Topic: Need help centering the stereo image in Adobe Audition or iZotope RX2

[jj69]

I made a wonderful recording of a hard rock band with one minor issue - for some reason the lead vocals are not in the center of the soundstage.  They're too far to the right.  I used AT943 cards and I was actually sitting in the center of the room, so I'm not sure if the vocals were just mixed too loud in the right channel, or if something else happened. 

I tried amplifying the right channel, but I have to add around 3db to the left channel to center the vocals.  Unfortunately, that leaves ms with output levels that are much higher in the left channel. 

Is there an easy way to pan the stereo image in Adobe Audition or RX2? 

[bombdiggity]

I would probably cross mix your channels...  not pan. 

In AA (3 anyway, but it's more or less the same in all of them):

Effects > Stereo Imagery > Channel Mixer   

Adjust the balance of each channel of the new L/R to taste.  You may only need to bring some of the right into the left or reduce the original left and add some right to it at the same time.  There is a pretty much endless array of possibilities (or work on the right and leave the left alone, or adjust both). 

Make sure the channels are in phase (or close) before doing this and that your result isn't clipped.  You can then balance them (or do that in the mixing and amplify after that). 

The panning and all is that section too if some of those effects are more to your taste.  You can preview almost everything to decide what is working on your source. 

[jj69]

Thanks for the quick reply. How do I check if the channels are in phase? Can this be done in Audition 3.0?

[bombdiggity]

That is mainly a visual thing for me.  Basically zoom real tight in a number of different areas and make sure the peaks and valleys are in the same place on the timeline in both channels.  If you see one or the other channel consistently ahead (in terms of the same sounds appearing earlier on one channel than the other) you should correct that before mixing.  With stereo there are usually a variety of natural phase differences (part of the stereo image) spread between both channels without any obvious consistent bias toward one side or the other.  That's fine and to be expected.  Sometimes (typically if mic placement is a bit off for the position in the venue) one channel will be leading and the other trailing.  That should be corrected, especially prior to mixing the channels together (unless you like a lot of "reverb"). 

[jj69]

In that case, I don't think there's a phase issue.  Zooming in, the channels appear so inconsistent that there's are very few peaks that are similar.  The few I can find that are seem to be spot on. 

I have been trying different things with the channel mixer, but it's very frustrating.  I can move the lead vocal from the far-right closer to the center, but I can't seem to get it in the center at all. 

The best result I can get is to add "100%" of the right channel into the "New Left Channel."  Anything other combination I try leaves me with a mono (or almost mono) result. 

Any other suggestions? 

[bombdiggity]

I'd probably need to look at and hear a bit of it to make any substantive suggestion beyond the general techniques involved. 

I've mixed things with 2 totally independent channels to results I liked but it always depends on lots of factors. 

I generally vary the % of each on each channel.  It will become less "stereo" but that's the nature of it if one doesn't like the natural stereo effect of the source. 

[jj69]

Here's a five minute sample in the original 48k/24bit wav format:

http://www59.zippyshare.com/v/14014759/file.html

This is the raw recording, with no processing or editing at all.  Please let me know if there's something you (or anyone) can do with it. 

I'm just frustrated that there isn't a better tool out there.  I would think there would be a plugin designed for issues like this - i.e. something that can move the center of the  soundstage from right to left. 

Anyone know if there's maybe something in iZotope Ozone? 

[bombdiggity]

That helps...

This recording is substantially out of phase so where you were hearing the vocals all in the right I was actually hearing everything in the right (this is essentially how one hears a severe phase shift - the leading channel is all you really hear/perceive, the trailing channel is basically dead). 

So mixing wouldn't fix your issue and most likely isn't necessary (at least not to my ear). 

My quick estimate is the right is ahead about 15 samples, though I didn't play with it much to try to fully verify that.  It seems to vary between 11 and 17 samples but 15 was looking like a fairly consistent offset.  You can actually try several different offsets on your sample by varying the starting point of the right channel where I note the specific point below and go with the one that lines up best:

So combine the entire set in Audition if it is in more than one part.  Convert your scale to samples.  Copy the right channel only, then with the right channel only selected go to the point 15 samples or so in and right click on the right channel only > Mix paste > replace.  This will move the entire right channel back by the number of samples you pasted it at and get them in phase with each other. 

Now you should hear both channels more or less in alignment and that very annoying effect you had will be gone.  There is substantially different content between channels but it seems quite a bit better after that.  I sort of suspect your left mic wasn't pointed in the same orientation as the right (or you were seated or standing somewhat sideways in orientation to the stage/sound sources). 
   
You also have a substantial DC offset in the sample.  Correct that by highlighting the entire thing (both channels) > effects > amplitude and compression > amplify/fade (process) > center wave preset

Your left channel is nominally about 1.5 dB weaker than the right.  That will also make the left seem overpowered by the right.  I hear no reason not to boost it to even them.  Of course when you're done with everything the whole thing needs somewhere over 12 dB to get the levels up near 0 dB, so you can balance them by amplifying them each once separately to par (which may be a better process). 

Here's the scratch sample with those quick adjustments so you're hearing apples to apples between channels (not with the final amplification though):

http://we.tl/1KgHOf0D6E

HTH. 

[jj69]

THANK YOU.  This is fantastic!  I just wish I knew how you spotted the phase issue.  I tried to zoom in in Audition, and I couldn't spot it.  Maybe I'm not looking for the right thing? 

You are probably correct that one of the mics was likely off axis, as this was a stealth recording with the mics clipped to my shirt.  You just try to do the best you can.  The left channel is usually down 1db - 2db in most of my recordings.  I assume there's an issue with my gear, but I don't know if it's the mics, the battery box, or the Sony M10.  I usually just boost the left channel as needed. 

At this show, I was seated almost exactly in the center, but in the sixth row (which was probably a bit too close to the PA).  Surprisingly, I did listen to another recording of the same show, and the sound is very much the same, favoring the right channel.  Is it possible there was an issue with the house PA that could cause this result, or was it more likely mic positioning? 

[bombdiggity]

Glad that seems to have solved it to your satisfaction.

Mainly I just hear out of phase immediately, though have become much more attuned to it once I started to realize what the impacts were.  It is easiest to hear in cans, which is often how I listen.  If one channel is completely out of the picture that's always the issue. 

Apparently you can hear it too but just didn't know what exactly it was you were hearing... 

To verify it is best to go to the start of a song or pick particular points where there are large changes in volume/stops and starts/etc.  To determine the offset try a number of points (as some variation in a stereo recording is natural) and don't be afraid to experiment a bit with trial and error if you're not sure.  Identifiable events are easier to pick up and pick out.  With mono recordings (which are often out of phase) it is very easy.  Stereo is trickier.  It happens more often than we may think, but usually is a relatively small shift.  This one was more dramatic, though I've heard worse.   

The difference in recorded level is probably the mics if consistent.  Those capsules are not matched, so a dB or two would likely be normal. 

Maybe something was wrong in the house.  They could have had some sort of wiring or signal processing issue.  If it happens more than once in the same venue or is present in multiple recordings of the same show I'd suspect that. 

[jj69]

Thanks again for your input!  You really saved this recording. 

Just one follow-up question.  You stated that "the right is ahead about 15 samples."  As a solution, you suggested that I should: "Copy the right channel only, then with the right channel only selected go to the point 15 samples or so in." 

This is exactly what I did, and with great results. 

Logically, however, this doesn't seem to make sense.  If the right channel is 15 samples ahead of the left, then why does pasting it another "15 samples in" solve the problem?  Wouldn't that make the right channel even further ahead of the left?  If it's already 15 samples ahead and I paste it another 15 samples ahead, then shouldn't the result be 30 samples ahead, thereby leading to an even worse result? 

I'm sure I'm missing something simple here, but something seems backwards??? 

[bombdiggity]

Thanks again for your input!  You really saved this recording. 

Just one follow-up question.  You stated that "the right is ahead about 15 samples."  As a solution, you suggested that I should: "Copy the right channel only, then with the right channel only selected go to the point 15 samples or so in." 

This is exactly what I did, and with great results. 

Logically, however, this doesn't seem to make sense.  If the right channel is 15 samples ahead of the left, then why does pasting it another "15 samples in" solve the problem?  Wouldn't that make the right channel even further ahead of the left?  If it's already 15 samples ahead and I paste it another 15 samples ahead, then shouldn't the result be 30 samples ahead, thereby leading to an even worse result? 

I'm sure I'm missing something simple here, but something seems backwards???

Sorry if it wasn't clear but you probably did it right. 

At least in how I think about it "Ahead" on the editing timeline is relatively left.  "Behind" is relatively right.  You want the same event happening at the same point in both channels.  Whichever channel the sound is occurring earlier on will be relatively left of center, the one that is behind relatively right of center.   Essentially you are eliminating the differential or variance between the two channels.  Once you know the distance you then move the leading channel back/in/right or can move the trailing channel ahead/up/left... 

If the right channel is ahead 15 samples you need to move it back in time (to the right on the timeline) to correct it.  So by going "in" 15 samples and pasting you are delaying it 15 samples.  You actually can't go the wrong way if you're starting from zero since you can't paste anything before zero. 

An alternative approach would be to copy the left channel from 15 samples after zero and paste that starting at zero. 

I usually move the one that is ahead back in time but nothing says you can't take the opposite approach.   

[page]

Mainly I just hear out of phase immediately, though have become much more attuned to it once I started to realize what the impacts were.  It is easiest to hear in cans, which is often how I listen.  If one channel is completely out of the picture that's always the issue. 

Apparently you can hear it too but just didn't know what exactly it was you were hearing...

All of this. In general, there are lots of effects and quirks that once you realize what you're hearing, become really easy to spot. Phase issues are one of them.

[bombdiggity]

^

Yes.  Hearing it is the first step.  Identifying the problem behind what you're hearing the second.  Solving it the third (at least in post, if not in person).  Learning the background behind it and getting to a point of quick recognition seems a much longer process. 

In this case the OP was hearing it and it turned out to be pretty easy to solve.  It takes time to become fully attuned but at least we have forums like these to discuss/learn in. 

[jj69]

Indeed. Thanks so much for saving this recording !