What Hi Fi - 16bit vs 24bit

Posted by: MartinCA on 05 April 2010

FYI, there's an interesting feature in What HiFi this month. They get three people to blind-test 16 bit and 24 bit versions of 3 tracks downloaded from the internet from Naim/Linn sites. They play these through a Naim Uniti off a memory stick.

The testers picked the 24 bit track (blind) as better 8/9 times.

Maybe not an extensive detailed test over time, but more evidence to suggest that downloads will replace CDs for audiophiles at some stage in the future?? It would be nice if there was a bit more mainstream 24 bit music available!
Posted on: 08 April 2010 by Eric Barry
I don't know what files they used, but the files offered as compatibility tests by Naim are not comparable, having much different eq.
Posted on: 09 April 2010 by jfritzen
quote:
Originally posted by Joe Bibb:
quote:
Originally posted by pcstockton:
oh.... i get it. That isn't an inherent problem of DVD-A as a format though right?

I have some live DVDs with corresponding 2 Channel 24/96. Also some of the "classic album" series'.

The American Beauty DVD-A is very nice.


Indeed, far from a problem with the format, more the usual botched introduction. Not enough titles, non compatible players with other formats, no real marketing etc. etc.


I agree and want to add: no prominent and influential supporters either, like Karajan in the 80s for the compact disc.

But the main reason is that most consuments of compact disc audio do not care for the additional benefits of Hi-Res audio, be it via download or via DVD-A. CD is good enough for most people and they do not believe that Hi-Res will sound any better.

Perhaps the difference is more audible in symphonic music? I listened to Tilson-Thomas' 24 bit recording of Mahler's 6th symphony yesterday and it was an impressive experience. It will be difficult to listen to 16 bit Mahler again.
Posted on: 19 April 2010 by Trevp
As I understand it, the number of bits simply determines the available dynamic range of digitally recorded music (it is the sampling frequency that determines the frequency response). As most popular music is currently mastered to only use a fraction of the dynamic range available from 16 bits, I can't understand the improvements that will be gained from 24 bit recordings. Even with classical music, if the full dynamic range of a 24 bit recording was utilised, most systems would not be able to reproduce it, and it would not be pleasant if it could (do you REALLY want a symphony orchestra in your living room? - I suspect the neighbours would have some complaints even if you lived in a detached house.). You would need an amplifier capable of several kilowatts output and speakers to match if you wanted to listen to the full dynamic range of an orchestra and hear the quietest passages at a realistic volume (the crescendos would be enormously loud without dynamic compression).
Posted on: 19 April 2010 by pcstockton
quote:
Originally posted by Trevp:
I can't understand the improvements that will be gained from 24 bit recordings.


Give a listen to some 24/96 files, then let us know.
Posted on: 19 April 2010 by Trevp
quote:
Originally posted by pcstockton:
quote:
Originally posted by Trevp:
I can't understand the improvements that will be gained from 24 bit recordings.


Give a listen to some 24/96 files, then let us know.


I can understand that increasing the sampling rate can improve sound quality - the standard 16 bit sampling rate is 44.1 Khz so an increase to 96 Khz should improve resolution, I just can't see the need to increase the number of bits at the same time when the dynamic range from 16 bits is more than adequate. File sizes would be a little more manageable with fewer bits.

All the best,
Trev
Posted on: 19 April 2010 by pcstockton
For what its worth, I hear a bigger change in SQ from 16/44 to 24/48 than I do from 24/48 to 24/96.

I think the bit depth brings the most to the party.

Also, it is the sample rate that really bumps up the size of a file. 192k files are monstrous, while 24/48 arent that much bigger than 16/44.

Lastly, I dont think the 16bit depth allows for a sampling rate as high as 96k. I could be wrong but have never seen it.

Basically, I think you are correct. 24/48 is just fine. It keeps the files size down and makes the most of what you have.

-patrick
Posted on: 19 April 2010 by 0rangutan
Patrick,

I agree entirely and there is another significant benefit of sticking with 24/48 - the bandwidth is just low enough for the Uniti's 802.11g wireless to stream these without buffering. 24/192 causes occasional buffering, making it unusable over wireless.

If you are looking for decent 24/48 material, this just happens to be the format chosen by B&W's Society of Sound service. Well worth checking out as ~£34 gets you a Hi Res album per month plus access to download their existing back catalogue (not many titles, but good quality).

John
Posted on: 19 April 2010 by pcstockton
Get your hands on the Beatles USB. Awesome.
Posted on: 19 April 2010 by Naijeru
quote:
Originally posted by Trevp:
You would need an amplifier capable of several kilowatts output and speakers to match if you wanted to listen to the full dynamic range of an orchestra and hear the quietest passages at a realistic volume (the crescendos would be enormously loud without dynamic compression).

Not really. You'd only need enough power to hear the full dynamic range of an orchestra from the best seat in the hall, not every seat in the hall. That requires significantly less power.
Posted on: 19 April 2010 by ferenc
Few day ago I had a chance to try a new DAC which shows not only the sample rate, but the bits as well on its display. It was surprising to see that even Foobar 2000 (running on my Macbook, Win 7 Ultimate 64 bit, latest HiFace driver, using wasapi mode) shows 24 bit while playing 24 bit files, the actual bits were truncated to 16 bit even in case of an original 24 bit recorded file. We could get 24 bit out only if there was a processing switched on and the volume control is less than 0 dB max. I tried the same system through the Toslink out, and it was the same truncated 16 bit out instead of 24 bit. Very strange, I do not know if it is a problem with my system/installation or a known phenomenon. Anyway I use mainly different Macs, so it is not a big deal, but probably worth to share the experience.

I tried iTunes on the Mac and was playing 24 bit files as 24 bits. Same for using Amarra and Pure Vinyl, however Pure Vinyl had a kind of stabilization time (roughly a minute) where the displayed bit numbers were jumping between 10,11 and 24 bits randomly. After roughly a minute it settled and showed the actual bits as iTunes.

It was quite interesting.
Posted on: 03 May 2010 by Mward2205
quote:
Originally posted by Trevp:
quote:
Originally posted by pcstockton:
quote:
Originally posted by Trevp:
I can't understand the improvements that will be gained from 24 bit recordings.


Give a listen to some 24/96 files, then let us know.


I can understand that increasing the sampling rate can improve sound quality - the standard 16 bit sampling rate is 44.1 Khz so an increase to 96 Khz should improve resolution, I just can't see the need to increase the number of bits at the same time when the dynamic range from 16 bits is more than adequate. File sizes would be a little more manageable with fewer bits.

All the best,
Trev




Difference between analog (top) and digital (bottom):

24 bits make the lower "sine wave" look much more like the upper one, meaning the digital recording is much more like an analog one - which is the ultimate goal here.