[headr]

We are all witnessing a huge "The emperor has no clothes" moment in online headphone reviews. Every reviewer who swore up and down that the HD 800S had audible distortion in the bass is now put on notice.

I do wonder how many people convinced themselves that the HD 800 bass was absolutely distortion free, and hence since the HD 800 S sounded different that they felt that must mean the HD 800 S was distorted.

Sounding different does not mean the sound is distorted. I suspect many people were simply not used to hearing the bass clarity of the HD 800 S in as resolving and detail oriented a headphone as the HD 800 S, and thus leapt to the conclusion that the bass presentation was distorted simply because it was not something they were used to in a headphone context.

Time to go back and look at all the reviewers who claimed distortion. I am starting my list...

 

[Whazzzup]

Well thank goodness the witch hunt will leave me out. I just waited for hd800S to be released in canada and bought it on Boxing Day sale. The 800S has been my go to ever since

 

[Mediahound]

We are all witnessing a huge "The emperor has no clothes" moment in online headphone reviews. Every reviewer who swore up and down that the HD 800S had audible distortion in the bass is now put on notice.

I do wonder how many people convinced themselves that the HD 800 bass was absolutely distortion free, and hence since the HD 800 S sounded different that they felt that must mean the HD 800 S was distorted.

Sounding different does not mean the sound is distorted. I suspect many people were simply not used to hearing the bass clarity of the HD 800 S in as resolving and detail oriented a headphone as the HD 800 S, and thus leapt to the conclusion that the bass presentation was distorted simply because it was not something they were used to in a headphone context.

Time to go back and look at all the reviewers who claimed distortion. I am starting my list...

There are reviewers (or a reviewer) who seem to be consistently wrong on things.

Not to toot my own horn or anything, but note I am not one who ever stated (or believed or heard) that the HD800S had more bass distortion than the HD800:

 

[wldcohso]

Yeaaaahhh booyiieeeee My Sennheiser HD 800S are in route to me, should have them by Saturday. My second time purchasing these... thanks to all the myth busters regarding this wonderful piece of art.

 

[headr]

There are reviewers (or a reviewer) who seem to be consistently wrong on things.

Not to toot my own horn or anything, but note I am not one who ever stated (or believed or heard) that the HD800S had more bass distortion than the HD800:

If you are a reviewer who correctly recognized that there wasn't bass distortion in the HD 800 S, now is the time to toot your own horn. It isn't easy to have an opinion in this field that goes against the grain, or that seemingly goes against people's widespread (and incorrect) assumptions. Sometimes having the fortitude to stand up and say things that are plainly true, especially when they go against the echo chamber, takes gusto.

In addition to a "naughty" list I encourage people to go back and look for those brave souls who dared say they didn't hear bass distortion with the HD 800 S. These are the honest voices we should be listening too...

 

[jude]

We are all witnessing a huge "The emperor has no clothes" moment in online headphone reviews. Every reviewer who swore up and down that the HD 800S had audible distortion in the bass is now put on notice.

I do wonder how many people convinced themselves that the HD 800 bass was absolutely distortion free, and hence since the HD 800 S sounded different that they felt that must mean the HD 800 S was distorted.

Sounding different does not mean the sound is distorted. I suspect many people were simply not used to hearing the bass clarity of the HD 800 S in as resolving and detail oriented a headphone as the HD 800 S, and thus leapt to the conclusion that the bass presentation was distorted simply because it was not something they were used to in a headphone context.

Time to go back and look at all the reviewers who claimed distortion. I am starting my list...

Again, we've only measured two HD800's, and they were early production units. We need to measure more. I think there's a brand new, unopened HD800 on its way; and another local used HD800 that has had no driver mods, but was modified to directly wire the drivers (which I'll make sure to note about that headphone in the measurement).

The only firm conclusions I've come to so far:

• The two Sennheiser HD800's we have measured do not have H2 that's lower than the H3 as shown in InnerFidelity's FFT, and do not measure (in terms of THD) that different from the two HD800S's we've measured (the most recent one being a brand new, previously unopened, unused HD800S yesterday).
  
  
• Our measurements seem consistent with feedback from Sennheiser.

Going back to @daltonlanny's previous post (which I responded to): I think the following article is one everyone interested in audio measurements should read. It was written by Dan Foley of Audio Precision (a frequent visitor to Head-Fi HQ, and one of several of my audio measurement mentors for over two years). Dan is also the current president of the Association of Loudspeaker Manufacturing & Acoustics (ALMA International). Here's a link:

Test and Measurement: “I Can Hear It. Why Can’t I Measure It?”

(The article was originally posted in audioXpress Focus on Test & Measurement Special issue, March 2016.)

Please read the whole thing, but here are some relevant excerpts from the article:

...Dan Foley discusses how audible distortion can be difficult to measure and explores the instrumentation noise floor’s impact on what can be measured. As he explains, instrumentation noise floor is an overlooked contributor for the reason we are not able to measure a particular audible “distortion,” especially with low sound pressure level (SPL) signals.

In many cases, the reason given as to why a particular audible “distortion” cannot be measured is that the measurement equipment doesn’t have sufficient signal processing capabilities, such as measurement bandwidth or Fast Fourier Transform (FFT) resolution. In this article, we will explore an overlooked contributor — the instrumentation noise floor, which can have a significant impact on what can be measured, especially with low sound pressure level (SPL) signals.

Every part of the measurement signal path introduces noise, beginning with the microphone capsule and its corresponding preamplifier. The IEC 711 couplers that are used in Head-and-Torso Simulators (HATS) test manikins and equivalent products typically incorporate a 0.5” microphone with a 12.5 mV/Pa sensitivity or −38 dB referenced to 1 pascal.

Given that 1 pascal is equivalent to 94 dBSPL, the lowest SPL one can potentially hear around 3.2 kHz is −5 dBSPL. At 94 dBSPL the voltage from the microphone is 12.5 mV. At −5 dBSPL, the voltage level will be 99 dB below this value, which is 141 nanovolts!

Instrumentation noise should be 10 dB below the level one wants to measure. In this case, the noise level in this frequency range would need to be 47 nanovolts, but it does not mean the instrumentation noise floor over the entire audio band needs to be 47 nanovolts. But, the lower the noise floor is from 20 Hz to 20 kHz, the better you can measure these extremely low voltages.

It's worth a read, and my point in bringing it up is that I do not believe you can turn to most of the DIY headphone measurement rigs (photos of which I included in my last post) to have a level of precision that's going to get us closer here.

We'll post more HD800 measurements as the other units come in.

 

[csglinux]

If you are a reviewer who correctly recognized that there wasn't bass distortion in the HD 800 S, now is the time to toot your own horn. It isn't easy to have an opinion in this field that goes against the grain, or that seemingly goes against people's widespread (and incorrect) assumptions. Sometimes having the fortitude to stand up and say things that are plainly true, especially when they go against the echo chamber, takes gusto.

In addition to a "naughty" list I encourage people to go back and look for those brave souls who dared say they didn't hear bass distortion with the HD 800 S. These are the honest voices we should be listening too...

I think there's a bit too much conspiracy theory creeping in here. The nature of measurements is that they have errors, and they can be really difficult to quantify. Personally, I never trust error bars, because you can't possibly know what the errors were in the error bars.

I don't think anybody is saying the HD800S is free of distortion in the bass. Jude's measurements have simply shown (meticulous job, BTW @jude!) that the levels of H2 from a 40 Hz sine wave appear to be comparable between HD800 and HD800S. The addition of the SuperDupont resonator mod might have been enough to slightly elevate overall levels of error in the low frequencies, simply because they would now have a greater contribution to the overall sound pressure level.

I've long pondered on how one might specifically add 2nd harmonic distortion in a passive driver component. It may be possible using some cleverly engineered anisotropic materials (that favor resonance at a particular mode), but I'm not sure how you'd prevent that from having unpleasant artifacts all the way up the frequency spectrum. In any event, Sennheiser seem to have quashed that idea themselves. I don't think Innerfidelity committed any terrible sin here - they saw something on their measurements and speculated. We all have the right to be mistaken. It's possible (by design or happy accident) that both HD800 and HD800S have a mild version of Tacoma Narrows at some low frequency. Here's a little video to show what can happen when you trap energy in a particular resonant mode:



Bose implement an acoustic version of this with their waveguides - long ports tuned to trap energy at a particular low frequency that gives that nice satisfying boom that all the kids like.

So I wonder whether the only real difference between HD800 and HD800S is that resonator mod? I can understand Sennheiser not wanting to reveal their secrets. But sometimes the secret is that there is no secret

 

[Whazzzup]

Its black magic voodoo type stuff

 

[ubs28]

In another thread here at Head-Fi, someone said that harmonic distortion was deliberately added to the Sennheiser HD800S (versus the HD800). That statement was only one of many (made in the past couple of years) stating the same thing. I had also assumed this to be true, only because it has been repeated so many times here and elsewhere that it's become a generally accepted truth about the HD800S. But is it, in fact, true?

Last summer, while doing measurements of several headphones, I measured the Sennheiser HD800 and Sennheiser HD800S on the GRAS 45BB-12. As I had also assumed the aforementioned theory to be true, I fully expected to see higher 2nd order harmonic distortion with the HD800S versus the HD800. However, any increase in harmonic distortion was minor (HD800S versus HD800), and I didn't see anything in the distortion product ratio plots that supported the theory / assumption. Further, when I brought up this topic with Sennheiser's Axel Grell in London (at CanJam), he was not surprised by our measurement results, as he said there was no deliberate distortion added to the HD800S.

Things got busy at the show, and I hadn't given that specific topic a lot of thought since, until I saw it stated as fact again the other day in another thread, where a few of us then discussed it briefly (see quotes immediately below):

---

So I decided to examine this further here, with the HD800 and HD800S we have, and using a completely different measurement head fixture than I used when I did the measurements last summer (details of the measurement system used can be found at the bottom of this post). Here's what I did:

---

• Guided by the FFT plot at InnerFidelity, where we first read about this theory / assumption, I first I did a quick FFT plot of the HD800 (16,000 points, no averages), playing a 40 Hz sine wave. I ended up with a plot that looked somewhat like this:
  
  
  
  As you can see, that's not particularly detailed, so I took advantage of our audio analyzer's ability to do extremely detailed FFT's, opting for 1.2 million points, and doing three averages. Setting it to 1.2 million points provides a much more detailed plot, and the averages help remove random noise (better revealing the actual noise floor, too). Without having moved the HD800 from the above plot, here's the FFT re-run with 1.2 million points with three averages:
  
  
  
  And here are the two overlaid:
  
  
  
  NOTE: The examples above are actual measures of the Sennheiser HD800, but these were just quick preliminary measurements done without turning off the building's heating / ventilation systems. Even though we do employ a lab-grade acoustic and vibration isolation enclosure, the measurement gear is obviously very sensitive, so we still turn off the HVAC systems when we're doing audio measurements to lower the noise floor. As a result, the example measurements above may have a higher noise floor than the measurements that follow (which were done with the building's HVAC systems turned off).

• Again, the above measurements were just done as examples of two different FFT resolutions. Now let's get to the actual comparisons of our HD800 and HD800S. Following are FFT spectrum measurements from the first seatings of the Sennheiser HD800 and Sennheiser HD800S, set to 1.2 million points and three averages, and playing a 40 Hz sine wave. I made the plot lines a bit thicker (and different colors) to try to make the two plots easier to distinguish from one another:
  
  
  
  Here (below) is the same plot with cursor lines showing the difference at 80 Hz (80 Hz being the second harmonic of the 40 Hz sine wave fundamental tone, since we're examining the theory that the Sennheiser HD800S's richer sound is due to (as quoted above) deliberately added 2nd harmonic overtones/distortion):
  
  
  
  The difference between the Sennheiser HD800 and Sennheiser HD800S at 80 Hz (H2) is only 2.783 dBSPL. Keep in mind that the center of this 2.783 dBSPL difference is occurring around 45 dBSPL lower than the fundamental tone. While the Sennheiser HD800S unit we have here does (to my ears) sound like other HD800S's I've heard (richer than the Sennheiser HD800), I do not believe this measured difference at 80 Hz suggests deliberately added 2nd harmonic overtones/distortion (to the HD800S versus the HD800).
  
  
• I decided to do a second seating of both headphones, also increasing the size of the headband for both headphones by one click on each side. Here (below) are the FFT spectrum plots from the second seating:
  
  
  
  Again, this time (below) with cursor lines showing the difference between the two in this seating at 80 Hz (the second harmonic of the 40 Hz fundamental test tone):
  
  
  
  The difference in this plot at 80 Hz is lower than the previous one, measuring only 1.673 dBSPL. As above, the center of this difference is about 45 dBSPL below the 40 Hz fundamental tone.

---

Again, keep in mind we only have one Sennheiser HD800 and one Sennheiser HD800S here at this time. I may measure more. (NOTE 2018-01-02: I did measure another HD800 later, and you can see it by clicking on the following link: Additional Sennheiser HD800 FFT measurements.)

Based on discussions with Sennheiser and the above measurements of the Sennheiser HD800 and HD800S we have on hand, it does not appear there was any deliberate addition of 2nd order harmonic overtones/distortion to the HD800S as has been the common theory for the past couple of years.

If you have any other suggested tests you'd like to see performed on these two units here that you think might generate different results, let me know. That said, based on the THD measurements I posted in this thread earlier, and other measurements we've performed on these two headphones (also on the GRAS 45BB-12), I'm not surprised by the above results.

The measurements in this post were made using:

• GRAS 45CA Headphone Test Fixture with:
  • GRAS RA0401 High Resolution Ear Simulators
  • GRAS KB5010/KB5011 Anthropometric Pinnae for 45CA
• Audio Precision APx555 Audio Analyzer
• Audio Precision APx1701 Transducer Test Interface
• Herzan custom acoustic and vibration isolation enclosure

---

NOTE 2018-01-02 00:56 EST: I measured a second Sennheiser HD800 and (with those measurements) offer a possible explanation for how this theory of deliberately-added higher 2nd order distortion in the Sennheiser HD800S may have come about. Long story short, our measurements so far do not support the assumption that Sennheiser added 2nd order harmonic distortion to the HD800S, which is also consistent with feedback from Sennheiser. You can find out more at the following link: Additional Sennheiser HD800 FFT measurements, as well as a closer look at the measurement that started this assumption nearly two years ago.

Glad to know my post started this. But according to my ears, there is a difference in the bass between the HD 800 and HD 800 S.

But do this following test:
1. Measure the frequency response of the HD 800
2. Measure the frequency response of the HD 800 S
3. EQ the HD 800 to exactly match the HD 800 S in terms of frequency response.
4. Listen if the HD 800 sounds exactly the same as the HD 800 S.

If this is not possible, then do the following.
1. Send the HD 800 to Sonarwork and let them calibrate a profile for the HD 800
2. Send the HD 800S to Sonarwork and let them calibrate a profile for the HD 800 S.
3. Listen if the HD 800 sounds exactly the same as the HD 800 S using the calibrated profiles provided by Sonarwork.

If there is still an audible difference, it would point that Tyll his measurements maybe correct. If not, then yours are more correct.

 

[amartignano]

Again, it's the HD800 measurement in the FFT I'm most interested in, for a couple of reasons. First of all, I think it's the one from which this discussion originated. I'm not sure which of the HD800 serial numbers it was (as both of the ones from that post's measurement booklet do show as having been modified, one of them several times). Second, that FFT shows the HD800 as having higher H3 than H2, and it's labeled "HD 800 DP Mod" (see the red and white arrows below), and this (H3 > H2) seems to be the reason for the biggest difference (and the resultant theory).



As I've said before, I am not saying the HD800 in this FFT was modded -- I'm only pointing out that it was marked as "HD 800 DP Mod," so I think it reasonable to point out.

If you look at the HD800 that is being modded (in the video) with the resonator mod, the HD800 had seen at least one prior round of modifications (as mentioned around 6:45 in the video), and as is clear from the view around the driver:



Here's one that has had no driver or damping modifications performed on it (S/N 00342):



Again, while the FFT in question says "HD 800 DP Mod," I can not confirm whether or not that measurement reflects a modified HD800, a stock-and-never-before-modified HD800, or a modified-and-restored-to-stock HD800. Only InnerFidelity can clarify that.

As I said, so far we have measured two HD800's, both of which have noticeably higher H2 than H3 at varying drive levels. More are incoming, and we'll measure them all -- and perhaps we'll come across one that also has higher H3 than H2.

As for the other InnerFidelity HD800 measurements you linked to, I noticed something in common with all of them (whether modified or unmodified HD800's), in varying degrees, but never insignificant. In all of them there are large spikes at 200 Hz and 2 kHz. Here's what I'm talking about (arrow pointing to the spikes in question):




I am rather certain those are not from the headphone, but from the system, and I do not know how (if at all) whatever is causing those spikes would affect the THD readings. Also, they do not appear on the FFT.

---

Almost all of the other measurements I have seen have come from DIY measurement systems, most probably having been generated on one of the systems in the many DIY measurement system photos shown below (and in the following post): DIY headphone measurement systems



While I think it's great that people are building rigs for measuring headphones, when it comes to trying to figure this out, I do not believe those systems are up to the standards of precision that should be expected when measuring the THD of a headphone like the Sennheiser HD800 for anything other than personal information or casual discussion.

Tyll explained the peaks in the THD measurements on Innerfidelity. The peaks are artifacts caused by the system switching ranges.
It's explained here: https://www.innerfidelity.com/conte...d-total-harmonic-distortion-plus-noise-part-2

 

[elNan]

can we forget about graphs and enjoy some good music in these amazing cans!?

 

[casper3127]

I usually buy my cans so as to, just and simply, listen to music right out of them. So easy and -usually- enjoyable.

 

[exdmd]

InnerFidelity has gone completely bonkers. Next month Tyll will find a problem with the Clears and take them off the WoF. I am just going to keep enjoying my HD800S.

 

[Snowpuppy77]

In another thread here at Head-Fi, someone said that harmonic distortion was deliberately added to the Sennheiser HD800S (versus the HD800). That statement was only one of many (made in the past couple of years) stating the same thing. I had also assumed this to be true, only because it has been repeated so many times here and elsewhere that it's become a generally accepted truth about the HD800S. But is it, in fact, true?

Last summer, while doing measurements of several headphones, I measured the Sennheiser HD800 and Sennheiser HD800S on the GRAS 45BB-12. As I had also assumed the aforementioned theory to be true, I fully expected to see higher 2nd order harmonic distortion with the HD800S versus the HD800. However, any increase in harmonic distortion was minor (HD800S versus HD800), and I didn't see anything in the distortion product ratio plots that supported the theory / assumption. Further, when I brought up this topic with Sennheiser's Axel Grell in London (at CanJam), he was not surprised by our measurement results, as he said there was no deliberate distortion added to the HD800S.

Things got busy at the show, and I hadn't given that specific topic a lot of thought since, until I saw it stated as fact again the other day in another thread, where a few of us then discussed it briefly (see quotes immediately below):

---

So I decided to examine this further here, with the HD800 and HD800S we have, and using a completely different measurement head fixture than I used when I did the measurements last summer (details of the measurement system used can be found at the bottom of this post). Here's what I did:

---

• Guided by the FFT plot at InnerFidelity, where we first read about this theory / assumption, I first I did a quick FFT plot of the HD800 (16,000 points, no averages), playing a 40 Hz sine wave. I ended up with a plot that looked somewhat like this:
  
  
  
  As you can see, that's not particularly detailed, so I took advantage of our audio analyzer's ability to do extremely detailed FFT's, opting for 1.2 million points, and doing three averages. Setting it to 1.2 million points provides a much more detailed plot, and the averages help remove random noise (better revealing the actual noise floor, too). Without having moved the HD800 from the above plot, here's the FFT re-run with 1.2 million points with three averages:
  
  
  
  And here are the two overlaid:
  
  
  
  NOTE: The examples above are actual measures of the Sennheiser HD800, but these were just quick preliminary measurements done without turning off the building's heating / ventilation systems. Even though we do employ a lab-grade acoustic and vibration isolation enclosure, the measurement gear is obviously very sensitive, so we still turn off the HVAC systems when we're doing audio measurements to lower the noise floor. As a result, the example measurements above may have a higher noise floor than the measurements that follow (which were done with the building's HVAC systems turned off).

• Again, the above measurements were just done as examples of two different FFT resolutions. Now let's get to the actual comparisons of our HD800 and HD800S. Following are FFT spectrum measurements from the first seatings of the Sennheiser HD800 and Sennheiser HD800S, set to 1.2 million points and three averages, and playing a 40 Hz sine wave. I made the plot lines a bit thicker (and different colors) to try to make the two plots easier to distinguish from one another:
  
  
  
  Here (below) is the same plot with cursor lines showing the difference at 80 Hz (80 Hz being the second harmonic of the 40 Hz sine wave fundamental tone, since we're examining the theory that the Sennheiser HD800S's richer sound is due to (as quoted above) deliberately added 2nd harmonic overtones/distortion):
  
  
  
  The difference between the Sennheiser HD800 and Sennheiser HD800S at 80 Hz (H2) is only 2.783 dBSPL. Keep in mind that the center of this 2.783 dBSPL difference is occurring around 45 dBSPL lower than the fundamental tone. While the Sennheiser HD800S unit we have here does (to my ears) sound like other HD800S's I've heard (richer than the Sennheiser HD800), I do not believe this measured difference at 80 Hz suggests deliberately added 2nd harmonic overtones/distortion (to the HD800S versus the HD800).
  
  
• I decided to do a second seating of both headphones, also increasing the size of the headband for both headphones by one click on each side. Here (below) are the FFT spectrum plots from the second seating:
  
  
  
  Again, this time (below) with cursor lines showing the difference between the two in this seating at 80 Hz (the second harmonic of the 40 Hz fundamental test tone):
  
  
  
  The difference in this plot at 80 Hz is lower than the previous one, measuring only 1.673 dBSPL. As above, the center of this difference is about 45 dBSPL below the 40 Hz fundamental tone.

---

Again, keep in mind we only have one Sennheiser HD800 and one Sennheiser HD800S here at this time. I may measure more. (NOTE 2018-01-02: I did measure another HD800 later, and you can see it by clicking on the following link: Additional Sennheiser HD800 FFT measurements.)

Based on discussions with Sennheiser and the above measurements of the Sennheiser HD800 and HD800S we have on hand, it does not appear there was any deliberate addition of 2nd order harmonic overtones/distortion to the HD800S as has been the common theory for the past couple of years.

If you have any other suggested tests you'd like to see performed on these two units here that you think might generate different results, let me know. That said, based on the THD measurements I posted in this thread earlier, and other measurements we've performed on these two headphones (also on the GRAS 45BB-12), I'm not surprised by the above results.

The measurements in this post were made using:

• GRAS 45CA Headphone Test Fixture with:
  • GRAS RA0401 High Resolution Ear Simulators
  • GRAS KB5010/KB5011 Anthropometric Pinnae for 45CA
• Audio Precision APx555 Audio Analyzer
• Audio Precision APx1701 Transducer Test Interface
• Herzan custom acoustic and vibration isolation enclosure

---

NOTE 2018-01-02 00:56 EST: I measured a second Sennheiser HD800 and (with those measurements) offer a possible explanation for how this theory of deliberately-added higher 2nd order distortion in the Sennheiser HD800S may have come about. Long story short, our measurements so far do not support the assumption that Sennheiser added 2nd order harmonic distortion to the HD800S, which is also consistent with feedback from Sennheiser. You can find out more at the following link: Additional Sennheiser HD800 FFT measurements, as well as a closer look at the measurement that started this assumption nearly two years ago.

Such an excellent post. Thank you so much for clearing this up. I do not hear any distortion in the bass on my HD800S. Now when I switched my amp from an Asgard 2 to a Ragnarok (balanced cable) the bass did tighten up considerably giving more definition and detail.

 

[MacedonianHero]

We are all witnessing a huge "The emperor has no clothes" moment in online headphone reviews. Every reviewer who swore up and down that the HD 800S had audible distortion in the bass is now put on notice.

I do wonder how many people convinced themselves that the HD 800 bass was absolutely distortion free, and hence since the HD 800 S sounded different that they felt that must mean the HD 800 S was distorted.

Sounding different does not mean the sound is distorted. I suspect many people were simply not used to hearing the bass clarity of the HD 800 S in as resolving and detail oriented a headphone as the HD 800 S, and thus leapt to the conclusion that the bass presentation was distorted simply because it was not something they were used to in a headphone context.

Time to go back and look at all the reviewers who claimed distortion. I am starting my list...

This urban myth has perplexed me actually when others mentioned it. Both offer outstanding bass definition and clarity.