N
Newbie dude
New member
What is DC[ Direct current I'm assuming?]offset?
SouthSIDE Glen
independentrecording.net
DC (yes, this does mean direct current) offset is when there is a DC "bias" current introduced in the signal somewhere along the chain. This current can be positive or negative, but it is a low-voltage DC current nonetheless.
These bias currents are usually - but not always - introduced by cheaper computer sound cards. They visuallly manifest themselves in audio editing software by a base line in the wavefor display that rides above or below (depending of positive or negative voltage) the "center line" of the display. For instance a small bias current can result in the editor recording a flat line at -15dB instead of the "-infinity" centerline. This "offset" cuts down on the usable headroom available for the wavefrom.
A "DC Offset" rectifier circuit (or software plug-in) acts to remove this flat line "offset" and return the flatline to center.
G.
These bias currents are usually - but not always - introduced by cheaper computer sound cards. They visuallly manifest themselves in audio editing software by a base line in the wavefor display that rides above or below (depending of positive or negative voltage) the "center line" of the display. For instance a small bias current can result in the editor recording a flat line at -15dB instead of the "-infinity" centerline. This "offset" cuts down on the usable headroom available for the wavefrom.
A "DC Offset" rectifier circuit (or software plug-in) acts to remove this flat line "offset" and return the flatline to center.
G.
B
biggestmuff
New member
SouthSIDE Glen said:
Okay, but how does this affect what I hear? Why is DC offset a bad thing? How does it hurt a recording? Is there anything positive about a DC offset?
mshilarious
Banned
biggestmuff said:
Glen said:
You will have less dynamic range before digital clipping. There is nothing positive about it.
SouthSIDE Glen
independentrecording.net
There is nothing positive about it whatsoever. It is inadvertantly added "noise" that is not part of the original signal, and therefore completly undesired.biggestmuff said:
Here's how it hurts in the editor:
Let's say that because of a rogue DC bias current introduced by the el cheapo sound card on your laptop computer (not uncommon), your zero-signal flatline on your laptop editor's waveform display is riding at, say, -25dBFS positive. This means that when you hit record on a totally silent channel, you'll see it recording a flat line riding at -25dBFS, floating above the centerline of your waveform display where it normally should be. The result will be that on the positive side of the waveform, you now only have 25dB of dynamic range before peaks in your recording go into clipping. This is less than half the dynamic range of an old vinyl album pressing. All that fancy digital, computerized gear and you'll have all the dynamic range of a telephone call.
Here's a real world physical effect:
Let's say you record that bias current on that silent channel. No noise, no music, just let that floating flatline run. Then you go ahead and play back that recording. You won't hear a thing except maybe a little pop or click at the very beginning. But look at the speaker elements on your studio monitors: when that little pop happens at the beginning, the woofer pushes out (that's what's making that little popping noise), but it never goes back in. It stays in that out position because your system is now reproducing that DC current that was recorded; the constant voltage is being sent to the speaker element as a 0Hz (DC) signal at a level corresponding to the level of the bias offset. Now, with that speaker being constantly pushed out and not resting naturally, just how good do you think any music pushed through that "biased" speaker is actually going to sound? Certainly not as good as if the current were not there. Also, by stressing the speaker that way, one is increasing one's chances of sending a harmful overload signal to the speaker because it's dynamic range is now just as comprimised as the editor's is.
Hope this helps explain it further.
G.
mshilarious
Banned
Hopefully the designer of the power amp had the sense to block DC at the input 
noisewreck
New member
Having a high-pass filter set around 10-15Hz will effectively eliminate the DC offset.
One thing I don't understand is how sometimes I end up getting DC offset on mixdowns or submixes. I always take care so that my individual tracks don't have DC offset, however, after mixing, at times DC offset gets introduced. Is it possible that compressors and other FX might be responsible for this? Please note that I mix in the box using Cubase SX 2. The EQs and compressors of choice are the UAD plugs.
One thing I don't understand is how sometimes I end up getting DC offset on mixdowns or submixes. I always take care so that my individual tracks don't have DC offset, however, after mixing, at times DC offset gets introduced. Is it possible that compressors and other FX might be responsible for this? Please note that I mix in the box using Cubase SX 2. The EQs and compressors of choice are the UAD plugs.
SouthSIDE Glen
independentrecording.net
Just educated guess/speculation here, but I wonder if what you're getting is true DC bias or just extreme LF coloration that may be building up across tracks and causing what appears to be an "offset", but is really a floor of subsonic mud.noisewreck said:
Where it would actually be coming from or why, I don't know. But this guess would both explain why a low shelf would get rid of it - theoretically (I would think) a frequency filter should have no effect on a DC signal - and would also explain why you see nothing on individual tracks but do after summing.
G.
bennychico11
...
SouthSIDE Glen said:
here's some pretend rep points for you Glen
I tried to give you some but it says I need to spread it around before I give you anymore (even though I think it's been awhile since I've rep-ed ya)
props for this explanation. crystal clear as usual.
SouthSIDE Glen
independentrecording.net
Thanks a bunch Benny. I wish my recordings were as clear as my writings, sometimesbennychico11 said:here's some pretend rep points for you Glen
I tried to give you some but it says I need to spread it around before I give you anymore (even though I think it's been awhile since I've rep-ed ya)
props for this explanation. crystal clear as usual.
. It's funny you mention the points thing, because I owe you and another regular here some props as well for the same reason.
G.
noisewreck
New member
Thanks for that reply Glen. It makes perfect sense. It also explains why I get a pop at the beginning of the track after I "remove" the DC offset
I'm trying to figure out what would cause the pileup though, as usually the only two things occupying the lows (below 100Hz) are pretty much the kick and bass/subbass, and I take care to make sure they don't occupy conflicting frequencies, including using a spectrum analyzer. Could a massive subbass (think Drum and Bass tune) that centers around 40-60Hz cause that?
I usually tend to high-pass my other sounds that don't need the lows, specifically so I don't get a massive buildup, eating into the headroom, especially since I tend to use instruments that alias quite strongly and that alone can introduce some serious unwanted low freq mess.
So, I understand what you're saying, which also means I need to look further into my mixdowns (not that they're any good anyway
).
Sorry, did I just hijack the thread?
I'm trying to figure out what would cause the pileup though, as usually the only two things occupying the lows (below 100Hz) are pretty much the kick and bass/subbass, and I take care to make sure they don't occupy conflicting frequencies, including using a spectrum analyzer. Could a massive subbass (think Drum and Bass tune) that centers around 40-60Hz cause that?
I usually tend to high-pass my other sounds that don't need the lows, specifically so I don't get a massive buildup, eating into the headroom, especially since I tend to use instruments that alias quite strongly and that alone can introduce some serious unwanted low freq mess.
So, I understand what you're saying, which also means I need to look further into my mixdowns (not that they're any good anyway
).Sorry, did I just hijack the thread?
B
bleyrad
New member
noisewreck said:
You need to keep in mind that the DC offset analysis provided by digital editors has no way of actually seeing if DC is offset. It uses some simple math to (essentially) find the average zeroing value of the waveform. The thing is, only pure waves are always totally symetrical. This means that normal music will almost always show some DC offset when analysed in this matter, even when there is none at all.
In other words, unless the offset is significant and/or you know you have a DC problem in your gear, you are usually better to do nothing about it as you will likely actually INTRODUCE a real DC offset if you go about "fixing" it.
Also keep in mind that practically all modern recording interfaces/soundcards will totally block any incoming DC before A/D conversion. Even cheap ones.
And practically all D/A's as well as speaker power amps will block DC (which will damage the voice coil). So the "cone staying in one place test" doesn't usually work in the real world - and if it does, you are damaging your speakers.
Sound quality wise, all a (real, rare) DC offset will do in the digital realm is give you a bit less headroom on that one side of the waveform. You can't "hear" a DC offset in any way and it will not affect overall fidelity except for this one factor.
A DC offset which limits your dynamic range to 25db (as in an example someone else mentioned) would be a sign that God Is Angry At You. Seriously, you would probably have to intentionally bypass the DC protection circuity and jumper-cable a car battery to your soundcard input in order for something that drastic to happen.
If any DC miraculously managed to squeak through in the real world it's almost never more than a half a db or so - totally unnoticable in terms of headroom.
SouthSIDE Glen
independentrecording.net
I have to say that I have seen several laptops - and one or two PCs with mobo-based sound cards - that do indeed introduce this much offset. In fact the integrated soundcard on my current Toshiba laptop does this. I don't recall just how much it measures out to, but it is quite signifigant (at least -25 to -30dBFS positive, if not higher).bleyrad said:
I, of course, do not actually use this sound card for pro recording
. But in playing around with it and the DC Offset plug in Adobe Audition, I find that if I record some silence at the beginning, letting the offset plug work on the flatline. it brings it to zero fine, of course. When the music begins there is no noticable or appreciable variance in the offset with the live data. Of course, as you say, since the plug is "synthesizing" (i.e. guessing at) the zero point, some error has to be expected, but it's probably quite minimal.But the bottom line is, if you do have a device that giving you that bad of a bias signal - and trust me, it does happen with the cheap interfaces - it's best to take that device out of your chain altogether (i.e. not use it) rather than use the offset plug, though in an emergency or other pinch, the plug can get you by.
G.
SouthSIDE Glen
independentrecording.net
Yeah, both you and blyrad have mentioned this, and they will for a pure DC signal. You guys are right that my example would not really happen in the real world with a decent amp. It was more for illustration than anything.mshilarious said:
However, the effect on the loudspeaker when it's playing music biased by a DC voltage remains the same, I believe. Correct me if I'm wrong on this, but when playing back music with a DC bias in it, in effect, the DC current is similar to a "carrier signal" with the waveform modulating it's amplitude. With that modulation, the amp circuit will not necessarily recognize a DC signal, but rather it will simply see a modulated waveform that just so happens to have more positive energy than negative energy. It will send this signal to the loudspeaker which will do it's best to reproduce it. It will do so by spending its overall average amount of time being pushed out. Or put another way, it will be pushed out farther than it is pulled in; its motion will be asymmetrical in relation to it's rest position. The added stress on the loudspeaker is still there.
G.
punkin
Univalve & Avatar Speaks
You guys are beating this subject up rather nicely. I just wanted to add that some time ago, I thought I was seeing a DC component on some tracks I was laying down but found out later that it was really the mic pre when used with one certain mic.
This only happend when the high-pass filter was engaged. When inspecting the track it looked "off-center" but found later that in fact, the positive side of the signal was simply not getting as much amplification. I fiddled with a wave editor to try and fix it and found that it really made things worse. I followed this up by putting a DC volt meter to the pre-output and found nothing there. Just some sort of weird signal smooshing thing going on.
Great stuff guys...
This only happend when the high-pass filter was engaged. When inspecting the track it looked "off-center" but found later that in fact, the positive side of the signal was simply not getting as much amplification. I fiddled with a wave editor to try and fix it and found that it really made things worse. I followed this up by putting a DC volt meter to the pre-output and found nothing there. Just some sort of weird signal smooshing thing going on.
Great stuff guys...