Tim S Kemp wrote:
Stewart Pinkerton wrote:

Absolutely not the case. Given good implementations of all three, the
sound will be *identical*. BTW, it's more like 93-94 dB for properly
implemented 16-bit digital, because of that essential dither.


The problem with all this is you're saying 16, 24, 32 bit, 44, 48, 96, 192
khz will all sound the same.

For an end user yes. Id say that 16/44 and 24/96 might just sound
noticeably different but beyond that no way.

So 8 bit 44khz is fine too? or 12 bit?

no. 8 bit is not good enough. 12 is approaching it but not quite.

Don't get me wrong, I'm not after a "my bits are better than your bits"
assault here, and much of the stuff I listen to is compressed (in both
senses of the word) anyway. But it is just damned wrong to say that the
difference between a 24/96 and a 16/44 recording is none at all.

To the end user, no its not better. if you're mixing the sound it helps
to have a more detailed source to begin with so as not to lose
definition during processing.

I'm sure a few moments with some test gear would show the output will not be
identical. Personally I'm not inclined to even bother trying.

audibly indistingushable and identical arent the same.

Hi,

In message , "Dave Plowman (News)"
writes
In article ,
tony sayer wrote:
As for the picture, it's different, but is it better????

Not in my opinion, against good analogue PAL...

I tend to notice the lack of detail in flesh tones.

We did some acceptability tests at work recently for MPEG video codecs.
During the 'training' phase, most subjects were understandably very
vocal about the 'shimmering' green tones in the grass on a football
field, and almost all rated it as being objectionable.

Nobody (out of 20 subjects) noticed that in the same piece of video, a
player's leg had become separated at the knee, and flown off as he
kicked the ball. The leg subsequently disappeared a third of the way
across the pitch, due to a very dodgy motion prediction algorithm.
People were amazed when it was pointed out to them; they just didn't see
it.

Unfortunately, it doesn't seem to take much critical viewing time to
learn to spot digital video compression artefacts, and once you start to
see them, there's no going back (at least in my case). I'm continually
critical of Sky Digital video quality, especially on the lower bitrate
channels.

As you point out, flesh tones are a prime culprit. Any areas of subtly
varying tones, especially with green content, seems to suffer this
problem.

--
Regards,
Glenn Booth

On Tue, 2 Nov 2004 20:23:09 +0000, Glenn Booth
wrote:

Hi,

In message , "Dave Plowman (News)"
writes
In article ,
tony sayer wrote:
As for the picture, it's different, but is it better????

Not in my opinion, against good analogue PAL...

I tend to notice the lack of detail in flesh tones.

We did some acceptability tests at work recently for MPEG video codecs.
During the 'training' phase, most subjects were understandably very
vocal about the 'shimmering' green tones in the grass on a football
field, and almost all rated it as being objectionable.

Nobody (out of 20 subjects) noticed that in the same piece of video, a
player's leg had become separated at the knee, and flown off as he
kicked the ball. The leg subsequently disappeared a third of the way
across the pitch, due to a very dodgy motion prediction algorithm.
People were amazed when it was pointed out to them; they just didn't see
it.

Unfortunately, it doesn't seem to take much critical viewing time to
learn to spot digital video compression artefacts, and once you start to
see them, there's no going back (at least in my case). I'm continually
critical of Sky Digital video quality, especially on the lower bitrate
channels.

As you point out, flesh tones are a prime culprit. Any areas of subtly
varying tones, especially with green content, seems to suffer this
problem.

The problem with the gross errors you pointed out - dismemberment
being one of them - is that once pointed out, you see them for ever
afterwards and they cease to be unimportant.

d
Pearce Consulting
http://www.pearce.uk.com

On Tue, 2 Nov 2004 18:09:06 -0000, "Tim S Kemp"
wrote:

Stewart Pinkerton wrote:
Absolutely not the case. Given good implementations of all three, the
sound will be *identical*. BTW, it's more like 93-94 dB for properly
implemented 16-bit digital, because of that essential dither.

The problem with all this is you're saying 16, 24, 32 bit, 44, 48, 96, 192
khz will all sound the same.

For most people under most circumstances they will sound the same.

So 8 bit 44khz is fine too? or 12 bit?

8 bits is certainly not adequate; a properly implemented 12 bit system
wouldn't be too bad, but not really high quality. (The BBC used to use
a 13 bit 32KHz distribution system for FM radio and the quality was
often better than the alternatives available at the time.

Don't get me wrong, I'm not after a "my bits are better than your bits"
assault here, and much of the stuff I listen to is compressed (in both
senses of the word) anyway. But it is just damned wrong to say that the
difference between a 24/96 and a 16/44 recording is none at all.

There should certainly be a measurable difference, but it is pretty
doubtful that there will be an audible difference and if there was it
would be pretty small.

What we are trying to do is capture and store all the information that
the ear is capable of resolving. In terms of dynamic range it is
generally accepted that the range between the threshold of hearing at
the ears most sensitive frequencies and the threshold of pain is just
over 120db; his would be represented by a bit depth of 20. However,
ther are significant practical limts as well; there are very few rooms
with a low noise level. If you have a room with a noise level of 30 db
above the threshold of hearing you have a very quiet room. Also the
maximum output level available from commonly available speakers will
be about 110db in a room, if you're lucky. With these adjustments you
get a dynamic range of 80 db which can be adequately represented with
14 bit, so the 16 bits of the CD format has ample spare dynamic range.
Also, the threshold of hearing increases dramatically at higher
frequencies and not quite so dramatically at lower ones which means
that the dynamic range requirements at the frequency extremes are much
reduced.

The other capability of the ear is bandwidth. The highest audible
frequency is generally taken to be 20Khz, but it does vary between
individuals. If you're over 25 it's likely to be lower than that,
especially if you've spent any amount of time in clubs! If you're very
young it might exceed 20Khz by a small amount. A frequency of 20Khz
will be correctly reproduced with a 44Khz sampling rate, a 48Khz
sampling rate will be able to reproduce 24 Khz, which should
accommodate the most bat eared human.

Bill

In article ,
Chris Isbell wrote:
You must have either a poor tuner, or a dodgy multipath'ed signal, or a
duff transmitter.

Can't say I've noticed anything amiss..until the other day;(

You could well be right. However, I believe that NICAM uses
compression - dynamically changing the range covered by the 11 bits
according to the signal level. I had therefore tended to put down the
effects I can hear on transients to NICAM compression artifacts.

TV NICAM uses 11 bit companded (which equates to 14 bit), but IIRC the
NICAM distribution system to radio transmitters ain't quite the same. I'll
see if I can dig out details.

--
*I went to school to become a wit, only got halfway through.

Dave Plowman London SW
To e-mail, change noise into sound.

Glenn Booth wrote:

Unfortunately, it doesn't seem to take much critical viewing time to
learn to spot digital video compression artefacts, and once you start
to see them, there's no going back (at least in my case). I'm
continually critical of Sky Digital video quality, especially on the
lower bitrate channels.

My favorite digital artefact is the gold brand / tattoo thing on teal'cs
head on stargate sg-1 which seems to have a life of its own on sky and never
appears to be attached to his head. On Channel 4 it's fine...

--
"Get a paper bag"

"Tim S Kemp" wrote in message

Stewart Pinkerton wrote:
Absolutely not the case. Given good implementations of all three, the
sound will be *identical*. BTW, it's more like 93-94 dB for properly
implemented 16-bit digital, because of that essential dither.

The problem with all this is you're saying 16, 24, 32 bit, 44, 48,
96, 192 khz will all sound the same.

Just the facts.

So 8 bit 44khz is fine too?

It's pretty hissy sounding.

or 12 bit?

Remarkably good but a tad hissy.

Don't get me wrong, I'm not after a "my bits are better than your
bits" assault here, and much of the stuff I listen to is compressed
(in both senses of the word) anyway. But it is just damned wrong to
say that the difference between a 24/96 and a 16/44 recording is none
at all.

Why? It's well known that nobody can hear the difference in a proper blind
test/ You've got problems with reality working the way it does?

I'm sure a few moments with some test gear would show the output will
not be identical.

That's because test gear can be tremendously more sensitive to technical
differences than the human ear.

Personally I'm not inclined to even bother trying.

Sign of a closed mind, no?

Remember, I did bother trying, and I make it as easly as possible for
anybody who is so inclined to do so for themselves.

"Stewart Pinkerton" wrote in message

On Tue, 2 Nov 2004 12:03:59 -0000, "Keith G"
wrote:

"Iain M Churches" wrote

snip 'my knob is my subwoofer' bleed-through

One often hears of people changing from an SS to a
valve amp. I have never heard of anyone going the
other way.

Neither have I.......

Sure you have. Me, Arny, and every other sensible audiophile past 50
years old - we *all* started with valves, because there was no choice.
Heck, the first 'hi-fi' amp I ever built was a 'flea power'
single-ended valve design! :-)

My first self-designed, self-built hi fi amp was push-pull 6V6's. My second
was push-pull 6BQ5s. I progressed from there to 7591s and EL 34s. I did
EL84's some place along the way.

I then progressed to 2N2147s, then 40410s, then 2N5630 and 2N6030

On Tue, 2 Nov 2004 18:09:06 -0000, "Tim S Kemp"
wrote:

Stewart Pinkerton wrote:
Absolutely not the case. Given good implementations of all three, the
sound will be *identical*. BTW, it's more like 93-94 dB for properly
implemented 16-bit digital, because of that essential dither.

The problem with all this is you're saying 16, 24, 32 bit, 44, 48, 96, 192
khz will all sound the same.

Yes. Not one single person has yet been able to show proof that, *all
other things being equal*, this is not the case.

So 8 bit 44khz is fine too? or 12 bit?

It seems to be the case that above 12-13 bits, we can't tell a
difference. Interestingly, that's also consonant with known theory.
BTW, the *only* audible difference with a properly encoded 8-bit
signal, is a higher noise floor. If there's another difference, then
you didn't do it right!

Don't get me wrong, I'm not after a "my bits are better than your bits"
assault here, and much of the stuff I listen to is compressed (in both
senses of the word) anyway. But it is just damned wrong to say that the
difference between a 24/96 and a 16/44 recording is none at all.

Sez hoo? Show me one single example of a level-matched blind test
which shows that two such recordings made from the same master, sound
different. Very easy to do this if you move slightly to 24/88.2, or
24/176.4 for the original.

I'm sure a few moments with some test gear would show the output will not be
identical. Personally I'm not inclined to even bother trying.

Try it with a signal bandlimited to 20kHz and with a dynamic range of
less than 90dB, i.e. any analogue master tape. There will be *no*
difference, either measurable or audible. That *you* can't be bothered
to discover the truth, is not *my* problem.
--

Stewart Pinkerton | Music is Art - Audio is Engineering

On Tue, 2 Nov 2004 18:05:19 -0000, "Tim S Kemp"
wrote:

Stewart Pinkerton wrote:
On Tue, 2 Nov 2004 07:54:40 -0000, "Tim S Kemp"
wrote:

Stewart Pinkerton wrote:

Why would you suppose that anyone *could* tell the difference?

The increased sample rate for one?

Irrelevant, since 44.1k samples/sec captures the entire audible range.

hmmmm

Asleep, or got something to say?

--

Stewart Pinkerton | Music is Art - Audio is Engineering