Soundboards -- Part 1: What's wrong with my piano?

I received  a call recently from a woman complaining about the sound of her piano.  She asked if I could come and take a look to see what was going on.  The piano, it turns out, was an older Steinway O from about 1910.  The tone was a dead giveaway (literally):  short, distorted attack, rapid decay, plinky sounds from the treble section, whining and buzzy sounds in the mid range, distorted and harsh bass response.  But I went through the routine, took several measurement of crown and bearing, listened carefully to all the sections, examined the rib and soundboard glue joints.  In the end I told that I was sorry but the the soundboard was not functioning properly and would need to be replaced.”  “Oh”, she replied and then after a thoughtful pause, “but won’t that change the sound?”    

For the piano playing public (and some in the industry as well) there is no shortage of mythology and misinformation about soundboards.  It is often referred to as the soul of the instrument and there are some (even in the trade) who, surprisingly, think that you should never change it, no matter what.  While that view contributes to piano shop on the left bank romanticism, it’s also misinformed.   So let’s explore what the soundboard does, how it does it, why a healthy soundboard is necessary for good quality tone, and why older doesn’t necessarily mean better.     

When a string is struck by a hammer in the piano the energy of the blow is transferred to the string and the string vibrates.  There is no energy gain, in fact there is a net loss (as there always is with a transfer of energy) but the string because of its relatively high tension does maintain a certain level of energy.  But the vibrating string alone doesn’t have enough mass to move enough air to be heard adequately.   So, in a piano, the energy from the vibrating string is transferred to the soundboard via the string’s coupling to the bridge which in turn is connected to the soundboard.  This transfer also comes with a certain amount of energy loss, but there’s still enough energy so that the soundboard itself begins to vibrate.  Since the soundboard has much more mass and area than the string, it moves the air much more efficiently, and as such, you can hear it.  That doesn’t mean that the soundboard is an amplifier, not in the true sense of the word.  The energy from the string isn’t enhanced or boosted in any way, in fact, all along the route, hammer to string to soundboard, there is net energy loss.   What happens is that the soundboard actually creates its own sound, a new sound, the sound of the soundboard vibrating.  The string is actually still there vibrating (until the damper comes down and stops it) and remains the source of the energy, but the sound you hear is not the string (except perhaps vary faintly), rather, it is the soundboard itself producing this new and unique sound.   You can think of the soundboard, then, not as an amplifier but rather as a speaker, technically a transducer.

This distinction is important because the soundboard’s ability to vibrate efficiently at various frequencies plus its ability to handle the energy inputs and dissipate them in a controlled way plus the unique characteristics of the design and materials used will to a great degree determine the quality and character of the sound produced ). Think of your stereo system for a minute.  A good one has various types of speakers, low range (woofers), mid range and high range (tweeters) that are best suited to produce various frequencies necessary for the wide range of musical demands.  The soundboard is no different except that you can’t really load up the soundboard with woofers, mid-range speakers and tweeters.  Everything must come from a single piece of shaped wood or really a composite structure referred to as the soundboard assembly.

The failure of a soundboard is not unlike a blown speaker.  When the soundboard is created it is designed (hopefully) to achieve the varied, requisite stiffness (and mass but we won’t complicate this discussion with that aspect) in order to best produce the range of frequencies and at an appropriate volume necessary for a balanced sound through the scale.  Because we’re dealing with wood, the process can be somewhat unpredictable and unstable.  Traditional methods of creating soundboard stiffness are often the most unstable.  Let’s look at that process briefly. 

For the soundboard to work properly it is formed into a shallow dome shape often referred to as crown.  This small dome functions somewhat like a spring made of wood.  When the dome is compressed by virtue of downward pressure exerted by the strings, the soundboard stiffens and by controlling that downward pressure you can control the stiffness in different sections to a degree.  But the formation the crown is a process which employs different methods depending on the piano builder.  One method employed involves first drying the soundboard panel (that large flat piece of wood that is visible under the strings) down to a very low moisture content (about 4.5% which equates to 80 degrees  and about 20% relative humidity).  When much of the moisture has been driven out of the panel the ribs are glued onto the back or underside or back  (crawl underneath a grand piano or look at the back of an upright and notice the ribs which are like small beams running perpendicular to the grain of the soundboard panel).  Then the panel is allowed to take on moisture again.  Of course, the panel wants to expand but is restricted by the ribs on one side so the soundboard panel wanting to expand is compressed and the ribs are bent into a curve.  That curve is known as the crown and this process is known as compression crowning.  Some technician’s, including myself, employ different methods for crown formation.  These involve shaping the supporting ribs into a preformed radius matching the targeted crown radius.  This obviates the need for compression crowning.  The panel is dried down but to a much higher level (around 6 - 6.5%) which puts the panel under much less compression stress, is more predictable and more stable.  It also lessens the inevitable damage which will arise when the wood cells are compressed beyond their elastic limits.  Whichever method is used, for prolonged soundboard health it is important to maintain the soundboard at a constant humidity level.  Extreme low humidity and high humidity, or the fluctuation back and forth between the two as is often seen in Midwestern or East Coast climates can stress the panels elastic limits and create cracks and pressure ridges in the panel. 

Crown formation is important for the panel to be able to withstand the downward pressure which comes from the strings and to control the stiffness in various parts of the soundboard assembly.  Without the requisite stiffness will have a more difficult time producing the various frequencies needed especially in the upper end of the piano.  A treble section which lacks adequate stiffness will tend to be percussive and have a short duration.  Tenor and bass sections of the piano are more forgiving.  Maintenance of the crown is equally important if the tone is to remain stable over time.  Sadly, compression crowning methods are probably the least predictable and the least stable.  That’s because the thin soundboard panel is under extreme stress during the crowning process and afterward.  The wood cells in the panel are often compressed beyond their elastic limits (they get crushed) and the strength of the wood panel is compromised.  When you add to that additional pressure that is exerted by the strings bearing down on the whole assembly, the length of time that the crown can be maintained can be quite short.  Sometimes within even a few months it is possible that the crown will start to change or collapse.  Over a longer period of time with seasonal changes in humidity and cycles of expansion and contraction the panel undergoes more stresses.  Eventually weakened areas where cellular damage has taken place can form what are called pressure ridges and these ridges eventually form cracks.  While the crack itself is not a problem per se, it is an indication that the panel has been stressed beyond the elastic limits of the wood.  While a crack or pressure ridge is not necessarily a reason to condemn a soundboard it is a sign.  If the stiffness is compromised, the soundboard’s ability to produce and modulate various frequencies and higher levels of energy input begins to deteriorate and along with it the tone. 

So looking at our piano in question this is what has happened.  Damage to the panel has weakened it to the degree that it is no longer able to hold crown (assuming it had some to begin with) and the requisite stiffness across the scale has been compromised.  The attack phase has become very percussive and with more forceful playing quite distorted especially in the mid tenor and bass.  Rib separations  (glue joint failure) are causing a lot of buzzing and high pitched whining.  The treble section has a very sharp and short attack with poor development of upper partials and with extremely shortened sustain.  The very high end of the piano has a very short plink sound virtually lacking any pitch recognition.   These are classic symptoms of a failed soundboard. 

So the answer to her question was, yes, replacing the soundboard would change the sound.  But, then, that’s the idea. 

**********

Electronic Versus Aural Tuning

The other day I received two interesting phone calls.  One was from a person saying they were looking for someone to tune their piano but would only hire them if they were a strict aural (by ear) tuner, the other person said they would only hire someone if they used and electronic tuning device (ETD).  So what's all the fuss about tuning by ear versus using an ETD?

Before I delve into the details, I can summarize by saying that while preferences and tuning styles vary I think there is a general consensus among technicians that aural tuning skills are required no matter whether you use an ETD or not and that ETDs can be valuable tools in terms of consistency, refinement, problem solving and under special circumstances which I will discuss later.  The answer that I gave the customers was, I use both. 

Let's begin by looking at what happens during a routine tuning, one done by ear and one done using an ETD.  The aural tuner begins by tuning a single note, usually A4 (above middle C) to a pitch reference such as a tuning fork.  The tuner then proceeds from that note to set what is called the temperament octave which functions as the reference for the rest of the piano.  The temperament octave (in equal temperament) is divided into 12 equal semitones by tuning a sequence of alternating fourths and fifths in a one octave confined circle of fifths, in effect.  Other intervals such as thirds and sixths are used to check the accuracy of these fourths and fifths which are themselves altered (or tempered) from pure just slightly in order to achieve relative consonance in each key.  While the explanation for this is beyond the scope of this piece of writing, suffice it to say that the tuner must manipulate the pitch of each note to a very fine degree in order to achieve an equally tempered reference octave.  From there the tuner tunes down through the bass using octaves and carefully checking those octaves using various reference intervals and then similarly up through the treble.  During this part of the process (or after depending on the the tuner's style) the unisons, or multiple strings which are part of each single note, are tuned to complete the tuning.  

The user of the ETD functions a bit differently.  Using the ETD (there are various types but most operate this way) the technician takes measurements of several sample notes in the piano and then calculates an ideal tuning curve (how the whole piano should be tuned) based on those samples. Different pianos do have different tuning requirements and the calculated tuning curve will vary depending on the piano.  The aural tuner, by the way, also faces these differences but deals with them more on the fly, as it were.  Different ETDs take more or fewer samples depending on how they are programmed.  Once the samples are taken and the curve is calculated, the tuner then is free to choose how to proceed. Each note on the piano can be dialed in and then tuned until the ETD indicates that you have hit your target pitch which it does with a visual display.  Some tuners, after calculating the theoretical tuning curve, simply start at the lowest note and tune up through the scale relying on the accuracy of the calculated curve.  Others start as they might doing an aural tuning which is in the middle of the piano and work outward. Similar to the aural tuner, the user of the ETD tunes the unisons either as they go or afterwards.  Generally, as we shall see, the ETD user tunes the unisons by ear in the same manner as the aural tuner.  

So the issues can be divided into several different categories.  Setting the reference note, temperament setting, the pre-calculated versus the on the fly aural tuning, unisons, and other factors.  We'll skip setting the reference note because aside from issues like tuning fork temperature and the mechanics of setting the first note from different types of pitch references they produce little difference in the ultimate outcome of the tuning.  So let's start with unisons.

Unisons

Even for tuners who rely more on ETD's than strictly aural tuners, about one third of the piano must be tuned by ear.  Most notes in the piano have three strings tuned in unison (60+ on average), about 20 have two strings and the remaining have one.  So typically at least 80 of the notes require you to tune two or three strings in unison even if you use the machine to set the first of those strings.  Unison tuning is typically done by ear even by those who use ETDs because, quite simply, it's more consistent and accurate.  While tuning the entire piano requires creating a uniform temperament and adjusting the intervals (octaves, fifths, etc.) so that they are consonant with each other, most people who respond to whether a tuning was good or not are responding to the how clean and stable the unisons sound.  To produce clean and stable unisons, the ear is more efficient and probably more accurate.  Part of the reason has to do with the anomalies of string dynamics and when in the strings vibrating phase the machine reads the pitch.  In short, the strings pitch can change somewhat during the vibrating phase so depending on where in the phase you measure (or read the ETDs output) this can cause slight variances in pitch accuracy.  Since the unison tolerances are so low (even a small deviation from perfectly in unison is easily detectable even by a non tuner) using the ear will tend to mitigate the possibility of reading errors from the machine and produce cleaner unisons.  Moreover, in the bass section of the piano where more partials (overtones) are audible, you sometimes need to make judgement calls about which partials you focus on as hand wrapped bass strings can produce less than perfect partial matching even when the strings are supposed to be identical.  For those reasons, unisons are almost always tuned by ear even by the most ardent users of ETDs.  So with respect to unisons, there is really no difference between aural and ETD tuning, both require the tuning of unisons by ear. 

Advantage:  Aural tuning but not really fair since both ETD and aural tuners tune the unisons this way.  We'll call it a draw.

Temperament 

Setting the temperament probably offers the greatest intellectual and mechanical challenge for the aural tuner.  The temperament octave represents the model octave from which the rest of the piano is tuned.  It is usually done from F3 to F4, but often from A3 to A4.  In equal temperament (we'll save the discussion of historical temperaments for another day) the idea is to divide this model octave into 12 exactly equal semitones.  The process is that which offers the greatest challenge for those learning how to tune because it requires the precise manipulation of various intervals to varying degrees, a sequence of interval checks (to make sure you are on the right track), and a few judgement calls.  No interval in a piano is tuned pure, or "just", with the exception of the unisons.  All intervals require some slight modification.  For example:  fifths are tuned slightly narrow, fourths slightly wide, major thirds are wide which means that minor sixths are slightly narrow, etc.  The reason for this has to do with string physics and something called inharmonicity which means, in short, that the upper harmonics produced by a vibrating string do not have perfect mathematical relationships to the fundamental tone (or first partial).  They're off a bit to the sharp side. How much basically depends on the diameter and tension in the string thus different pianos have different amounts of inharmonicity.  Moreover, different sections of the same piano have different amounts of inharmonicity as well.  Thus, if you tune a circle of perfect fifths within one octave (and you can do this by alternating between fifths and fourths) when you get back to the note you started on but one octave higher it will be quite sharp when compared with your starting note one octave lower.  In order to get things to come out right and have that last note in the circle of fifths sound consonant with the note you started with you will have to tune the fifths slightly narrow.  Not a lot, hardly perceptible, but enough so that that last note matches your beginning note when compared as an octave.  This is the procedure known as "tempering" and is the reason the central octave is called the temperament octave.  At least that's the simple explanation and as far as I'll go for now.  More when we get to octave tuning.

So the aural tuner has to tune this circle of fifths with the exact precision required to get all the intervals: fourths, fifths, sixths, thirds, seconds and sevenths, to sound right and in all keys.  It does offer a challenge and most people learning to tune spend the most time practicing this particular sequence, wrestling with the smallest of controlled manipulations and various checks to insure that they accomplish their goal. 

The ETD user, on the other hand, simply calculates the inharmonicity in that section of the piano, pushes a button and the ETD will calculate a perfect sequence of 12 notes miraculously divided into 12 equal parts.  A skilled aural tuner can, of course, tune a very accurate sequence and end up with a beautifully crafted temperament octave.  But having done it both ways I have to say that measured over time and lots of tunings with pianos in various states, the ETD is more consistent and more accurate, especially if you factor in speed and (let's be honest) fatigue.  How important is a perfectly crafted temperament octave?  We'll that's a discussion for another day.   

Advantage:  ETD

Octaves--Leaving the Temperament Octave

Leaving the temperament octave there are a variety of approaches.  Most aural tuners proceed up or down tuning octaves and checking these octaves using various other intervals also used in the temperament octave such as fifths, fourths, thirds and sixths (and some wider intervals as well such as tenths, twelfths, double octaves, etc.) to insure a smooth progression going down as well as going up through the rest of the piano. 

The ETD user has a slightly different approach.  Having measured some number of sample notes in various places in the piano, the ETD calculates a theoretical tuning curve for the rest of the piano based on those measurements.  Here's where things get a bit tricky.

Let's revisit inharmonicity for a moment.  You recall that inharmonicity refers to the degree to which the upper partials of a string don't have perfect mathematical relationships to the fundamental or first partial.  What does this mean actually?  Each vibrating string produces a sequence of partials and they are referred to in the following way:  the 1st partial is the fundamental note itself, the 2nd partial is one octave above that, the 3rd partial is a fifth above that, the 4th partial is a fourth above that (which is now two octaves above the first partial), the 5th partial is a major third above that, the 6th partial is a minor third above that (or now two octaves plus a perfect fifth) and so on.  Every vibrating string produces this partial sequence and these frequencies are all audible when you play a single string (if you listen carefully). 

Since all notes produce this same sequence that means that for any two different notes played there may (or may not) be some partials which the two notes have in common.  Those partials that the two notes have in common are called "coincident partials".  For example, the note C1 (the lowest C on the piano) and the note C2 (one octave higher) have in common the first partial of C2 and the second partial of C1.  Moreover, they also have the second partial of C2 and the 4th partial of C1; also the 3rd partial of C2 and the sixth partial of C1.  Octave have lots of coincident partials.  Other intervals, like minor seconds, have none.

So why is this important?  Because when you are tuning octaves together what you are listening for is often a particular coincident partial.  The octave, when tuned to a particular set of coincident partials is thus referred to by the name of those partials, for example a 2:1 octave would be one in which you are comparing the 2nd partial of the lower note with the 1st partial of the higher note.  So what, you say.  Well you recall the bit about inharmonicity, right?  As it turns out the measured amount of inharmonicity varies depending on which partial you measure it on!  In effect, the higher you go up the partial chain, the more out of tune the partials get.  

What this all means is that depending on which type of octave you are tuning you will tune it differently and the relationship between the two notes will change ever so slightly.  Not all octaves are created equal.  As an aural tuner you will select the type of tuned octave based on either your own preferred style and/or on what sounds right.  An ETD user will select a programmed "style" for the tuning curve which will calculate octaves based on the settings in that particular style.  It may be the best for that piano, but it may not.  Moreover, as I mentioned, inharmonicity can vary within an individual piano and it may be necessary to deviate from the theoretical curve in various parts or the piano, most notably between the tenor and bass.  Here, it's necessary for the ear to guide you and there are some judgment calls to be made.  The ETD can be of assistance because it can quantify what you are hearing.  You can set the the ETD to focus in on a particular partial that you are targeting and get a reading from both notes that you are comparing.  That may help you make a decision in conjunction with what you are hearing as to the best way to proceed in that particular section.  Nevertheless, because pianos can be somewhat unpredictable this way, the ear must be the final judge of what sounds right.  

Advantage:  Aural tuning with an assist from an ETD to outline a theoretical tuning curve and be able to visualize what's happening in a particular section of the piano. 

Other Situations 

This is getting long so let's wrap it up.  There are cases where using an ETD has distinct advantages.  When encountering a piano that is significantly off pitch and requires a preliminary pitch correction before one can attempt to fine tune the piano the ETD is extremely useful.  Pianos don't tolerate much of a pitch change and remain stable.  That's because any significant change in pitch creates a moving change in overall tension as you progress through the scale.  This causes, in effect, the piano to move back in the direction it came from by a factor of about 30%.  So if you are pulling the pitch of a piano up from some amount flat, you need to first over shoot your target by about 30%.  When you are finished with the first tuning the piano will have settled back very close to your target and you can begin, then, the second tuning and the process of fine tuning the instrument.  The ETD can be very useful in measuring and calculating the required amount of "overpull" such that you end up as close as possible to the target pitch.  It's much harder to do accurately by ear. 

Another area where the ETD has an advantage is when tuning more than one piano in unison.  A calculated and memorized tuning curve can then be easily transferred to one other or (88 other pianos for that matter) without having to worry about comparing notes on the two different instruments.  A problem, sometimes, if you don't have very long arms.

Similarly, if you are working with a recording studio and the pianist wants to go back some time later to edit, or cut-in over the existing recording, it will be critical that the piano is tuned exactly as it was for the edit to blend perfectly.  By storing the exact tuning used in the ETD you can easily duplicate it with a great deal of accuracy. 

Advantage: ETD

Summary

So you can see that both aural skills and the intelligent use of the ETD have their own distinct advantages and can be of great benefit in producing consistent and quality tunings.  The ETD, like anything else, is simply a tool.  When considering hiring someone be less concerned with what they use and more concerned with the quality and stability of the tuning you get no matter how it's accomplished.

(If you got through this, congratulations.  It's far more than I intended to write but I hope it helps.)

**********