Skip to main content

Topic: Question about harmonic distortion in headphones (Read 474 times) previous topic - next topic

0 Members and 1 Guest are viewing this topic.
Question about harmonic distortion in headphones
I have the Sennheiser HD650 being run with the JDS Labs ODAC and O2. According to various measurements all over the internet, the HD650 along with the HD600, HD580, as well as the new HD800S, all have a rising slope of 2nd Harmonic distortion starting from about 100 Hz downward, ending at 20 Hz where it usually reads over 1-2% distortion at 90 dB. This is also common with many other dynamic driver headphones such as the Beyerdynamic DT 880 and T1, and the AKG K701. Usually in many popular planar magnetic headphones from Hifiman and Audeze, this distortion slope downward is lessened or just non-existent. This got me thinking, perhaps the distortion may also be a result of not just the driver type, but the pad type, as almost all those dynamic headphones I listed have velour pads, while most planar magnetic headphones have leather/pleather pads that seal better. That's just a guess, however, I still don't understand why there is significantly more 2nd order harmonic distortion in the bass on many of these dynamic drivers. Does this type of distortion really detract that much from perceived sound quality, and is it reason to go with a different driver type such as planar magnetic or electrostat? I know both planar and electrostat headphones are described as being "faster" and cleaner, but is that due to frequency response mostly or does this distortion actually affect the bass that significantly? I don't really understand much about the audibility of harmonic distortion in general, so any clarification on this subject would help.

Edit: also, I have the Klipsch X10, which is a BA driver IEM, and I can hear audible distortion in the 1 Khz area, even though graphs on innerfidelity show the distortion levels to be consistently under 1% at that frequency. I'm not sure if what I am hearing is actually harmonic distortion, but it doesn't sound like a pure tone at 1 Khz. How is it I can hear distortion in the 1-2 Khz range on an IEM under 1%, yet I can't hear the distortion at 20-100 Hz on the Sennheiser, which is over 1%?
  • Last Edit: 05 January, 2017, 05:16:42 PM by achelgeson1

  • DVDdoug
  • [*][*][*][*][*]
Re: Question about harmonic distortion in headphones
Reply #1
Quote
This got me thinking, perhaps the distortion may also be a result of not just the driver type, but the pad type, as almost all those dynamic headphones I listed have velour pads, while most planar magnetic headphones have leather/pleather pads that seal better. That's just a guess, however, I still don't understand why there is significantly more 2nd order harmonic distortion in the bass on many of these dynamic drivers.
I can't think of any reason the pads would cause or contribute to the distortion.   The seal can affect the bass (if they are closed/sealed headphones).

Quote
Does this type of distortion really detract that much from perceived sound quality, and is it reason to go with a different driver type such as planar magnetic or electrostat?
I'm not an expert on the audibility of distortion, but in general you shouldn't hear 1 or 2% harmonic distortion, especially with normal music, which contains lots of harmonics.   Other non-harmonic distortions tend to be more problematic.    And, a single-number for harmonic distortion doesn't tell you the intensity/distribution of the various harmonics.

I never hear distortion unless the amplifier or speakers are over-driven, or if something is defective.

Quote
I know both planar and electrostat headphones are described as being "faster" and cleaner, but is that due to frequency response mostly or does this distortion actually affect the bass that significantly?
I'd guess it's frequency response.  "Faster" and "cleaner" are not specs.    To me, these are just audiophile nonsense...  Words that give an impression and seem to have a meaning, but might mean different things to different people, and they can't be measured or quantified.      I assume  speed is related to transient response, but mathematically you can derive frequency response from impulse response, so they are related (and frequency response can  be measured & specified).    "Clean" would seem  to relate to distortion, but boosted high frequencies might also be described as sounding "clean" or "clear", or we often say "bright".   An Aphex Aural Exciter effect might make the sound "cleaner", but it's actually adding harmonics...

Quote
Edit: also, I have the Klipsch X10, which is a BA driver IEM, and I can hear audible distortion in the 1 Khz area, even though graphs on innerfidelity show the distortion levels to be consistently under 1% at that frequency. I'm not sure if what I am hearing is actually harmonic distortion, but it doesn't sound like a pure tone at 1 Khz.
We'd only be guessing about what you're hearing, but it's possible that it's non-harmonic distortion.      It's also possible, but unlikely, that your headphone amplifier (whatever you're driving them with) is distorting with the IEMs but not with the headphones.   Or, it might just be a perception you're getting with IEMs.   Speakers, headphones, and IEMs are all "different experiences" and it's not always easy to describe the differences.
  • Last Edit: 05 January, 2017, 05:54:22 PM by DVDdoug

Re: Question about harmonic distortion in headphones
Reply #2
Thanks for the reply. So, I guess I probably shouldn't worry too much about distortion, at least harmonic distortion. When it comes to the Klipsch X10 however, I think you're right that what I am hearing is not harmonic distortion. I'm also hearing the same distortion out of my iPod, computer jack, and O2 amp, so it's not the amp. It's also only noticeable while playing sine waves, not listening to music. My guess is the balanced armature design in the Klipsch is not done very well, and there may be something in there vibrating when it's not supposed to. I hope this problem is only present in my particular IEM, and not in other BA driver IEMs, because I was planning on getting the Etymotic ER4SR.

  • eric.w
  • [*]
Re: Question about harmonic distortion in headphones
Reply #3
I'm not so sure harmonic distortion in balanced armature IEM's is inaudible, at least on sine waves.

This measurement of the ER4SR shows ~0.5% THD at 1kHz, and it's almost all 3rd harmonic:
http://clarityfidelity.blogspot.ca/2016/09/etymotic-research-er4sr-iem.html
(That doesn't seem unusual, looking at other BA IEM's on Speakerphone and Rin Choi's blog)

Here is an experiment I tried: In Audacity, generate a 1kHz sine wave, then a 3kHz sine wave in a separate track. Select the 3kHz track and amplify it by -46dB (attenuation factor for 0.5% THD according to this calculator: http://www.sengpielaudio.com/calculator-thd.htm ). Now play the two tracks together, and mute/unmute the 3kHz track. Spoiler: the -46dB 3kHz tone should be very easy to hear, even at low volume levels.

Did I mess up anything, or should that simulate the ER4SR's distortion when playing back a 1kHz tone?
Blind testing with a music sample would be more interesting... anyone know of a tool that can apply distortion to a music sample, according to a 2nd/3rd/4th/etc. harmonic distortion % vs. frequency graph?
  • Last Edit: 11 January, 2017, 12:32:47 AM by eric.w

Re: Question about harmonic distortion in headphones
Reply #4
I'm not so sure harmonic distortion in balanced armature IEM's is inaudible, at least on sine waves.

This measurement of the ER4SR shows ~0.5% THD at 1kHz, and it's almost all 3rd harmonic:
http://clarityfidelity.blogspot.ca/2016/09/etymotic-research-er4sr-iem.html
(That doesn't seem unusual, looking at other BA IEM's on Speakerphone and Rin Choi's blog)

Here is an experiment I tried: In Audacity, generate a 1kHz sine wave, then a 3kHz sine wave in a separate track. Select the 3kHz track and amplify it by -46dB (attenuation factor for 0.5% THD according to this calculator: http://www.sengpielaudio.com/calculator-thd.htm ). Now play the two tracks together, and mute/unmute the 3kHz track. Spoiler: the -46dB 3kHz tone should be very easy to hear, even at low volume levels.

Did I mess up anything, or should that simulate the ER4SR's distortion when playing back a 1kHz tone?
Blind testing with a music sample would be more interesting... anyone know of a tool that can apply distortion to a music sample, according to a 2nd/3rd/4th/etc. harmonic distortion % vs. frequency graph?

Wow, yeah that experiment worked for me. Using the HD650, I can clearly hear it. I am most definitely hearing 3rd order harmonic distortion in the Klipsch X10. It's so easy to hear too, probably since it's an odd harmonic and not even. That worries me, since the ER4SR shows above 0.5% 3rd harmonic distortion around 1 kHz. Well, I guess I'm sticking with dynamic driver IEMs...