Phase problem

I've discovered something very strange in my setup.  I'm inclined to think it's the result of my Behringer board (which will be replaced as soon as I can afford it), but I can't tell.

When I record in dual mono, the high frequencies are out of phase.  It's clearly visible on sibilance when looking at the phase analysis window.  The problem comes when I record in edit view in mono.  The frequencies cancel out...Much like a de-esser.  The result is a file that is much more dull than it should be.  It's not terrible, but a close listen reveals loss of presence.

To test, I recorded a dual mono file...Inverted one channel and averaged.  I was left with something that sounds exactly like the sidechain of a software de-esser.

This problem didn't arise until I put in my E-mu 0404 card.  However, I was previously using a consumer card, which may not have had good channel separation, so it may have been cancelling and narrowing in the soundcard itself.

So, have you seen this before?  Do you think it's more likely that this is the mixer or the soundcard?

Also, is there a way to specify a channel when recording in mono in edit view?  Or does it always record a L/R average?

It's funny.  I certainly thought I was boosting the high-end more than I needed to, but the final product always sounded like it was lacking.  Now I know why!

Thanks!

Emmett

So, have you seen this before?  Do you think it's more likely that this is the mixer or the soundcard?

No, I've never come across a mixer that did this - well not blatantly, anyway.

I've been trying to think of a easy method of checking the output of the mixer which doesn't involve the soundcard, and I have to say that short of using external test equipment, I haven't come up with one. Well, not a dead cert one where you could quantify the error, anyway.

It does seem a little odd though. If it was the mixer, it would mean that the master stages (ie everything after the panpot contain an interchannel relative timing error, which would certainly create a phase problem. Now whilst that's possible, it's actually quite hard to achieve (even for Behringer!) Which mixer is it?

What you could try is the following:

Feed the stereo outputs of the mixer directly to a stereo amp with a mono button, and feed a mono signal to one input channel, which you pan across both outputs. It should appear dead centre. Hit the mono button. If the sound quality suddenly changes, then there's either a path-length difference or a phase inequality between the two master channels. If everything is working correctly, you should be able to make small adjustments to the pan pot so that hitting the mono button on the amp makes no difference to what you can hear at all.

Checking the soundcard is in some ways easier. Just split a single signal into two, and feed it to both inputs. If the results don't cancel (or at least reduce to a very low level) when you do an invert paste, then something's happening within the card that shouldn't be.

Also, is there a way to specify a channel when recording in mono in edit view?  Or does it always record a L/R average?

It sums the two channels - there's no way to alter this.

The other thing that occurred to me is that if you used RightMark's free analyser twice, first on the soundcard alone with just a loopback, and then with the mixer introduced into the loop, that any differences the mixer introduced would show up in the second results when compared to the first ones. Unfortunately, the one thing that RightMark doesn't test is anything to do with timing. But it will indicate very clearly and graphically any changes in the crosstalk figures, which might be interesting to compare. Also, if you get it to save the analysis files, a careful comparison using Audition would indicate timing differences if there were any, I'm sure.

It's not one of those Behringer mixers with built in digital effects by any chance?

Thanks for the responses!  I've narrowed it to the card.  I switched cables and ran a split output from my (mono) mic processor.  Results were exactly the same.  When summed to mono, highs above 8kHz or so attenuate by about 1.5-2dB...Certainly enough to be irritating, but not enough to really cause a major problem.  Here's a file you can examine...Just a short wav file:

http://www.newaudioproductions.com/Clients/personal/phasetest.wav

Do you think this would be corrected if I used the digital input on the card?  My next mixer will have S/PDIF outs, so that may be an option.

Thanks!

Emmett

We had a problem like this with some Aureol Vortex based cards from Turtle Beach.  I can't remember the exact details, except there were phase problems when the signal was mono'd somewhere within the card.

I got as far as Aureol (think that's how you spell it) sending me an SDK and and Aureol T-shirt. On the day the package reached this side of the Atlantic, the news came through that Aureol had gone bust and was no more.

That kind of removed any incentive that I had to look at the sdk.

I know this is no help, but it's good to reminisce.

Do you think this would be corrected if I used the digital input on the card?  My next mixer will have S/PDIF outs, so that may be an option.

Well, all of this does rather confirm my suspicions that it is  hard to achieve this sort of problem in a mixer, even if it's a Behringer one. But there is also the intriguing possibility that it isn't the card at all - the driver could achieve this all on its own if it was so disposed to...

And if this is the case, I don't think that using the S/PDIF input will make any difference to the results. I don't have an 0404 to try this with, but I do have a couple of its more grown-up but clearly related siblings - I will attempt to find out whether this is a generic trait or not. The answer to fixing it is obvious though - just disconnect one input when recording in mono into EV.  It's probably worth dropping E-Mu and email about it though - just to see what their response is. Or there's that unofficial Emu forum that a couple of their staffers frequent - that's generally been the best way to get some sort of response from them in the past.

That kind of removed any incentive that I had to look at the sdk.

Look on the bright side - you still got a free T-shirt .

Do you think this would be corrected if I used the digital input on the card?  My next mixer will have S/PDIF outs, so that may be an option.

Well, all of this does rather confirm my suspicions that it is  hard to achieve this sort of problem in a mixer, even if it's a Behringer one. But there is also the intriguing possibility that it isn't the card at all - the driver could achieve this all on its own if it was so disposed to...

And if this is the case, I don't think that using the S/PDIF input will make any difference to the results. I don't have an 0404 to try this with, but I do have a couple of its more grown-up but clearly related siblings - I will attempt to find out whether this is a generic trait or not. The answer to fixing it is obvious though - just disconnect one input when recording in mono into EV.  It's probably worth dropping E-Mu and email about it though - just to see what their response is. Or there's that unofficial Emu forum that a couple of their staffers frequent - that's generally been the best way to get some sort of response from them in the past.

I updated the driver before I posted.  Didn't do anything to help the problem.  I've taken this to the E-mu forum, so hopefully I'll hear back!  I've had the card for about a year, and I've always known the problem was there...It's just bugging me more these days.  I really attributed the problem to the mixer until yesterday.

Thanks for the help!

Emmett

Justin from E-mu pinned it down.  One channel is delayed by ONE sample...Who would've guessed that one sample could make that much difference?!

Emmett

Justin from E-mu pinned it down.  One channel is delayed by ONE sample...Who would've guessed that one sample could make that much difference?!

Well, at 11kHz or thereabouts, that represents a 90 degree phase shift, so I wouldn't have said that it was that surprising...

Have you tried what Justin suggested? Personally, I don't think it's likely to make any difference, but you never know...

Steve,

The workaround does indeed work and it does sound 100% better.  And that's great and all...But it's still a workaround that shouldn't be needed.  Yes, 95% of my stuff recorded via the analog ins is mono.  But there's still a small percentage that needs to be stereo.  I'd sure like to be able to do that properly, without a bunch of extra work!

Emmett

Please let us in on the secret. What did you have to do to get round the problem?

Oh, it's really very simple.  I've simply disabled the right channel and panned the left channel to 0.  So anything that goes through the analog ins will be left channel only.  It works for the moment, but won't work forever.

Emmett

Oh, it's really very simple.  I've simply disabled the right channel and panned the left channel to 0.  So anything that goes through the analog ins will be left channel only.  It works for the moment, but won't work forever.

In other words it hasn't really fixed the problem at all!