Head-Fi.org › Forums › Equipment Forums › Sound Science › The Most Important Spec Sheet: The Human Ear
New Posts  All Forums:Forum Nav:

The Most Important Spec Sheet: The Human Ear - Page 6

post #76 of 85
Quote:
Originally Posted by bigshot View Post

We're looking for a general threshold of audibility for non-linear distortion. Is there anything in there like that?

Doubt you'll get it.  There is really no such a threshold or audibility because of the wide variations in non-linear distortion and the resulting products.  The audibility threshold would have to be represented in at least a 3 axis graph, perhaps several.  You just can't spec that kind of thing with a single number like 2%, you need so many more qualifiers.  And that would apply just for the audibility of distortion using sine wave signals.  Once you throw in complex music as the stimulus, it gets even tougher.  

post #77 of 85

Audibility for who, me or my dog bat?

 


Edited by grokit - 6/23/13 at 8:18pm
post #78 of 85
Quote:
Originally Posted by grokit View Post

Audibility for who, me or my dog bat?

 

How about audibility for a dog-faced bat?

post #79 of 85
Thread Starter 
How about a range of best to worse case in the type of distorition usually found in modern amps and players? Broad strokes. Real world.
post #80 of 85
Quote:
Originally Posted by bigshot View Post

We're looking for a general threshold of audibility for non-linear distortion.

I did a worst-case test once, and for me the threshold is about -70 dB. I mixed two sine waves, 100 Hz and 3 KHz, and once the 3 KHz component was more than 60 dB softer it was inaudible under all listening conditions. So I added 10 dB for safety or possibly better ears and decided -70 dB or 0.03 percent is a reasonable threshold. If the total distortion is 0.03 percent or lower, I doubt it will be audible to anyone under any circumstances at any volume, no matter the makeup of the specific distortion.

--Ethan
Edited by EthanWiner - 6/24/13 at 9:07am
post #81 of 85
Thread Starter 
The specs I've been including here are based on listening to music, which is what we all ultimately are doing. What figure do you think would be safe with audibility in music?

I suppose I could put a second set of specs alongside for audibility in test tones.
Edited by bigshot - 6/24/13 at 9:24am
post #82 of 85
Quote:
Originally Posted by EthanWiner View Post


I did a worst-case test once, and for me the threshold is about -70 dB. I mixed two sine waves, 100 Hz and 3 KHz, and once the 3 KHz component was more than 60 dB softer it was inaudible under all listening conditions. So I added 10 dB for safety or possibly better ears and decided -70 dB or 0.03 percent is a reasonable threshold. If the total distortion is 0.03 percent or lower, I doubt it will be audible to anyone under any circumstances at any volume, no matter the makeup of the specific distortion.

--Ethan

That's an excellent setup for a worst case scenario (fundamental five octaves away from distortion product which falls in the middle of the range of greatest audibility) and its results are in congruence with some of the other tests which have reported detection of distortion components at around -50db. Still it's an exceedingly unlikely example, so much so that I question the need for the extra 10db safety or the inclusion of this data point in the first post. Especially when tests (like the Lee/Geddes study which can be found on a link from the previous page) show that with some program material even distortion products which are even less than 20db down may be inaudible. Maybe something like a '-20 to -40db typical, -40 to -60db synthetic worst case scenario' disclaimer would work.


Edited by anetode - 6/24/13 at 11:15pm
post #83 of 85
Quote:
Originally Posted by bigshot View Post

The specs I've been including here are based on listening to music, which is what we all ultimately are doing. What figure do you think would be safe with audibility in music?

Well, music consists entirely of sine waves, so my test is equally relevant for music! biggrin.gif

---Ethan
post #84 of 85
Quote:
Originally Posted by anetode View Post

Still it's an exceedingly unlikely example, so much so that I question the need for the extra 10db safety or the inclusion of this data point in the first post.

Of course I agree. I make the point in my Audio Expert book that the top 20 to 30 dB is what matters the most.

I estimated conservatively to avoid someone in a forum being able to post an example that (barely) refutes my numbers. This kind of stuff comes up often when people claim to be able to hear jitter, or the benefits of dither. So I'll spot people that first 70 dB, but they have to provide hard proof for the last 26 dB. biggrin.gif

--Ethan
post #85 of 85
Thread Starter 
I think those types of people are just grabbing at straws to back up their incorrect theories. It's not fair to the vast majority of people who just want to know what matters to give them figures designed for people counting angels on the heads of pins.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › The Most Important Spec Sheet: The Human Ear