve it.  Is the rising distortion in the lower midrange/bass region (assuming its accurate) composed of the former, or the latter? Who knows, and I think it's very important to make that determination before making any assumptions.  Intermodulation distortion (a problem for wideband drivers such as used in headphones, where a single driver is responsible for playing the entire frequency range) is also generally not tolerated well, and even small amounts can be easily audible.  A design like the Pioneers which incorporates a copper shorting ring atop the pole piece of the motor generally will generally reduce intermodulation distortion significantly, and obviously that is not something that is captured in Tyll's measurements.
 
So it does not surprise me in the least to find that despite the rather limited information presented in Tyll's distortion measurements, that you find these headphones to sound quite good.

 

[LFC_SL]

Got to answer my own question with Pioneer reps. Third party cables will not have to locking notch with the mmcx connectors but should still be fine without it