Head-Fi.org › Forums › Equipment Forums › Sound Science › 24bit vs 16bit, the myth exploded!
New Posts  All Forums:Forum Nav:

24bit vs 16bit, the myth exploded! - Page 107

post #1591 of 1840

one thing to consider is most dacs and adcs have anywhere from 1 to 3 bits of error. so while a 16 bit dac may be sufficient on paper in reality it may come up a bit short.  it is like building a bridge to the minimum spec for the intended load.  a little wind and a heavy truck and you have Tacoma Narrows.

post #1592 of 1840
Well we are subjective beings somewhat, objective measures don't measure up to our own bug dsp chips which are our brains.. Large numbers have positive connotations so that's why very large resolutions appear when they actually do not provide any actual improvement.

Either way, economics, marketing and loads of other social and psychological factors come into play when we value anything. That is why we have DSD and DXD and whatnot because you can't just separate the ideal from the practical. Everything comes as a whole instead of discrete parts that can be scrutinized individually.

Just my two cents
post #1593 of 1840
Quote:
Originally Posted by groovyd View Post
 

one thing to consider is most dacs and adcs have anywhere from 1 to 3 bits of error. so while a 16 bit dac may be sufficient on paper in reality it may come up a bit short.  it is like building a bridge to the minimum spec for the intended load.  a little wind and a heavy truck and you have Tacoma Narrows.


How many DAC chips are only 16 bits now?  Giving you the benefit of the doubt, 24 bits minus 3 bits is 21 bits to do 16 bit data.  Seems plenty sufficient.  The great majority of current DACs all the way out to the analog output exceed the 96 db needed for clean accurate 16 bit reproduction.

post #1594 of 1840
Quote:
Originally Posted by Lespectraal View Post

Well we are subjective beings somewhat, objective measures don't measure up to our own bug dsp chips which are our brains.. Large numbers have positive connotations so that's why very large resolutions appear when they actually do not provide any actual improvement.



Either way, economics, marketing and loads of other social and psychological factors come into play when we value anything. That is why we have DSD and DXD and whatnot because you can't just separate the ideal from the practical. Everything comes as a whole instead of discrete parts that can be scrutinized individually.



Just my two cents

 



The question of needing a 24-bit DAC for 16-bit playback is not subjective in this context. For DAPs which use digital volume control 24 bit DAC is essential to maintain good SNR. For USB DAC purpose, ASRC also helps to reduce jitter which requires 24 bit DAC even for 16 bit playback.
post #1595 of 1840
Quote:
Originally Posted by headwhacker View Post

Quote:
Originally Posted by Lespectraal View Post

Well we are subjective beings somewhat, objective measures don't measure up to our own bug dsp chips which are our brains.. Large numbers have positive connotations so that's why very large resolutions appear when they actually do not provide any actual improvement.



Either way, economics, marketing and loads of other social and psychological factors come into play when we value anything. That is why we have DSD and DXD and whatnot because you can't just separate the ideal from the practical. Everything comes as a whole instead of discrete parts that can be scrutinized individually.



Just my two cents

 



The question of needing a 24-bit DAC for 16-bit playback is not subjective in this context. For DAPs which use digital volume control 24 bit DAC is essential to maintain good SNR. For USB DAC purpose, ASRC also helps to reduce jitter which requires 24 bit DAC even for 16 bit playback.
I guess I was being too general there in my post.

Of course there exceptions, like the bit depth in the case or digital volume control. I'm just talking about high resolution audio in general, like needlessly high sampling rates and dual-mono dac chip arrangements and whatnot.
post #1596 of 1840

agreed - was just bringing up the bit error point as a comparison against when DACs were 16 bits in saying that you may not be really getting 16 bits worth of signal.

post #1597 of 1840
Quote:
Originally Posted by groovyd View Post
 

agreed - was just bringing up the bit error point as a comparison against when DACs were 16 bits in saying that you may not be really getting 16 bits worth of signal.

 

Unless you have a player from the 80's/early-90's you are likely experience a DAC that's some variation of oversampling+deltasigma. The linearity error in those tends to be effectively less one bit equivalent of 24bits. The SNR is pushing 115-120dB or so. In effect, you are getting the error floor of >16bit DAC even when you are playing 16bit data.

post #1598 of 1840

I don't know, the lack of dynamics may not be a bad thing for most ppl, as most ppl nowadays listen to music while commuting, jogging, driving etc. With a small dynamic range like less than 5dB, all you need is for the music level to be 5dB over the noise level so you don't get part of it drowned out by the ambient noise. The ambient noise level in the city where I work in averages around 83dB and peaks at around 96, if I were wearing non-noise isolating ear buds I'd have to have the music level as high as 88dB to even hear it properly. Can you imagine if the dynamic range were higher, like 15dB? So I think it is a kind of dilemma how loud should music be mixed at. Probably best solution would be a low dynamic version as standard and a special high dynamic audiophile version in 24/192? Most cinema goers would know what high dynamic music sounds like, as most cinema sounds can have a 30dB + dynamic range.

 

Quote:

Originally Posted by harmonix View Post
 

I get what you are saying - IMHO it's not just the recording and mastering chain either. Alot of modern day rock/pop musicians don't play with real dynamics.

How many times have you heard musicians that can only play f, ff and fff ? Really annoying.

 

I think electric instruments and amplification are the root cause of alot of this - maybe people should learn to play unplugged but then alot of things would really sound crappy but hey that may not be a bad thing.

post #1599 of 1840

Funny thing you mentioned movies. Everybody i know have always complained about the sounds in movies, the weaker sounds tend to drown out in the ambient noises ( often times talking ).

But the explosions and music can be very very loud. In cinemas they crank the volume so high that it sometimes feels like you can get permanent hearing damage from the Peak sounds.

This is actually one aspect that in my opinion should be discussed, can higher dynamic range lead to increased cases of hearing damage if 24 / 96 is made standard?

Most adults will know to limit the volume so no damage is done, but what about the young people?

While i think higher dynamic's can serve some people, i think those people are very small percentage of music listeners over all.

Like Danz03 pointed out, there are situations where high dynamic range would work to hinder music listening quite a bit.

post #1600 of 1840

Dynamic range that would actually benefit from 24-bit resolution (and not just for marketing reasons) would indeed not be practical for most people listening at home under typical conditions. The noise floor of CD quality audio even with the simplest dithering at a loud 110 dB peak SPL is about 14 dB (A-weighted), which is normally not noticeable, especially when there is any music to mask it.

 

The problem with the "loudness war" is not just that the dynamic range is compressed, but it is done beyond the point of diminishing returns, when the last few dB of extra loudness is achieved at the cost of severe and audible distortion.

post #1601 of 1840
Quote:
Originally Posted by Notus View Post
 

Funny thing you mentioned movies. Everybody i know have always complained about the sounds in movies, the weaker sounds tend to drown out in the ambient noises ( often times talking ).

But the explosions and music can be very very loud. In cinemas they crank the volume so high that it sometimes feels like you can get permanent hearing damage from the Peak sounds.

This is actually one aspect that in my opinion should be discussed, can higher dynamic range lead to increased cases of hearing damage if 24 / 96 is made standard?

Most adults will know to limit the volume so no damage is done, but what about the young people?

While i think higher dynamic can serve some people, i think those people are very small percentage of music listeners over all.

Like Danz03 pointed out, there are situations where high dynamic range would work to hinder music listening quite a bit.

 

 real noisy environments as danz03 suggested, or the opposite, when you want to listen very very quietly, then music with low dynamic range might help. I can't imagine listening to some great symphonies in the subway, I would have to listen to deafening levels to hear most of the sounds,else I would miss 2/3 of the symphony.

 

but for "usual" listening, a big dynamic is actually much less fatiguing to the ears. what ruins your ears isn't only the pressure level of the sound(unless you play it at unreasonable levels) but also the duration of that pressure. when you have some album with 8db of DR(hello metallica...), you'll listen to that level all the album, and the quiet parts will never really be quiet.

when you have something with a lot of dynamic, you'll set the volume control for the loudest parts higher, yes! but you'll also have a lot more moments when the pressure from the music will be lower than the minimum on metallica. ups and downs(high DR) are more natural and ok than constant moderate-to-high volume levels(low DR). I'm making a general statement here so there must be some point where such DR is better for the ears than such. but overall, big DR(who are we kidding, 60DB on most tracks would be a dream come true) is considered better for the ears.

 

and for all of the above, I'm talking actual dynamic range of the music, not dynamic range available because it's 16 or 24bit. it's not like there was a lot of tracks out there that actually use 16bit of dynamic. so I wouldn't worry too much about 24/96 becoming a standard.

post #1602 of 1840

The purpose of 24 bit audio is to provide the music producers with some headroom for mixing and mastering.  Downsampling a recording to 16 bit red book doesn't cause any audible loss.  16 bits can handle all the dynamic range we can handle.  If anyone is interested in doing a bias controlled listeing test, there is a feature for Foobar that can make that easy for you.

post #1603 of 1840
Quote:
Originally Posted by blades View Post
 

The purpose of 24 bit audio is to provide the music producers with some headroom for mixing and mastering.

 

I am curious to know why would music producers bother with 24 bits if 16 bits is sufficient?

post #1604 of 1840
The extra headroom allows them to mix parts with adjusted gains and apply affects while keeping the noise floor at negligibly small levels. It also simplifies the recording process: the overall level of the recording can be kept low enough to avoid clipping while there is sufficient dynsmic range to record everything without the noise floor being a problem. With 24 bits, it is easy to keep the entire recording within the dynamic range without needing the careful setup and sound check required when using 16 bit or analog recording equipment.

Cheers
post #1605 of 1840

so basically engineering of music should be done in 24 bits so that we get a quality 16 bits out.  i would agree a well mastered 24 bit track then quantized down to 16 bits is really all a human ear can perceive.  the biggest issue in recording quality is the effects of not mastering a track properly.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › 24bit vs 16bit, the myth exploded!