R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: [1] 2  All   Go Down

Author Topic: DC offset - some questions.....  (Read 5481 times)

zakco

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 286
DC offset - some questions.....
« on: July 17, 2004, 04:57:42 PM »

Hi there.

I have a couple of questions regarding DC offset:

1. How much/little (if any) is acceptable/normal in a final mix of a 24 bit 44.1 wav file?

2. I have a current project that shows DC offset of 0.29% and 0.31% (L&R) in the final mixed wav file.

I went back to the multitrack project and checked EVERY audio file in the project and NOT ONE of them showed ANY DC offset. (0%)

No outboard gear was used during mixdown and therefore, no D/A/D conversion took place after tracking.

How is this possible?
Could it be introduced by plugins?

I can't hear any audible problems with the mix....should I be concerned?

Thanks,

-Zak Cohen

Level

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1811
Re: DC offset - some questions.....
« Reply #1 on: July 17, 2004, 05:25:20 PM »

The power supply in your workstation can cause this. In some acoustic recordings, I have seen DC offset occur, even with solo instruments. As long as it is less than 0.5%, it certainly is not a real cause for concern. When mixed with many other instruments, DC offset is something I have heard to be described as a "natural occurance".

It is a cause for concern if it climbs out of hand when rendering.

How does the track sound?

Some mastering engineers will not allow any "measurable in the statistic window" DC offset. I for one have experimented with and without DC offset correction at both ends of the loudness scale and as I mentioned, if you hear no audible artifacts, I would not be overly concerned. Anyone that says "any DC offsent is bad" has not recorded acoustic instruments enmass before!

What did we do before it was easily measurable?
Logged
http://balancedmastering.com

"Listen and Learn"
---Since 1975---

Jakob Erland

  • Full Member
  • ***
  • Offline Offline
  • Posts: 203
Re: DC offset - some questions.....
« Reply #2 on: July 18, 2004, 04:18:47 AM »

Level wrote on Sat, 17 July 2004 22:25

What did we do before it was easily measurable?


Before this was easily measurable, we did not have the problem, as none of the classic recording techniques had frequency response down to zero Hertz. Modern digital has, and thus the problem.

If you do an "in-the-box" mix of a number of audio files that by them selves has no DC offset, and your mix turns out having some offset, then there's simply something wrong with your mix algorithm. As it's pure math, powersupplies has no saying in this.

Jakob E.
Logged
..note to self: don't let harman run your company..

bobkatz

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2926
Re: DC offset - some questions.....
« Reply #3 on: July 18, 2004, 11:55:22 AM »

zakco wrote on Sat, 17 July 2004 16:57

Hi there.

I have a couple of questions regarding DC offset:

1. How much/little (if any) is acceptable/normal in a final mix of a 24 bit 44.1 wav file?




I simply judge it like this: If start and stop causes clicks or pops, there's too much DC offset. It's that simple.

Quote:



No outboard gear was used during mixdown and therefore, no D/A/D conversion took place after tracking.

How is this possible?
Could it be introduced by plugins?




I imagine it could be introduced by plugins.

Quote:



I can't hear any audible problems with the mix....should I be concerned?




If starting and stopping does not produce audible clicks or pops, you should not be concerned, in my opinion.

BK
Logged
There are two kinds of fools,
One says-this is old and therefore good.
The other says-this is new and therefore better."

No trees were killed in the sending of this message. However a large number of
electrons were terribly inconvenienced.

Level

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1811
Re: DC offset - some questions.....
« Reply #4 on: July 18, 2004, 02:44:44 PM »

Jakob, the reason I brought up power supply issues, I had an digital console once that was direct coupled and the output was presenting DC offset. It was in the analog stages of this desk, nevertheless, the culpret was found. Upon repairing the PS, the problem went away. That simple. I thought I would include this unique experience. Others may experiece it. I also have had direct instuments introduce DC offset.
Logged
http://balancedmastering.com

"Listen and Learn"
---Since 1975---

zakco

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 286
Re: DC offset - some questions.....
« Reply #5 on: July 18, 2004, 03:16:23 PM »

Hello, and thanks for the replies....

I've done a little bit of troubleshooting on this and various other current projects and now that I'm more aware of the issue I'm seeing a few things:

1. Audio files that show no measurable DC offset in Cubase SX sometimes show small amounts when analyzed in Wavelab (for example: 0.0016%) Somehow, the two applications are measuring differently. Also, Wavelab displays DC offset as both a percentage and a db level. For example, it will often show DC offset at -144 db but 0%. Are these two measurements contradictory?

2. I am getting small but audible clicks when stopping and starting playback. I was attributing this to Steinbergs global plugin delay compensation that was introduced in V.2 of both Nuendo and Cubase SX but Bob's comments suggest that these are possibly symptoms of DC offset.

However.....these clicks are still present after removing DC offset  using both Cubase and Wavelab's DC removal processing. So, perhaps the clicks I'm hearing are relating to something else (audio drivers/buffer settings/PDC etc...)

So here I am, not really any further ahead but painfully aware of   the clicks during stop/starting.

I finally have a day with no clients in the studio, so I'm going to do a little testing with various sources and see what I can come up with.

I really appreciate everyones attention to my little problem and any other comments/suggestions would be welcomed of course.

Thanks,

-Zak Cohen



zakco

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 286
Re: DC offset - some questions.....
« Reply #6 on: July 18, 2004, 03:52:35 PM »

Just to give an idea of scale in a practical sense:

A typical, heavily compressed, modern rock production is showing DC offset at about 0.02% at the final master stage. That is after approximately 2-4 db of limiting. In other words, this is the maximum DC offset that will be present in the final product.

Is this a typical result?

-Zak

bobkatz

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2926
Re: DC offset - some questions.....
« Reply #7 on: July 18, 2004, 05:04:16 PM »

zakco wrote on Sun, 18 July 2004 15:16



However.....these clicks are still present after removing DC offset  using both Cubase and Wavelab's DC removal processing. So, perhaps the clicks I'm hearing are relating to something else (audio drivers/buffer settings/PDC etc...)




Hmm, In that case, wouldn't you be getting clicks like that when you start/stop any source?

DAWs are particularly problematic when you start/stop because they can introduce cross fades or rampup/rampdown.

Perhaps if you dub this to a hard disc recorder and start/stop it you can eliminate the variables.
Logged
There are two kinds of fools,
One says-this is old and therefore good.
The other says-this is new and therefore better."

No trees were killed in the sending of this message. However a large number of
electrons were terribly inconvenienced.

ssltech

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 4780
Re: DC offset - some questions.....
« Reply #8 on: July 19, 2004, 11:04:47 AM »

-144dB is as low down as a 24-bit converter can go (and about 20dB below the real noisefloor!) so can be regarded as zero percent, being as low as can be measured.

Clicking can be expected if you start during a signal that has high LF content, so I think perhaps Bob's suggestion should be re-tried diring a silent pasage -either before the downbeat or after the tail... then tell us if it clicks. If you get a click during silence, you have DC. (an inaudible signal that only gives away it's presence by it's commencement or interruption)

BTW, Much respect to bobkatz. -Bob, I should drop round again soon... Also, Level, it was a great surprise to meet you with Mike the other day... I haven't replied to your PM in the other forum yet... been busy as hell, but I'll certainly let you have a copy of the 'classical Brass' project as soon as it's edited and shaped up a bit! ... which is all a bit premature since the record window isn't till late August!!!

BTW I had this problem once with a Neotek console (no blocking caps and a hefty DC offset on most outputs) and a Synclavier (input converters actually do record down to DC...) people thought that the signal they were seeing on the meters was HF noise... it was actually the other end of the spectrum!

Very Happy Wink

Keith
Logged
MDM (maxdimario) wrote on Fri, 16 November 2007 21:36

I have the feeling that I have more experience in my little finger than you do in your whole body about audio electronics..

Level

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1811
Re: DC offset - some questions.....
« Reply #9 on: July 19, 2004, 03:40:12 PM »

Keith, it was very pleasant meeting you as well! Keep up the splendid work and I look forward to checking out the 'classical Brass' !!

Cheers to you!

Bill
Logged
http://balancedmastering.com

"Listen and Learn"
---Since 1975---

zakco

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 286
Re: DC offset - some questions.....
« Reply #10 on: July 24, 2004, 12:13:19 AM »

OK, sorry about the delay in getting back to this thread...anyways, after doing a little more homework, I have discovered:

1. Cubase SX and Wavelab use different methods to calculate DC offset. In my experience, they never seem to get the same results. I'm assuming that the DC offset that shows up in the final mixes is a cumulative effect of very low level offset on individual tracks combining. Is this a reasonable assumption?

2. I only get stop/start clicks during sections that contain substantial program material. Quiet or silent passages don't produce clicks at all.

So....I guess was getting all worked up about nothing.

I appreciate everyone's input. Thanks.

-Zak

bbalin

  • Newbie
  • *
  • Offline Offline
  • Posts: 8
Re: DC offset - some questions.....
« Reply #11 on: July 27, 2004, 09:59:14 AM »

Sorry for jumping in here as the thread is nearly finished.

I did some tests regarding dc offset some time ago.
What I found out :
- All DAW applications I could test use a rather primitive approach to DC offset calculation. They simple assume that in the long run positive and negative going voltages must be equal. This way they can catch constant offsets but not wandering offsets.
I had a drum clip with wandering offset which after processing with Wavelab showed perfect figures but the offset was not removed at all.
- DC offset sometimes is introduced by plugins e.g. dynamics processing or transient modulators. Typicalla the DC offset from these devices is changing according to the signal applied. One example would be the digitalfishphones plugins. Take care of DC offsets when using this type of plugins.

Boris Balin
Logged

Jakob Erland

  • Full Member
  • ***
  • Offline Offline
  • Posts: 203
Re: DC offset - some questions.....
« Reply #12 on: July 27, 2004, 11:21:37 AM »

Strictly technically, wandering offsets are NOT DC offsets, but low-frequency content.

DC removal is best understood as a lowcut filter, and thus only possible if you define a frequency below which it works. Frequencies higher than this cutoff frequency won't be affected, so these "sub-audio" frequencies will still affect the signal..

Jakob E.
Logged
..note to self: don't let harman run your company..

St. Domingo's

  • Newbie
  • *
  • Offline Offline
  • Posts: 5
Re: DC offset - some questions.....
« Reply #13 on: July 27, 2004, 02:16:59 PM »

Keith,

'classical Brass'

As in the Liverpool interpretation of "a brass", or the more normal interpretation of "brass".

Either way, sounds interesting!
Logged
Matt Ward,
Tech Support,
Metropolis - Group,
London.

Any opinions expressed on this forum are my own and are in no way the opinion of my employers.
 
matt.ward@metropolis-group.co.uk
www.metropolis-group.co.uk

Immanuel Kuhrt

  • Newbie
  • *
  • Offline Offline
  • Posts: 46
Re: DC offset - some questions.....
« Reply #14 on: September 08, 2004, 03:43:12 AM »

Jakob Erland wrote on Tue, 27 July 2004 17:21


DC removal is best understood as a lowcut filter, and thus only possible if you define a frequency below which it works. Frequencies higher than this cutoff frequency won't be affected, so these "sub-audio" frequencies will still affect the signal..

Jakob E.



I often wondered, if I need to use a DC blocker before using a lo-cut, when working in digital - if the presence of DC would make the lo-cut act in unwanted ways?

But if I understand this right, then using a lo-cut will eliminate the need of a DC blocker, and the DC will not harm the lo-cut algorithm?

If someone would tell me if I am right or wrong, it would be highly appreciated.

Immanuel
Logged
Disclaimer - I ain't no pro
Pages: [1] 2  All   Go Up
 

Site Hosted By Ashdown Technologies, Inc.

Page created in 0.08 seconds with 17 queries.