[castleofargh]

 

  It would be more accurate to say "some" distortion is audible, as I think it would not be too hard to create samples where 0.02% distortion is inaudible in blind testing.

 
.02% is below -70dB. I seriously doubt that any human being would ever be able to discern that under music. Ethan Winer in his Audio Myths seminar takes the absolute worst kind of noise and takes it down under music a little bit at a time. At the point the noise reaches -40dB or so... which is 1%... it starts to become impossible to hear.
 
If you are interested in finding out how the numbers being quoted actually relate to what your ears can hear, I recommend his two videos at youtube. He does several tests that are eye opening (and ear opening too!)
 
AES Audio Myths Seminar: http://youtu.be/BYTlN6wjcvQ
AES Damn Lies Seminar: http://youtu.be/Zvireu2SGZM

it's hard to give a general value I think. first because as mentioned, there is distortion, and distortion. but also because some headphones have distortions going up with loudness, and others where it's the opposite. so 2 headphones with the same distortion readings at 90db might not have the same at all on the quiet part of a song.
now I know that 1% is the generally accepted lvl for headphones, but I did tend to dislike most headphones with 1% disto in the medium frequencies. before knowing about the measurements for most, but I'll admit that I might have been biased for a few others where I already knew about the distortion measurements and might have pre-disliked the headphone for it. still I think 0.5% would be a more realistic value to stop worrying. and about bass distortion, not only it doesn't bother me, I actually seem to enjoy it a lot on some headphones where it rises very high in bass and sub bass. so there is definitely hearing distortion, and being bothered by it, just like with tube amps.

 

[bigshot]

Harmonic distortion follows the dynamics of the music by definition, doesn't it? If it is constant through silent parts it would be considered noise. Citing the Ethan Winer noise test I linked to above... If constant noise becomes extremely difficult to hear below -40dB except in complete silence, I would expect that harmonic distortion following the dynamics of the music would be even more difficult to hear, no matter what kind of sound the distortion is.
 
I think the generally accepted "acceptable levels" of THD you see quoted in audiophile forums have had the worst case of worst case scenarios applied, and then bumped down a few notches more "just to be safe" several times. ("OH! If this guy says .05% is good, I WANT .01% just to be safe!")
 
I honestly can't see anyone needing averaged THD ratings below 1% unless they are in a studio situation and are going to do a lot of processing of the signal. Decent amps and CD players are so far below that now, it doesn't really matter anymore. And speakers can have distortion levels of 3 or 4% THD and still sound darn good. Distortion just isn't generally an issue, except with really cheap transducers.

 

[Strangelove424]

These studies (at the bottom) have it cited pretty high, mostly above 1%. http://www.audioholics.com/room-acoustics/human-hearing-distortion-audibility-part-3
 
Do you guys know if harmonic distortion sounds the same on each play through? Or is there enough variation in it to trigger a slight difference between two playbacks? (trying to track down the source of noise/distortion in difference files I'm making in the cable thread)

 

[swspiers]

   
 
I think the generally accepted "acceptable levels" of THD you see quoted in audiophile forums have had the worst case of worst case scenarios applied, and then bumped down a few notches more "just to be safe" several times. ("OH! If this guy says .05% is good, I WANT .01% just to be safe!")
 
I honestly can't see anyone needing averaged THD ratings below 1% unless they are in a studio situation and are going to do a lot of processing of the signal. Decent amps and CD players are so far below that now, it doesn't really matter anymore. And speakers can have distortion levels of 3 or 4% THD and still sound darn good. Distortion just isn't generally an issue, except with really cheap transducers.

It's easy to buy into this.  Heck, I did it for over twenty years, and I have certification in studio engineering (albeit from 1984.)
 
It's one of the insidious aspects of the subjective mindset- there are numbers attached, so it must be scientific.  It takes discipline and intent to even begin to wonder if the numbers are meaningful in any way

 

[bigshot]

  Do you guys know if harmonic distortion sounds the same on each play through? Or is there enough variation in it to trigger a slight difference between two playbacks? (trying to track down the source of noise/distortion in difference files I'm making in the cable thread)

 
Should be the same unless it is the result of something heating up and acting differently hot than cold.
 
What kind of amount of distortion are you finding? Are you doing any processing of the files?

 

[Strangelove424]

   
Should be the same unless it is the result of something heating up and acting differently hot than cold.
 
What kind of amount of distortion are you finding? Are you doing any processing of the files?

Yeah, I had to downsample from 88.2 to 44.1 to use in DiffMaker, and then DiffMaker automates level normalization, waveform matching, and number of other parameters to extract the A-B.
 
esldude had this to say: "In any case, been thinking on it more and can tell you why you  hear remnants of the music if you amp it up by 50 db or more.  When you normalize files that are identical other than some noise, the noise causes the normalizing process to get the signal portion lower in the noisier file.  What this means when you difference them is instead of getting only noise and none of the original signal, you get some noise and a very low level residual of the original signal which if amped enough you hear."
 
I originally was thinking it might have been caused upstream of the cable, like if the reviewer's amp was distorting the source file (it was a very badly compressed sample) then that might pop up in all the samples. But if esldude is right, it's just a byproduct of the differencing process. Regardless, the noise differences were so slight between the cables and so far below the audible noise floor that for all practical purposes there was no difference.  

 

[bigshot]

Today, I got a Spears and Munsill calibration disk from Amazon Prime and sat down to do to my projector and blu-ray player what I've done with my sound system. Amazingly enough, my eyeballed settings were VERY close to calibrated standard. The only thing I was off on was sharpness. (I need glasses!) The brightness, contrast and color all ended up in the same spot that I had them in. It's interesting to try out a systematic calibration on something other than audio. This disk is very well organized and made the process very easy. It has audio calibration tools too. I'm going to look those over, but I am too satisfied with my current settings to mess with them.

 

[Grave]

There is an obvious difference between headphones with 1% distortion and 0.1% distortion. That's like the difference between what audiophiles call "low-fi" and hi-fi" right there.

 

[swspiers]

Biggie: What TV do you have?

 

[swspiers]

There is an obvious difference between headphones with 1% distortion and 0.1% distortion. That's like the difference between what audiophiles call "low-fi" and hi-fi" right there.

Agreed. That decimal point is right there, sure enough!

 

[Grave]

If you think all headphones sound the same in terms of distortion that is just sad.

 

[castleofargh]

  If you think all headphones sound the same in terms of distortion that is just sad.

if you put it on a scale of signal variation instead of what's nice or not, audible or not. then 1%distortion is nothing compared to a +1DB boost in FR somewhere.
-1db would be what a 90% variation? ^_^
that's why I clearly understand people who care mostly about FR and only check that the rest is below audible/disturbing level.
to me the main differences, between headphones come from the driver tech, if it's closed or opened, often I seem to favor larger drivers but I can't really say what impact that has on sound so let's dismiss it here. and really what it comes down to for me is when it rolls off and how fast? just like speakers. for IEMs I'm resigned to have a crappy frequency range, but for fullsize headphones I really do care mostly about FR and roll off. and roll off could really be counted as FR so here you have it, I care really just for FR ^_^.
 
 

  Harmonic distortion follows the dynamics of the music by definition, doesn't it? If it is constant through silent parts it would be considered noise. Citing the Ethan Winer noise test I linked to above... If constant noise becomes extremely difficult to hear below -40dB except in complete silence, I would expect that harmonic distortion following the dynamics of the music would be even more difficult to hear, no matter what kind of sound the distortion is.

 I understand what you mean, if the distortion creates a signal at some frequency -40below the original signal you expect a proportional response with music.
but look at Tyll's measurements and how much the distortions values can change from a test signal at 90db to a test signal at 100db.
 I take what I often see with ortho, that getting louder seems to make the distortions go down. if I get 0.3% at Xkhz at 90db from a headphone, then what lvl of distortion am I to expect from a music signal at 40db on a quiet part of a song(quiet not silent)?
that was the meaning of my post, is the harmonic distortion created still close to 0.3% of the original signal? so 50db under the original 40db loud music(we can as well say nonexistent)? or does the distortion grow massive in proportion to the original signal?
if the distortion values can go from 0.1 to 0.5 from a 10db variation in the test signal, I'm really curious to know what can result of something 40 or 50db lower. (the fun part is that if the distortions values are massive, it makes the high res fans and their "96db of dynamic isn't enough", even less credible ^_^).
 
here is a random example of what I mean http://www.innerfidelity.com/images/AudezeLCD2sn53211704circa2012.pdf
 plz someone call the Tyll police so that he can come save me, and tell us what we should expect from a signal at 50 or 60DB instead of the usual 90 and 100db distortion readings. and if distortion stays proportional or if the driver control can sometimes just crumble under and below some values?

 

[bigshot]

Biggie: What TV do you have?

I have a Epson 7500 UB with a ten foot screen.

 

[bigshot]

  If you think all headphones sound the same in terms of distortion that is just sad.

The ones over 1% THD I would check out before buying. They might not be good. But response is generally much more of an issue than distortion.

 

[bigshot]

  but look at Tyll's measurements and how much the distortions values can change from a test signal at 90db to a test signal at 100db.

 
100dB is very loud.