1. Register NOW and become part of this fantastic knowledge base forum! This message will go away once you have registered.

Understanding phase........again...

Discussion in 'Recording' started by vinniesrs, Jul 3, 2003.

  1. vinniesrs

    vinniesrs Active Member

    I just wanted to give everyone an update on the results of my little experiment. I found that by using varying distances you could drastically change the sound of anything safely, in a mono situation. Or, if you want to mic in stereo you can use a pair of mics for each side(4 in total l
    ). This is basically a way of pre-equing a track. A specific interruption of phase at a certain frequency to diminish the amplitude on the recorded track. Try it. It works, and once learned is easy and practical. Just use your ears to find what you dont like, then toss on some cans and use a little knowledge as a guide.
    It also made for one hell of a demo to my assistant engineer. :D


    (Dead Link Removed)

    [ July 03, 2003, 11:05 AM: Message edited by: Kurt Foster ]
     
  2. Ethan Winer

    Ethan Winer Active Member

    Steve,

    > by using varying distances you could drastically change the sound of anything safely <

    "Safely" is a bit optimistic. Whenever one source arrives through different paths that are then combined, the result is comb filtering. Sometimes that filtering sounds good, but often it does not. I think of it as unintended EQ that sometimes gives a good result.

    --Ethan
     
  3. vinniesrs

    vinniesrs Active Member

    Ethan, by safely I meant only that if you recorded in mono then the sound you achieved through this method would not create problems in the mix. The purpose of my experiment, was to see if I could predict the frquency at which phase canellation would occur. I realize that this would not be one specific frequecy, but rather a range with the center at 180*.

    I tried the same effect in stereo by combining 2 mics on each side and applying the same math. It is much more complex, and somewhat impractical, but none the less it was a fun exercise.

    As I see it, since an eq operates by interrupting phase at a certain frequency this should work too. (sounds like it did.) Also since q is a ratio, theoretically then you should be able to vary the q of such an effect by changing the distances of the two mics. Both in reationship to the source, and to each other.

    Ethan, if you could please explain the errors inherent with my exercise. I realize I am not a genious here, I am just trying to find interesting ways to get sounds, and to increase my understanding of the powers that be.
    :c:
     
  4. Ethan Winer

    Ethan Winer Active Member

    Steve,

    > if you recorded in mono <

    Just to be clear, you mean using two mikes that are mixed to mono before the recorder's input, right?

    > then the sound you achieved through this method would not create problems in the mix. <

    It might not create surprises when you do a mono check, but phase cancellation is what it is, and it can definitely cause an affected sound quality.

    > > since an eq operates by interrupting phase at a certain frequency this should work too. <

    Yes, equalizers work by shifting phase, but they do it in a more controlled manner than what you describe. And the phase shift of a given single EQ stage changes only that one band. When you use time delay instead of plain phase shift you create a comb filter, and that affects the start frequency as well as all higher frequencies.

    > you should be able to vary the q of such an effect by changing the distances of the two mics. <

    No, the Q of a typical filter is changed by applying feedback within the circuit.

    I'm not arguing that mixing two or more mikes to mono can't give good results. Just that there are more direct ways to EQ a source.

    --Ethan
     
  5. Kurt Foster

    Kurt Foster Distinguished Member

    Steve,
    I have to side with Ethan on this.. IMO multiple mics smear the sound and deliver less impact or punch, because of the pahasing issues. When ever possible I prefer to use just one mic, choosing mic and mic pre to color the tonal content. Also I will attempt to alter the sound at the source (tune the drum , change the settings on the amp) before resorting to using eq.
     
  6. vinniesrs

    vinniesrs Active Member

    Well now. THis was the info I was looking for. I didn't consider time delay. You got me.
    Quote by ethan:
    "Just to be clear, you mean using two mikes that are mixed to mono before the recorder's input, right?"

    Yes.
    Quote by ethan:
    "It might not create surprises when you do a mono check, but phase cancellation is what it is, and it can definitely cause an affected sound quality."

    Okay, considering the time delay, I can see your point. :p
    Quote by ethan:
    "No, the Q of a typical filter is changed by applying feedback within the circuit."


    The thought here was that the distance between two mics would determine varying degrees of phase at certain frequencies, thus determining the amount of cancellation. Again, I have overlooked the time factor.
    Quote by ethan:
    "I'm not arguing that mixing two or more mikes to mono can't give good results. Just that there are more direct ways to EQ a source."

    I will be the first to admit that my knowledge of electronic design and function is quite basic. I do understand basically how an eq works, but the main motivating thought here is the difference between a good eq, and a bad one.

    In response to Kurts comment about dealing with a sound at the source, that is always my first area of attention, and where I spend the bulk of my time. I play, and understand the setup and tuning of drums, bass, guitar, most wind instruments, and miscellaneous percussion instruments.

    I have found my best results to be yeilded from single mic setups, and it is certainly faster than playing with these elements.

    I guess the whole purpose of this for me was to determine if the results yeilded would be better than that of an eq. I left out a few important factors, and I'm grateful that you pointed them out for me. This still doesn't change the fact that I was able to acheive the desired results I set out for. Also the amount of time is directly related to the distance. 1.13ft per 1ms. At that rate you should be able to have differences in distance from mic A to mic B of a little over 4 ft before most people would notice. Every 10.5 inches would be 360* of phase at 1khz.

    Now the question! Why would you guys use mulit-mic setups after having said all of these things? The only other application I have used this type of setup for would be to a or acd a few mics to save time, and get the right mic placed for the job. I currently have only a couple of pre's to select from, which makes that part easy.

    Also I have tried using two mics on a kick, or snare with good results, but not stellar by comparison to a single mic setup. This was another motivator for me, to maybe find the key to successful, and consistent two mic setups.

    To sum It all up I have found a great way to illustrated the effects of phaseand time(Hmm.)to both my apprentice and to myself. I will say that the combination of the practical and theoretical here has been great, and also the input of the other pro's in this forum. :cool:

    Thanks again for humoring a whim.
     
  7. Kurt Foster

    Kurt Foster Distinguished Member

    Steve,
    Usually less is more.. but nothing is set in stone. I will often use a mic on the top and bottom of a snare. But I check to make sure that they are in phase. I also usually gate the bottom mic , keying it off of the top mic. This helps keep the kick out of it and avoids the out of phase issues with that.

    I also have used multi mic set ups on some guitar amps. Kenny "Blue" Ray's rig is a perfect example.. He uses open backed amps most of the time so I will throw up a couple of 421's on the front, left-right and the put a U87 sideways behind the amp in figure 8 pattern. After checking phase, this not only yields a stereo image but also provides a great deal of a sense of "depth" to the sound. Try it sometime. It's killer!

    I regards to eq, (I can tell you are looking for alternatives) one trick I like is the use of a compressor with an eq through the side chain. By boosting certain frequencies, you can achieve a cut in that same region.. eq without the phase shift and time smear usually associated with eq circuits....
     
  8. realdynamix

    realdynamix Well-Known Member

    :confused: Ethan, all this time I thought phase shift was the undesirable result of the filter components rather than the actual method used. Some analog EQ makers boast of having little or no phase shift. I am confused, do you mean comb filters?

    --Rick
     
  9. Ethan Winer

    Ethan Winer Active Member

    Steve,

    > the main motivating thought here is the difference between a good eq, and a bad one. <

    I'm sure I'll get flamed for this, but as far as I'm concerned any decent EQ you care to use is fine, hardware or software. There are only a few things that distinguish any audio component:

    1. Frequency response
    2. Signal to noise ratio
    3. Distortion

    [I hope I didn't forget anything obvious!]

    There are lots of types of distortion, and lumping them together is only for brevity. Same for noise, because modulation noise on analog tape comes and goes with the signal and so is very different from "normal" noise. And the same for frequency response, because severe anomolies like ringing are more damaging than an overall deviation from flat. But it's important to keep the basics in mind, because these really are the only ways in which any audio gear differs.

    > I guess the whole purpose of this for me was to determine if the results yeilded would be better than that of an eq. <

    I don't see how it could be. I never understood why people will pay $1000 more for a microphone that has a pleasing presense boost. You can do exactly the same thing with EQ! It's true that frequencies can be lost or severely attenuated due to poor mike placement, and in those cases it's difficult or impossible to recover them later with EQ. But otherwise, EQ from mike choice is identical to EQ in a box. And EQ from mike placement is the same as EQ in a box that also affects the proportion of reflections captured. So mike placement also changes the ambience too.

    > This still doesn't change the fact that I was able to acheive the desired results I set out for. <

    Absolutely! There's no one right way to do anything in audio. As they say, "If it sounds good, it is good."

    > Every 10.5 inches would be 360* of phase at 1khz. <

    Yeah, but with a different amount of shift at all other frequencies. So unless you're recording a theremin sustaining a single note that never changes, you can't count on a given distance to equal some set amount of phase shift.

    > Why would you guys use mulit-mic setups <

    Only two reasons I ever use two mikes: For stereo, and to capture ambience. Even if you plan to have a guitar amp end up as mono, you can get a big sound by mixing a close-up mike with another mike that's farther away. Personally, I always use two separate tracks, so I can pan the close and far mikes somewhat differently if I want.

    --Ethan
     
  10. Ethan Winer

    Ethan Winer Active Member

    Rick,

    > all this time I thought phase shift was the undesirable result of the filter components <

    Yes, and that's a huge misconception. As proof that phase shift is absolutely benign, think about the finest recording you have ever heard. If it used any EQ, that EQ applied phase shift.

    > Some analog EQ makers boast of having little or no phase shift. <

    I don't think that's possible. Some digital EQs make that claim, though I've never heard one.

    > I am confused, do you mean comb filters? <

    All EQ - except the digital ones that use special trickery - work by combining a direct and phase-shifted version of the same signal. The whole notion of "time smear" caused by phase shift is just wrong. To me, "smear" is too much ambience, or disjointed reflections that confuse the stereo image. Sometimes people report hearing what they assume are phase changes when they boost an EQ. But I'm sure what they're really hearing is slight comb filtering that was already present, and is just made more obvious by the very act of applying EQ!

    I'm not sure what you mean by comb filters. Those are created with either a large amount of phase shift [phaser effects] or with straight time delay [flanger effects]. The only real difference is effects that use phase shift have fewer teeth in the comb, so to speak.

    --Ethan
     
  11. Kurt Foster

    Kurt Foster Distinguished Member

    Not a flame, but I submit that the kind of EQ lift that one would get from using a C12 has a different character than that of an EQ, partly because of the lack of phase anomalies. And different eq's will exhibit different qualities even with exactly the same settings. A 1k cut on a Neve/Amek eq with a given q factor is going to sound different than the same settings on a Speck ASC.. not that one is better than the other, the will just have a different color or flavor. Mics, pres, eq’s and compressors all have different sounds.

    I remember about ten years ago, Julian Hirsh was embroiled deeply in a controversy where he stated that there would be no difference in the sound of different power amps. Well we all know that there is a big difference of a Peavey CS400 and a Bryston amp. The same thing applies to mics, pres comps etc. All the more reason to have a bunch of different types on hand. And not just high end stuff either. Perhaps at some point, that Mackie pre is going to be the best thing for a part. It's all part of the pallet of colors we have available to us in the world of audio and just when you think something is worthless, it will turn out to be just the thing for a certain application..
     
  12. Ethan Winer

    Ethan Winer Active Member

    Kurt,

    > Not a flame, but I submit that the kind of EQ lift that one would get from using a C12 has a different character than that of an EQ, partly because of the lack of phase anomalies. <

    I can handle very hot flames, as long as things remain civil, so never worry about that.

    And I submit that the only thing that defines "character" is the amount of EQ, its frequency, and its Q. I mean, what else is there?

    > different eq's will exhibit different qualities even with exactly the same settings ... A 1k cut on a Neve/Amek eq with a given q factor is going to sound different than the same settings on a Speck ASC. <

    Yes, this is the crux of it. I can't see how that's possible, unless maybe you're close to overloading an internal stage or something like that. Again, what would vary from one EQ to another?

    > Mics, pres, eq’s and compressors all have different sounds. <

    To be sure, but whatever is different is definable and can be easily understood. Not much about audio basics has changed in a many years.

    > Perhaps at some point, that Mackie pre is going to be the best thing for a part. <

    I love my Mackie 1202, and it's not even one of the newer VLZ models. I enjoyed very much the recent report I saw the other day about the blind test where all those famous engineers chose the Mackie as best. Never underestimate the truth-revealing powers of a blind test!

    --Ethan
     
  13. Kurt Foster

    Kurt Foster Distinguished Member

    Actually the reason I sited those two EQs is I happen to have a pair of both of them here. I can very easily record a comparison and post it. I assert that there will be a difference. I have heard this difference termed as the "transfer function..
     
  14. Guest

    Mic coloration vs. EQ:

    (Gee, Ethan, I still have the bruises from the last time we tangled on this one...)

    I suppose in theory you COULD painstakingly create an EQ that duplicates a microphone's frequency response, but that would only work for a source at a fixed volume, distance, and angle from the microphone. The mic's presence peaks, etc. dynamically change as sound hits it from different angles and distances (pressure gradients vs. velocity gradients as Stephen explained). So, for instance, the early reflections of a small recording space would be EQ'd by a mic quite a bit differently than the direct signal.If you could come up with an EQ that could dynamically model all those factors, that would be one pretty sophisticated (and expensive) EQ . Vocalists would develop whole new sets of skills (from working a mic to working an EQ...)

    I would guess that's a lot of the reason why mic modellers are only marginally useful, if at all, in replacing the real thing.

    But then again, what the hell do i know...
     
  15. Ferd Berfel

    Ferd Berfel Active Member

    Assuming a "real-world" environment:

    1) PHASE is the measure of specific point on a waveform relative to a specified reference point (i.e., a time or a point in space).

    2) PHASE SHIFT is a term usually used to describe a change in a waveform's phase (see #1...a relative measure).

    3) PHASE SHIFTING will always occur when a signal is delayed in time.

    4) In an echoic acoustic space (i.e., *NOT* an anechoic space) signal delay is always created whenever there can be more than one path between the sound source and the sound receiver (e.g., ear or microphone).

    5) A COMB FILTER is created whenever a signal is combined with a delayed version of itself.

    6) A COMB FILTER is characterized by regular and periodically alternating peaks and valleys at various frequencies in a system's frequency response.

    7) In an echoic acoustic space, a large number of comb filters are naturally created whenever a source's original acoustic wave can combine with a reflected version of that wave. This filtering occurs whether one is using zero, one or an infinite number of microphones.

    8) A frequency-selective audio filter is a device that can increase or decrease the size of one frequency (or group of frequencies) relative to others.

    9) In audio systems, time delay (& hence, phase shifting...see #3 above) is a NECESSARY component to realizing a filter.

    10) All audio filters, whether analog or digital, achieve their aim through specific application of phase-shifting/time-delay.

    Regards,
    Ferd
     
  16. Guest

    One of Ferd's points brings up a question I always meant to ask:

    Regardless of what they call it, can there really be a "phase-linear" EQ, and if so, how does one get filtering without phase?
     
  17. Ferd Berfel

    Ferd Berfel Active Member

    Yes, there really can be a "phase-linear" EQ (actually, "linear-phase"). Many of the "standard" filters used for EQ sections (etc.) have a phase response that varies with the inverse of a term (or squared term). You could call this "hyperbolic-phase". Linear-phase filters, however, shift the signal's phase(relative to the input signal) changes in proportion signal frequency as it passes through the filter.

    Linear-phase response filters are implemented with what's called a "FIR" filter ("FIR" stands for Finite Impulse Response). One of the characteristics of this filter is the proportional phase change (with frequency) mentioned above, due to the fact that the FIR implementation utilizes a delay line as its foundational mechanism.

    This does NOT mean that phase-shift does not occur in a linear-phase filter. In all causal filters, we need phase-shift to do the job!

    Regards,
    Ferd
     
  18. Duardo

    Duardo Guest

    The Nighpro EQ's (which are analog) do...I'm not sure how they do it, but they definitely sound different from any other analog EQ I've ever heard...

    I'm not sure exactly how it's possible either, but it certainly is...most of the analog EQ's I've heard sound very different from one to the next, even with relatively minor boost or cut. Are you saying that all EQ's should sound the same with the same settings?

    [quoter]I enjoyed very much the recent report I saw the other day about the blind test where all those famous engineers chose the Mackie as best. [/quote]

    I didn't see that report...did they choose the Mackie EQ's or preamps? Preamps wouldn't necessarily suprise me, but the EQ would...

    -Duardo
     
  19. realdynamix

    realdynamix Well-Known Member

    :) :D

    The product mentioned does have some, though small, amount of phase shift.

    Read the specs

    --Rick
     
  20. Ethan Winer

    Ethan Winer Active Member

    Kurt,

    > I can very easily record a comparison and post it <

    Sure, that would be great. If you post a short mono Wave file - not MP3! - of both equalizers processing pink noise that would be the most useful. Then I could analyze both and see what's different.

    I have no doubt that two EQs set the same could sound different. But I don't see how that difference could be due to anything other than variations in the panel labeling. That is, you set both to have a Q of 1.0 at 500 Hz but one is actually giving you a different Q and a different frequency. It doesn't take much deviation to make an audible change.

    > I have heard this difference termed as the "transfer function" <

    I have no idea what transfer function means in that context. Transfer function is a general term that simply means the difference between what you put into something versus what comes out.

    --Ethan
     

Share This Page