Talk:Balanced audio/Archive 1

Archive 1

Article needs a topic sentence

This article begins

Balanced audio connections are extremely important in sound recording and production because...

What's needed is a definition. I don't know enough about the subject to make one (that's why I came here!). Jer ome 19:49, 28 March 2007 (UTC)

Hi, I made modifications and added a Basic Setup chapter just for that reason, but a user that edits this page a lot called Binksternet seemed not to like my changes. Maybe I'll add it back later. The article is not very well structured now. Robijn (talk) 00:08, 4 December 2012 (UTC)

Incorrect

The article states that differential signalling is responsible for the noise rejection, but this is a common misconception. The noise will couple into both lines regardless of whether there's a differential signal on the lines or not. The feature that's important to common mode rejection is the impedance of the lines to ground.

This is the basis of "impedance balancing" or "quasi-balanced" connections (which we really need to cover in this article), in which the tip and ring are both terminated with the same impedance, but the ring goes to ground instead of an inverted source.

See Talk:Differential_signaling#Accuracy and Talk:Balanced_line#Accuracy for parallel discussion. — Omegatron 18:13, 22 May 2007 (UTC)

I'm a little rusty on the differences between near-field inductive and capacitive coupling, and far-field electromagnetic coupling:

Twisted pairs are primarily for rejecting inductive interference, right? You are minimizing the loop area so that a changing magnetic field induces less, and alternating the loop direction so that a changing magnetic field that induces equally through two adjacent loops causes equal and opposite currents, which cancel out. Does this technique have any effect on other types of interference?

Twisted pairs reduce inductive interference, but also cut down on capacitive crosstalk. --Heron 17:48, 26 May 2007 (UTC)

That's another reason for differential signalling, though not as applicable to audio? Audio is relatively low frequency and signals almost always travel alone and shielded, but in other cables, twisted pairs are routed next to each other. Differential signalling minimizes the EM from the pairs so that they don't couple into each other or into other devices. — Omegatron 15:56, 31 May 2007 (UTC)

And impedance balancing is more for capacitive interference? A changing electric field near a cable would create a similar effect on both lines, it would seem. Does impedance balancing also affect the other types of interference?

The mechanism of crosstalk depends on frequency. At LF it is mostly capacitive, and at HF the inductive effect takes over. --Heron 17:48, 26 May 2007 (UTC)

Shielding is more for electromagnetic interference, right? From RF and such? It's a type of Faraday cage, and so protects from electrostatic fields but not magnetostatic fields, and has varying effect on higher frequencies? See Talk:Faraday_cage#Mesh_size.3B_explanation.3B_magnetic_fields for the same discussion in a more general context. — Omegatron 15:12, 24 May 2007 (UTC)

Shielding is also used to reduce crosstalk (both capacitive and inductive). --Heron 17:48, 26 May 2007 (UTC)

Hmmm... Crosstalk between the two wires in a balanced audio connection would just mean a decrease in level? Though it might vary with frequency.

An ungrounded shield or chassis actually increases crosstalk between things that are inside it, though, because they couple to each other capacitively through the shield. — Omegatron 16:20, 31 May 2007 (UTC)

}} — Omegatron 21:28, 24 May 2007 (UTC)

And the ability to interface with a shorted pin 2 is actually somewhat important, because of the pin 2/3 = hot problem. — Omegatron 04:29, 25 May 2007 (UTC)

The article seems OK now in the area of differential versus balanced. (I'm ignoring the disputed section on amplifier internals, since I don't know much about that.) I found some notes from a seminar on EMC, and added some comments to your points above. Unfortunately it seems that you can't easily separate E, M and EM when it comes to protecting against them. --Heron 17:48, 26 May 2007 (UTC)

I think the text under Differential signalling is incorrect. Indeed, quasi-balanced lines reject noise just as well if the receiver has a differential input. However, it is not fully balanced since no current is flowing through the negative line, resulting in no electromagnetic field. Now the electromagnetic field of the positive line will not be cancelled. After a long distance the negative line will start to carry signal as well. If the negative line is not terminated with the correct resistances at both ends, this will result in signal degradation. I propose to remove this chapter (since its contents is already covered earlier) and introduce a chapter on quasi-balanced lines. Also the term impedance-balanced at the very start of the article is incorrect. It is impedance-matched balanced line. Sometimes impedance-balanced is used as a synonym for quasi-balanced. Robijn (talk) 11:20, 8 January 2012 (UTC)

Needs clarification about phase

Please forgive the newbie question, but it seems to me that "The two wires form a circuit carrying the audio signal; one wire is in phase with respect to the source signal, the other wire is 180° out of phase." is misleading because a) 180° out of phase is not the same as polarity reversal unless you are dealing with perfect sine waves and b) as stated later in the "Differential signalling" section later in the article and also in Balanced_line#Balanced_and_differential that signal symmetry is not required in order to balance a line. Can someone clarify this? AdamSpiers (talk) 12:51, 30 May 2013 (UTC)

I don't understand how you think that perfect sine waves differ from other audio signals in the way they respond to polarity reversal, or to phase changes including 180° shift. I think that both respond the exact same way.

It is true that signal symmetry is not required for all forms of balanced audio. However, signal symmetry is the goal of fully differential balanced circuit design, the kind of circuit with active components driving each wire. Binksternet (talk) 13:01, 30 May 2013 (UTC)

Please read http://www.soundonsound.com/sos/apr08/articles/phasedemystified.htm - 180° phase shift means the wave has been shifted right or left by half a wavelength, which is in general completely different to polarity reversal (i.e. reflection about the X axis)! For example imagine a wave which is a perfect sine wave above the X axis, but sawtooth below it. If you phase shift that 180°, it will not cancel itself out! This additionally means that AFAICS "inverting" is ambiguous and can either mean signal inversion or a 180° shift.

Your second comment confirms my point, which is that the sentence I critiqued misleadingly implies that all forms of balanced audio use signal symmetry, whereas in fact it is only used in a subset of cases. That sentence appears above the section on Differential signalling, and as such conveys the impression that it is generally applicable, whereas in fact it is only applicable to differential mode. AdamSpiers (talk) 13:24, 30 May 2013 (UTC)

The Sound On Sound article does not discuss a wave signal which is a pure sine at positive voltages but a triangle wave at negative voltages.

Phase is relative to frequency. If a signal is shifted 180° at all frequencies then it is indistinguishable from polarity reversal. This signal will behave the same whether it is pure sine or more complex.

Let me know if my tweak to the article text addresses your concern. Binksternet (talk) 16:09, 30 May 2013 (UTC)

Yes, his sine/sawtooth example was ill conceived. However, his original objection was sound. The point of the article is that it is nonsensical to describe a signal as "shifted 180 degrees at all frequencies". To "shift" a signal means to delay it with respect to the reference signal, and as the article explains, any fixed delay time results in different cancellation at different frequencies. Polarity reversal is a different animal, and it's a binary +/- thing. It is not measured in degrees. The takeaway here is that sound engineers may be used to describing the situation with a certain vernacular, but Wikipedia should take note if that differs from the actual physics of sound or electromagnetism. 137.254.4.7 (talk) 15:19, 24 September 2014 (UTC)

Your understanding is flawed. If you correctly apply a "180 degree phase shift" a complex waveform you get a waveform that looks like the original, only inverted. Yes, if you simply a delay equal to half the cycle period of the fundamental, then you are correct - it does not appear "inverted". But that is not correctly applying a 180 degree phase shift to the entire signal, just to one component! If you do a Fourier analysis on it and then construct the waveform with each component delayed by half of its period, you will get an inverted version of the original. This is what Binksternet meant by "at all frequencies". Jeh (talk) 03:33, 1 March 2015 (UTC)

I would agree that "inverted" is probably more intuitively understood by the lay reader, but "180 degree phase shift" is not incorrect. Jeh (talk) 09:25, 1 March 2015 (UTC)

"Eight-channel analog balanced audio connectors like ADAT use DB25 connectors...

"Eight-channel analog balanced audio connectors like ADAT use DB25 connectors, which can also carry up to 16 digital channels."

This sentence makes no sense.

ADAT connectors are toslink fiber optic using a digital signal.

The TASCAM analog DB25 standard is what they are really talking about. I'm going to edit the sentence to make sense. Change it back to the old wrong sentence if you want.

• edit

I just deleted the sentence. It did not contribute any information.

Popularity relative to digital?

Could someone please add some info about the relative popularity of this system relative to (i) unbalanced audio and (ii) digital audio? I'm particularly interested to know how much digital audio has displaced analog in recording studios. Analog audio might be going the way of vinyl, in which case we need to say so. --Heron (talk) 10:53, 13 October 2016 (UTC)

In studio use, balanced connections are still ubiquitous. USB microphones won’t replace XLR anytime soon. If we can get some good sources, I’m sure it will be a nice addition to the article. Gutten på Hemsen (talk) 00:11, 12 June 2021 (UTC)