OT: Sampling sound at 48 kHz rather than 44.1 kHz

[NY]

In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage was,
given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.

Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.

What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

[Woody]

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage was,
given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.
Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.
What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

44.1kHz is/was the sample rate for CDs, whereas 48kHz was more about DVD
multichannel sound.
Accepted most people would have difficulty hearing above about 15kHz but
there is something in the comment about harmonics. Having said that you
would be looking for third or fourth harmonics of maybe 8kHz or more
which is into the land of absorption by furniture/carpets etc. Another
way of looking at it is that a violin actually generates more sound on
the second harmonic than it does on the fundamental being played!!

Unless you find CD sound unacceptable just live with it I would say!

[Rupert Moss-Eccardt]

Post by Woody

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage was,
given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.
Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.
What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

44.1kHz is/was the sample rate for CDs, whereas 48kHz was more about DVD
multichannel sound.
Accepted most people would have difficulty hearing above about 15kHz but
there is something in the comment about harmonics. Having said that you
would be looking for third or fourth harmonics of maybe 8kHz or more
which is into the land of absorption by furniture/carpets etc. Another
way of looking at it is that a violin actually generates more sound on
the second harmonic than it does on the fundamental being played!!
Unless you find CD sound unacceptable just live with it I would say!

"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

And, we do know that we need to sample at twice the rate of the range
we do want to hear (ignoring the overtones)

[Java Jive]

Post by Rupert Moss-Eccardt
"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

Scientific provenance for this claim?

Post by Rupert Moss-Eccardt
And, we do know that we need to sample at twice the rate of the range
we do want to hear (ignoring the overtones)

Yes.

[Rupert Moss-Eccardt]

Post by Java Jive

Post by Rupert Moss-Eccardt
"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

Scientific provenance for this claim?

It was something that I came across in the 80s but I can't find a
citation now. As I recall, the experiment was to have people to simply
report if they noticed anything changing. Not about hearing but
sensing. It is a bit like subsonics.

[Jim Lesurf]

Post by Java Jive

Post by Rupert Moss-Eccardt
"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

Scientific provenance for this claim?

Post by Rupert Moss-Eccardt
And, we do know that we need to sample at twice the rate of the range
we do want to hear (ignoring the overtones)

Yes.

Nyquist-speaking we should say 'more than' by some practical amount to
allow for filters not being brickwalls and recordings being of finite
duration. But that's a fairly fine point.

Jim

[Brian Gregory]

Post by Rupert Moss-Eccardt
"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

No.

Even if it looks a different shape on an oscilloscope, and all you've
done is filtered out harmonics that are above the upper limit of
hearing, it will sound exactly the same.

[Roderick Stewart]

On Fri, 12 Apr 2024 17:18:38 +0100, Brian Gregory

Post by Brian Gregory

Post by Rupert Moss-Eccardt
"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

No.
Even if it looks a different shape on an oscilloscope, and all you've
done is filtered out harmonics that are above the upper limit of
hearing, it will sound exactly the same.

It's lower frequencies that you can sometimes hear when they aren't
there, not higher ones. A mixture of harmonics without the fundamental
can sound as if the fiundamental is present, even though it isn't.

Rod.

[NY]

Post by Roderick Stewart
On Fri, 12 Apr 2024 17:18:38 +0100, Brian Gregory

Post by Brian Gregory

Post by Rupert Moss-Eccardt
"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

No.
Even if it looks a different shape on an oscilloscope, and all you've
done is filtered out harmonics that are above the upper limit of
hearing, it will sound exactly the same.

It's lower frequencies that you can sometimes hear when they aren't
there, not higher ones. A mixture of harmonics without the fundamental
can sound as if the fiundamental is present, even though it isn't.

Presumably due to "beating" (heterodyning). So if you have two sine
waves of 2000 Hz and 2100 Hz, it can sound as if there is a third one at
2100-2000 = 100 Hz.

[Roderick Stewart]

Post by NY

Post by Roderick Stewart
On Fri, 12 Apr 2024 17:18:38 +0100, Brian Gregory

Post by Brian Gregory

Post by Rupert Moss-Eccardt
"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

No.
Even if it looks a different shape on an oscilloscope, and all you've
done is filtered out harmonics that are above the upper limit of
hearing, it will sound exactly the same.

It's lower frequencies that you can sometimes hear when they aren't
there, not higher ones. A mixture of harmonics without the fundamental
can sound as if the fiundamental is present, even though it isn't.

Presumably due to "beating" (heterodyning). So if you have two sine
waves of 2000 Hz and 2100 Hz, it can sound as if there is a third one at
2100-2000 = 100 Hz.

No. A beat signal will be at the sum and difference frequencies of the
original two, and not necessarily nicely related to either of them.

The 'missing fundamental' effect occurs where you have a sequence of
frequencies that are multiples of some lower frequency that is not
present, and at the amplitudes that they would have had if they were
harmonics of it. This can result in you 'hearing' the full sound with
its fundamental even though the fundamental is not there. This is how
harmonic rich sounds can sometimes appear bigger and beefier than they
really should when reproduced by small loudspeakers that can't manage
the lowest frequency components.

Rod.

[Brian Gregory]

Post by Roderick Stewart
On Fri, 12 Apr 2024 17:18:38 +0100, Brian Gregory

Post by Brian Gregory

Post by Rupert Moss-Eccardt
"Harmonics" sort of cover it but it is possible for people to detect a
change in tone way about the range of hearing but they don't actual
"hear" the notes.

No.
Even if it looks a different shape on an oscilloscope, and all you've
done is filtered out harmonics that are above the upper limit of
hearing, it will sound exactly the same.

It's lower frequencies that you can sometimes hear when they aren't
there, not higher ones. A mixture of harmonics without the fundamental
can sound as if the fiundamental is present, even though it isn't.

But if the mixture of harmonics are all above the range of human hearing
you won't hear anything. Your brain will have no information to use to
fill in the supposed fundamental.

[Jim Lesurf]

Post by Brian Gregory
But if the mixture of harmonics are all above the range of human hearing
you won't hear anything. Your brain will have no information to use to
fill in the supposed fundamental.

Our sense of 'pitch' (not quite the same as frequency) also gets less as
you go to higher frequencies.

In reality, the problems with *early* digital/CD audio were down to poor
ADCs, DACs, processors, etc. Also to some witless behaviour by those using
the kit to make/process recordings. Plus many very early digital recordings
weren't at 44k1 and get resampled badly for the early CD releases, etc.

In terms of what you can choose/use now, a basic laptop and around a
hundred quid USB ADC/DAC can do far better than those early examples.
Simply due to better implimentation of what is required.

Jim

[David Woolley]

Post by Brian Gregory
But if the mixture of harmonics are all above the range of human hearing
you won't hear anything. Your brain will have no information to use to
fill in the supposed fundamental.

You are assuming a totally linear system. The transducers and
electronics will have defects, and so will the mechanical parts of the
ear. They will produce intermodulation products, which are in the
audible range.

[Jim Lesurf]

Post by David Woolley

Post by Brian Gregory
But if the mixture of harmonics are all above the range of human
hearing you won't hear anything. Your brain will have no information
to use to fill in the supposed fundamental.

You are assuming a totally linear system. The transducers and
electronics will have defects, and so will the mechanical parts of the
ear. They will produce intermodulation products, which are in the
audible range.

Yes. However decent DACs in items like CD players shouldn't output much in
the way of tones > 22kHz. Thus the scope for that causing intermod products
at much lower frequencies should be pretty limited. And if the DAC is that
poor then it may futz up 48k material as well.

Alternatively a nonlinear cartridge in ye olde LP replay may well create
rather more garbage at HF - then 'enhanced' by HF nonlinearities in some
speakers. But not really relevant to the title of this thread...


Jim

[Brian Gregory]

Post by Brian Gregory
But if the mixture of harmonics are all above the range of human
hearing you won't hear anything. Your brain will have no information
to use to fill in the supposed fundamental.

You are assuming a totally linear system.  The transducers and
electronics will have defects, and so will the mechanical parts of the
ear. They will produce intermodulation products, which are in the
audible range.

The effects of non linearity in the ear must be small or there would be
little point in putting effort into making good quality audio equipment.

[Jim Lesurf]

Post by Brian Gregory
The effects of non linearity in the ear must be small or there would be
little point in putting effort into making good quality audio equipment.

Curiously, the ear itself is actually very *nonlinear* in its reponse. Yet
our brain 'reprocesses' the results and converts it (for most of us) into
seeming 'linear'.

However this process tends to be based on a 'learning' process often at
unconscious levels. One reason why people can come to 'prefer' some things
which we know have been altered in a nonlinear manner.

Jim

[David Woolley]

Post by Brian Gregory
The effects of non linearity in the ear must be small or there would be
little point in putting effort into making good quality audio equipment.

Because the same distortion is present for the live source.

[Brian Gregory]

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage was,
given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.
...

Agreed. It sounds like the person making the argument is talking out of
their arse.

In practice the frequency response of a 44.1kHz sampled digital
recording isn't perfectly flat up to 22.05kHz where it suddenly cuts
off. Instead there will be some kind of filtering that makes the
response tail off and reach a good level of attenuation by 22.05kHz.
There will be some level of attenuation for frequencies a little below
22.05kHz and some level of alias frequency generation for frequencies
above 22.05kHz.

For recordings made in a studio with expensive equipment it should be
close but in other cases like a home made recording done with a cheapish
PC sound card probably not.

Anyhow there might be some difference within the audible range just by
increasing the sample rate from 44.1kHz to 48kHz.

[David Paste]

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD).

I've no doubt it will do if you believe it to.

I asked what the advantage was,

Post by NY
given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.

All just my anecdote, so it's worth what you've paid me for it:

I have an old book, from the late '60s or early '70s about how to set up
your own hi-fi. It mentions reproduction of 16 kHz. This set me
thinking...

I ripped some CDs and in Audacity applied a low-pass filter at 16 kHz.
Through good headphones there was practically no difference. Maybe my
ears?

Did a ropey DIY test of hearing, I can hear up to about 17 kHz, 20 kHz at
loud volumes, most unpleasant.

So I then took the test tracks and applied a high pass filter at 15, 16,
and 17 kHz to see what is actually up there. Not a lot, as it turns out.
And of curse, why would there be? If I am listening to music, I'm
listening to music, not incidental artefacts of the instruments / voices /
harmonics at higher frequencies.

I then ripped some songs from youtube in mp4 and opus formats. The opus
files don't roll off at 16 kHz whilst mp4 ones do (also, mp4 are 44.1k and
opus are 48k but that is practically irrelevant for this). Really nothing
to miss. Do some more tinkering in audacity to hear what the differences
are, and honesty I can't say there is any, or if there are, it's probably
encoding artefacts between the two.

"Fact is" that most hi-fi recorded material these days is dependent on the
recording competence and mastering than the encoding. I'm always happy to
be proven wrong but I suspect any actual benefits will be rendered null by
a combination of factors including biology, replay equipment, listening
space, amongst others.

Recording at higher sampling rates is a benefit for editing. Playback is
more or less perfect at 44.1.

Post by NY
Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.

Doesn't matter at all. It would be so quiet that it's lost to noise. If
the 20-ish kHz signal loud enough to be heard as part of the music, it
will be horrible to listen to.

Post by NY
What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

No idea, but I always thought NICAM sounded alright when I was younger ;)

[Smolley]

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage was,
given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.
Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.
What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

https://imgur.com/a/sZSfiL0

[Smolley]

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage was,
given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.
Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.
What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

https://imgur.com/a/Jiq3aeX

[Jim Lesurf]

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage
was, given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.
Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.

This may provide some info albeit from a different context that has been
driven by the beliefs some have about this topic.
https://www.audiomisc.co.uk/MQA/questions/QuriousAssumptions.html

Post by NY
What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

Some people can hear sounds above c22kHz. But this ability isn't universal.
Not is the presence of such components in music, etc, that make an audible
difference.

44.1k was adopted by Philips/Sony as convenient for other reasons at the
time. 48kHz gives more 'elbow room' for being able to record and playback
whilst giving good results up to c20kHz. So became the pro/boradcasting
choice later on. (Ignoring the BBC's use of NICAM. :-) )

Curiously, all too many Hi-Fi enthusiasts have liked BBC R3 via FM when
criticising 44.1k as being too low. Go Figger as the Americans say.. 8-]

Jim

[Phil_M]

Post by Jim Lesurf

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage
was, given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.
Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.

This may provide some info albeit from a different context that has been
driven by the beliefs some have about this topic.
https://www.audiomisc.co.uk/MQA/questions/QuriousAssumptions.html

Post by NY
What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

Some people can hear sounds above c22kHz. But this ability isn't universal.
Not is the presence of such components in music, etc, that make an audible
difference.
44.1k was adopted by Philips/Sony as convenient for other reasons at the
time. 48kHz gives more 'elbow room' for being able to record and playback
whilst giving good results up to c20kHz. So became the pro/boradcasting
choice later on. (Ignoring the BBC's use of NICAM. :-) )
Curiously, all too many Hi-Fi enthusiasts have liked BBC R3 via FM when
criticising 44.1k as being too low. Go Figger as the Americans say.. 8-]
Jim

I remember listening to both FM and DAB in a proper listening room many
years ago, and found I liked the FM version better as the DAB seemed to
have no "body" to it somehow. I think the DAB bit rate was 320kHz. I
tend to listen to R3 on FM at home and DAB in the car.

Phil M

[Jim Lesurf]

Post by Phil_M
I remember listening to both FM and DAB in a proper listening room many
years ago, and found I liked the FM version better as the DAB seemed to
have no "body" to it somehow. I think the DAB bit rate was 320kHz. I
tend to listen to R3 on FM at home and DAB in the car.

BBC FM is level compressed by their Optimod units in a way that DAB and
iPlayer are not. This shows up in things like piano on R3 FM having added
'warmth' due to the compression raising the level of the 'tail' of piano
chords relative to their initial peak.

People get used to this and tend to like it as at home it brings out more
audibility of the experience.

IIRC I did some comparison measurements at the time. Someone sent me a
digital capture of DAB as well to compare. These days I use iPlayer 320k
as the reference as it is closer to what gets sent out to be distributed
for FM, DAB, etc.

Jim

[Brian-Gaff (bed 2)]

I feel that compression is really all over the place these days. Add that to
lossy encoding and it means I've seldom heard much good quality for years
however its delivered. A friend of mine was critical of my system being too
loud then too quiet on an old CD made by Denon, and to me the dynamics
seemed pretty much what you would hear at a concert. I fear that people have
become accustomed to compression and lack of detail and find the more
realistic recordings wrong. Brian
--
From the Bed of Brian Gaff Reply address is active
Remember, if you don't like where I post
or what I say, you don't have to
read my posts! :-)

Post by Jim Lesurf

Post by Phil_M
I remember listening to both FM and DAB in a proper listening room many
years ago, and found I liked the FM version better as the DAB seemed to
have no "body" to it somehow. I think the DAB bit rate was 320kHz. I
tend to listen to R3 on FM at home and DAB in the car.

BBC FM is level compressed by their Optimod units in a way that DAB and
iPlayer are not. This shows up in things like piano on R3 FM having added
'warmth' due to the compression raising the level of the 'tail' of piano
chords relative to their initial peak.
People get used to this and tend to like it as at home it brings out more
audibility of the experience.
IIRC I did some comparison measurements at the time. Someone sent me a
digital capture of DAB as well to compare. These days I use iPlayer 320k
as the reference as it is closer to what gets sent out to be distributed
for FM, DAB, etc.
Jim
--
Please use the address on the audiomisc page if you wish to email me.
Electronics
https://www.st-andrews.ac.uk/~www_pa/Scots_Guide/intro/electron.htm
biog http://jcgl.orpheusweb.co.uk/history/ups_and_downs.html
Audio Misc http://www.audiomisc.co.uk/index.html

[Jim Lesurf]

I'd agree. The BBC 320k streams *can* sound superb. But this depends on who
'balanced' what was fed into the encoder, etc. All too often level
compression is applied assuming no-one is listening 'seriously'.

And the 'outsourcing' of many programmes means they get into "louder is
better" treatment by commercial companies.

Add to that many BBC items on iPlayer now bet "toped and tailed" with
promos for other items - often at wildly different levels to the main
programme. With no-one overall checking this and requiring a consistent
treatment of loudness.

Now all producers, few directly employed engineers.

Jim

Post by Brian-Gaff (bed 2)
I feel that compression is really all over the place these days. Add
that to lossy encoding and it means I've seldom heard much good quality
for years however its delivered. A friend of mine was critical of my
system being too loud then too quiet on an old CD made by Denon, and
to me the dynamics seemed pretty much what you would hear at a concert.
I fear that people have become accustomed to compression and lack of
detail and find the more realistic recordings wrong. Brian

[Vir Campestris]

Post by NY
In a Facebook discussion, someone has said that sampling at 48 kHz gives
better sound quality than 44.1 kHz (CD). I asked what the advantage was,
given that humans cannot hear above about 20 kHz (and that is for
children: the limit decreases with age). And he replied that it is for
the harmonics.
Am I going loopy? Does it matter whether the frequency above 20 kHz is a
fundamental or a harmonic (eg 3rd harmonic of 8 kHz)? If it's above the
range of human hearing it's not detected and it's as if it was never
present. OK, I realise that human hearing isn't a flat response with a
brick-wall above 20 kHz.
What is the real reason than some sound formats sample at 48 rather than
44.1 kHz. I gather that the exact value 44.1 (as opposed to 44, for
example) is governed by bandwidth of TV systems, given that high-quality
sound was recorded at one time on videotape as a pseudo-video signal.
But roughly 44 kHz is enough to give 20 kHz (by Nyquist) with a bit of
leeway for low-pass filters which need to make sure there is no
discernable signal above (sampling frequency) / 2 to avoid the dreaded
aliasing.

Coming into this late, surely the problem isn't when you play a 44.1
sampled sound at 44.1, or a 48 at 48, but when you resample the one into
the other?

Andy

[Jim Lesurf]

Post by Vir Campestris
Coming into this late, surely the problem isn't when you play a 44.1
sampled sound at 44.1, or a 48 at 48, but when you resample the one into
the other?

That is, indeed, quite an 'awkward' process to perform well as the rates
are low (in terms of the required audio bandwidth) and 'close' in terms of
finding any common factor.

In practice good digital convertors will tend to do it via upsampling the
input to a much higher rate and then downsampling it. At least in terms of
the process of computation. e.g. use a sinc filter with a high sample rate
and loads of taps.

That was between hard and impossible back when CD was launched. But
nowdays you can do it with 'sox' in real time. :-)

Jim