SOUND ON SOUND

[Hugh Robjohns]

In the 80s a hit record wouldnt sound like a hit record does today.

That much is true... but the point I was making is that 'precision' time-alignment is not essential in creating a 'hit'. Moreover, just because the modern tools allow us to do something like that, it that doesn't mean we should, or that we need to!

I am trying to understand if having a kick and bass that are time aligned with a lot of low end in each, does that not cause phase cancellation?

Cancellation occurs if signals of a similar frequency are at diametrically opposite stages of their phase rotation, and not if they are more or less in the same phase/polarity.

Obviously precise time-alignment of bass and kick will give a rock-solid sense of timing, but it will also make it harder to distinguish between those instruments and it will almost certainly generate much higher initial transients resulting in reduced headroom and/or clipping and a need for additional dynamic control to reign them back in.

I guess it depends on the musical genre, but the way live musicians play against each other is a critical element of what makes great music, and it's often the subtle inconsistencies of timing and timing offsets that create the feel and groove of a track. Precision time alignment is likely to destroy that.

[armans]

The confusion of these terms comes from the notion that (considering a sinusoidal signal), a phase shift of 180 degrees looks exactly like a polarity inversion.... but it really isn't because a phase shift of 180 degrees inherently involves a time delay equivalent to half the wavelength whereas a polarity inversion is instantaneous and doesn't involve a (frequency-dependent) delay!

So, a polarity inversion is an inversion of the signal voltage whereas a phase shift involves a frequency-dependent delay.

Ok I think I understand. And of course engineers use both phase shifts or polarity inversions to get better correlation between their audio right? In the case of a snare drum where you have two mikes top and bottom, you usually flip the polarity of one but you could also use a phase shift to align the two tracks as well and essentially be achieving the same thing right?

Going back to me Op. I am not convinced yet :) Let's say you have two bass tracks: Di + Amp. It is very important that these two recordings are in phase right? So why should this be so different between the overall bass sound and the kick drum if they occupy the same frequency range?

[armans]

Time aligning the initial transients will give you a hit, but unless you're going to tune each kick to each bass note you're quickly going to lose that alignment

This is what I did in my experiment so yeah it is a lot of work if you had to do it for a whole track but...

so you're going to have a very spiky bottom end with the pitched notes lacking consistency. I think you're also going to have a very indistinct lower octave as it won't be clear what instrument is the source.

I disagree here, the bass line sounded much better when I aligned the transients to the kick drum and the waveforms both worked together. The bass just became part of the kick drum sound and sounded like one unit.... I wouldnt need to EQ it or anything to be honest, it sounded much better. I will post some clips tomorrow.

[armans]

I haven't looked at the video, but what the person means is that assuming the bass guitar and bass both play a note at exactly the same time then if they're recorded out of polarity it means that the note from the guitar if played solo would cause the speaker cone to move forwards while the kick note would cause the cone to move backwards. Combine the two at the same instant and they will, at least to a degree, cancel each other out.

What you're talking about is the two notes being played almost together, but with the same polarity. Because they're not exactly together there will be some phase interaction meaning the two notes produce some artefacts. Aligning them by eye would cancel that, but actually loses, IMO, some of the dynamic of the music.

This is from the video:

He (Jacquire King) points out that while the phase relationship between elements is essential on a multi-miked source, the polarity of their combined sound should be positive, so that the first transient pushes the speaker out, rather than pulling it back. If the polarity gets accidentally reversed, it will start by pulling instead of pushing the speaker.

If the bass and kick are hitting together and one is pushing, and the other is pulling, it will have a "nullifying effect," according to King. In a case like that, he says, you'll end up "chasing your tail" trying to figure out how to make things sound like they're sitting correctly together, and you'll have difficulty getting the low end to glue together. It's an issue that he pays a lot of attention to.

So my question, is: What does that all mean then? To have the speaker push and pull at the same time the only way to do it is to align the transients so that they push and pull together.

[Hugh Robjohns]

Again, there's a confusion about polarity and time alignment. The first is fundamentally important. The second is a matter of taste.

The quotation is quite right about matching polarity of similar signals occurring at similar times. It's the same thinking for matching the polarity of signals from top and bottom snare mics, or kick and overheads... Or bass and kick.

You simply flip the polarity and listen for which condition gives the most solid, loudest sound, indicating that the two signals are of the same polarity.

You can then, if you want, time-align the initial transients to improve the apparent timing and 'togetherness'... But tere are pros and cons to that as previously discussed.

[Mike Stranks]

So my question, is: What does that all mean then? To have the speaker push and pull at the same time the only way to do it is to align the transients so that they push and pull together.

It means you must be careful of polarity and ensure any inverse polarity is intentional. Polarity can become inverted by incorrectly wired plugs/cables or in a piece of kit. I've reviewed a preamp for SOS which inverts the polarity of the signal. Not a good idea to then feed an (unknown) inverted signal into a mix.

The speaker will not 'push and pull at the same time'. Assuming level is identical (and ignoring frequency) a mix of a polarity inverted signal and a non-inverted signal will result in no output - ie the speaker cone will not move.

We use polarity invert on snare because the top mic is recording the skin as it moves away from the mic, while the bottom mic is recording the skin as it moves towards it. We need the polarity invert to prevent the mics output being nulled in the mix.

[Hugh Robjohns]

... you usually flip the polarity of one but you could also use a phase shift to align the two tracks as well and essentially be achieving the same thing right?

No. You use a polarity inversion because the signals are already time aligned (assuming sensible mic placements) but one skin is moving in with the stick impact and the other is moving out. A phase shift inherently involves a frequency-dependent time delay, which is not appropriate in that situation.

Let's say you have two bass tracks: Di + Amp. It is very important that these two recordings are in phase right? So why should this be so different between the overall bass sound and the kick drum if they occupy the same frequency range?

The bass and DI are essentially the same signal -- certainly the same fundamental frequencies, and you're trying to compensate for the acoustic time delay of the mic in front of the speaker, relative to the faster arrival of the direct electrical signal. So you effectively phase shift the electrical signal to time shift it back to alignment with the mic signal.

In contrast, Kick and bass are different fundamental frequencies and different signals entirely. So you can align the transients, but there is no static phase relationship between he two signals because they are different frequencies...

[armans]

I haven't looked at the video, but what the person means is that assuming the bass guitar and bass both play a note at exactly the same time then if they're recorded out of polarity it means that the note from the guitar if played solo would cause the speaker cone to move forwards while the kick note would cause the cone to move backwards. Combine the two at the same instant and they will, at least to a degree, cancel each other out.

Ok so the bass and kick drum just need to have the same polarity? Even if they are not time aligned? I will be honest, I have never paid attention to the polarity of my recordings in this way. Should I?

[blinddrew]

I disagree here, the bass line sounded much better when I aligned the transients to the kick drum and the waveforms both worked together. The bass just became part of the kick drum sound and sounded like one unit.... I wouldnt need to EQ it or anything to be honest, it sounded much better. I will post some clips tomorrow.

OK, I think we're aiming for different things then. Most people, well, most people here maybe are trying to make them both work together but not be part of the same thing.

[armans]

... you usually flip the polarity of one but you could also use a phase shift to align the two tracks as well and essentially be achieving the same thing right?

No. You use a polarity inversion because the signals are already time aligned (assuming sensible mic placements) but one skin is moving in with the stick impact and the other is moving out. A phase shift inherently involves a frequency-dependent time delay, which is not appropriate in that situation

Ok. Could you be so kind as to give me some examples of when a phase shift would be more appropriate than flipping the polarity? I record acoustic guitar with a spaced pair and I also record cajon with 2 mikes. Would flipping polarity or phase shifting the tracks be more appropriate in these cases?

[Hugh Robjohns]

Ok so the bass and kick drum just need to have the same polarity? Even if they are not time aligned? I will be honest, I have never paid attention to the polarity of my recordings in this way. Should I?

If you have two similar signals occuring at similar times, it is generally essential that they have the same polarity and it's always worth checking by flipping the polarity of one or other. If they are in opposite polarities the sound will be weak, thin, and quiet. When they are in similar polarities the sound will be stronger, bassier and louder -- and that's what you want.

[blinddrew]

... you usually flip the polarity of one but you could also use a phase shift to align the two tracks as well and essentially be achieving the same thing right?

No. You use a polarity inversion because the signals are already time aligned (assuming sensible mic placements) but one skin is moving in with the stick impact and the other is moving out. A phase shift inherently involves a frequency-dependent time delay, which is not appropriate in that situation

Ok. Could you be so kind as to give me some examples of when a phase shift would be more appropriate than flipping the polarity? I record acoustic guitar with a spaced pair and I also record cajon with 2 mikes. Would flipping polarity or phase shifting the tracks be more appropriate in these cases?

Ok, so take your acoustic guitar spaced pair, if it sounded a little weak to you and on zooming in you saw that the two wave forms weren't time aligned, you'd do so and bring them into phase.
Now consider the cajon, perhaps you use a front mic and a rear mic? On these you'd probably* flip the polarity because a beat on the front will push the two diaphragms in different directions.

* 'Probably' because it doesn't always sound better, especially when you fit it in the rest of the mix.

[CS70]

Ok so the bass and kick drum just need to have the same polarity? Even if they are not time aligned? I will be honest, I have never paid attention to the polarity of my recordings in this way. Should I?

No, no particularly. :) You can try and see what happens, but that's it, outside the specific case below.

It's easy to understand if you think of two sine waves of same periodicity, plotted against X and Y axes, where Y is amplitude and X is time. Look a the points X where the sine waves start (i.e. Y=0) and where the crest and valley are.

The waves are relative phase if they start at the same X: if summed, the resulting sine wave will have higher amplitude (i.e. be "louder" if the wave is transduced by a speaker).
If the two X are as far as they can get (i.e. the peak point of the first wave is where the second one has its valley), they are in opposite phase (or "out of phase") - the sum will have shorter amplitude (i.e. be "quieter") and even cancel out completely if the waves have the same amplitude.

For all the other relative starting point position, they will be a bit in phase and a bit out of phase - depending on how you prefer to look at it.. if the starting points are near to each other, they're "mostly in phase", in they are near the "opposite" position, they're "mostly out of phase".

To change the relative phase, you simply delay one of the two waves a little bit - nudging it in time.

Obviously for more complex waves, and even sine waves with just different periodicity, they will be almost always in this gray area and things will shift all the time from "mostly in phase" to "mostly out of phase" very fast - in other words, a normal interference pattern which results in specific timbres.

"Polarity" is just where's the north and where's the south. The Ø button simply flips north with south, so if you push it on a channel, you will flip the corresponding wave, changing the timbre of the summation.

Of course in rare cases where the sound waves are very similar (say bottom and top mic for a snare, the snare component in the overheads and the snare mics etc) and they happen to be mostly out of phase (it will never be exactly so of course unless you duplicate a recorded sound), that button is a quick way of nudging one of the two to get them mostly in phase. It's just a shortcut to get in the ballpark, instead of inserting a delay (which, I guess, used to be a quite complex thing with desks and physical inserts).

Otherwise, you're just gonna change the interference pattern - aka the timbre.

It can be nicer, it can be worse - it all depends on your taste. It's worth trying to see how it changes, but it's nothing you have to do or otherwise critical. If you have timbres you like, you're all set.

[CS70]

I disagree here, the bass line sounded much better when I aligned the transients to the kick drum and the waveforms both worked together. The bass just became part of the kick drum sound and sounded like one unit.... I wouldnt need to EQ it or anything to be honest, it sounded much better. I will post some clips tomorrow.

Nothing surprising - you have an infinite palette of summed waves when you keep one fixed and nudge the other, since there's an infinite (real) number of time points between a crest and a valley. Some you'll like better than others.

The challenge with phase is only when you have the same wave occurring in three or more tracks, containing a mix of components in phase and not. Then you're screwed, because if you nudge the second channel to get some components with the first, you will get other components out of phase in the third.

That's why it's critical to check phase at recording stage when more than two mics capture the same sound (typical examples are overheads and snare and of course any sufficiently strong bleed between multiple mics when recording a band).

[armans]

If you have two similar signals occuring at similar times, it is generally essential that they have the same polarity and it's always worth checking by flipping the polarity of one or other.

Ok thank you. What exactly do you mean by "similar times" and how similar is "similar"? Lets say I record my bass without editing, the notes will be very inconsistent in relation to the kick drum so what difference does it make if the polarity is one way or another since every note relative to the bass drum will be different?

I understand that when recording a snare drum you would hear the difference immediately when flipping the polarity because it is the same thing recorded with two microphones but in the example above, Mr King is not talking about the same polarity on one recording done with two mikes, he is talking about having the same polarity on two separate sounds/instruments. So, flipping the polarity will have a varied effect because in a normal un-edited recording of a bass player playing a bass line, each bass note would be in a slightly different position relative to the kick drum.