"Jim Lesurf" wrote in message
...
In article , David Looser
wrote:
"Jim Lesurf" wrote in message
...
In article , David Looser
wrote:
"Arny Krueger" wrote Or this?

http://gizmodo.com/#!5768446/why-24+...-bad-for-users


The link seems to be broken.

That was my experience. But I managed to find the article by using the
search box on the page and asking for "24 bit bad".

Yes, that's what I did in the end, only I searched on just "24 bit".

I entirely agree with the author of that article, I can see no point in
24bit for audio intended for domestic listening. The only difference
between 16bit and 24bit is the S/N ratio, and who needs a higher S/N
ratio than the 90dB or so of 16bit?

That depends. :-)

In theory, people making 44.1k/16 recordings for CDDA will take care to
avoid any clipping or level compression and keep the mean and peak levels
well clear of 0dBFS. They will also carefully downconvert from the
'master'
recordings at higher rate and larger sample depth and employ an optimum
choice of downsampling method, dithering, and noise-shaping. You will then
play the result using a DAC with excellent reconstruction filtering that
does no damage to the inband signals whilst killing the out-of-band
garbage.

In practice there are measurable signs this often doesn't happen. Mainly
due to the laziness, idiocy, or sheer arrogance of those making CDs. e.g.
the obsession with "louder is better".

Choosing something like 96k/24 may avoid such severe downsampling
processes. There is a lot more 'elbow room' for avoiding problems in
downsampling if you are going from, say, 192k/24 to 96k/24 than right down
to 44.1k/16. It also makes the job of the DAC easier as the reconstruction
filtering doesn't have to be as 'brick shaped' to get you good results to
above 20kHz without aliased crap.

Secondly, just as some will doubtless use '96k/24' to sell the same
rubbish
again, others will use it as a flag that the buyer is serious about sound
quality. e.g. the way some of the early dual releases were 'better' on
DVD-A or SACD because they had been processed to give different results.
The assumption being that the DVD-A or SACD buyer did *not* want sounds
compressed to death, etc.

So in theory it probably should make no difference. But in practice it
may.
Note that I have been talking about small specialist companies like
Chandos
and Linnrecords. Even from their CDs you can tell they take care. The
irony
I pointed out was that this may mean you have *less* reason to expect
their
'high rez' files to sound better. :-) But it may point a direction for
others.

I can't comment as yet on any 'inherent' audible differences. Not heard
any
96k/24 beyond a few brief test files. And my hearing probably isn't
'golden
eared' anyway. But it may well give audibly better results with excellent
setups in some cases. However my interest is to reduce the processes the
biz uses to furtle up the transfer from original recording to the item you
buy. These processes - in theory - don't matter much. But in practice I
suspect they do matter in many cases.

As usual, though, the basic two rules of the market will fight it out.

A) You can only buy what someone will offer for sale.

B) commercial 'success' then hinges on what items on offer people actually
buy, and how much they will pay.

Think of it this way. A cheap and crappy 'mp3 download' output for the
mass
market with a low cost-per-item. (Probably mostly pop music.) Then in
parallel a high quality 'hi rez' output which sells in much smaller
quantities - but with a higher markup - aimed at those who want much
better
sound quality for content that justifies it. (Probably specialist music
types like Classical, Jazz, etc.)

There are *some* upward pressures here. The BBC 320k stream is a nice
example. As are Chandos, Linn, and a select few others. But how this will
turn out, I have no idea.


I read through your post carefully, Jim, looking for the bit where you
explained *why* a domestic listener might need an S/N ratio greater than
90dB, but failed to find it. Instead I read a load of stuff about people
taking care, and the possibility of the *label* "24-bit" being used to imply
an "audiophile" quality recording. Oh, and there was also some stuff about
the possible advantages of 96k vs 44.1k, but neither the original article,
nor my comment, addressed *that* issue.

I don't dispute that, under ideal circumstances, some listeners (those with
younger ears than mine) might detect a marginal improvement in going from
44.1k to 96k. But as to 24-bit? nah!

David.