Hi.

Today my intern asked me a question that I did not have a clear answer
for (this happens now and then). While I was describing various menus
and their definitions and functions in Pro Tools he asked me what DC
Offset was.

"Good Question" I said. "It has something to do with very low
frequency hum that can be introduced into a signal from power
sources", I said.

He then asked me what that menu select did in PT and I quickly changed
the subject. "Hey..look at that bird out the window!"

God forbid I should look that up. Or even know this by now.

Could anyone please ellaborate on DCoffset.
Is this below human hearing?
Is this purely a restoration issue?

Thanx

BP.

Robert Pascarella wrote:

Hi.

Today my intern asked me a question that I did not have a clear answer
for (this happens now and then). While I was describing various menus
and their definitions and functions in Pro Tools he asked me what DC
Offset was.

"Good Question" I said. "It has something to do with very low
frequency hum that can be introduced into a signal from power
sources", I said.

He then asked me what that menu select did in PT and I quickly changed
the subject. "Hey..look at that bird out the window!"

God forbid I should look that up. Or even know this by now.

Could anyone please ellaborate on DCoffset.
Is this below human hearing?
Is this purely a restoration issue?

DC = Direct Current. DC is 0 Hz. It just sits there at some voltage
level with respect to signal return. For audio if that level is 0 Volts
then you don't worry about such things as wasting energy heating voice
coils in speakers. If it is not 0 Volts then you can see intersting
things like woofers suddenly moving to a new position in the basket and
staying there when your signal is applied. Too much of this and you lose
dynamic range amongst other ills. It's like having a battery in series
with the AC component of the signal.

bobs

Bob Smith
BS Studios
we organize chaos
http://www.bsstudios.com

So when I am in logic audio and I check for DC offset, and it says 0.003
percent found, and I remove it, does it effect my audio at all by removing
any of the audio signal also.

in article , Bob Smith at
wrote on 8/14/03 12:58 AM:



Robert Pascarella wrote:

Hi.

Today my intern asked me a question that I did not have a clear answer
for (this happens now and then). While I was describing various menus
and their definitions and functions in Pro Tools he asked me what DC
Offset was.

"Good Question" I said. "It has something to do with very low
frequency hum that can be introduced into a signal from power
sources", I said.

He then asked me what that menu select did in PT and I quickly changed
the subject. "Hey..look at that bird out the window!"

God forbid I should look that up. Or even know this by now.

Could anyone please ellaborate on DCoffset.
Is this below human hearing?
Is this purely a restoration issue?

DC = Direct Current. DC is 0 Hz. It just sits there at some voltage
level with respect to signal return. For audio if that level is 0 Volts
then you don't worry about such things as wasting energy heating voice
coils in speakers. If it is not 0 Volts then you can see intersting
things like woofers suddenly moving to a new position in the basket and
staying there when your signal is applied. Too much of this and you lose
dynamic range amongst other ills. It's like having a battery in series
with the AC component of the signal.

bobs

Bob Smith
BS Studios
we organize chaos
http://www.bsstudios.com

"Robert Pascarella" wrote in message


Today my intern asked me a question that I did not have a clear answer
for (this happens now and then). While I was describing various menus
and their definitions and functions in Pro Tools he asked me what DC
Offset was.

"Good Question" I said. "It has something to do with very low
frequency hum that can be introduced into a signal from power
sources", I said.

Wrong. DC offset is not low frequency, its DC. DC means that the frequency
is zero.

Could anyone please elaborate on DCoffset.

DC offsets are mistakes that sometimes creep into recordings because of
human, equipment or software errors.

Is this below human hearing?

It is until you start chopping audio up and matching it up with audio that
lacks the identical same DC offset. Then you get tics and pops at the edit
points.

Is this purely a restoration issue?

The problem with DC offset is that it creates tics and pops when you paste
in audio that lacks the identical same DC offset. Rather than trying to
match DC offsets, its easier to just eliminate them.

Robert Pascarella wrote:

Could anyone please ellaborate on DCoffset.
Is this below human hearing?
Is this purely a restoration issue?

DC offset occurs when the long-term average of a waveform overall (taking
into account both its positive and negative excursions) is something other
than exactly 0. This can occur because some signals themselves are not
perfectly symmetrical, but in the most common recording situations it
occurs simply because most A/D converters aren't designed specifically
to "center" their output exactly around 0--because very little harm is
actually caused by small, relatively constant amounts of offset, and if
it is a problem in a particular situation, it can always be removed.

In fact in the old days, some converters deliberately introduced small
amounts of DC offset because it was thought that this would improve the
sonic performance of D/A converters. And perhaps back then it did.

DC offset can become a problem particularly when you're editing material
that has it in varying amounts--i.e. if take 4 of a piece has a DC offset
value that's markedly different from the DC offset value of take 2 and
you have to make an abrupt transition from one to the other. Under some
circumstances a thump or an "unidentifiable something wrong" can occur at
the transition point, giving away the fact that an edit was made. Or it
can simply be unaccountably difficult to find a good edit point.

Actually it takes quite a bit of DC offset to restrict the dynamic range
of a system significantly, because in linear PCM the value of one LSB
is large only near the zero (crossing) point; closer to full scale, one
LSB represents only a tiny fraction (e.g. 1/10000 or less IIRC) of a dB.

In article writes:

So when I am in logic audio and I check for DC offset, and it says 0.003
percent found, and I remove it, does it effect my audio at all by removing
any of the audio signal also.

The functional way that DC offset is removed (regardless of exactly
how the math is done) is by filtering out everything below a very low
frequency, usually 5 Hz or lower. So, yes, if you actually recorded
anything down that low, it will affect it. In practice, it won't.

DC offset used to be a bigger problem than it is with today's A/D
converters. The two main consequences of having significant DC offset
are a loss of symmetrical headroom (by the amount of the offset), and
the tendency to generate a click at an edit point when you think
you're cutting at the zero crossing (but are actually cutting at zero
plus the DC offset.

It's a problem that you shouldn't worry about until you find a
problem.



--
I'm really Mike Rivers - )

"Richard Crowley" wrote in message
...
"Geoff Tanner" wrote ...
Is there really a software dc offset adjustment on the
outputs of a protools system?

I would think that any decent audio software would have
a function to adjust DC offset.

I would have guessed that they would have had dc servo
amps specifically to clamp the outputs at 0v and so negate
the need for dc blocking capacitors.

Few pieces of audio equipment outside "golden-ears" audiophile
circles are DC-coupled end-to-end. The first capacitor or trans-
former in the path will eliminate any DC-offset.

Another good reason to always fade-in and fade-out when
finishing a mix. You CAN hear the "click" if you leave the
DC-offset in a CD track without fading in and out.


Hi

I'm still puzzled why the D to A has to have a client accessible DC offset
adjustment and why the amplifier should not be set permanently to zero dc
offset. What possible advantage is there is having the audio riding on a dc
level... other than biasing an electrolytic capacitor you would fit there to
zap the offset?

Geoff
www.auroraaudio.net

"Geoff Tanner" wrote in message
ink.net

I'm still puzzled why the D to A has to have a client accessible DC
offset adjustment

We're talking legacy equipment,here. Things like the old 3M digital recorder
ca. late 70s, early 80s.

Modern equipment lacks just about any kind of adjustment that clients or the
production line can foul up.

and why the amplifier should not be set permanently
to zero dc offset. What possible advantage is there is having the
audio riding on a dc level... other than biasing an electrolytic
capacitor you would fit there to zap the offset?

Once upon a time some converters treated the transition from positive to
negative signal voltages as a special case. Like most special cases, there
could be some pathological behavior associated with the transition at zero.
The idea of adding an offset was to move this pathological behavior to
signal levels where it was a smaller percentage of the total signal.

On a mostly-irrelevant note, in telecommunications there was a similar
concept called "sealing current". The effects of incidental contact
rectification was reduced by biasing the line so there never was a change in
the direction that current flowed through it.

In article . net writes:

I'm still puzzled why the D to A has to have a client accessible DC offset
adjustment and why the amplifier should not be set permanently to zero dc
offset. What possible advantage is there is having the audio riding on a dc
level... other than biasing an electrolytic capacitor you would fit there to
zap the offset?

No advantage at all, which is why it's adjustable - to zero. Things
drift and servos aren't always perfectly accurate. Get 0 VDC to your
speaker output and you don't need an output transformer or capacitor.
You don't want to put DC on a speaker, at least not much, for very
long.



--
I'm really Mike Rivers - )

Rick Thomas wrote:

So when I am in logic audio and I check for DC offset, and it says 0.003
percent found, and I remove it, does it effect my audio at all by removing
any of the audio signal also.

Depends on the method. I've seen it done with a highpass
filter with a high enough cutoff that there was phase shift
into the audible range and I've seen it done by just
averaging all the samples in the track and then subtracting
that from every sample in which case the effective cutoff is
so low that there are no artifacts at all.

To see what is happening for a particular implementation, do
a removal on a single sample pulse in a few seconds of
silence. If what remains is still a single sample pulse
with just a slight level shift everywhere then it is the
latter method. If you see a change in the shape of the
pulse it is the former.


Bob
--

"Things should be described as simply as possible, but no
simpler."

A. Einstein

in article znr1060863536k@trad, Mike Rivers at wrote on
8/14/03 10:13 AM:


In article
writes:

So when I am in logic audio and I check for DC offset, and it says 0.003
percent found, and I remove it, does it effect my audio at all by removing
any of the audio signal also.

The functional way that DC offset is removed (regardless of exactly
how the math is done) is by filtering out everything below a very low
frequency, usually 5 Hz or lower. So, yes, if you actually recorded
anything down that low, it will affect it. In practice, it won't.

DC offset used to be a bigger problem than it is with today's A/D
converters. The two main consequences of having significant DC offset
are a loss of symmetrical headroom (by the amount of the offset), and
the tendency to generate a click at an edit point when you think
you're cutting at the zero crossing (but are actually cutting at zero
plus the DC offset.

It's a problem that you shouldn't worry about until you find a
problem.

Ok, Ive noticed that when I trim an audio file to get rid of all the lead in
time and end time I allways end up with a click at the beginning of the
track. I guess this must be the DC offset. Wow, it just goes to show no
matter how long youve done something you allways learn something new. From
now on I will remove Dc offset from every single track before editing.



--
I'm really Mike Rivers - )

This is great, guys. I want to thank you for these comments.

So, anything at 0hz (DC) would imply that the signal polarity never
changes. There are NO cycles per second. Thus meaning that any two
signals with these requirements would match perfectly. No pops or
ticks.

Anything higher would be considered an Alternating Current or AC?

If something (like a bad A/D) changes the value from 0hz, you may hear
those pops and ticks inedits due to the "unmatched frequencies"?

By applying the offset, are we restoring the signal to 0hz?

Thanx again

BP.

"Mike Rivers" wrote in message
news:znr1060863536k@trad...

The functional way that DC offset is removed (regardless of exactly
how the math is done) is by filtering out everything below a very low
frequency, usually 5 Hz or lower. So, yes, if you actually recorded
anything down that low, it will affect it. In practice, it won't.

I don't know if there's any practical purpose for frequencies that low, but
I did make a recording with functional low frequencies, once. As an
experiment, I recorded a five-second copyright notice, stretched it out to
a couple of minutes and mixed it in with some tracks. It was too low to
hear, but when the final product was played back at high speed, the notice
became audible.

dtk

"Rick Thomas" wrote in message
...
in article znr1060863536k@trad, Mike Rivers at wrote
on
Ok, Ive noticed that when I trim an audio file to get rid of all the lead
in
time and end time I allways end up with a click at the beginning of the
track. I guess this must be the DC offset. Wow, it just goes to show no
matter how long youve done something you allways learn something new. From
now on I will remove Dc offset from every single track before editing.

You don't need to remove DC offset every time (if at all most times
nowadays). Just make sure the cut points happen at a zero crossing. Most all
editing software has an option to make sure that all edits have zero
crossing.

"Robert Pascarella" wrote ...
So, anything at 0hz (DC) would imply that the signal
polarity never changes. There are NO cycles per second.

Yes.

Thus meaning that any two signals with these requirements
would match perfectly. No pops or ticks.

Few people would consider 0Hz (dead silence) to be a "signal".

Anything higher would be considered an Alternating Current or AC?

Yes. But 0Hz is no signal at all (at least in the audio sense).

If something (like a bad A/D) changes the value from 0hz, you may hear
those pops and ticks inedits due to the "unmatched frequencies"?

You seem to be confusing Hz (AC frequency, cycles-per-second)
with Volts (signal amplitude, or "strength"). Not the same in any way!

Hz measures the "length" of the waveform (in cycles/second).
Most would say it measures the "frequency".

While V measures the "height" of the waveform (in Volts or
milli-volts, etc.

By applying the offset, are we restoring the signal to 0hz?

No. You are restoring the signal to 0 volts offset.

Yes. Bob didn't explain it in a way that you might SEE on the screen, but
look at the way things are described. We have ZERO dBFS (Full Scale), which
is the absolute maximum any signal digitally can obtain, and we have ZERO in
the sense of the visible scale representing the point between positive bits
and negative bits. In a 16 bit environment that would be between 32767
and -32768 in the computer ( and we talk about 16 bit because that's the CD
Redbook Standard). If a signal is centered directly between both extremes
then it is the normally recorded file. If it starts somewhere higher than
ZERO (not dBfs) then it's due to DC offset. So while one channel is playing
along fine at the right reference level, the other is starting higher, still
can't exceed the max ZERO dBFS, and consequently brings less dynamic range
into the picture and insinuates noise. Calculating for DC offset subtracts
the noise and re-establishes dynamic range.

Now if Logic Audio is good enough to discern .003 percent, then I'd say NO,
you're not going to hear the difference. 3 thousandths of a percent is
pretty negligable. However, if you do a lot of mixing with DC offset being
consistent within the tracks being mixed, you ain't going to get it out from
the mix. If you have this problem then you need to go through and calculate
the DC offset for all offending tracks before you mix the material. And an
even better exercise is to find out why any DC offset is being induced into
your system. That's a totally different subject.

--


Roger W. Norman
SirMusic Studio
Purchase your copy of the Fifth of RAP CD set at www.recaudiopro.net. See
how far $20 really goes.




"Rick Thomas" wrote in message
...
So when I am in logic audio and I check for DC offset, and it says 0.003
percent found, and I remove it, does it effect my audio at all by removing
any of the audio signal also.

in article , Bob Smith at
wrote on 8/14/03 12:58 AM:



Robert Pascarella wrote:

Hi.

Today my intern asked me a question that I did not have a clear answer
for (this happens now and then). While I was describing various menus
and their definitions and functions in Pro Tools he asked me what DC
Offset was.

"Good Question" I said. "It has something to do with very low
frequency hum that can be introduced into a signal from power
sources", I said.

He then asked me what that menu select did in PT and I quickly changed
the subject. "Hey..look at that bird out the window!"

God forbid I should look that up. Or even know this by now.

Could anyone please ellaborate on DCoffset.
Is this below human hearing?
Is this purely a restoration issue?

DC = Direct Current. DC is 0 Hz. It just sits there at some voltage
level with respect to signal return. For audio if that level is 0 Volts
then you don't worry about such things as wasting energy heating voice
coils in speakers. If it is not 0 Volts then you can see intersting
things like woofers suddenly moving to a new position in the basket and
staying there when your signal is applied. Too much of this and you lose
dynamic range amongst other ills. It's like having a battery in series
with the AC component of the signal.

bobs

Bob Smith
BS Studios
we organize chaos
http://www.bsstudios.com