View Single Post
Old Sep 10, 2012, 06:32 PM
quintin3265's Avatar
quintin3265 quintin3265 is offline
Senior Member
Join Date: Mar 2009
Location: State College, PA
Posts: 193

Originally Posted by LiquidAcid View Post
I agree that 192Khz music doesn't make much difference over 96Khz. The Library of Congress agrees on that one, as it archives music at 96. But the difference between 16/44 and 24/96 is very plain and despite the math, can't be explained away with graphs and figures.

I read the article, but it seems to focus far too much on test tones and noise as opposed to real music. It may be true that test tones show little difference, but if you compare the DVD-audio version of Santana's "Shaman" with the CD version, even if you listen to lossless files of both, there is no comparison. Even if I turn the speaker system to "pure direct," and turn it down low, you can still hear the difference.

One reason I believe I hear this is that I have a TX-NR1009 connected to Polk TSI-400 towers - with a signal to noise ratio of 110dB. The signal-to-noise ratio in most motherboard integrated audio is usually in the 85-90dB range, or 100 times less. 89dB is also the limit of 16-bit resolution, coincidentally, so people who pipe 24-bit audio into cheap earbuds using a standard soundcard will be disappointed.

I also wonder when these listening tests are performed if the people listening know what to listen for. People who don't know how audio compression works are less likely to be bothered by compression artifacts. Similarly, people who spent years working in loud factories are included in these "random" samples. I would like someone to commission a study of people who take a hearing test first, and then eliminate people from the study who have tinnitus or hearing loss. This group would be much more likely to show a correlation than the random group that includes hearing issues.

But finally, I think that most magazines that study this issue miss the idea that the Nyquest theory applies to conscious hearing. It is true that most people can't hear sounds that are above 20Khz or so - but that doesn't mean that they cannot perceive them in some other way the scientists are not measuring. Retaining these overtones in the recording may cause people to feel "better" while listening without being able to say why or how - or 16-bit music might make them feel subconsciously as if something is "wrong." Studies that ask people to identify sampling frequency wouldn't be able to pick that up because they only ask people if they "hear" anything differently.

This is an interesting issue that there seems to be a lot of debate over - but I still think that once Apple starts offering 24-bit music, there will be a switchover.
Now you can embed your songs in forum posts and webpages just like this image! Click the image to find out how!