[lewdslewrate]

Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?
Will the recorder simply see an audio stream regardless of the
differing bit rates?

It's happening in 2 days so I am keen to sort it. thanks in advance.

[Scott Dorsey]

lewdslewrate wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?
--scott

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

[Mike Rivers]

On 11/27/2011 10:49 AM, lewdslewrate wrote:

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1.Will my audio be compromised?

Relative to what? You'll be truncating the data, but it
might not really matter, depending on what you're recording.
And don't bother to tell us. There's no way of guessing.

But if this is what you have to do, then do it. Do you have
any other options between those two units? Like maybe analog
out of the preamp to analog in to the recorder? Or have you
used up all the analog inputs?

Will the recorder simply see an audio stream regardless of the
differing bit rates?

Almost certainly. And it won't sound as bad to you as it
will sound to Scott.

It's happening in 2 days so I am keen to sort it. thanks in advance.

Well, you have a day to try it out. Better do it and get an
idea of whether it'll work for you or if you'll have to hit
your client up for more money to rent some other gear.


--
"Today's production equipment is IT based and cannot be
operated without a passing knowledge of computing, although
it seems that it can be operated without a passing knowledge
of audio." - John Watkinson

http://mikeriversaudio.wordpress.com - useful and
interesting audio stuff

[Arny Krueger[_4_]]

"Scott Dorsey" wrote in message
...
lewdslewrate wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?

I just checked the Focusrite 828 user manual and there is no word length
setting in sight.

Ditto for the AW2400

That all said, it is likely that the source has enough residual noise for
self-dithering of the truncation by the noise in the signal itself. Not my
preference, but not a solid red flag.

[lewdslewrate]

On Nov 28, 12:54*pm, "Arny Krueger" wrote:
"Scott Dorsey" wrote in message

...

lewdslewrate wrote:
Hi. *I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 *to *Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. *(X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). *Will my audio be compromised?

Yes. *You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. *Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?

I just checked the Focusrite 828 user manual and there is no word length
setting in sight.

Ditto for the AW2400

That all said, it is likely that the source has enough residual noise for
self-dithering of the truncation by the noise in the signal itself. *Not my
preference, but not a solid red flag.

all this Jogged my memory....i installed a digital card into my unit
and on the bit of paper that shows how to fit (it)...is a list of
various user options. 24/20/16 bit depth can be chosen by moving a
jumper. I lost the paper but focusrite promptly sent me a download.
nothing wasted here though...i understand dithering more than i did
before. so thank you.

[Les Cargill[_4_]]

Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?
--scott


But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

--
Les Cargill

[hank alrich]

Les Cargill wrote:

Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?
--scott


But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

True, though once truncation artifacts are introduced, will dither
decorrelate them?

--
shut up and play your guitar * http://hankalrich.com/
http://www.youtube.com/walkinaymusic
http://www.sonicbids.com/HankandShaidri

[Arny Krueger[_4_]]

"hank alrich" wrote in message
...
Les Cargill wrote:

Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16
bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?
--scott


But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

True, though once truncation artifacts are introduced, will dither
decorrelate them?

Nope.

The way to justify less concern over an undithered truncation is to find
more than 1 LSB of noise in the source. Usually, a studio or live 16 bit
recording has 3-4 of the LSBs filled with noise.

However, if there is dither available, then it should be used.

[Scott Dorsey]

In article ,
Les Cargill wrote:
Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?

But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

Yes, but once the truncation has happened, you can't fix it. You need to
dither _before_ truncating.
--scott

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

[Les Cargill[_4_]]

hank alrich wrote:
Les wrote:

Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?
--scott


But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

True, though once truncation artifacts are introduced, will dither
decorrelate them?



My understanding is that it will.

--
Les Cargill

[Les Cargill[_4_]]

Arny Krueger wrote:
"hank wrote in message
...
Les wrote:

Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16
bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?
--scott


But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

True, though once truncation artifacts are introduced, will dither
decorrelate them?

Nope.



Suppose I do this odd 24-16 bit truncation thing on the raw tracks.
I run them through a mix, with dither at the end.

That should completely mask this, especially if I add 1 bit of noise
shaping. And I always do.


The way to justify less concern over an undithered truncation is to find
more than 1 LSB of noise in the source. Usually, a studio or live 16 bit
recording has 3-4 of the LSBs filled with noise.


But the noise is still pre-truncation, so...

However, if there is dither available, then it should be used.



--
Les Cargill

[Les Cargill[_4_]]

Scott Dorsey wrote:
In ,
Les wrote:
Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16 bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?

But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

Yes, but once the truncation has happened, you can't fix it. You need to
dither _before_ truncating.
--scott


I never can get that straight.

The only 24 but thing I have is the driver for a Lightpipe card,
driven by a ... 16 bit device. It delivers 24 bit samples, with all 8
LSB set to zero...

--
Les Cargill

[Scott Dorsey]

Les Cargill wrote:

True, though once truncation artifacts are introduced, will dither
decorrelate them?

My understanding is that it will.

It will not.

There is a discussion in the FAQ about how dither works, it's really
kind of ingenious. You can think of it as introducing errors into the
word length reduction process... errors which turn corellated distortion
into uncorrellated noise (which is a great improvement). It's a thing
that has do be done before the word length is reduced.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

Les Cargill wrote:

Suppose I do this odd 24-16 bit truncation thing on the raw tracks.
I run them through a mix, with dither at the end.

That should completely mask this, especially if I add 1 bit of noise
shaping. And I always do.

Dither does not mask errors, dither changes errors in the truncation
process from one form to another.

"Masking errors" is totally unrelated to the dither process.

The way to justify less concern over an undithered truncation is to find
more than 1 LSB of noise in the source. Usually, a studio or live 16 bit
recording has 3-4 of the LSBs filled with noise.

But the noise is still pre-truncation, so...

Yes, that's Arny's point. I don't think it's necessarily valid, but it
does reduce the concern.

However, if there is dither available, then it should be used.

This isn't 1985 any more, there's really no reason to deal with any of that
stuff anymore.
--scott

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

[Les Cargill[_4_]]

Scott Dorsey wrote:
Les wrote:

True, though once truncation artifacts are introduced, will dither
decorrelate them?

My understanding is that it will.

It will not.

There is a discussion in the FAQ about how dither works, it's really
kind of ingenious. You can think of it as introducing errors into the
word length reduction process... errors which turn corellated distortion
into uncorrellated noise (which is a great improvement).

*That* part I got

One of these days, I am going to have to really dig out why
the ordering matters. It just doesn't come up often enough.
Since it's going to be one of those "it's the error spectrum"
things...

It's a thing
that has do be done before the word length is reduced.
--scott

--
Les Cargill

[Mark]

On Nov 28, 7:17*pm, Les Cargill wrote:
Scott Dorsey wrote:
Les *wrote:

True, though once truncation artifacts are introduced, will dither
decorrelate them?

My understanding is that it will.

It will not.

There is a discussion in the FAQ about how dither works, it's really
kind of ingenious. *You can think of it as introducing errors into the
word length reduction process... errors which turn corellated distortion
into uncorrellated noise (which is a great improvement).

*That* part I got

One of these days, I am going to have to really dig out why
the ordering matters. It just doesn't come up often enough.
Since it's going to be one of those "it's the error spectrum"
things...

It's a thing
that has do be done before the word length is reduced.
--scott

--
Les Cargill

Well no one has asked this yet so I will...

A lot depends on how hot the 24 bit recording is...

If the original 24 bit recording is recorded at a low level relative
to Full scale and some of the top bits are not active then the
truncation will be more harmful.

If the original 24 bit recording is recorded at a high level relative
to full scale and the top bits are all active, then truncating will be
not much worse than making a 16 bit recording in the firs place.

Dither will be critical if the original 24 bit recording is low and
the top bits are not active, in which case when you truncate you are
going to have something worse then a 16 bit recording.

Bototm line is to try it....

Don't destroy the original until you get it all sorted out and if
worse comes to worse it can be recovered after you make the new
recording.

Why not load the 24 bit recording into a DAW and do what needs to be
done instead of trying to transfer it to another machine.


Mark

[Anahata]

On Mon, 28 Nov 2011 17:55:48 -0800, Mark wrote:

If the original 24 bit recording is recorded at a high level relative to
full scale and the top bits are all active, then truncating will be not
much worse than making a 16 bit recording in the firs place.

You may be forgetting that the 16 bit A/D process should be properly
dithered by (usually) the addition of controlled analog noise. Pure
truncation of a 24 bit recording to 16 bits will be worse than a 16 bit
recording with dithering.

Dither will be critical if the original 24 bit recording is low and the
top bits are not active, in which case when you truncate you are going
to have something worse then a 16 bit recording.

All you are saying is that if your signal is hot enough some people might
not hear the distortion, but it doesn't alter the fact that it's always
better with dither than without.

--
Anahata
--/-- http://www.treewind.co.uk
+44 (0)1638 720444

[Mike Rivers]

On 11/28/2011 7:17 PM, Les Cargill wrote:

One of these days, I am going to have to really dig out why
the ordering matters. It just doesn't come up often enough.

By that time, you'll have upgraded to 21st Century gear and
you won't have to worry about these things.



--
"Today's production equipment is IT based and cannot be
operated without a passing knowledge of computing, although
it seems that it can be operated without a passing knowledge
of audio." - John Watkinson

http://mikeriversaudio.wordpress.com - useful and
interesting audio stuff

[Arny Krueger[_4_]]

"Les Cargill" wrote in message
...
Arny Krueger wrote:
"hank wrote in message
...
Les wrote:

Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16
bit/
44.1. (X4 mic's will go direct into recorder) (I need the track
count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither
down
to 16-bit?
--scott

But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

True, though once truncation artifacts are introduced, will dither
decorrelate them?

Nope.

Suppose I do this odd 24-16 bit truncation thing on the raw tracks.
I run them through a mix, with dither at the end.

That should completely mask this, especially if I add 1 bit of noise
shaping. And I always do.

Masking and decorrelation are two different things.

I submit to you with all due respect that there is a very good possibility
that your recordings are noisy enough (room tone 1 LSB) that they are
effectively dithered at the point of truncation from 24 bits to 16.

If they aren't, I'm wondering how it is that you have somehow managed to
reduce room tone and mic preamp and other pre-truncation noise to more than
93 dB below peak levels. Under the most contrived and ideal circumstances I
and whole bunch of other recording engineers never get noise better
than -87 dB, while -75 dB dB is a very good number indeed. Even -65 dB can
be a pretty good sounding recording.

The way to justify less concern over an undithered truncation is to find
more than 1 LSB of noise in the source. Usually, a studio or live 16 bit
recording has 3-4 of the LSBs filled with noise.

But the noise is still pre-truncation, so...

Exactly. Just because you don't intentionally add the noise, doesn't mean
that it isn't already there.

[Arny Krueger[_4_]]

"Les Cargill" wrote in message
...
hank alrich wrote:
Les wrote:

Scott Dorsey wrote:
wrote:
Hi. I await Yamaha response but maybe it's a generic question?

I will output x8 mic channels via ADAT @ 24 bit/44.1 from Focusrite
ISA828 to Yamaha AW2400 multi track recorder set to record at 16
bit/
44.1. (X4 mic's will go direct into recorder) (I need the track count
so cannot set recorder to 24 bit). Will my audio be compromised?

Yes. You're putting 24 bits into the lightpipe, the machine at the
other end reads 16, so the last 8 are being truncated. Truncation is
bad.

Does either the Focusrite or the Yamaha have the ability to dither down
to 16-bit?
--scott


But dither will happen eventually... unless the raw tracks are printed
to media for distribution...

True, though once truncation artifacts are introduced, will dither
decorrelate them?



My understanding is that it will.

I've tried this experimentally many different ways, and it doesn't work. All
the books and papers say it doesn't work. So, I don't feel bad about
failing. ;-)

[Arny Krueger[_4_]]

"Les Cargill" wrote in message
...
Scott Dorsey wrote:
Les wrote:

True, though once truncation artifacts are introduced, will dither
decorrelate them?

My understanding is that it will.

It will not.

There is a discussion in the FAQ about how dither works, it's really
kind of ingenious. You can think of it as introducing errors into the
word length reduction process... errors which turn corellated distortion
into uncorrellated noise (which is a great improvement).

*That* part I got

One of these days, I am going to have to really dig out why
the ordering matters.

The short answer is that ordering doesn't matter in linear systems, but it
can and often matters greatly in nonlinear systems. Converting 24 bit data
to 16 is a highly nonlinear process. QED.

[Arny Krueger[_4_]]

"Mark" wrote in message
...
On Nov 28, 7:17 pm, Les Cargill wrote:


Well no one has asked this yet so I will...

A lot depends on how hot the 24 bit recording is...

That's not a question, it is a declaration. It is also wrong.


If the original 24 bit recording is recorded at a low level relative
to Full scale and some of the top bits are not active then the
truncation will be more harmful.

Top bit usage doesn't matter. What matters is the nature of the data, and
this is true even if only the bottom few bits are active or even all the top
bits are active.

If the original 24 bit recording is recorded at a high level relative
to full scale and the top bits are all active, then truncating will be
not much worse than making a 16 bit recording in the firs place.

This completely misplaces the proper logical emphasis.


Dither will be critical if the original 24 bit recording is low and
the top bits are not active, in which case when you truncate you are
going to have something worse then a 16 bit recording.

What is true is that a 24 bit recording that is recorded at a very low level
will effectively strip off the noise that would normally give you the
desired dithering decorrelation effect. For example, if the top 12 bits are
not used (peak levels about 60 dB below FS) then the normal room tone and
electronic noise of mics and mic preamps will be below the LSB. Obviously,
peaks 60 dB below FS is a pretty severe situation - a grotesque mistake.
Change that to where the top 8 bits are not used (still very bad - peaks
about 40 dB below FS) and enough of that low level noise will still get
through to decorrelate the quantization process.