[[email protected]]

I was wondering what sample rates engineers are mixing and tracking
at these days. I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What is
the advantage if any to tracking and mixing to 96k before go to a cd?
Glenn.

[Paul Stamler]

wrote in message
...
I was wondering what sample rates engineers are mixing and tracking
at these days. I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What is
the advantage if any to tracking and mixing to 96k before go to a cd?

If you're mixing on an analog console, you might as well track to 96k. It
won't hurt anything, storage space is cheap, and it ought to please any
passing bat.

If you're mixing in the box, recording to 44.1k means no sample rate
conversion, which is a good thing in my book. I've been told that if you're
doing a lot of in-the-box EQ in the top octave it's useful to record at 96k
to avoid computation errors causing artifacts. Don't know if that's really
true, or just urban legend, but since I don't do a lot of that kind of EQ
work, I've never done the comparison.

I normally record to 44.1k for everything except mic checks and distortion
testing.

Peace,
Paul

[Anahata]

On Sat, 25 Oct 2008 06:59:06 +0000, Paul Stamler wrote:

I've been told that if
you're doing a lot of in-the-box EQ in the top octave it's useful to
record at 96k to avoid computation errors causing artifacts.

Any non-linear distortion (clipping would be a gross example, but
compression also does it) creates harmonics over 20kHz which can be
aliased back into the audible band at 44.1KHz S/R.
At 96kH S/R, all harmonics up to about 76kHz will not be aliased. That's
got to make some difference.

I guess that accumulated rounding errors in the EQ processing also
introduce some noise that gets aliased to lower frequencies in the same
way.

--
Anahata
-+- http://www.treewind.co.uk
Home: 01638 720444 Mob: 07976 263827

[Karl Engel]

Do the brick wall filters that supposedly prevent aliasing distortion on the
way in at 44.1 not also work on the way out?

(Obviously revealing my weaknesses in understanding here, but doesn't the
DAC have a way of removing the numbers that represent very high frequency
content which might be misrepresented as lower frequency distortion during a
44.1 conversion?)


--
"Anahata" wrote in message
et...
On Sat, 25 Oct 2008 06:59:06 +0000, Paul Stamler wrote:

I've been told that if
you're doing a lot of in-the-box EQ in the top octave it's useful to
record at 96k to avoid computation errors causing artifacts.

Any non-linear distortion (clipping would be a gross example, but
compression also does it) creates harmonics over 20kHz which can be
aliased back into the audible band at 44.1KHz S/R.
At 96kH S/R, all harmonics up to about 76kHz will not be aliased. That's
got to make some difference.

I guess that accumulated rounding errors in the EQ processing also
introduce some noise that gets aliased to lower frequencies in the same
way.

--
Anahata
-+- http://www.treewind.co.uk
Home: 01638 720444 Mob: 07976 263827

[Mike Rivers]

wrote:
I was wondering what sample rates engineers are mixing and tracking
at these days.

Whatever the customer asks for. If I'm making the decisions, I record at
44.1 kHz, 24-bit in the studio or if I have a full setup in the field,
16-bit if I'm making a casual recording.

I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What is
the advantage if any to tracking and mixing to 96k before go to a cd?

Given your sources and projects, really none. Some plug-ins work better
at 96 kHz, particularly noise removal tools, but few mics have any
usable response above 20 kHz, as do few speakers. No point in burning up
disk space on stuff that you can't hear. Better to just not record it
than to have it recorded, not be able to hear that it's getting in the way.


--
If you e-mail me and it bounces, use your secret decoder ring and reach
me he
double-m-eleven-double-zero at yahoo -- I'm really Mike Rivers
)

[Scott Dorsey]

wrote:
I was wondering what sample rates engineers are mixing and tracking
at these days. I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What is
the advantage if any to tracking and mixing to 96k before go to a cd?

No real advantage other than that you can put "96K MIX" on the cover of
the album.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

Karl Engel wrote:
Do the brick wall filters that supposedly prevent aliasing distortion on the
way in at 44.1 not also work on the way out?

Yes, although in the modern era we use oversampling to avoid having to
actually implement brickwall filters in the analogue world.

(Obviously revealing my weaknesses in understanding here, but doesn't the
DAC have a way of removing the numbers that represent very high frequency
content which might be misrepresented as lower frequency distortion during a
44.1 conversion?)

A hardware sample rate converter, though, will filter high frequencies in
the process of reducing the sample rate. The data sheet for the AD1890
goes through all the math.

Presumably an equalizing filter will also have internal stuff to deal with
this, to prevent it from creating aliasing. I know the Oxford filters do.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[[email protected]]

On Oct 25, 4:23*am, Anahata wrote:
On Sat, 25 Oct 2008 06:59:06 +0000, Paul Stamler wrote:
I've been told that if
you're doing a lot of in-the-box EQ in the top octave it's useful to
record at 96k to avoid computation errors causing artifacts.

Any non-linear distortion (clipping would be a gross example, but
compression also does it) creates harmonics over 20kHz which can be
aliased back into the audible band at 44.1KHz S/R.
At 96kH S/R, all harmonics up to about 76kHz will not be aliased. That's
got to make some difference.

I guess that accumulated rounding errors in the EQ processing also
introduce some noise that gets aliased to lower frequencies in the same
way.

--
Anahata
-+-http://www.treewind.co.uk
Home: 01638 720444 * * * * Mob: 07976 263827

If I were to record at 96k and mixdown to 96k, it would sound better
than tracking and mixing at 44k? I remember trying 96k a few years
ago. Then I was told you were better off to pour from a larger cup.
That is record at 96k then funnel down to 44k.
I mix in the box. No analog in the path.
I don't like suprises. a few years ago when I decided to track at
44k my logic was why have a 96k clear playback that changes everytime
you bounce down to a cd. Of course this logic could be wrong? that's
why I am writing this.
Glenn.

[[email protected]]

On Oct 25, 9:17*am, Mike Rivers wrote:
wrote:
I was *wondering what sample rates engineers are mixing and tracking
at these days. *

Whatever the customer asks for. If I'm making the decisions, I record at
44.1 kHz, 24-bit in the studio or if I have a full setup in the field,
16-bit if I'm making a casual recording.

* I have been tracking in pro tools at 44k 24 bit for

the last few years. Mostly pop, rap and blues bands this year. What is
the advantage if any to tracking and mixing to 96k before go to a cd?

Given your sources and projects, really none. Some plug-ins work better
at 96 kHz, particularly noise removal tools, but few mics have any
usable response above 20 kHz, as do few speakers. No point in burning up
disk space on stuff that you can't hear. Better to just not record it
than to have it recorded, not be able to hear that it's getting in the way.

--
If you e-mail me and it bounces, use your secret decoder ring and reach
me he
double-m-eleven-double-zero at yahoo -- I'm really Mike Rivers
)

Let me ask this. If I were to record let's say my blues band. If I
recorded a song and tracked it at 96k and mixed it to a 96k master.
(not a cd). Then I recorded the same song at 44k and mixed to a 44k
master. Do you think I would hear a difference between the two?

Glenn.

[philicorda[_6_]]

On Sat, 25 Oct 2008 15:39:47 -0700, glennerd1 wrote:

snip
Let me ask this. If I were to record let's say my blues band. If I
recorded a song and tracked it at 96k and mixed it to a 96k master. (not
a cd). Then I recorded the same song at 44k and mixed to a 44k master.
Do you think I would hear a difference between the two?

Try it. There are some variables like how your particular converters
sound at 96k vs 44k, and how the plugins you commonly use are affected.
I'd be interested in the results.

Years back I started recording a band's album at 96k, but found the
computer was not up to it when the overdubs started piling up. So I
converted the entire multitrack to 44k. It did not sound as good, but I
have a strong suspicion that the sample rate converter (Nuendo 1.0) was
responsible.

[Mike Rivers]

wrote:

Let me ask this. If I were to record let's say my blues band. If I
recorded a song and tracked it at 96k and mixed it to a 96k master.
(not a cd). Then I recorded the same song at 44k and mixed to a 44k
master. Do you think I would hear a difference between the two?

I would certainly think you'd hear a difference between the two, but not
necessarily because of the difference in sample rate, but because you
played it differently the two times.

As to whether there would be a difference in quality, that depends on th
equipment that you used. If you used the same mics and they were kind of
cheap mics, no, you wouldn't notice any difference. If you used great
mics in a great room and went from the cheap sound card that came with
your computer to a high grade A/D and D/A converter, and were monitoring
on wide bandwidth speakers, you probably would notice a difference. If
you have an M-Audio USB interface and just threw the switch from 44.1 to
96, you probably wouldh't notice a difference.

Or maybe you would. My editor at Pro Audio Review loves recording at 96
kHz even if he'd just strumming his guitar at home. He says it sounds
better than 44.1, but he has really good converters, mics, preamps, and
speakers. But he also knocks a lot of the new portable recorders because
they won't record at 96 kHz. I think he's nuts, but what makes him happy
is what's important.

I once attended a demonstration that my dealer set up which was actually
a demo of the Sennheiser MKH-800 mic. The idea was that the mic has
usable frequency response up to 40 kHz so this is (finallly!) a
justification to record at 2X sample rate. He had a singer with an
acoustic guitar set up in a good studio with a crossed pair of the mics.
He had set up two sets of Lavry Blue A/D and D/A converters connected
back-to-back, one set for 44.1 kHz, the other for 96 kHz. I don't
remember what mic preamp he was using, but it was a good one. Monitors
were some large PMCs. He had a switch box set up so that we could hear
the mic preamp output directly, the preamp through the 44.1 kHz
converter chain, and through the 96 kHz converter chain. Levels were
calibrated and matched accurately, and we set to listening.

I was actually surprised that I could hear a difference, but it wasn't
in the high end as you'd expect, the low end sounded a little better
with the 96 kHz converters than the 44.1 kHz converters. The direct
preamp output sounded best. It also surprised me that I could hear a
difference between the straight preamp output and going through those
high class converters. But honestly it wasn't a big difference. It was
very small. The difference between the Lavry converter at 44.1 kHz and,
say, an M-Audio converter would have been much more noticeable than
switching sample rates with the Lavry. But there's probably a 10 times
difference in price.

Now, just why is it that you've asking this question? Are you thinking
about buying a new audio interface? If so, what? Or did you flip the
switch on the gear that you have now, not hear any difference, and want
to know whether there's something wrong with your ears, your monitors,
or are you concerned that someone else would hear something that you
don't? Or have you done no experienting yourself yet?



--
If you e-mail me and it bounces, use your secret decoder ring and reach
me he
double-m-eleven-double-zero at yahoo -- I'm really Mike Rivers
)

[Mike Rivers]

wrote:

I don't like suprises. a few years ago when I decided to track at
44k my logic was why have a 96k clear playback that changes everytime
you bounce down to a cd. Of course this logic could be wrong? that's
why I am writing this.

Well, OK, you've sort of answered the question I asked in my previous
post. 10 years ago, sample rate conversion was either absurdly expensive
or not very accurate, so there was no advantage, and maybe even a
disadvantage to recording at 96 kHz if the recording was going to stay
digital forever and only to to a 44.1 kHz medium. Today sample rate
conversion is no big deal and there are people who think that SOME 96
kHz recordings converted to 44.1 kHz sound better than SOME recordings
made at 44.1 kHz.

But how are you, why do you care, and can you afford to care enough to
really do something that makes a difference? If you've got a string of
hits and just want to spend some more money in the studio because you
can, sure, go for it. Record both ways and pick which you like better.
You might not like the same process on all songs. But if you're
recording your band in your garage and selling your music on an MP3
download service or even CDs at gigs, sheesh! don't bother. Put your
money into better mics or better preamps or better A/D converters or
monitors, or work on the acoustics of your studio and control room.

Sure, disk space is pretty cheap, but then you have twice as much to
back up. It's just not worth it in the general case.



--
If you e-mail me and it bounces, use your secret decoder ring and reach
me he
double-m-eleven-double-zero at yahoo -- I'm really Mike Rivers
)

[Scott Dorsey]

wrote:

Let me ask this. If I were to record let's say my blues band. If I
recorded a song and tracked it at 96k and mixed it to a 96k master.
(not a cd). Then I recorded the same song at 44k and mixed to a 44k
master. Do you think I would hear a difference between the two?

You might well hear a difference between the two, but I wouldn't place
any money on which one sounded better.

Some converters sound a lot more neutral at 44.1 than at 96. Others do
not.

Paul Stamler notes some issues with doing processing in the digital domain
at the lower sampling rate. If your EQ software is properly implemented,
it's not an issue... but a lot of software isn't, and it's entirely possible
yours might sound better with radical EQ if you're using the higher sampling
rate. It's also possible it won't.

Of course, if you're mixing with an analogue console and using analogue
EQ and compression, all of those arguments are moot anyway.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[[email protected]]

On Oct 25, 9:55*am, (Scott Dorsey) wrote:
Karl Engel wrote:
Do the brick wall filters that supposedly prevent aliasing distortion on the
way in at 44.1 not also work on the way out?

Yes, although in the modern era we use oversampling to avoid having to
actually implement brickwall filters in the analogue world.

(Obviously revealing my weaknesses in understanding here, but doesn't the
DAC have a way of removing the numbers that represent very high frequency
content which might be misrepresented as lower frequency distortion during a
44.1 conversion?)

A hardware sample rate converter, though, will filter high frequencies in
the process of reducing the sample rate. *The data sheet for the AD1890
goes through all the math.

Presumably an equalizing filter will also have internal stuff to deal with
this, to prevent it from creating aliasing. *I know the Oxford filters do.
--scott
--
"C'est un Nagra. *C'est suisse, et tres, tres precis."

Let me ask another question. If I were to record , let's say my blues
band. If I recorded a song at 96k and mixed it to a 96k master. (not a
cd) Then if I tracked and mixed that same song ( another take) using
44k tracking and mixed down to a 44k master. Do you think I could hear
a difference?
Glenn.

[Paul Stamler]

wrote in message
...

Let me ask this. If I were to record let's say my blues band. If I
recorded a song and tracked it at 96k and mixed it to a 96k master.
(not a cd). Then I recorded the same song at 44k and mixed to a 44k
master. Do you think I would hear a difference between the two?

Yes, but a pretty small one if my experience is representative. Not nearly
as much as the difference between a 16 bit and 24 bit recording.

Peace,
Paul

[Peter Larsen[_3_]]

wrote:

Let me ask this. If I were to record let's say my blues band. If I
recorded a song and tracked it at 96k and mixed it to a 96k master.
(not a cd). Then I recorded the same song at 44k and mixed to a 44k
master. Do you think I would hear a difference between the two?

There is a clear, unambiguous answer to this question: perhaps. What matters
in this context is - in my religion - the initial AD conversion, some
converters - such as my Midiman Delta cards - provide better sounding
samplings when running at 96 kHz. But is darn bulky to work with, so it
might make more sense to convert to 48-32 (as pr. an AES recommendation I
recall reading in some context) before mixing, if mixing in the box.

There is a counterpoint, some real world dynamic microphones have strange
things going on above 20 kHz, narrow 10 dB peaks for instance. There is
some kind of point in tracking at 64 or even 96 if you use recent Schoeps or
DPA gear tat produces usable, clean output to 30+ kHz on violins, but there
could also be a very good point in favour of using 48 kHz in case you
sprinkle a mist of dynamics all over a rock band.

"What do you solve?" is a very good question to ask oneself prior to doing
something, some of the time there really is not a problem to solve by "doing
it right".

Glenn

Kind regards

Peter Larsen

[Anahata]

On Sat, 25 Oct 2008 22:44:13 +1100, Karl Engel wrote:

Do the brick wall filters that supposedly prevent aliasing distortion on
the way in at 44.1 not also work on the way out?

No, because the aliasing products below 20kHz are already present in the
digital domain once the distortion has happened.

A lot of people don't seem to understand this.

--
Anahata
-+- http://www.treewind.co.uk
Home: 01638 720444 Mob: 07976 263827

[Scott Dorsey]

Anahata wrote:
On Sat, 25 Oct 2008 22:44:13 +1100, Karl Engel wrote:

Do the brick wall filters that supposedly prevent aliasing distortion on
the way in at 44.1 not also work on the way out?

No, because the aliasing products below 20kHz are already present in the
digital domain once the distortion has happened.

The filters on the way in are anti-aliasing filters, that prevent high
frequencies from getting into the converter and causing aliasing.

On the other hand, the filters on the way out are "reconstruction filters"
(I know some folks hate that term but it's unfortunately standard terminology)
which remove high frequency components generated spuriously in the D/A process.
If they are not removed, they are apt to cause distortion in the analogue
electronics after the fact.

A lot of people don't seem to understand this.

The FAQ has a really good explanation of how all this stuff works.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[HKC]

wrote:
I was wondering what sample rates engineers are mixing and tracking
at these days. I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What is
the advantage if any to tracking and mixing to 96k before go to a cd?



There is one big (ish) advantage and that is the latency is lowered when
increasing the samplerate. Latency is based on samples and if you go from 48
to 96 a simple calculation would show that if 64 samples means 1.5 ms at 48
KHz it would become 0.75 at 96 KHz.
Some of the good soundcards like the rather recent RME Raydat is reported to
work very well at 32 samples (and no safety-buffer, the card is a PCIe) so
the big issue when it comes to latency these days is the latency that comes
from the AD/DA conversion which again gives you another thing to look for
apart from the converters actual sound, how fast are they.
The really fast ones are below 10ms which means that it's now possible to
get below 1 ms from source to headphones if you have a system like that
(could be an entire RME system like Raydat and the QS converters or the
Apogee Symphony card with the AD/DA16x converters).
Soundwise I doubt that you will find much improvement going from 44.1 to 96
but during tracking it's pretty cool.

PS I would probably stay in 88.2 to stay clear from too much conversion.....

[Mike Rivers]

HKC wrote:

There is one big (ish) advantage and that is the latency is lowered when
increasing the samplerate.

That used to be the case, but it isn't always any more. We had this
discussion a while back (I think Arny came up with a reasonable
explanation) but it's not uncommon for devices to have just a tiny bit
less latency when doubling the sample rate than the half that you'd expect.

PS I would probably stay in 88.2 to stay clear from too much conversion.....

It's another fallacy that integer ratios of sample rates are easier to
convert than odd ratios. No matter what the ratio of output to input
sample rate, you still have to actually resample (mathematically) or
you'll get the wrong result. You can't just take every other sample of
an 88.2 kHz audio stream and call it a 44.1 kHz stream.



--
If you e-mail me and it bounces, use your secret decoder ring and reach
me he
double-m-eleven-double-zero at yahoo -- I'm really Mike Rivers
)

[Eeyore]

wrote:

I was wondering what sample rates engineers are mixing and tracking
at these days. I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What is
the advantage if any to tracking and mixing to 96k before go to a cd?

Well, I know some people using 192kHz and they say it DOES sound 'better'.

Graham

[Eeyore]

philicorda wrote:

glennerd1 wrote:

Let me ask this. If I were to record let's say my blues band. If I
recorded a song and tracked it at 96k and mixed it to a 96k master. (not
a cd). Then I recorded the same song at 44k and mixed to a 44k master.
Do you think I would hear a difference between the two?

Try it. There are some variables like how your particular converters
sound at 96k vs 44k, and how the plugins you commonly use are affected.
I'd be interested in the results.

Years back I started recording a band's album at 96k, but found the
computer was not up to it when the overdubs started piling up. So I
converted the entire multitrack to 44k. It did not sound as good, but I
have a strong suspicion that the sample rate converter (Nuendo 1.0) was
responsible.

Never could fathom why 88.2kHz never caught on. Such problems would be hugely
minimised.

Graham

[Arny Krueger]

"Anahata" wrote in message
et...
On Sat, 25 Oct 2008 06:59:06 +0000, Paul Stamler wrote:

I've been told that if
you're doing a lot of in-the-box EQ in the top octave it's useful to
record at 96k to avoid computation errors causing artifacts.

As things evolve, this is only a problem if you screw up and drive the eq
itself into clipping.

Any non-linear distortion (clipping would be a gross example, but
compression also does it) creates harmonics over 20kHz which can be
aliased back into the audible band at 44.1KHz S/R.

There is no nonlinear distortion in a digital filter unless you screw up and
overdrive it. Digital filters generally have far more dynamic range than
analog filters, so it is unlikely that it will clip unless someone really
screws up.

Harmonics due to clipping that get folded back are much smaller than the
harmonicsdue to clipping that naturally fall in the passband. Therefore
foldback is second-order effect, and minor compared to the primary problem,
which is clipping.

When you clip an audio signal, fidelity goes out the window no matter what
the sample rate is.

At 96kH S/R, all harmonics up to about 76kHz will not be aliased. That's
got to make some difference.

It makes a very bad situation a little less bad.

I can see only one case where this is actually a real-world problem, and
that would be a digital EFX unit that is designed to have nonlinear
distortion. It would be up to the designer of this filter to use whatever
means he needed to use, in order to get the intended operation of the unit.

Digital consoles are designed to be free of nonlinear distortion, and have
more than enough dynamic range that there should be no problem with clipping
within the confines of the digital console. Therefore this is not a serious
problem.

I guess that accumulated rounding errors in the EQ processing also
introduce some noise that gets aliased to lower frequencies in the same
way.

However, the accumulated rounding error is very small.

It seems like using a longer data word is a more productive way to
accomplish the same goal. If I go from 16 bits to 24 bits, rounding error is
tremendously decreased, but the data involved only increases by 50%. If I
increase the sample rate from 44.1 to 96, the amount of data increases by a
factor of well over 2.

[Arny Krueger]

"Eeyore" wrote in message
...

Never could fathom why 88.2kHz never caught on. Such problems would be
hugely
minimised.

If you look at Yamaha's digital consoles, you'll notice that their older
products sampled up to 96 KHz, but most of their more recent products only
sample up to 48 KHz.

[Chris Hornbeck]

On Mon, 27 Oct 2008 00:20:32 +0000, Eeyore
wrote:

philicorda wrote:

glennerd1 wrote:

Let me ask this. If I were to record let's say my blues band. If I
recorded a song and tracked it at 96k and mixed it to a 96k master. (not
a cd). Then I recorded the same song at 44k and mixed to a 44k master.
Do you think I would hear a difference between the two?

Try it. There are some variables like how your particular converters
sound at 96k vs 44k, and how the plugins you commonly use are affected.
I'd be interested in the results.

Years back I started recording a band's album at 96k, but found the
computer was not up to it when the overdubs started piling up. So I
converted the entire multitrack to 44k. It did not sound as good, but I
have a strong suspicion that the sample rate converter (Nuendo 1.0) was
responsible.

Never could fathom why 88.2kHz never caught on. Such problems would be hugely
minimised.

This choice *could* give an advantage in a hardware solution, where
the recording was done at the modern high-sample-rate-low-bit-depth
native to modern A/D converters, followed by anti-aliasing filtering
for conversion to a "working" sampling rate of, say, 88.2K
samples/sec.

Storage and all manipulations then occur at the "working" sampling
rate, after which another low-pass filter, now called a
"reconstruction" filter, and decimation, precede D/A.

The only thing required for this special case is that *no other*
sampling rate can be allowed. Only the integer relationship -
no problem in a hardware device, and even a (technical) advantage.

In a general purpose machine I can see this being a very different
problem, requiring sample rate muliplication engines.

Much thanks, as always,
Chris Hornbeck

[[email protected]]

On Oct 24, 10:42*pm, wrote:
I was *wondering what sample rates engineers are mixing and tracking
at these days. *I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What is
the advantage if any to tracking and mixing to 96k before go to a cd?
Glenn.

For in the box stuff.. i like 48k 16 bit. 48k gets you over a lot of
the aliasing probs. Outside the box, 88.2 makes more sense if you want
to get all the harmonics. The difference from 96k, and 88.2k is not
worth the sonics. Down converting to 44.1 from 88.2 on your own will
sound better, and the math is cleaner. Cutting at 16 bit might sound
silly to you , but bit depth will be more affected by tracking style.
If you are packing it with compression, leaving lots of head room, or
clipping your converters to get it up to the top. How you use your
tools will determine the sound more than the confines of the medium.
Keep experimenting, and use the things on the side of your head.

Paul Logus

[Scott Dorsey]

wrote:

For in the box stuff.. i like 48k 16 bit. 48k gets you over a lot of
the aliasing probs.

What aliasing problems?

Outside the box, 88.2 makes more sense if you want
to get all the harmonics.

What harmonics?
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[[email protected]]

On Oct 25, 8:19*pm, wrote:
On Oct 25, 9:55*am, (Scott Dorsey) wrote:





Karl Engel wrote:
Do the brick wall filters that supposedly prevent aliasing distortion on the
way in at 44.1 not also work on the way out?

Yes, although in the modern era we use oversampling to avoid having to
actually implement brickwall filters in the analogue world.

(Obviously revealing my weaknesses in understanding here, but doesn't the
DAC have a way of removing the numbers that represent very high frequency
content which might be misrepresented as lower frequency distortion during a
44.1 conversion?)

A hardware sample rate converter, though, will filter high frequencies in
the process of reducing the sample rate. *The data sheet for the AD1890
goes through all the math.

Presumably an equalizing filter will also have internal stuff to deal with
this, to prevent it from creating aliasing. *I know the Oxford filters do.
--scott
--
"C'est un Nagra. *C'est suisse, et tres, tres precis."

Let me ask another question. *If I were to record , let's say my blues
band. If I recorded a song at 96k and mixed it to a 96k master. (not a
cd) Then if I tracked and mixed that same song ( another take) using
44k tracking and mixed down to a 44k master. Do you think I could hear
a difference?
* * * * * * * * *Glenn.- Hide quoted text -

- Show quoted text -

Sorry again about the double posting. I was unable to get any updates
on my computer after about 11:30 Saturday night. I don't know if this
site was down? I guess it was my system.At any rate( drum crash. ha
ha) thanks for the info. I will record this week at 44k and 96k. Last
time I looked into it. I could hear more highs in a Dat recorded at
48k than one at 44k. They told me when I bought the Pro tools HD
system to buy the 192 interface even if I recorded only at 44k or
48k. Why? They said it had better converters than the Digidesign 96k
interface. Yea that's right even the unde rpaid over worked guys at
Guitar Center can get it right,once in a while.
Glenn.

[Scott Dorsey]

wrote:
Sorry again about the double posting. I was unable to get any updates
on my computer after about 11:30 Saturday night. I don't know if this
site was down? I guess it was my system.

This is not a "site" this is a Usenet group. Your usenet provider may
have been down. There are thousands of Usenet servers around the world,
all run by different organizations, and all set up a little bit differently.

At any rate( drum crash. ha
ha) thanks for the info. I will record this week at 44k and 96k. Last
time I looked into it. I could hear more highs in a Dat recorded at
48k than one at 44k.

Let me guess... you did this using an SV-3700? The SV-3700 was notorious
for poorly designed filters, and they used the exact same filters at both
44.1 and 48 ksamp/sec rates. Some of the early DAT machines did a lot of
shameful corner-cutting like this... and it resulted in bad sound overall
at any sample rate.

They told me when I bought the Pro tools HD
system to buy the 192 interface even if I recorded only at 44k or
48k. Why? They said it had better converters than the Digidesign 96k
interface. Yea that's right even the unde rpaid over worked guys at
Guitar Center can get it right,once in a while.

That's entirely possible. The newer interface may sound better even at
44.1 than the older interface... but if so, it's not because it supports
a higher sample rate, it's because it's a better-sounding interface.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Blind Hog]

Eeyore wrote in
:

wrote:

I was wondering what sample rates engineers are mixing and tracking
at these days. I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What
is the advantage if any to tracking and mixing to 96k before go to a
cd?

Well, I know some people using 192kHz and they say it DOES sound
'better'.

I've been out of the group for a while, so forgive me if I'm repeating.

I track at CD quality. 16 bits and 44.1K capture everything my
microphones and preamps can send.

My first action in the editing software is to expand to 24 bits. My
last action before burning a CD is to convert back to 16 bits. I can
hear the difference after some math-intensive processes like reverb and
noise reduction if I don't alias-out the lowest bits. But I hear no
benefit from frequencies above 22K.

No, my ears aren't as good as they were at 16, but the 16 year olds I
make CDs for sometimes haven't complained yet.

[Geoff]

Blind Hog wrote:
Eeyore wrote in
:

wrote:

I was wondering what sample rates engineers are mixing and tracking
at these days. I have been tracking in pro tools at 44k 24 bit for
the last few years. Mostly pop, rap and blues bands this year. What
is the advantage if any to tracking and mixing to 96k before go to a
cd?

Well, I know some people using 192kHz and they say it DOES sound
'better'.

I've been out of the group for a while, so forgive me if I'm
repeating.

I track at CD quality. 16 bits and 44.1K capture everything my
microphones and preamps can send.

My first action in the editing software is to expand to 24 bits. My
last action before burning a CD is to convert back to 16 bits. I can
hear the difference after some math-intensive processes like reverb
and noise reduction if I don't alias-out the lowest bits. But I hear
no benefit from frequencies above 22K.

Why would you record at 16 bits and 'expand' to 24 bits - why not just
record at 24 ?

geoff

[Blind Hog]

"geoff" wrote in
:

Blind Hog wrote:
Eeyore wrote in
:

wrote:

I was wondering what sample rates engineers are mixing and
tracking at these days. I have been tracking in pro tools at 44k
24 bit for the last few years. Mostly pop, rap and blues bands this
year. What is the advantage if any to tracking and mixing to 96k
before go to a cd?

Well, I know some people using 192kHz and they say it DOES sound
'better'.

I've been out of the group for a while, so forgive me if I'm
repeating.

I track at CD quality. 16 bits and 44.1K capture everything my
microphones and preamps can send.

My first action in the editing software is to expand to 24 bits. My
last action before burning a CD is to convert back to 16 bits. I can
hear the difference after some math-intensive processes like reverb
and noise reduction if I don't alias-out the lowest bits. But I hear
no benefit from frequencies above 22K.

Why would you record at 16 bits and 'expand' to 24 bits - why not just
record at 24 ?

geoff

I could.

But recording at 16 bit gives me more record time/tracks on location.
And it practically impossible to record outside the studio and get more
than 60 dB dynamic range above the noise floor. As long as I have any
kind of handle on my gain and trim settings, 16 bits does fine.

Also, 16 bits copies from the field machine into the editing machine a
third faster.

If I was recording directly to the editing machine (which is both large
and fast) there would be absolutely no advantage. My location recorder
(small and light) is not so capable.

The point is that 24 bits is only valuable during editing, and I hear no
advantage to higher sampling rates at any time as long as the output
medium is no better than CD quality.

If I'm recording for video, the same holds for 48K.

[Geoff]

Blind Hog wrote:


I could.

But recording at 16 bit gives me more record time/tracks on location.
And it practically impossible to record outside the studio and get
more than 60 dB dynamic range above the noise floor. As long as I
have any kind of handle on my gain and trim settings, 16 bits does
fine.

Also, 16 bits copies from the field machine into the editing machine a
third faster.

If I was recording directly to the editing machine (which is both
large and fast) there would be absolutely no advantage. My location
recorder (small and light) is not so capable.

The point is that 24 bits is only valuable during editing, and I hear
no advantage to higher sampling rates at any time as long as the
output medium is no better than CD quality.

If I'm recording for video, the same holds for 48K.


What recording device are you using ? Obviously available media/memory is a
restriction with it for you.....

24 bits would allow less fretting over level-setting optimisation, and
obviate the necessity for extra processing should ypou feel the need to
convert.

I can't imagine that you would be getting any benefit from converting to 24
bits before any processing. But that depends, I guess, on how primitive your
editing app is. The processing 'bitdepth' does not necessarily have to have
a fixed relationship with source media file bit-dept.

geoff

[Arny Krueger]

"geoff" wrote in message
...

24 bits would allow less fretting over level-setting optimisation, and
obviate the necessity for extra processing should you feel the need to
convert.

In the past 3 years I've made over 500 live recordings for small-run
distribution using a 16 bit CD recorder. The redbook CD's dynamic range of
16 bits, being some 30 dB greater than the dynamic range of the live
performances I record, causes me zero concern over level setting.

I figure that if you can't easily set levels within 20 dB (still leaving 10
dB or more headroom), you simply don't have any business doing professional
or even serious recording.

BTW, I do any post-processing @ 16 bits. About 150 of those recordings get
quite a bit of manual leveling after the fact. The leveling comes about
mostly due to considerable spill of acoustic sources mostly spoken word. So
the board tape levels being essentially differential levels, are very often
wrong.

The other 350 recordings are delivered as recorded, and I do fiddle with
levels so that their peak levels are adequate for playback in a classroom
situation.

[Mike Rivers]

geoff wrote:

What recording device are you using ? Obviously available media/memory is a
restriction with it for you.....

Flash memory recorders are pretty popular today. Sure, you can get 32 GB
memory cards, but not every recorder accommodates them, nor does every
budget. I use 16-bit in the field myself. It has always been good enough
and 24-bit isn't better given the available dynamic range.

24 bits would allow less fretting over level-setting optimisation, and
obviate the necessity for extra processing should ypou feel the need to
convert.

Converting is no problem and non-intrusive - it just adds a string of
zeros to the end of the word. As far as level setting, you still don't
want to be careless. By recording at too low a level, when you boost it
up to "normal" in post processing, you'll also be boosting the preamp
noise. It's not a big thing with most modern preamps, but it's better to
do it right the first time.

I can't imagine that you would be getting any benefit from converting to 24
bits before any processing. But that depends, I guess, on how primitive your
editing app is.

True. Most modern DAW applications use as many bits as they need (up to
a reasonable limit of course) and then output the word length that you
want. But if the application works with 24-bit input, you can leave the
word length reduction to the bitter end and know where you stand
throughout the processing chain. It's not a harmful operation, doesn't
take much time, and it makes they guy feel better about what he's doing.
Why do you insist that he's wasting his time or bits?



--
If you e-mail me and it bounces, use your secret decoder ring and reach
me he
double-m-eleven-double-zero at yahoo -- I'm really Mike Rivers
)

[HKC]

PS I would probably stay in 88.2 to stay clear from too much
conversion.....

It's another fallacy that integer ratios of sample rates are easier to
convert than odd ratios. No matter what the ratio of output to input
sample rate, you still have to actually resample (mathematically) or
you'll get the wrong result. You can't just take every other sample of
an 88.2 kHz audio stream and call it a 44.1 kHz stream.


But you will stay in tune when mixing files like maybe samples and loops
coming from CDs with recorded files.....

[Peter Larsen[_3_]]

geoff wrote:

Why would you record at 16 bits and 'expand' to 24 bits - why not just
record at 24?

My question would be why only use 24 bits as work-format rather than 32
bits, headroom is nice. As for the recording .... not much point in wasting
system time and resources writing 8 additional 1's, at least in some
contexts. Otoh ... the relaxed level setting of the longer wordlength can be
very useful. There is no way to meaningfully argue this without defining the
recording context.

geoff

Kind regards

Peter Larsen

[Scott Dorsey]

HKC wrote:
PS I would probably stay in 88.2 to stay clear from too much
conversion.....

It's another fallacy that integer ratios of sample rates are easier to
convert than odd ratios. No matter what the ratio of output to input
sample rate, you still have to actually resample (mathematically) or
you'll get the wrong result. You can't just take every other sample of
an 88.2 kHz audio stream and call it a 44.1 kHz stream.

Ahh, but you CAN if you have first run a low-pass filter to make sure there
is no information above 22.05 KHz in the dataset. This is called "filtering
and decimation" and it's the easiest possible sample rate conversion but of
course only works on integral multiples.

But, it's true that unless your software is specifically designed to do this,
it will probably use the standard resampling algorithm (which you can find
in Numerical Methods in Fortran or on the AD1890 applications notes).
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Geoff]

"Peter Larsen" wrote in message
...
geoff wrote:

Why would you record at 16 bits and 'expand' to 24 bits - why not just
record at 24?

My question would be why only use 24 bits as work-format rather than 32
bits, headroom is nice.

As I referred to, depends on how primitive the editor is. Most use way more
that fixed 24 bit processing these days. The source media bit depth should
be irrelevant.

geoff

[Robert Orban]

In article ,
says...

There is no nonlinear distortion in a digital filter unless you screw up and
overdrive it. Digital filters generally have far more dynamic range than
analog filters, so it is unlikely that it will clip unless someone really
screws up.

Virtually all digital filters have some sort of nonlinear distortion. The IIRs
typically used in equalizers are subject to limit cycles and truncation noise
due to finite length arithmetic. There are various techniques (like error
feedback) for minimizing these issues, but it is very rare to find a digital
filter in the real world that in fact adds dither before every truncation,
which is what would be required to linearize the filter.

These problems all increase as the poles of the IIR filter get closer to the
unit circle in the z plane. This is a strong argument for using as low a
sample rate as you can get away with because doing so minimizes the noise in
any filters. In audio, this includes EQ as well as crossovers in multiband
compressors. In FIR filters, there is a similar advantage to minimizing the
sample rate because it reduces the length of the filter required to produce a
given target response.

If you are doing nonlinear processing, like compression and limiting, then you
can get aliasing in the sidechains and also when the audio is multiplied by
the sidechain signal. So here the advice is different -- use a sampling rate
that is high enough to render any such aliasing inaudible.