Picture this: you’ve spent hours recording and capturing the perfect guitar tone from your amp. But come mix time you take a listen back to your tracks and they sound totally thin and unrecognizable. You swear the mics sounded great by themselves when you set them up, but now when they are all playing together in the mix your tone has been completely obliterated! Where did it go? We’re about to find out…
Graham Cochrane is a creative and versatile producer, engineer, songwriter, and worship leader based out of Tampa, FL. He provides remote mixing, and mastering services to clients worldwide.
As the founder of The Recording Revolution, Graham's articles and tutorials have been featured worldwide by Avid, PreSonus, Slate Digital, Editors Keys, and Behringer to name a few, reaching readers in over 40 countries.
For more information and samples of Graham's work, please visit www.grahamcochrane.com.
Comments
get well soon
Thanks
This makes me laugh, as it reminds me of the exact same situation I was in when I first started recording. I was recording my Line 6 amp and had an SM57 on the front and a 57 on the back capturing the low end of the speaker cone. This was the first time ever that I discovered phase. I was baffled when I was combining both mics and thinking “why does this sound so bad?! My guitar tone sounded fine just a minute ago!”
Definitely a great tip. I’d say phase is probably the most important thing when recording… that along with not clipping the signal on the way in and the quality of the instrument itself.
Here is a funny trick I learned at the homerecordingshow.com when multimiking a guitar amp: place the first microphone wherever you like, or wherever sounds best, then when placing the second mic search for the spot where the two mics sound horrible together, then flip the phase and you have a really great sound. It’s just easier to hear an awful sound, you get what I mean?
Very interesting idea.
This is super helpful!
Once again, I knew that this happens and this issue exists but had no idea how to fix it or what was going on theoretically. Thanks again Graham!!
Yes get well soon!!
Hi Graeme.
I’m assuming it’s possible for two tracks (same performance, different mics) to be out of phase by something other than “complete opposite”?
If I was to try plotting a simple illustration of what you dealt with, this is what I come up with (Example 1):
http://www.wolframalpha.com/input/?i=sin%28x%29%3B-sin%28x%29
Basically two sine waves (easier to graph than a distorted guitar!) but perfectly out of phase with each other.
But what about this (Example 2):
http://www.wolframalpha.com/input/?i=sin%28x%29%3Bsin%28x%2Bpi%2F2%29
This is basically two sine waves out of phase by 90 degrees. In this case, no matter how you invert phase on either track, you’ll always end up with bands during the wave where one signal is positive and the other is negative, therefore resulting in some degree of cancellation.
In this plot, I move them ALMOST completely out of phase with each other (Example 3):
http://www.wolframalpha.com/input/?i=sin%28x%29%3Bsin%28x%2Bpi*0.8%29
In this case, inverting phase would definitely improve the mix, but it would still have a small degree of cancellation.
In this plot, I move them only SLIGHTLY out of phase with each other (Example 4):
http://www.wolframalpha.com/input/?i=sin%28x%29%3Bsin%28x%2Bpi*0.2%29
In this case, inverting phase would make things worse, but again there is a small degree of cancellation.
Although toggling phase on a track will give you two options (and you select the better of the two with less cancellation), would it not help to also “shift” one of the tracks so that the waveforms are in perfect alignment?
I speculate possible reasons as being:
- Mics at different distances
- Mics at different orientations
- Incorrectly wired/patched leads
- Different A/D latency per mic (if using multiple interfaces)
I’m not very hands-on these days, but I still enjoy learning/thinking about this stuff. Thanks for a great video, which inspired me to think/write this!
P.S. When looking at the WolframAlpha links, there is a section called “Parametric Plot” under the regular “Plots”. This is quite good for visualizing how much time is spent “in” or “out” of phase.
Top-Right means both signals are positive = Good
Bottom-Left means both signals are negative = Good
Bottom-Right or Top-Left means one signal is positive, other is negative = BAD
My Example 1 plot is worst case scenario where it just ends up with a line through the BAD sectors. Flipping phase would make it all good!
My Example 2 plot is equally spread out between good and bad. Flipping phase would make no difference.
My Example 3 plot is mostly bad, with a bit of good thrown in. Flipping phase would make it mostly good, with a bit of bad thrown in (therefore recommended).
My Example 4 plot is mostly good, with a bit of bad thrown in. Flipping phase would make it mostly bad, with a bit of good thrown in (therefore not recommended).
Marty – Cool use of WolframAlpha
And your analysis is correct.
The phase (aka polarity) button Ø is equivalent to swapping pins 2 and 3 on an XLR cable. Everything that was positive voltage becomes negative, and vice versa. As a result, the signal is 180º out of phase across the entire waveform (all frequencies). As your examples 1-4 show, when the sound is out of alignment but not fully reversed, inverting one signal relative to a second often makes then sound “better” together (closer to in-phase), albeit not perfectly matched.
An alternative method is “time alignment” — delaying one track a few samples compared to another so that the peaks and valleys line up. If you have a near and far mic on a single source (like a guitar cabinet), this is the more accurate way to line up the signals. But as it’s a bit more tedious to set up, many audio pros try the phase/polarity button first to see if that gets the signals “close enough”.
A third approach, which is more “by ear”, is to use EQ and crossover circuits. Any EQ or crossover not specifically advertised as “linear phase” will change phase relationships of various frequencies. Applying different EQs or crossovers on 2 tracks of the same signal can produce comb filtering — or conceivably reduce existing filtering on certain frequencies based on phase or time alignment issues.
You can find hardware and software solutions that employ each of these methods, and can replicate them with built-in features on most DAWs.
Thanks for the additional info. You just blew my mind with thinking about how EQ changes the relationship as well!!! I’m going to dedicate some thinking time to that…
Okk, now I know what a phase problem is and how to overcome it..Definitely going to double mic my guitars from now on, without phase issues..
That’s Awesome..
Or you could just send the signal through a virtual amp modeler on a sidebus and mix it in like Graham explained in another video. That way you won’t have any cancellation issues. Worked for me. Used a 4×12 low emulation to give guitars a fatter sound on my album.
[...] Links: Graham Cochrane – How Not To Destroy Your Guitar Tone [...]
Thanks for the explanation, Marty.
Your post really helped me out with understanding phase scope, another great tool to get your phase right!
I just couldn’t get the hang of it, but now I can finally use them (:
Glad it helped! Just wish I had more time in my life so I could actually apply some of this in practice. Alas, I haven’t recorded anything in a few years due to other projects taking priority. Soon tho!!!