I create a lot of drum and instrument parts from samples using Audacity. I often layer multiple samples (the same, similar, or different samples) in multiple tracks. I then export them into my DAW. Once in my DAW I often find out that there is phase cancellation. Ocassionally I can correct this by inverting one of the channels, but often I can’t. I have some plugins designed for fixign phase problems, but these plugins all seem to be geared to fixing phase cancellation between two or more different audio signals, not one single audio file that has phase cancellation “cooked in.”
So, what can I do to check for phase cancellation before I export? If I know there is phase cancellation before export I can try to bump one or more samples back and/or forth a millisecond/sample or two. Or flip channels. But if I don’t know until after export, I might not be able to correct it (either because I didn’t save the project, or because it’s simply not worth the effort).
Open a new Audacity project and import the WAV file.
Mix it down to mono, then listen to it. If there is phase cancellation in the original, the mono version of the music will go noticeably quiet.
I’m not talking about anything that drastic. With really bad phase issues I can just tell that with no meters at all, as I’m creating in Audacity.
I’m talking about a reading of slight phase cancellation in my DAW. Sorry, I don’t know how to quantify it. I just know that anything center or right is good, and anything left of center = some phase issues.
Also, it’s not just a left vs right thing. I’m sometimes layering 2, 3, 5, 8, 12 different stereo samples over one another. So there are multiple spots where the right channel can phase cancel the right channel, the left channel can phase cancel the left channel, left on right, right on left etc. I really want a real-time monitoring option while I’m in the creation process. I have that in my DAW, but at that point I;ve already exported, and it’s a pretty big time suck to go and blindly try to tweak something in Audacity, re-export and then try again.
Is there really not a plugin to do this in real-time with Audacity?
But I will try your mix to mono suggestion to see if it can serve as a better diagnostic.
I guess that what you really want is something like this: https://www.youtube.com/watch?v=ZG_Vu32c0ZA
That was done on Linux using Jack Audio System and Jack Meterbridge. I don’t know if that can be done on Windows.
I don’t think there’s any automatic plug-in or process that can detect or fix that. The nature of audio (and video) production is that it generally requires human listening and human judgment. If it’s a timing issue (say a 1 millisecond delay between two identical samples) you’ll get a comb filtering effect that can only be fixed by re-aligning the samples.
The best solution would be to avoid layering identical samples. i.e. If two drummers each play a snare drum at (approximately) the same time, there is enough natural variation between the two drums and the timing that you’ll avoid phasing issues. But if you electronically duplicate the same snare hit, aligning it exactly will sound exactly like turning-up the volume of one recording. If they are not exactly aligned, you’ll get that “phasey” comb filtered sound. That’s why true double tracking is better than an automatic double tracking effect (Well… unless you like the sound of automatic double tracking).
I have that in my DAW, but at that point I;ve already exported, and it’s a pretty big time suck to go and blindly try to tweak something in Audacity
From what you’re describing, much of this simply can’t be fixed in the stereo mix. i.e. If you’ve got two out-of-phase drum hits in the same channel, that can’t be fixed. Or if one drum-hit is out-of-phase in the left & right channels and the rest of the recording is OK, you can’t fix that without messing-up the other “layers” or the other parts of the recording.
I’ve been wanting a phase meter for centuries, but yes, you are simply begging for trouble using identical tracks like that. If you play your cards right, you can get a frequency response like attached (comb filter) rather than straight, flat lines. That can sound like playing into a wine glass or rain barrel.
It’s the reason desk microphones can be a bad idea. The voice reflections from the desk can create hollow, phasing effects. Because they are actually phasing errors.
Just to clarify, I’m not layering two identical samples directly on top of each other. As you point out, that’s pointless. When the same (or similar) sample is repeated in the same track, they’re spaced out by an eight note, a quarter, a half etc – some musically /rhythmically relevant spacing. No phasing issues there. And when they’re on two different tracks, they’re not always the exact same sample. Actually, they’re usually not. It’s often two different samples of different velocities of a cymbal hit, for example. I’ve been getting phasing issues in two non-identical samples, fired at non-identical times (i.e. at least a 32nd note apart). I know this doesn’t fit the typical understanding of why one gets phasing issues, but this is what’s going on. Even if I do use two identical samples on two different tracks (two different tracks in Audacity, which eventually gets rendered to just one track for use in my DAW), they would never have the exact some onset. They would be offset by a rhythmically relevant unit. With really short samples, it’s rarely a problem, almost never in fact. Where it becomes an issue is with things with a long sustain, like cymbals, tuned percussion, samples of an individual Rhodes organ note, or a synth chord etc.
I don’t want to fix it in the stereo MIX. I want to be able to fix it as I’m creating the instrument part from individual samples. And I do that when I have audible cancellation or that “squeaking” you can get when samples rub against each other in weird ways (which has gotta be some sort of comb filtering, right?). I just bump one or more tracks back or forward a few ms, or I invert phase, or I flip channels, and I’m usually good to go. But I’m making parts that are sometimes up to 8 different tracks that all get rendered down to a single track for export. So it’s hard to really listen for really minor losses of audibility. It’s the really SUBTLE phase issues that I want to address. The stuff that I can’t hear, but which nonetheless cause some amount of loss of audibility. If there is phase cancellation, even fairly minor (inaudible even), it means there is acoustic energy in my mix that does not equate to audible sound, which means potential loss of headroom, and I think we can all understand why that is something that should be avoided when possible. Which is why I would love to see phase metering in Audacity. But if all y’all are telling me it aint possible, I have some thinking to do.
I realise this is not exactly what you are after, but you might take a look at the Prefix VST effect from VoS, which - amongst other things - has some nifty phase correction tools you might find useful: https://varietyofsound.wordpress.com/2011/05/16/prefix-1-0-out-now/
Almost makes me wish I wasn’t a 64 bit purist. Looks awesome. I’m on Studio One 2, which is 64 bit native, and I don’t want to mess around with bit bridging.
But could I use this with Audacity? I have a suspicion the answer is nooooo…
But it looks like a great mixing tool - EQ/bandpass, trim, phase, gate all in one. I always load a fader (that goes down to negative infinity, which is important for how I mix/automate) on every channel and a separate bandpass filter on most channels, so a tool like this would allow me to use one plugin instead of two! But there are other things that do that as well, so I’ll keep looking for the niftiest mixtool out there. I’m thinking of buying this: http://www.kvraudio.com/product/xtrim-by-sknote/details
