24bit vs 16bit, the myth exploded!

[SilverEars]

  Listening fatigue is due to response spikes. The lack of super high frequencies won't cause it, but having too much up there will.

Would fatigue also degrade the quality of the sound perceived also?  I notice this when I listen to fatiguing phones, as over time it sounds more distorted.

 

[tdog]

  I'm going to play the devil's advocate here (please note this is honest curiosity and I'm trying to further my somewhat limited knowledge).
 
What about the arguments that one can "feel" a difference in the music at 24 bits?  One example would be greater listening fatigue when comparing 16 to 24.  This type of difference wouldn't be apparent in a quick ABX test, but only after extended listening periods.

 
I have no problem with the theoretical possibility that 24-bit audio might "feel" better than 16-bit, i.e. the perception is fairly subconscious. However, of the many listening test that have been conducted (thousands or only hundreds? I've lost count) there would be a result of preference for 24-bit, and _every_ test I have seen or been a part of has shown that people are guessing (as long as the material is properly dithered to 16-bit etc).
 
 

Would it be possible that the lesser noise of 24 bits over 16 bits (even though slight) could be perceived, maybe even on a subconscious level, and would lead to a better listening experience?

 
It is possible that, given the right program material and a pristine listening environment, quantisation distortion or noise is just audible on undithered 16-bit audio. However it is highly dependent on the program material, and only if it's undithered.
 
None of this stops me buying and enjoying hi-res music files, and as an audio professional I always record in hi-res for the obvious reasons (headroom on production, easier anti-aliasing filtering etc), but as a delivery format, 16-bit is just fine and completely transparent if it is done properly... (of course, it's not always done properly, but that is hardly the fault of the format).

 

[bigshot]

  Would fatigue also degrade the quality of the sound perceived also? 

 
Not if the spike is narrow enough, or is out of the range of human hearing.

 

[bigshot]

  It is possible that, given the right program material and a pristine listening environment, quantisation distortion or noise is just audible on undithered 16-bit audio. However it is highly dependent on the program material, and only if it's undithered.

 
What sort of audio is undithered when it's bounced down? I've never heard of anyone doing that.

 

[SilverEars]

   
Not if the spike is narrow enough, or is out of the range of human hearing.

So you're talkin about sibilance.  Tsss, Sssss.  That thick spike?

 

[bigshot]

Not necessarily. A spike anywhere in the core frequencies can be irritating, but not catastrophic if the band is narrow enough. Sibilance is just one of these sorts of things.

 

[castleofargh]

I cannot stand spikes on the frequencies. a bad crossover with complete silence somwhere, I might not even notice, and when I do, I'm just sad that I'm missing something. that's it, no aggressive sounding from it.
but 1 spiky bump of more than 5db somewhere and you can be sure I will dispise the headphone and find it fatiguing even if I don't know why beforehand(the 1khz/5khz being my nemesis).
now I'm even more focused on how they smoothed out the frequency response graphs than on how it is compensated. I've been tricked by golden ears measurements that way they make everything so smooth that you know nothing about how it really is. Tyll is my favorite for that very reason.

 

[cjl]

Where is this lesser noise?

I've never seen evidence showing there is more noise in 16 bit....there is evidence that higher res is actually noisier than Redbook

If your audio chain isn't designed to deal properly with ultrasonics then your subconscious is far more likely to perceive the distortions created in the chain

There's absolutely no question that a properly implemented 24 bit playback chain has a better SNR and lower noise than 16 bit. The ultrasonics you are talking about have nothing to do with bit depth - high sample rate is what allows for ultrasonic content (and, as you said, potential distortion). Now, you could make a very good argument that the SNR and noise levels in a modern 16 bit setup (usually around -100dBFS or better) are good enough that it doesn't really make any difference anyways. Certainly every blind test implemented using reasonable audio samples has indicated that this is the case. However, the basic premise of the original statement (that 24 bit can have lower noise and better SNR) is definitely true.

 

[bigshot]

For the purposes of listening to music in a home audio situation, there is absolutely no difference in noise levels between 16 and 24 bit.

 

[headdict]

For the purposes of listening to music in a home audio situation, there is absolutely no difference in noise levels between 16 and 24 bit.

In a noisy on-the-go situation there is even much less difference.

 

[bigshot]

Much less than none at all!

 

[RazorJack]

  I'm going to play the devil's advocate here (please note this is honest curiosity and I'm trying to further my somewhat limited knowledge).
 
What about the arguments that one can "feel" a difference in the music at 24 bits?  One example would be greater listening fatigue when comparing 16 to 24.  This type of difference wouldn't be apparent in a quick ABX test, but only after extended listening periods.
 
Would it be possible that the lesser noise of 24 bits over 16 bits (even though slight) could be perceived, maybe even on a subconscious level, and would lead to a better listening experience?

 
Feeling, or perceiving, how exactly, with which of the five senses?
 
This has nothing to do with science. 
 
That said, I think people feeling things is what keeps 99,9% of Head-Fi and the audiophile community thriving, and there's nothing wrong with that.

 

[headdict]

Feeling, or perceiving, how exactly, with which of the five senses?

This has nothing to do with science. 

That said, I think people feeling things is what keeps 99,9% of Head-Fi and the audiophile community thriving, and there's nothing wrong with that.

Feeling can be subject to scientific studies. Does anybody know whether extended double blind testing has ever been done to capture potential long-term effects of higher bit rates?
Fully agree on the importance of feelings (and money!) for the audiophile community. Beats science anytime.

 

[cjl]

  For the purposes of listening to music in a home audio situation, there is absolutely no difference in noise levels between 16 and 24 bit.

Absolutely. That's a very different claim than the claim that 24 bit is noisier than 16 bit though.

 

[Rooster81]

  There's absolutely no question that a properly implemented 24 bit playback chain has a better SNR and lower noise than 16 bit. The ultrasonics you are talking about have nothing to do with bit depth - high sample rate is what allows for ultrasonic content (and, as you said, potential distortion). Now, you could make a very good argument that the SNR and noise levels in a modern 16 bit setup (usually around -100dBFS or better) are good enough that it doesn't really make any difference anyways. Certainly every blind test implemented using reasonable audio samples has indicated that this is the case. However, the basic premise of the original statement (that 24 bit can have lower noise and better SNR) is definitely true.

Thanks for answering the question.