Phasing??? wha the #^^@%#^&%!!!
- Thread starter fistula
- Start date
Mutant
I hate that supercow !
You don't need hundredths of inches of accuracy to achieve in phase. Both mics on the grill = in phase. The fact that I was getting about 30 samples per foot of difference would indicate hotter temperatures, not Arctic. Again, I was looking for ballpark figures, not some labcoat entrenched Aes proposal a-weighted tested standard on graph paper.
With one mic on the grill and one mic pulled back a foot, the time delay would be the very thing causing the phase difference. Same source = same frequencies, so that's ruled out. Peaks and valleys on one waveform would be delayed by a number of samples compared to the other waveform, though I was looking at the zero crossing points of both wavs to compare when I got the above figure of about 30 samples/foot.
At 70 degrees, sound would be about 1130'/sec. 44100 divided by 1130 is about 39 samples of time difference per foot, which is about 0.9ms, something that Mood Bender also came up with. One peak on one wave would be 39 samples delayed than the other. You might not be aware of the sample count when the mouse click lands on a spot when editing, but it should be around that number - for those figures - if you were ever so inclined. And if you're like most guys, a single sm57 is going to do it, and this isn't going to apply, of course.
All that said, I'm starting to like pulled back condensors and close up dynamics again, without even bothering to look at the phase, and keeping it out of phase.
That's not true - unless you have identical mics holding identical positions you're going to have different responses, and if you have any regard at all for high frequencies you're going to want to have things aligned properly to not have a comb-filtered effect. That's why I don't pull out the yardstick when it comes time to record - 1mm of difference can completely throw you off (just try the Fredman test if you don't believe me) in the vast majority of situations.
Again, unless you have identical mics on the same source in the exact same position you're going to have different sounds picked up and *that* is what we're looking out for. Your recording fidelity is not what you're trying to get aligned properly, the source frequencies have to be together - if it was your recording settings, downgrading from 96k to 44.1k would throw everything out of phase and any phase anomalies would be solved by lowering your fidelity. Complete and utter bullshit.
Jeff
Mood Bender
Member
Actually I was using two sm57s for my testing. But I wouldn't feel the need to align two different mics if they were the same distance from the source. Like the 421, which I've been getting into again. I wouldn't even stress over 1mm of difference. You can even hear when the phase alignment is good enough if you flip the phase of one mic and position the mics to where they cancel out each other the most. No big deal. Personally, I prefer mics to be straight on instead of one angled like the Fredman technique, but I've heard people make it sound good.
Seems like you are not grasping the concept of sampling frequency as a measurement of time, not just a fidelity issue. Instead of milliseconds, it's possible to count the samples in smaller fragments of time, this is the only thing I refered to, not the quality of audio as far as fidelity. 1 ms = 441 samples.
I grasp the sample count timing perfectly, but I do not consider it to be how you should be aligning sound sources. This is something found by anyone who tried the stuff in the thread about the Clayman sound - tiny changes in position do matter, especially when mics are as close. My point is that when you're grabbing different parts of the sound from a source (as you would invariably be with multiple mics) you need to physically align them in a way that reduces summation anomalies - 'both mics on grill' can be just as out of phase with each other as anything can. When it comes to phasing, sample rate is not of concern until you get to the point where you're chopping off significant amounts of your signal, and that's the problem I have with that you seem to be saying.
Jeff
Jeff
If sound is made of waves, then shouldn't my speakers get wet?
Yup, that's what I thought. Trying to impress people with your technical-speak... and you didn't think you'd get caught!
Amateurs...
Yup, that's what I thought. Trying to impress people with your technical-speak... and you didn't think you'd get caught!
Amateurs...
JBroll, it doesn't seem like you're upset. I'm not upset. We are just debating, right? No harm done, people, we're just talking. Disagreements can be diplomatic. Emotional, but civil.
I don't think sample count is necessarily how one should align phase either, nor did I ever state that. Remember, I stated that it is the waveform itself, the peaks, dips or zero-crossing point that should be aligned with the other waveform. Dragging and droping, whatever. The sample count is just a byproduct of a particular time, for a particular sound, for particular mics, in a particular postition. However, noting the sample time difference is fundamental in understanding why things are phasing in the first place, and the byproduct of fixing that is shifting the time. In "Modern Recording Techniques", the author states that: "phase shift is a term that describes the amount of lead or lag in one wave with respect to another." Time delay, pure and simple. My case was time with respect to samples. Someone else might just look at the pictures of the wavs on the screen and drag and drop and not even care about how many samples, but that's whats going on behind the scenes, like it or not.
As for both mics on the grill being just as out of phase as anything, that's just simply not true. The fact that they cancel each other out almost completely when flipping the phase on one of them is a testimony to that. I implore you to test for yourself if you don't believe me. The same distance for both mics is the closest phase alignment you will get, barring editing.
I don't think sample count is necessarily how one should align phase either, nor did I ever state that. Remember, I stated that it is the waveform itself, the peaks, dips or zero-crossing point that should be aligned with the other waveform. Dragging and droping, whatever. The sample count is just a byproduct of a particular time, for a particular sound, for particular mics, in a particular postition. However, noting the sample time difference is fundamental in understanding why things are phasing in the first place, and the byproduct of fixing that is shifting the time. In "Modern Recording Techniques", the author states that: "phase shift is a term that describes the amount of lead or lag in one wave with respect to another." Time delay, pure and simple. My case was time with respect to samples. Someone else might just look at the pictures of the wavs on the screen and drag and drop and not even care about how many samples, but that's whats going on behind the scenes, like it or not.
As for both mics on the grill being just as out of phase as anything, that's just simply not true. The fact that they cancel each other out almost completely when flipping the phase on one of them is a testimony to that. I implore you to test for yourself if you don't believe me. The same distance for both mics is the closest phase alignment you will get, barring editing.
Mutant
I hate that supercow !
Just take any two mic recording and move one of the waves back and forward in your daw.
I'm not upset, don't worry.
The problem I have with using drag-and-drop is that we're aligning summations of waves and we can't even see all of them after they're combined - at best you could see when they both start and end if they're identical, but apart from that there's no telling what order of fresh hell you're getting into when you align them like that, so I'm with you there.
On the mic distance and phase, I didn't say they were - I said they could be. If you're getting identical signals from the mics, then you're just having a louder version of one mic and you might as well not even bother; if they're different (as they practically have to be in anything resembling reality) then 99 times out of 100 you're going to have something canceling something else out. I have tried it myself, and that is the subject of the Clayman thread next to this one - flat or angled, either way they're picking up such different parts of the sound from the different parts of the speaker that they're going to just be strange sometimes. I'd have to disagree about your same-distance blanket statement, as I don't think that's a safe thing to say, and I'll get back with some nastily out-of-phase mic-on-grill combinations if I wind up home at some decent hour when I can record.
Fistula, another thing you could try (if you have two similar mics) is pointing a mic at the point where the speaker's cone meets its dustcap and then sweeping another mic around the speaker - if you pan the mics on opposite sides, you won't hear them affecting each other, but if you put them both in the center you'll find strange interactions like comb filtering at some places and great-sounding combinations in a few others. I have some sample guitar clips laying around that I can just combine in and out of phase for you, but it'll be later today when I'm out of class. What Mutant said was right, though - you could even just take one individual clip, copy it into a second track, and move one forward and backward little bits at a time to see how they interact, and then flip the phase on one of the tracks and play with it again, and so on.
Jeff
The problem I have with using drag-and-drop is that we're aligning summations of waves and we can't even see all of them after they're combined - at best you could see when they both start and end if they're identical, but apart from that there's no telling what order of fresh hell you're getting into when you align them like that, so I'm with you there.
On the mic distance and phase, I didn't say they were - I said they could be. If you're getting identical signals from the mics, then you're just having a louder version of one mic and you might as well not even bother; if they're different (as they practically have to be in anything resembling reality) then 99 times out of 100 you're going to have something canceling something else out. I have tried it myself, and that is the subject of the Clayman thread next to this one - flat or angled, either way they're picking up such different parts of the sound from the different parts of the speaker that they're going to just be strange sometimes. I'd have to disagree about your same-distance blanket statement, as I don't think that's a safe thing to say, and I'll get back with some nastily out-of-phase mic-on-grill combinations if I wind up home at some decent hour when I can record.
Fistula, another thing you could try (if you have two similar mics) is pointing a mic at the point where the speaker's cone meets its dustcap and then sweeping another mic around the speaker - if you pan the mics on opposite sides, you won't hear them affecting each other, but if you put them both in the center you'll find strange interactions like comb filtering at some places and great-sounding combinations in a few others. I have some sample guitar clips laying around that I can just combine in and out of phase for you, but it'll be later today when I'm out of class. What Mutant said was right, though - you could even just take one individual clip, copy it into a second track, and move one forward and backward little bits at a time to see how they interact, and then flip the phase on one of the tracks and play with it again, and so on.
Jeff
Mutant
I hate that supercow !
I'm not sure if i undertood what you said here.
An oscilloscope is the right tool if you want to see everything in realtime - in our situation you would need just 3 of them to see both waves and their sum.
Mutant, I'm not talking about that, if we were to align things visually (which we already know to be a no-no) we'd have to be able to see all of the individual components of the wave before summation, and that's just not fucking happening.
Fistula, I've put together about a minute and a half of demonstration from a short riff from a Mnemic song (Door 2.12, I think) that was miced with a modded SM57 and an Audix i5. Here's the clip, and I'll ramble mindlessly about what I did to change things here and there. Thanks again to Gavin, for being helpful/generous/silly enough to share space for such nonsense as this, and here we go:
http://www.upload.celtiaproductions.co.uk/uploads/JBroll_-_Phase_Samples.mp3
The riff itself is only 8 seconds long or so, but I wanted to use something short because of size/length limitations (we're talking eight different 'sections' from this clip, and it's done at 320kb/s, so we can't get too long) and I think this gets chugs nicely enough to get the point across.
Anyway, the setup was simple - as I said before, SM57 and i5, right on the grill pointed at the two speakers on my amp, and it all went through identical modded-to-fuck-and-back ART Tube MP preamps and into an Audiophile 2496. It may be a bit 'boomy' for your taste because these weren't meant to be doubled up, but whatever. I also like mids - a lot - and I make no apologies for that. The clip is a teeny bit sloppy, as I was half asleep and almost forgot the thing, but whatthefuckever - on with reality.
The first four 'sections' (the first fifty-six seconds) are with the two files, which are/were aligned as they were out of the amp (for the record, the 57 needed flipping to be put in phase and that was taken care of at the preamp), one panned hard left and the other panned hard right. The first clip is just as things were, with nothing funny but the original phase correction. The second clip is the same but with the phase on one of them flipped (it sounds to me like this one is getting ripped inside out, but whatever) in the DAW. The third clip has a slight shift in one of them, not enough to be perceived as a 'shift' but enough to have even more mid and even less fizz - it was something obscenely small, like a twentieth of a millisecond if I'm not reading the DAW wrong, but it certainly makes a noticeable difference. The fourth clip is the same as the third but with the phase on one of them flipped - again, I get that ripped inside out thing, but the sound is still somewhat identifiable. Technically, the sounds aren't 'directly' interacting, as they will be when everything is smacked into mono, but they can certainly 'feel' funny even when they aren't really 'touching' each other... for some reason the flipped segments make my head feel funny when I loop them for too long, and I start seeing the bears again, and the smilies start talking to me (fuck, I hate those smilies), but that may not all be the phase issue.
The last four sections, from fifty-six seconds on, is with both tracks dead center. The first track is the two tracks together, one of them may be shifted slightly but I can't fucking tell through this headache; again, you can tell that they weren't meant to be doubled by how thumpy they are, but no complaining until someone pays me to do this - everything in the 'meat' is thickened up, perhaps to a fault, and the fizz is much more fun than what usually happens from one mic. The second track is those same tracks flipped: here we get all of the fizzle that was yoinked out of the first two, because rather than adding all of the meat and subtracting the fizzies we've killed our meat and now we're adding the Alka-Seltzer by the bottle... this is the exact opposite of the Fredman technique (and a bloody good sign of success, if I do say so myself, when I flip the phase switch) and looking at it through a frequency analyzer I found hardly anything in the 'meat' range of the guitar. The third clip is like the first, but the shift I did is doubled - I think it shows how well we can have a 'sweet spot' where the planets align and the bunnies give the children their Easter eggs and that bitch didn't take my good knives when she left and our sounds are lined up just right, and then shift things a tiny bit too far and sound like we put a blanket over everything. The fourth is the same as the third, but again with the phase switched and with a tiny hint of lower stuff involved - I might analyze this later, if I have a *lot* of candy and nothing else to do, but right now I'm not worrying about it.
Basically what I've tried to show is that, when properly used, phase cancellation and addition can be fun... but that proper use is a bitch and a half. Before these clips were put down a while back, I had been spending day after day looking for the happy spots on the speakers, and when I sat down to tinker with it tonight I found that after all of that time I was still off by the tiniest fraction of an instant from the gigantic mess of a Fredman setup I was after - goes to fucking show that no matter what you do some asshat (like me!) can come along, make the tiniest and most minute change, and make everything perfect, and all it came down to was having the patience to make that micrometer difference or that twentieth-of-a-microsecond delay or something obnoxious like that. You can get about five metric shittons of pain dropped right on your ass because your mics were off by a distance you couldn't even see to save your life, or every god of every religion could spend all of eternity blowing you for the joy you've brought to the world with your phase anomalies, and it's not entirely unlikely that something less dramatic than the previous two will occur in most situations (although in my experience that's not the case), so sit there, play with details, breathe on your mics if something isn't just right (actually, you'll look like an ass blowing on all of your mics that long, so hire an intern for credit and make her do it), sit there and bang randomly on the keys, making 'Ook! Ook!' noises until effective (or don't), and DON'T
EVER
be afraid to just hit the fucking mute button on a track that's annoying you. You won't hurt it's feelings, in the long run you're just giving it more time to go off and get hammered with its buddies if you don't use it. Honestly, I'm never going to use those two tracks together in any project, and it's almost 100% guaranteed that I'm not going to use this current setup for anything more than study and novelty, and the world probably couldn't care less how I got things to sound this way or that in the long run. Phase summation can be fun, but sometimes it's utter nonsense - sometimes the best sounding option is to just mute this or delete that and be done with it, other times the godly tone of your dreams only comes from eighteen mics and a piezo pickup duct-taped to a trash can lid, so using what little there is to learn from a lunatic on a message board just put shit together in a way that makes sense in at least a half-conscious stupor, don't worry too much about the numbers or the pictures if the sound isn't completely off, and go off and do your little victory dance when you get what you're happy with.
Jeff
Fistula, I've put together about a minute and a half of demonstration from a short riff from a Mnemic song (Door 2.12, I think) that was miced with a modded SM57 and an Audix i5. Here's the clip, and I'll ramble mindlessly about what I did to change things here and there. Thanks again to Gavin, for being helpful/generous/silly enough to share space for such nonsense as this, and here we go:
http://www.upload.celtiaproductions.co.uk/uploads/JBroll_-_Phase_Samples.mp3
The riff itself is only 8 seconds long or so, but I wanted to use something short because of size/length limitations (we're talking eight different 'sections' from this clip, and it's done at 320kb/s, so we can't get too long) and I think this gets chugs nicely enough to get the point across.
Anyway, the setup was simple - as I said before, SM57 and i5, right on the grill pointed at the two speakers on my amp, and it all went through identical modded-to-fuck-and-back ART Tube MP preamps and into an Audiophile 2496. It may be a bit 'boomy' for your taste because these weren't meant to be doubled up, but whatever. I also like mids - a lot - and I make no apologies for that. The clip is a teeny bit sloppy, as I was half asleep and almost forgot the thing, but whatthefuckever - on with reality.
The first four 'sections' (the first fifty-six seconds) are with the two files, which are/were aligned as they were out of the amp (for the record, the 57 needed flipping to be put in phase and that was taken care of at the preamp), one panned hard left and the other panned hard right. The first clip is just as things were, with nothing funny but the original phase correction. The second clip is the same but with the phase on one of them flipped (it sounds to me like this one is getting ripped inside out, but whatever) in the DAW. The third clip has a slight shift in one of them, not enough to be perceived as a 'shift' but enough to have even more mid and even less fizz - it was something obscenely small, like a twentieth of a millisecond if I'm not reading the DAW wrong, but it certainly makes a noticeable difference. The fourth clip is the same as the third but with the phase on one of them flipped - again, I get that ripped inside out thing, but the sound is still somewhat identifiable. Technically, the sounds aren't 'directly' interacting, as they will be when everything is smacked into mono, but they can certainly 'feel' funny even when they aren't really 'touching' each other... for some reason the flipped segments make my head feel funny when I loop them for too long, and I start seeing the bears again, and the smilies start talking to me (fuck, I hate those smilies), but that may not all be the phase issue.
The last four sections, from fifty-six seconds on, is with both tracks dead center. The first track is the two tracks together, one of them may be shifted slightly but I can't fucking tell through this headache; again, you can tell that they weren't meant to be doubled by how thumpy they are, but no complaining until someone pays me to do this - everything in the 'meat' is thickened up, perhaps to a fault, and the fizz is much more fun than what usually happens from one mic. The second track is those same tracks flipped: here we get all of the fizzle that was yoinked out of the first two, because rather than adding all of the meat and subtracting the fizzies we've killed our meat and now we're adding the Alka-Seltzer by the bottle... this is the exact opposite of the Fredman technique (and a bloody good sign of success, if I do say so myself, when I flip the phase switch) and looking at it through a frequency analyzer I found hardly anything in the 'meat' range of the guitar. The third clip is like the first, but the shift I did is doubled - I think it shows how well we can have a 'sweet spot' where the planets align and the bunnies give the children their Easter eggs and that bitch didn't take my good knives when she left and our sounds are lined up just right, and then shift things a tiny bit too far and sound like we put a blanket over everything. The fourth is the same as the third, but again with the phase switched and with a tiny hint of lower stuff involved - I might analyze this later, if I have a *lot* of candy and nothing else to do, but right now I'm not worrying about it.
Basically what I've tried to show is that, when properly used, phase cancellation and addition can be fun... but that proper use is a bitch and a half. Before these clips were put down a while back, I had been spending day after day looking for the happy spots on the speakers, and when I sat down to tinker with it tonight I found that after all of that time I was still off by the tiniest fraction of an instant from the gigantic mess of a Fredman setup I was after - goes to fucking show that no matter what you do some asshat (like me!) can come along, make the tiniest and most minute change, and make everything perfect, and all it came down to was having the patience to make that micrometer difference or that twentieth-of-a-microsecond delay or something obnoxious like that. You can get about five metric shittons of pain dropped right on your ass because your mics were off by a distance you couldn't even see to save your life, or every god of every religion could spend all of eternity blowing you for the joy you've brought to the world with your phase anomalies, and it's not entirely unlikely that something less dramatic than the previous two will occur in most situations (although in my experience that's not the case), so sit there, play with details, breathe on your mics if something isn't just right (actually, you'll look like an ass blowing on all of your mics that long, so hire an intern for credit and make her do it), sit there and bang randomly on the keys, making 'Ook! Ook!' noises until effective (or don't), and DON'T
EVER
be afraid to just hit the fucking mute button on a track that's annoying you. You won't hurt it's feelings, in the long run you're just giving it more time to go off and get hammered with its buddies if you don't use it. Honestly, I'm never going to use those two tracks together in any project, and it's almost 100% guaranteed that I'm not going to use this current setup for anything more than study and novelty, and the world probably couldn't care less how I got things to sound this way or that in the long run. Phase summation can be fun, but sometimes it's utter nonsense - sometimes the best sounding option is to just mute this or delete that and be done with it, other times the godly tone of your dreams only comes from eighteen mics and a piezo pickup duct-taped to a trash can lid, so using what little there is to learn from a lunatic on a message board just put shit together in a way that makes sense in at least a half-conscious stupor, don't worry too much about the numbers or the pictures if the sound isn't completely off, and go off and do your little victory dance when you get what you're happy with.
Jeff
Mutant
I hate that supercow !
???
True
???
Why ?
In Adobe Audition you can zoom all the way to the single sample level both horizontally and vertically (for me its fairly easy to recognize which sample from wave A is which in wave B).
Thank you - I really like the sound too, but I prefer it when it's just one mic (you can hear that in the first clips just turning one speaker off); I'm sure it would be shit in a mix but I'll have to find out later. I'll probably get around to editing that mess, probably should have it down to no more than half a screen, but if it got the point across, great.
Jeff
Jeff
