Vinyl and Absolute Polarity: Q&A, Part Two

Vinyl and Absolute Polarity: Q&A, Part Two

Written by J.I. Agnew

In Part One of this dialog between J.I. Agnew and reader/engineer Bob Lehman (Issue 112), they delved into the technical aspects of vinyl record manufacturing and absolute polarity. The dialog continues here. It’s longer than our usual article length, but the complexity of the subject requires going into detail – including addressing Einstein’s theory of relativity!

Bob Lehman: So, are these transducers [microphones, tape recorder heads, phono cartridges and disk cutter heads] in some way responding to the derivative of the air pressure, magnetic field, or physical displacement phenomena?

J. I. Agnew: For tape heads, e = N(dφ/dt). For phono cartridges, e = ds/dt, but internally this just translates to e = N(dφ/dt), since any displacement of the stylus causes the magnet or coil to move relative to each other, thereby changing the flux linking the coil turns, which generates an EMF (electromotive force or signal voltage) whose magnitude depends on the number of turns (N) of the coil and the flux swing (this applies the same way to moving magnet and moving coil cartridges, with slight variations for moving iron cartridges but it can be considered equivalent).

BL: If true, would it really matter in any practical way, since, with the possible exception of the case of  [a DC signal], under the principle of the Fourier theorem/series, any complex waveform can be expressed as the sum of the integral pure sine wave multiples of its fundamental frequency?

JIA: Both in mathematical terms (a Fourier transform) and in practical measurable/audible terms, although the frequency range of interest may be adequately covered (down to 10 Hz or so for vinyl records), the fact that the system response does not extend down to DC will cause issues with the phase response. In a mathematical Fourier transform, a complex wave is “transformed” into multiple sine waves, one for each frequency component of the original complex wave, each at a certain amplitude and with strictly defined phase relationships between them. If we attempt to reconstruct the original complex wave from these sine waves, at the correct amplitudes, but with altered phase relationships, the result would be a different waveform!

Errors at subsonic frequencies, such as 10 Hz, can cause phase anomalies well into the audible range, greatly reducing the system resolution and the realism in the low frequency range. It is worth remembering that the lowest C note produced by a 32-foot organ pipe has a fundamental frequency of 16.351 Hz and the low A on a piano has a fundamental frequency of 27.5 Hz.

BL: If the [above] assumptions are true, is that taken into account in any way in the design of the related mechanical, electrical and magnetic equipment? [Conversely], a DC signal applied to a DC-coupled electromagnetic loudspeaker can and will push the diaphragm in one direction and it will stay there, and if the listening room is 100% air-tight, a constant air pressure increase or decrease will be produced in it. Ah, yes, finally, we get to the connection of these questions to absolute polarity! I assume that a totally DC-coupled electrostatic loudspeaker would do the same thing, but I don’t believe that any exist, as most are transformer-coupled to the driving amplifier. Electrostatic headphones do of course exist, and are often directly driven by solid-state or tube drivers without coupling transformers, so I’m guessing that they could produce a constant diaphragm displacement if driven by a DC signal, if they are still otherwise fully direct-coupled.

JIA.: Most transducers and electronics used in sound recording and reproduction are designed to have cut-off frequencies low enough as to be effectively insignificant. A 0.5 Hz cut-off frequency on a microphone preamplifier is low enough to not affect frequency or phase in the audible range. The 10 Hz mass/compliance resonance of the tonearm and cartridge is not as low as would be ideal for perfect phase response, but is essential to filter out unwanted sounds and is well manageable with good design, to the extent of being almost insignificant on its own. It does become significant, however, as soon as someone carelessly designs a second 10 Hz cut-off into an amplifier or preamp. Anything intended for use with a turntable must have its cut-off at a much lower frequency.

The worst offenders are usually the loudspeakers, which very rarely do much below 30 Hz, with most low-cost bookshelf speakers having cut-off frequencies as high as 80 – 100 Hz! Needless to say, with such loudspeakers, any effects from the 10 Hz cut-off of the turntable would be entirely masked.

Loudspeakers with responses extending below 20 Hz and very respectable phase responses do exist, and they tend to be large and expensive. They are, however, absolutely essential in professional mastering facilities, where low-frequency phase error effects must be clearly audible to be adequately dealt with. In the current state of the industry, with minimal recording budgets and only a small percentage of the audience in possession of big loudspeakers that can actually reproduce low frequencies, very few recording and mastering facilities are actually equipped with proper monitoring facilities. This is not just limited to equipment design, but also studio design and the mentalities prevalent in the recording industry. The gap between good sound and horrible sound is widening!

Regarding DC pressurization, following on from the discussion in a previous question, while it could be theoretically possible, it would not necessarily be unconditionally beneficial to the listening experience. From an absolute performance standpoint, the electronics would probably all benefit from a response down to DC. Most transducers would not, as the microphones then would be acting more as barometers, given that the static atmospheric pressure due to gravity is much greater than the small modulations of air pressure that we perceive as sound. The cartridges would then act as displacement sensors, tape heads would act as compasses, and so on. It would be difficult to exclude the “environment” information, but otherwise keep the response to DC, to ensure a perfect phase response. There are much more practical ways of preserving a good phase response without needing everything to work down to DC, such as by careful design of cut-off points in transducers and by considering the entire recording and reproducing chain as one system, in which each individual component must work in harmony with all other components.

BL: What about a tape head and recording tape if everything is direct coupled (theoretically, even if not usually implemented that way). If presented with a DC signal, would the tape head create, and would the tape record, a constant magnetic field?

JIA: Tape itself is just ferromagnetic particles (on a substrate). In the presence of a permanent magnetic field, tape will become permanently magnetized (DC magnetization). This could even be done without a tape machine, just by bringing a permanent magnet close to the tape. It can also be done with an electromagnet, with or without a tape machine. The recording head is essentially an elaborate electromagnet, mounted to a tape machine. Recording DC magnetization onto tape is as simple as hooking up a battery to the recording head with the tape running (or even stationery, which would just magnetize the part of the tape in front of the head). DC-coupled recording electronics will achieve the same. It is simple to record DC to tape, but it cannot be played back with Faraday-principle repro heads. Such a tape recording could be played back with Hall-effect devices, but to no sonic benefit. In fact, most probably to great sonic detriment.

BL: And the same question [about presenting DC signals to] a disc cutter lathe – again, if 100-percent direct-coupled (even if in practice things aren’t usually implemented that way), would the cutter produce a “constant” displacement cut in the lacquer disc? (Aha, yes, your comments about viewing the cut disc with a microscope were what triggered these questions, at least at this time – I’ve wondered about them for a long time, since learning in college that the output of a coil in the presence of a magnetic field (e.g., a phono cartridge) is a function of the first derivative of the magnetic field, as opposed to all of the electronic components in the signal chain.

JIA: There are in fact examples of fully DC-coupled disk recording lathes, where even the cutter head was mechanically DC-coupled to the disk. A direct current to the drive coils would increase the depth of cut. Therefore, such a system could produce a constant amplitude (constant displacement) cut from DC up until you fry the coils at some high frequency as a result of gross thermal overload while trying to accelerate the cutting stylus to the velocity of light!

This is a pretty hard boundary, since even if we could create a cutter head which would not suffer from thermal overload at extreme velocities, as the cutting stylus would approach the velocity of light, it would suffer a relativistic mass increase (as per Albert Einstein’s paper, titled “On the Electrodynamics of Moving Bodies” published in 1905) which would most probably result in somewhat reduced output from the head, assuming that force and acceleration are not limited. (Note: Disregarding the fact that such velocities could never be accurately represented on 33-1/3 or 45rpm 12-inch records, electromagnetic induction effects only remain dominant while the velocity is much less than that of light. If the moving system of a phono cartridge would be made to approach the velocity of light, displacement current effects would need to be considered and very strange things would probably happen. Do not try this at home!)

Either that or the cutting stylus would travel in time, which might cause minor glitches in the time response of the system (or a gap in the time-space continuum).

BL: Again, if any of this matters, is it taken into consideration in the design of the related recording and/or playback equipment?

JIA: Playback turntables are never DC-coupled to the disk due to the existence of a tonearm (this applies the same to pivoted and tangential arms). But a lot of research went into where to place the mass/compliance resonance of the system, for the least amount of audible side effects.

Compared to turntables, there were very few manufacturers of disk recording lathes and only three or four types of cutter head suspension designs. The early ones were actually DC-coupled to the disk and had an excellent phase response, but were not easy to control for groove depth using the disk-cutting automation systems of the time. As market forces came to value longer playing duration per side much more than phase integrity, and at a time when very few loudspeakers offering decent low-frequency phase response existed, most of the industry moved to floating cutter head suspensions, where groove depth could be automatically controlled, along with the pitch, to assist with fitting more music onto a record side.

Not only would the buying public get more minutes per side for their money, but they could also have it cheaper! The most popular disk mastering lathes became the ones that were offering automation that allowed a far less experienced operator (read: someone who could be paid much less and be easily replaceable) to cut records, without accidentally destroying the incredibly expensive equipment involved and with a low disk-rejection ratio as well. The machines themselves were no longer designed for best sound, but to best benefit business. Everyone was happy – apart from those who cared about sound quality.

Today, 35 years after the commercial mass-manufacturing of disk mastering lathes ceased, the surviving systems in active use are all heavily modified with custom parts, a bit like hot rods! Some rebuilders went for maximum automation, others aimed for lowest cost, and a few went for ultimate sound quality. In this last category, all of the aforementioned issues are taken into consideration and improved parts are custom-made as one-off items, at considerable expense. This is a micro-world of its own; no major manufacturers want anything to do with this, as by now this kind of specialization relies entirely on a very few remaining, highly skilled and entirely irreplaceable individuals, complicated and time-consuming precision machining operations (which often cannot be automated) and a very small but rabidly dedicated niche market.

BL: Does the RIAA curve take any of this into account, or is that purely to manage the practicality of which frequencies can be cut at which amplitudes without overheating the cutting head, eating up too much groove spacing or causing too much pre- or post-echo, [along with] introducing too much noise, causing the playback stylus to jump out of the groove, [and so on]? Or might the recording and playback processes normally be complementary in some way [from a mathematical standpoint] such that, if derivatives are involved, an integration somewhere complements a derivative elsewhere?

JIA: The RIAA curve only serves to reduce the space requirements the record groove would take up at low frequencies and to improve signal-to-noise ratios at high frequencies. (In other words, the ratio of the intentional amplitude of playback stylus excursion due to groove modulation, versus the unintentional playback stylus excursions while tracing the molecular structure of the disk record material).

The RIAA pre-emphasis causes phase shift at the time of recording and the RIAA de-emphasis causes the exact opposite phase shift. The phase shifts cancel each other out and restore the original relative phase relationship. That said, the mathematical differentiation and integration (pre-emphasis and de-emphasis) are indeed complementary, but are entirely due to electronic circuitry, implementing the RIAA time constants, and are unrelated to the derivatives discussed as pertaining to transducers.

BL: When looking at a stereo disc from the front of the playback equipment (i.e., at the front of the phono cartridge), which of the two groove walls correspond to the left and right channels? You alluded to this in discussing that the cutting lathe microscope can be looking in one or the other direction, thereby possibly reversing the groove sides from what one might assume, but you didn’t define which channels are on which side from either orientation.

JIA: Looking at the front of the playback cartridge (not the wiring side), the left channel is towards the spindle, on the left groove wall, and the right channel is towards the edge, on the right groove wall.

BL: Despite the various existing standards that you cited [in Part One of this article], and the many variances [in these standards], are you in effect saying that, yes, absolute phase differences can be heard, at least under some conditions, but given the typical multiple microphones of different types and the plethora of other equipment used to make almost every recording, and [when you factor] in the disc (or tape) manufacturing chain and the playback equipment chain, there’s “practically” no hope of ever achieving any absolute phase consistency?

JIA: On the contrary, it is not actually that difficult to maintain phase consistency and absolute polarity, when working sensibly and with properly engineered equipment. But it is necessary to verify the operation of all equipment to be used in a professional audio facility and to keep it all properly calibrated and interconnected. Home systems are much simpler and the availability of products by manufacturers known for their high standard of engineering certainly helps.

BL: And even if there was a completely and rigidly controlled environment for the aforementioned variables, wouldn’t the different microphone distances from different instruments alter absolute polarity anyway, with either the popular “multi-track mono and mixer panning” approach used for most recording or perhaps even the minimalist two- or three-mic “natural” approach that is used for some elite audiophile recordings? (The highly-desired natural stereo imaging for the latter “natural” approach with real-time acoustic music depends mostly on the relative phase differences between the mics as they “hear” the different instruments, vocalists, room/hall ambience, etc., whereas the other approach depends largely on amplitude differences between the stereo channels via the multi-channel mixer pan pots.)

JIA: First and foremost, for a recording engineer to be able to assess the effects of microphone selection and positioning, to decide which technique would best convey the sound to the recording medium, a completely and rigidly controlled environment is an absolute requirement. If the room acoustics and the monitoring system in the studio do not provide adequate resolution of detail, everything else is totally up to good luck. There is no way for an engineer to measure or guess what the monitoring environment does not reveal. Especially when the target audience would be audiophiles with exceptional systems, the studio environment would need to be much more detailed and more revealing than any of the audiophile systems that the record would subsequently be enjoyed on. The engineer, in this case, would also need to have some experience listening to exceptional audiophile systems, to be able to fully appreciate how high the expectations can be.

As for microphone distances, they do affect what is captured, in terms of frequency/phase response, dynamics, reverberation, and stereophonic imaging, so their positioning and careful selection is critical.

Microphones do not only capture the sound from an instrument or ensemble, but also the sound of the room. Changing the position of the microphones or performers by even a few centimeters can make a huge difference.

Using the example of sitting in a concert hall, the sound of the instruments nearest to our ears arrives earlier than the sound of the instruments further away. This adds a sense of depth to the “soundstage” and enhances the sense of width, both in real life and in a reproduced recording. However, in all performance spaces, there are always seats that offer a better overall auditory experience than others. In positioning the microphones for a recording, the intention is usually to find the “best seat” and capture this experience.

When a coincident or spaced pair of microphones (or three as in the technique employed by Decca) are used, the intended absolute polarity is the one that the engineer or producer decided upon as offering the best representation of the performance, through the position of the microphones. It does not necessarily mean that all instruments will always produce an outwards loudspeaker cone motion for a positive pressure increase in front of the instrument! This is irrelevant, since the performance is not enjoyed by placing your head in front of a trombone and a violin at the same time! What matters in this case is the pressure decrease/increase in the hall, where the microphones are positioned.

This should then reproduce, on your system, something resembling what the engineer and producer heard when making the recording, which was presumably intended to sound as close as possible to what they heard in the performance space, prior to setting up any microphones. In multiple-microphone recordings, inaccurate positioning of the microphones relative to each other causes extreme phase cancellations, a confused, unstable stereophonic image and unnatural-sounding instruments. Microphones also need to be very carefully positioned with multi-miked recordings, but the approach is a bit different, since the multiple-microphone approach usually aims to create an experience that was not necessarily there to begin with, instead of just capturing what was there. This can be beneficial in cases where the original performance space’s acoustics are problematic, but when everything is as it should be in the hall, I personally prefer just a pair of microphones.

In rock and pop music, however, additional microphones are sometimes used for effect, to present a larger-than-life soundstage and enhance rhythmic impact. In such cases, the polarity of an individual microphone could be intentionally inverted to create the desired effect. Since this is done to create an artificial reality that was never really there, the “correct” absolute polarity is again whatever the producer intended, provided that it was possible to properly hear the result while the recording was being monitored in the studio.

If we all had exactly the same living room (and drove identical cars) with exactly the same loudspeakers in them, in exactly the same position, then we could probably skip the studio environment and just work on recordings in any living room, since we would hear exactly the same result that everyone else would also hear in their own living room.

If, in such a world, we wouldn’t have gotten arrested by the secret police and if music wouldn’t have been forbidden, along with reading, thinking, or defecting to the other side, where the sheer diversity in living rooms and loudspeakers would necessitate completely and rigidly controlled environments in which recordings were to be monitored. No need to prohibit reading, thinking or defecting. Market trends would have pretty much rendered the first two obsolete. As for the third, where would you go to? Perhaps this is why space colony concepts were popular with multi-millionaires, although recent global events may make them unpopular again, and hopefully steer things in a different direction, before we all end up having the exact same living room in space!


Header image courtesy of Wikimedia Commons/Evan-Amos.

Back to Copper home page