The most challenging, time-consuming, and fascinating part of making my tanpura album, Uncommon Drones, was getting the instrument in tune for each track. I’d like to share some observations about the tuning process in a way that might be helpful to anyone who wants to know the “backstage story” of the album, as well as to those readers who play the tanpura or have an interest in the broader topic of just intonation.
It takes a minute or two to get a tanpura acceptably in tune, but it can take hours of playing and experimenting to make the instrument truly sing. Not every instrument is equally responsive. I have played tanpuras that have a fairly straightforward acoustic “personality,” where once the instrument is correctly tuned, there’s nowhere to go from there, no refinements to be explored. But in Uncommon Drones, I play a magnificent large male vocal tanpura made by the late Hemen Sen, and when working with such an exquisitely responsive instrument, you realize that even inside the range of what could be considered a correct tuning, there are countless “interpretations,” countless possible varieties of sound that can be realized.
How is this so? The most common tuning for the four-string tanpura sets the first string at “Pa,” the fifth degree of the scale, while the second and third strings are tuned to the tonic above, and the last string is tuned to the tonic below. We know that the most acoustically pure fifth arises from a 3:2 frequency ratio – the Pa string, the fifth, vibrates at 1.5 times the rate of the low Sa string, the tonic. Wouldn’t it seem that the goal of tuning should be to achieve the closest possible fidelity to this “perfect” ratio, and if so, where is the room for variety, for multiple interpretations?
Part of the variety comes from the fact that there is no single way that an interval like the fifth, even at its perfect 3:2 ratio, actually sounds: what we hear is largely dependent on the timbre used to realize the interval. When you hear a tanpura play a fifth you are not hearing two sine waves in a specific relationship, but rather you are hearing an entire chorus of overtones generated by each string. The relative amplitudes of these interacting overtones can give the interval a brighter or a darker sound, a smoother and more consonant sound or a rougher and more dissonant sound.
If you are at all familiar with the concept of musical consonance and dissonance you may be surprised that I suggest that a perfectly tuned fifth, at an exact 3:2 ratio, could be anything but consonant – after all, the fifth is considered the purest and smoothest interval aside from the octave, and when fifths sound rough or jarring, we usually attribute this to their being slightly mistuned. In fact, if you look at the harmonic series above each note in a fifth, you see that some of the harmonics coincide, while others fall close together without actually matching. To give just one example, the bottom note in a fifth generates a 13/8 harmonic that is somewhere between a minor and a major sixth above the fundamental, while the upper note in a fifth generates its own 9/8 harmonic that is just a little higher than harmonic mentioned before – it is a major sixth above the bottom note. If these two close-but-not-exactly-matching harmonics happen to be very loud, a listener may well hear them as conflicting, even though the interval that gave rise to them is perfectly tuned. In most cases, the “overtone conflicts” within a consonant interval like the perfect fifth are not audible, because we often employ timbres where the higher overtones are very faint (and rapidly decaying) relative to the fundamental. However, the tanpura, particularly a male vocal tanpura tuned in the range of low C, is one musical context where certain overtones can be unusually strong and long-sustaining, and if the wrong pairs of overtones are emphasized, one can begin to hear discord even within a properly tuned “consonant” interval. This can be a distraction in the tuning process, since one might think the discord indicates that the tuning itself is wrong; in fact what’s “wrong” or undesirable is not the basic tuning, not the relation between the fundamental pitches of each string, but the “voicing,” the way various overtones in the tanpura’s upper sound spectrum are being emphasized.
Of course the tanpura does not offer the same kind of fine-grained control over the amplitudes of individual overtones that one would find in, say, a software additive synthesizer, but it does offer several mechanisms by which the player can influence the overtone content. First there is the feature probably unique to the tanpura, where cotton or silk threads are wedged beneath the strings and the bridge. Moving the thread up and down along the string adjusts the amount of buzzing or “jawari” that the string produces: usually one finds a specific point where the buzzing is strongest, and moving the thread away from that point in either direction reduces or eliminates it. The important thing to realize is that the thread position not only controls the intensity of the buzzing, but it can also cause specific overtones to come to the fore. When adjusting the threads one should not just listen for the “amount” of buzzing, one should listen for any overtones that gain prominence and consider how these prominent overtones interact with those coming from the other strings. If a specific overtone sounds too strong, an adjustment of the thread might soften it.
In some cases, a certain overtone might be pleasing in itself but might not work for the particular raga one intends to sing against the tanpura. The most common example of this is the way the Sa strings often produce a very strong and ringing Pa harmonic, and yet there are some ragas where Pa should be weak or absent, and so the first string of the tanpura is tuned to ma (the fourth scale degree) or Ni (the seventh) instead of Pa. In a tuning that attempts to de-emphasize Pa one might still find that Pa is quite present as a strong overtone coming from the Sa string; in such cases, even though the Pa harmonic can never be fully suppressed, one might wish to adjust the voicing so that this particular harmonic is a bit quieter. A similar situation arises with the Ga harmonic – the major third – which might not be welcome as a strong presence if the raga itself excludes the major third and instead includes the minor third.
Another way the strengths of the overtones can be influenced is by taking advantage of the reinforcement effect – the sympathetic resonance between strings. One finds that even in the range of of what sounds like a well-tuned fifth, the tension of one string can be nudged ever-so-slightly up or down, and when this is done, certain pairs of overtones between the strings will come into alignment and begin to reinforce each other, actually becoming louder through their interaction. The player can experiment with this “nudging” to try to emphasize the overtones that are most desired.
Finally, the strengths of different overtones can be influenced by the dynamics of playing – the position and angle and timing of the finger stroke. One may find that playing high or low on the string brings out certain sets of sounds while playing in the middle brings out another set. There are cases where playing too hard causes a kind of wobbling effect where the strings are temporarily detuned, and yet there are also cases where one hears “beating” and disagreement if the strings are played too softly, but when they are fully driven there is a kind of mutual-reinforcement where all the strings come into perfect agreement.
Though not a “feature” of the tanpura itself, the acoustics of the room where you play will also affect what aspects of the instrument’s sound are heard most prominently. And when making a recording, microphone type and position are hugely important. Any elements that are applied in audio post-processing, like EQ, compression, and reverb can also affect – sometimes in a dramatic way – which specific overtones come to the foreground. The speakers and other audio equipment used to play the recording can also have a dramatic effect, so much so that you might think you are hearing a different recording depending the playback system. While this variability can happen with any audio recording, it’s particularly apparent in the case of a tanpura recording where the overtone content is so rich and where individual overtones are sustained for so long that the listener can focus on them closely.
It is important to listen not just to how the tanpura sounds at one given moment but to observe how the sound evolves over time, as the strings are being struck and after all four have been struck. The tunings that seem to me as most beautiful are those that have the longest sustain, the longest and purest “ring.” In certain tunings one may find that a bit of discord enters not during the strokes but afterwards, during the sustain period as the sound decays, and this trailing discord can be hidden somewhat by playing faster, so that the strings are never idle for too long. (I am of two minds, though, as to whether this kind of masking through an increased tempo is ever truly worthwhile.)
We started off discussing the sources of “variety” in tuning – why there is no single correct way that a perfect fifth sounds on the tanpura. In addition to voicing concerns – the relative strength and weakness of different overtones – there is also the question of inharmonicity. Many people who first learn about the harmonic series and just intonation seem to think that when any string is struck, you hear the fundamental along with overtones at the mathematically prescribed frequencies of two times, three times, four times, five times the fundamental, and so on. However, many strings possess a degree of stiffness that prevents them from generating an ideal harmonic series, but causes some of the overtones to be skewed. Since the tanpura strings typically have different thicknesses and are played at different tensions, the degree of inharmonicity may vary from string to string. What this means is that even if one attempted to put all the strings in unison, it would be physically impossible to get all of the slightly skewed overtones to align perfectly, and there would always be some level of discord in the sound, detectable by those who know how to listen for it. In practice this means that when tuning a tanpura, one must sometimes choose whether to focus on aligning the fundamentals of different strings in a “perfect” ratio, or allowing the fundamentals to deviate from the perfect ratio in exchange for better alignment of the overtones. Each approach yields different sounds and contributes to the “variety” that exists even when tuning a standard interval. While the existence of inharmonicity may surprise some musicians, it is familiar to good piano tuners who sometimes “stretch” octaves to make sure the harmonics align – indeed, piano strings, particularly those on smaller instruments, are notable for their inharmonicity.
The tuning of a tanpura is also affected by environmental concerns. A change in room temperature or humidity can cause an instrument to slip out of tune. In many cases, an instrument will seem to “warm up” and hold its tune better the longer it has been played. An experienced player will know how to turn the pegs (for coarse tuning) and adjust the beads or swans below the bridge (for fine tuning) so they are least likely to slip, but some degree of slippage is unavoidable, and one finds an instrument is most secure not directly following an adjustment, but after it has remained in a certain configuration for some time. I have found that even doing so much as lifting the instrument and putting it back down can affect it, not causing any kind of radical detuning but still making a subtle and perceptible difference in how it sounds – handle it carefully and with reverence!
If the tanpura is to accompany a specific performer it must be tuned to that performer’s desired pitch, but when you consider the tanpura on its own, you may find that a certain instrument is most resonant and most stable at a certain pitch but not another. I sometimes experiment by damping all the strings except the low Sa string, and then playing that low Sa string at different pitches to see when the instrument sounds the most settled and the most resonant. At certain pitches, the low Sa string, even with the others strings damped, might seem to wobble and feel unstable, while at other pitches the string might sound perfectly “happy” and stable, and these differences can occur within a very narrow pitch range, well under a semitone.
Many of the factors we have discussed are not of critical importance when the tanpura is to assume a background role and to be heard from a distance, as happens in a traditional performance context. However, when recording the tanpura as I have done in Uncommon Drones, with a microphone placed close to the instrument, one can hear nuances – interactions among higher partials – that would be inaudible at a distance, and these nuances come to be very important in the overall feel of the recording. I could easily make a “decent” recording of the tanpura in any given tuning, but to really capture the instrument in its full radiance, I must pay careful attention to all the factors we’ve discussed; often this involves making multiple candidate recordings over many days with subtle variations in the parameters to capture the very best of what’s possible. This leads to the question of how one actually “tests” a given tuning to decide whether it’s right.
What I consider to be the most important method of testing a given tuning, and one that seems to be very rarely discussed, is to place one’s hand on the body of the instrument (for a Hindustani-style tanpura this would be the large resonating gourd) and feel the “texture” of the vibrations that unfold as you play. How smooth or rough do they feel to your hand? Often one finds that a good tuning can be detected by touch alone: when your hand feels a pattern of vibration that’s smooth, even, consistent, and free from “turbulence” or noticeable fluctuations in intensity – that’s when the instrument actually sounds best. This point is so important that I feel it deserves more than a paragraph but I don’t have much more to say about it; this is something that can’t be profitably discussed at length but which proves incredibly valuable in practice. Find a tuning that feels good to your hand!
The next most important, and rarely discussed method of testing a tuning is to sing against it. Ultimately, the tanpura is there to support the voice, so to tell whether it’s well-tuned, you can’t just listen to it, you have to actually sing against it and consider how the singing feels. Sometimes the tanpura might “sound” good to your ear but might not “feel” comfortable against your voice. And sometimes you might not notice particular aspects of the tanpura’s sound until you sing a note and your voice begins to interact with the sound – only then do you become conscious of the harmonics that relate closely to the note you tried to sing. It can be helpful to experiment with different vocal timbres as you do this, including trying some “overtone singing” if you are familiar with it.
I recommend two parts of the singing test. First, you should sing each note in the raga to be performed, starting with the most important notes (the tonic of course, and then the “vadi” and “samvadi”) and consider how comfortable it feels to sing the note. Does your voice seem to “lock” into a specific position for each note or do you feel uncertain whether the note belongs a little higher or a little lower? Does the note seem to resonate against the tanpura – in fact, does singing the note make the tanpura sound even better? – or does the note feel vague and dull? Second, you should pick an important note and continue singing it as you play each string of the tanpura. Notice how the striking of each successive tanpura string affects your voice. Does each string seem to reinforce the note you’re holding, or do different strings seem to pull or push your voice up or down a little? Ideally you would want a tuning where your voice is not pulled or pushed, but you can hold it perfectly steady as each string is struck without any temptation to move – and ideally this would be true when singing any of the most important notes in the raga.
Another method of testing a tuning is to play harmonics (by lightly resting a finger at specific locations on each string and then plucking with the other hand) and check whether these harmonics match. As mentioned above, the low Sa string produces a Pa harmonic, and you can check whether this harmonic matches the octave harmonic of the Pa string itself. The same principle can be used in less common tunings, for example when the Pa string is tuned up to shuddha-Dha, or the major sixth. In this case, the Pa harmonic of the Dha string should match the Ga harmonic of the low Sa string. While harmonics are a useful method for checking a tuning, I feel strongly that they should not be used exclusively, considering some of the factors mentioned above. The phenomenon of inharmonicity implies that one can never assume each harmonic is really where it “should” be. Furthermore, one may find that the harmonics of different strings seem properly aligned when the strings are softly plucked, yet when the strings are driven more forcefully the tuning changes and no longer seems perfect. One simply can’t assume that an alignment of harmonics at low playing force will translate into a resonant and agreeable tuning at higher force.
In tuning the tanpura there are also a few “gotchas” to be aware of. It’s critical that you get the two high Sa strings perfectly aligned. Often you might struggle finding the proper tuning for the outer two strings, when really the problem is that the two middle strings do not perfectly agree, and so there can be no proper position for the outer strings. What’s potentially confusing is that the two middle strings can sound quite acceptable, and even have a pleasing “chorus” effect, when they are ever-so-slightly mistuned from unison, and yet this mistuning will spell doom for the project of getting the outer strings in tune.
Another thing to be aware of is that the low Sa string can be the most difficult to adjust. It is the thickest string and admits the coarsest pitch control. Many people say that one of the upper Sa strings should be tuned as the very first step in the process, but I think a case could be made for getting the low Sa string to be stable before anything else is attempted, and then trying to leave the low Sa string alone while tuning the upper Sa strings to agree with it.
Ultimately, tuning the tanpura involves many different modalities of listening, sensing, and testing, and adjusting. One must consider how the tanpura feels to the hand, how it sounds to the ear, how well it cooperates with the voice singing against it, how well the harmonics on different strings seem to be aligned when played separately, and how all four strings sound together when fully driven at an ideal rhythm… and, how well all of these variables match the raga to performed.
If the tanpura is to be used as accompaniment in a concert scenario, of course it must be put in decent tune in not too much time. But in trying to record the tanpura up close, so that all its nuances are captured, one may find that circumstances simply don’t work out on a particular day, and the best thing to do is give up that day and try again later. Those times when the tanpura really radiates, when it seems to come alive, when simply touching a string seems to unleash a shimmering current of sound – those moments stand out as very special, something like a gift, when you are lucky enough to feel and to hear an uncommon resonance. It is in the pursuit of these magical moments that I have worked on the Uncommon Drones album and each track is the result of many hours, sometimes many days or weeks of tuning effort. I find that while it’s important to have an intellectual understanding of the challenges one faces in the tuning process, ultimately it’s intuition, sensation, and persistent experimenting that lead to the most glorious sounds.