Stereo Drift

[Neil Eckstein]

So here's a fun one.

Recently, and of course, randomly, I've been having a situation where my stereo field drifts to the left. At first, I thought that it was just me losing my mind, but when I pan L to R there is an obvious volume difference. Then, it just goes away. I've had the wherewithal on two of these occasions to quickly check the gain staging that Cubase is reading, as well as the DM's playback. There is no discrepancy--everything is normal.

The only reason I am posting this here, because I'm guessing most people will say it's a speaker issue and not the board, is because of another troubling issue I've been battling (I'm posting that as a separate thread). I'm starting to think the source of both problems could be a capacitor or something worse...

Let me know what you think--I know this is pretty vague and unsubstantial, but it's all I've got right now.

 

[skier]

In the analog domain, it could be a capacitor issue, but it could also be a resistor issue. Both have been known to change value as they're heated when they begin to fail. Generally, when one of these components causes such a problem (especially a resistor), it's usually part of a voltage divider, either through design or as the resulting effect. So, a voltage divider with a set of resistors could be employed, let's say, to provide a bias to a gain stage. A change in bias could lower or increase the gain affecting the output on one side. Similarly, a capacitor (especially an electrolytic one) meant to provide DC blocking might increase in impedance as it dries out. That increase in impedance could work with a resistor (or another drying capacitor) to form a voltage divider where none was intended. The problem is finding the culprit.

If you have some test equipment, are comfortable using it on complex gear, and are comfortable opening whatever device is the problem, you can try to isolate which stage is causing the problem, and then start testing individual components in that stage to find the one causing the symptom. This is all on the analog side of things; on the digital side, things aren't nearly as simple. And couple that with today's micro-miniaturization and the use of component networks instead of a discreet component and the service issue is harder yet.

I don't know what more to offer in that, the foregoing issues notwithstanding, I've not had the need to open my DM yet and hope to never have to do so. That said, I would continue to attempt to "suss out" the physical device first. It could be the desk; it could be a speaker, it could be a device between the two if you have any there, such as an external amp driving the speaker if the latter is passive, etc.

As you learn more, post it and we can try to help. Perhaps others have had a similar problem and can go to the cause more quickly because of their experience.

 

[Neil Eckstein]

Do you think this could be related to this problem?

http://www.tascamforums.com/threads/clicks-pops-and-squarewaves.4509/

 

[Arjan P]

You'll have to give more info: is the stereo drift also visible on the DM meters or only audible from the monitoring stage? If the meters (display and stereo output meter) show equal levels the problem is very likely to be in the analog output stage. If that is the case, try sending the same source to a different set of analog outputs (AUX out for instance) and see if it still happens.

 

[Neil Eckstein]

The stereo drift is not visible on the DM meters. It's not visible on ANY meters, including within Cubase.

 

[skier]

Neil, some questions:

Have you checked your routing to assure you don't have any unintentional other sources feeding your left channel?

And when the signal appears to drift to your left channel, is it truly a "drift" that you can here moving left to right? Or is it fine at one time and then you notice there's more of it in the left channel at another time?

I ask these questions because, if it seems to just appear there, it could be the result of another source adding to the left channel. The cause could be multiple sources and improper routing would be the natural cause. However, if you can actually here a drift moving toward the left channel, that either means that the source in the left channel is increasing in amplitude or decreasing in amplitude in the right channel. And that could imply a possible component failing as discussed above, or, it could be the result of automation in your DAW that increases the level of the drifting signal you're hearing directed to the left channel.

So, as a starting recommendation, take a thorough look through your routing sessions to see if anything beyond what you're expecting is enabled. And second, try to determine whether or not the perceived drift is truly a drift or something that you suddenly notice is louder on the left.

 

[Arjan P]

I think the possibilities skier names would all result in a visible drift in the DM's meters since they would occur in the digital domain and the meters take from that. I'd suggest trying a test as mentioned above - looks like an analog problem after the DA stage.

PS, somehow I have no quote link for the last message in a topic, otherwise I'd have used that.

 

[skier]

Arjan's comment regarding my suggestions is correct — the louder volume in the left channel should be seen on the left meter. However, you didn't state how much of a difference in level exists, only that it drifts. Therefore, my thinking is that this could be a slight difference that might not be readily noticeable on the meter because they obviously jump a few db following the dynamics of a song. But that does bring up other questions, and Arjan did allude to this in his first post: we, (or at least I) need more info:

1. How pronounced is this shift? Is it slight, or is it a large difference? You stated in the original description that "when I pan L to R there is an obvious volume difference". There is an obvious difference due to your panning? There should be. But I'm not clear what that means of the magnitude of the shift itself. Please explain.

2. How have you assured that the shift is not occurring within Cubase? You and we have assumed it's in the desk, and you've stated that the shift is not visible in the desk meters nor in those in Cubase. But if we're considering a failing component, that could be in many places.

3. How does the signal come into the desk? Is it via Firewire? Is it something else? I'm trying to get a handle on whether or not there are other aspects of the signal in the analog domain somewhere. So, please layout the signal path from Cubase on your computer through the desk to your speakers.

4. Do you have the ability to feed a stereo signal from some other source into the desk via the same route to see if the shift occurs from another source? This could implicate or exonerate the desk. In fact, if the problem is in the desk in one of the later stages, feeding a stereo signal in anywhere manifest the shift.

If you can eliminate or implicate a device in your signal chain, it would give you and us more to work with.

And Arjan, if I'm still missing some obvious things, please feel free to mention them. I, too, am getting older, ya know.

 

[cmaffia]

I'd check this first. Make sure your Pan Law in Cubase's project settings is set to 0dB.

http://www.musiciansfriend.com/resources/article/Panning-Laws-Revealed/m710671

I recall having panning weirdness when I first started using Cubase/DM3200 and it was due to Cubase's default Pan Law setting (equal power - which drops the center by -3dB). I would try the 0dB setting if you're not currently using it.

If it's already set to 0dB, then I'd try to figure out your monitoring/square wave-like sound/routing issue next and hopefully that resolves all your issues. If not then I'd proceed with the suggestions in this thread.

 

[Neil Eckstein]

The routing seems to be correct; and even if it wasn't, I don't think it would be so random, scarce, and extreme. I've wondered about possibly resetting to a default template because maybe there's something buried I'm not seeing (is there even a default reset somewhere?). However, as Arjan noted, I would see it in the meters.

In response to Skier:

1. It is a DRASTIC drift. What I meant by panning one way or the other is that I was testing to see if I was crazy or not--when I panned left (where the signal was drifting), the signal was normal volume, and when I panned right it was almost inaudible.

2. I am NOT convinced it isn't Cubase (or even windows!). The only reason I'm doubting the computer or the DM's internals is because I don't see any changes in the meters when it happens. If the meter I'm looking at is before some output component that I'm ignorant of, then it could be cubase or it could just be between the board and the speaker. I just don't know.

3. The signal comes in Firewire as stereo outs 1 & 2 from Cubase. Those are routed to ch's 47 and 48 (linked) as slot 1-1 and 1-2. Those two channels are "output assigned" to "Stereo."

4. I guess I could run another computer back to the board via firewire but I still wouldn't be certain if I could replicate the problem. I can't get it to happen intentionally!

I recognize that the fact that I can't reproduce the problem at will limits the ability to implicate. Trust me--it's pissing me off.

-----------

@cmaffia: How do I make those changes you suggested?

 

[cmaffia]

Project settings

 

[skier]

Sorry for the delayed response; I was, well, skiing up at Whiteface at Lake Placid.

1. [It is a DRASTIC drift. What I meant by panning one way or the other is that I was testing to see if I was crazy or not--when I panned left (where the signal was drifting), the signal was normal volume, and when I panned right it was almost inaudible. ]

- Ok, so when drifting, regardless of where set your pan levels, it appears that most of your stereo signal ends up in the left channel. When this happens, are you getting almost the entire stereo signal in the left channel, in other words, mono? Or, is the right channel signal missing? If the former, I currently have no reasonable scenario in my mind to explain that yet. But if the latter, it means that you're losing right channel and the shift is really the right channel amplitude decreasing for some reason.

- Also, do you see the right channel meter on the desk reduce its level? And/or, does the right channel level reduce on the Cubase stereo meter?

2. [I am NOT convinced it isn't Cubase (or even windows!). The only reason I'm doubting the computer or the DM's internals is because I don't see any changes in the meters when it happens. If the meter I'm looking at is before some output component that I'm ignorant of, then it could be cubase or it could just be between the board and the speaker. I just don't know.]

- I agree, we have insufficient info and that's why we're trying to eliminate possibilities to get to the true cause.

3. [The signal comes in Firewire as stereo outs 1 & 2 from Cubase. Those are routed to ch's 47 and 48 (linked) as slot 1-1 and 1-2. Those two channels are "output assigned" to "Stereo."]

- Well then, what happens after the signal leaves the desk? What does it go to? Does it go straight to a pair of active speakers? Or does it go through some other equipment, such as a monitor controller, or to an amp that feeds each speaker, or something else? Anything failing in the right channel after the desk will not be seen on either Cubase or the desk meters, so this is a particularly important question.

4. [I guess I could run another computer back to the board via firewire but I still wouldn't be certain if I could replicate the problem. I can't get it to happen intentionally!]

- This is a good idea if it's easy for you to do. Otherwise, why not wait till we explore the previous options. If we make little progress there, then it'll be worth the extra effort so as to pinpoint the cause as in the computer or in the desk.

 

[skier]

Neil,

One more thing to add to item 4 above:

You don't actually need the output of another computer to feed into the desk to determine whether or not most of the desk is causing the right channel decrease. All you need is an audio source of some kind and that should allow you to test, not all, but most of the path you're currently using. The source can be a CD player, a tablet, phone, iPod, etc. Just connect something playing music with a stereo output into some line inputs and see if the drift problem occurs.

You won't be testing the Firewire input, but I'm doubting the problem lies there. Once your signal is in the desk, route the inputs you're using as identically through the same stages to your output as possible. In other words, match as much of the signal path through the desk as possible to test the very same path the signal from Cubase is taking. (Sorry if I'm saying the same thing 3 different ways — I just want to be clear.)

Finally, let us all know what you find.

 

[cmaffia]

Curious as well if you checked the Cubase project's Pan Law setting and whether or not its set to 0dB.