A Gentle Introduction to South Indian Classical (Karnatic) Music – 3 of 4 [Sourced from Soc.Culture.Indian archives for non-commercial, educational purposes only] [ 1 of 4 | 2 of 4 | 3 of 4 | 4 of 4 ]

Ramesh Mahadevan


If frequency and related concepts like tone, scale and octaves form an important ingredient in music, the other equally important element is time and related items like speed, rhythm, meter etc. In fact, a musical piece is nothing but a source of sound emitting sound waves as a function of time. If you looked into the Western system of musical notation, (the ‘Staff notation’) you would have noticed that frequency is notated on the Y axis and time is given in the horizontal axis.

Let us forget about the frequencies etc for this chapter and see how a melody progresses in time. The first concept is ‘speed’. Any song, even ‘Jana gana mana’ and ‘Roop tera mastaanaa..’ has a prescribed speed. ‘Roop tera..’ probably lasts about four minutes and if you sing it much faster or slower, it might even sound funny. (You must have played some old records at a slow speed or fast speed and had a good laugh when you were small) The Western music and Hindustani music recognize various degrees of speed or tempo, all the way from very, very slow to ultra fast. However, in Karnatic music, we do not talk about ABSOLUTE speed – there is no connection to an external clock. This has often led to arbitrariness in speed when songs are performed. Some musicians become well-known for slow rendition of songs. And perhaps an equal number have become equally well-known for their fast rendition of perhaps the very same songs. History and tradition have been the guidelines for the ‘speed’ of a Karnatic piece. Since the concept is hardly invoked, we will not discuss it any further. We will not impose an Adagio (one of the many Western music tempos) or Vilambit (a slow speed used in Hindustani music) on Karnatic compositions.

Once the speed is chosen, Karnatic music is reasonably strict about keeping the speed constant – you should not slow down or speed up during the course of a song unless mandated. However, Karnatic musicians occasionally double and even quadruple their speeds relative to their fundamental speed during the course of certain pieces, just to build up the mood. In fact, lately, there have been songs rendered in ‘seven speeds’, much like a Hamilton Beach blender. The basic speed is referred to as ‘First kaalam’, literally meaning, first speed and correspondingly, when the baud rate is doubled it is called ‘second kaalam’ and when quadrupled, it is called ‘third kaalam’.

Let us now look at the other concepts such as rhythm and meter. Rhythm is probably the most fundamental aspect of music. Ancient civilizations beat their drums much before they made their harps and lutes. Repetitive sound patterns, such as the pulsation of the heart, are so primitive that everyone can relate to them. When we sing ‘Roop tera mastaanaa..’, even if it is just the first line, we feel the rhythm – sometimes even if we forget how exactly the tune went, we could still recall its rhythm. Amazing, isn’t it.

We can tap our feet or pat our thighs or clap as we sing ‘Roop tera mastaanaa..’ How many times did you pat or clap during the first line of the song? How about ‘Baa baa black sheep’ or ‘jana gana mana’? How many times did you ‘beat’ during the first lines of these pieces? What you have been doing by tapping or patting is simply ‘meter’-ing the song to count how long each line lasted – sort of counting minutes. Let us now look at why such metering is important.

Songs are ‘structured’ – they have lines, stanzas and melodic phrases, much like prose having paragraphs and sentences and words.

A good way to write ‘Roop tera.. ‘ will be:

Roop tera
Pyaar mera

Each of the above lines is a musical phrase or melody in itself and seems to last about TWO beats. By the time you finish singing the above four lines, you would have counted EIGHT beats. There is a cyclical repetition. At the beginning of the first (and the third, fifth and seventh) beat, a new musical phrase begins. If the time interval between your clapping or beats is T seconds, then each small melody lasts 2T seconds and the entire opening stanza lasts 8T seconds. In fact, if you went through the rest of this song, you will see that there are some longer melodies lasting 4T seconds.

Now, on to our next example:

Baa baa black sheep (Four beats)
Have you any wool (Three beats)
Yes sir, yes sir (Four beats)
Three bags full (Three beats)

Difficult to figure out. The melodic structure lasts either three beats or four beats. We can therefore conclude that the periodicity is seven beats, with a substructure of Four-beats and Three-beats. Or put another way, the melodic phrases last 7T seconds and even here, there seems to be two sub-melodies in one phrase lasting 4T and 3T seconds.
In the final analysis, we can say that ‘Roop tera mastaanaa..’ is set to a Four-beat cycle (Two beats would actually do, but two is too small a number to define a beat cycle) and ‘Baa baa black sheep’ to a Seven-beat cycle. The periodicity of the four beat cycle is 4T and that of the seven beat cycle is 7T. It might sound trivial, but we cannot hope to play ‘Roop tera mastana..’ using a 7T cycle (if we do so, we might end up chopping the melody in the middle to fit our beat cycles) and by the same token we cannot hope to play ‘baa baa black sheep’ using a four beat cycle.
Let us summarize our concept of beats and cycles and metering. At the beginning of each beat cycle, a new melody seems to emerge. (Melodic phrases emerge even in the middle of the cycle in both the songs; however, fresh phrases almost always begin with each new beat cycle) In fact, if ‘baa baa black sheep’ happened to be a Karnatic song, it is made up of nothing but melodies which are 7T (could even be 14T or 21T) long in time. In fact, the number 7T is some kind of a ‘characteristic time constant’ for this song. Therefore, we see that each song in Karnatic music is not only set to a Ragam, but is also composed of a series of melodies which are of specific duration to fit a beat cycle.
So if you are a music composer, you should select a beat cycle too. Let us say you selected a five beat cycle – your melodies then are constrained to last 5T or 10T or 15T etc long, but not 3T or 7T. This is precisely the basis for Talam. A Talam is essentially a beat cycle. The seven beat cycle is a seven beat Talam and so forth.

An advanced aside! – In general, a new melody NEED NOT begin JUST at the start of a Talam cycle. In fact, you can choose any point in the Talam cycle, (any beat, or even in between the beats) as the starting point to begin your new melodic phrase each time. This point in the Talam cycle is called ‘Eduppu’ in Tamil.

The simplest cycle has only three beats per cycle – called Roopaka talam. By the way, we do not consider a simple TWO beat per cycle as a Talam. Even though Four beats per cycle is a legitimate Talam, Karnatic music does not consider it. Instead, the popular ‘binary’ talam is interestingly an eight beat cycle , called Adi Talam. You would have sung the entire first four lines of ‘Roop tera..’ by the time you completed one cycle of the Adi Talam. A version of the five beat Talam is called Khanda Chapu. There are six beat Talams although sometimes you may be able to get by with just two cycles of the three beat talam, Roopakam. One version of the seven beat cycle, (which is appropriate for ‘Baa baa black sheep’) is called Misra Chapu.

Another quick aside: Misra Chapu and Khanda Chapu are quite popular in Karnatic music. ‘Misra’ in the context of Talam means seven and ‘Khanda’ means five and ‘Chapu’ denotes that the cycle is not uniform! For example, Khanda Chapu goes one-two, one-two-three – here there is a subdivision of the cycle itself into two parts. The first part with two beats last for a different duration compared to the second part with three beats. So the duration is 2T1 plus 3T2 and not just 5T!

There are more complex Talams than what I have mentioned thus far. There are complexities either due to a large number of beats per cycle (such as eleven, thirteen, umpteen) or due to their internal structure with subdivisions (like the ‘Chapu’ Talams) Imagine composing musical melodies which perfectly last a time cycle of thirteen beats! As improbable as it may sound, talented Karnatic musicians in fact choose to compose and perform in extremely complicated and long beat cycles of twenty plus beats per cycle. Legend has it that (of course, in ancient times when kings and queens had nothing to do but to get entertained) a musician sang in a 79 beat cycle and utterly humiliated a rival musician.

However, the vast majority of the compositions are in the simple Talams Roopakam and Adi. You can refer to erudite texts on the other less known Talams. In fact, the morphology of Talams rivals that of the Melakarta scheme and one can get very sophisticated.

The Talams are ‘put’ or ‘kept’ (that is, the beats are counted) by various time-honored ways. Even though at first it might look funny, the familiar way still is to tap one’s thighs and sweep with the right hand while squatted. Ask your musically inclined friend to teach you how to ‘put’ simple Talams like Adi and Roopakam. It is really a good idea to ‘put’ Talam when you listen to a song. You can not only learn ‘keeping talam’ in just a half an hour or so of trying, but also see that you are already a step above your friends who don’t know about it. In fact, although it can be quite annoying at times, you will see that many, many people in the audience would ‘put’ Talam in a concert and ‘participate’ – such feedback is often expected and appreciated by the artistes. In real concerts, the rhythm is kept also simultaneously by percussion instruments like Mridangam or Ghatam.


Thus far we examined (rather lightheartedly) the basic tools of Karnatic music like the Ragams and Talams. It is now time to go to our stereo systems and listen to Karnatic music with an entirely new mindset. You can start off with classical based movie songs or play instrumental music. Can you see the names of Talams and Ragams on the CD jacket? From the name can you figure out whether the Ragam is a Melakarta Ragam by looking up Table IV or a child Ragam? Do you know the keys used in the Ragam? More interestingly, do you have two songs in the same Ragam? If so, can you see the similarities? Can you correctly identify your favorite melodic phrases, which occur in a musical piece EACH TIME THEY OCCUR?

Karnatic music – the hard-core classical music, that is – is divided into two broad areas. The first one is the realm of pre-composed music and the second one is improvisation or creative music. In Indian terminology, the pre-existing compositions are called ‘Kalpita’ (literally meaning ‘that which is taught’) and the creative (improvisation) aspects are called ‘Manodharma’. Every performing artiste learns both the aspects of classical music. He or she not only has a repertoire of several (sometimes hundreds) pieces of well-known songs, but he or she also knows how to create music.

One easy way to improvise is to take an existing song and distort it! Just look at the way the American national anthem is sung by various artistes in their own unique way! But Karnatic music tradition has reasonably stringent guidelines on such ‘liberties’ and where to improvise.

In the first (‘kalpita’) part, the pre-existing compositions range from the very simple to the very complicated pieces. There are Bhajan type songs, dance songs, love songs and songs with a lot of vocal gimmicks. Students of music start with such pre-existing compositions, train their voice, assimilate the ‘moods’ of Ragams, be comfortable with Talams and finally the training wheels come off and they move into the territory of improvisation, where they start making their own melodies as well. Some really talented musicians have whopping careers as pure ‘composers’ whose songs are performed by other musicians. By and large, most of the famous Karnatic musicians have left their marks not only as performers but also as creators.

Most compositions in Karnatic music have three parts to their body. The first two lines of the song (sometimes just one) is called Pallavi. Like ‘Raghupathy Raghava Raja Ram…’ or ‘Roop Tera Mastana’ they occur over and over, especially after each stanza. Usually, the Pallavi is followed by two more lines (sometimes just one more). ‘Eeshwar Allah Tere Naam..’ in Raghupathy Raghava Raja Ram is an example. This portion is called Anu Pallavi. This is sung at the beginning for sure, but sometimes even during the end of the song, but not necessarily after each stanza. The stanzas of a song are called ‘Charanam’.

So, a song unfolds as follows:

Anu Pallavi
Charanam 1 followed by Pallavi
Charanam 2 followed by Pallavi
Charanam n followed by Pallavi
Anu Pallavi (optional)
end of song.

Typically, the Pallavi is set in lower tetrachord and in the lower octave (this is not a strict rule) and Anu Pallavi goes to the upper keys and to the next octave as well. Notice this in the song ‘Raghupathi Raghava.’. The Anu Pallavi, ‘Eeshwar Allah ..’ goes to higher frequencies.

Given all this, let us see how Karnatic music education is imparted to the students.


(Learning the simple elements – the ‘Kalpita’ aspects)
If you are a Karnatic music lover and if a good teacher happens to be around, try to LEARN Karnatic music formally. Seriously. You may not be gifted like some other people and your voice might sound like a vacuum cleaner (many famous Karnatic musicians have lousy voices) – and you may think you are ‘tone deaf’ or have no musical aptitude. But none of these should come in the way of your attempting to learn music. If people can be trained to learn foreign languages or to ski or to become software programmers, they can be eqully well trained in Karnatic music. Even if you don’t become a concert grade musician, you can learn enough to develop a deeper appreciation for music. Nothing like hands on experience.

In the olden (‘Gurukulam’ or ‘Gharaana’) days, music was handed down from generation to generation orally and the emphasis was heavily on memorizing and relentless practicing. (Even now practice is a major aspect of learning) These days, Karnatic music teaching has become more modern and streamlined and less painful for the students. Students are taught the theoretical and analytical techniques as well, instead of being forced to regurgitate what the Gurus teach.

The first thing that happens when you start to learn music is figuring out your characteristic octave. In Karnatic terminology this is referred to as ‘finding out one’s sruti – note that we are now using the word ‘sruti’ in a completely different sense. This is because: everybody has her or his signature octave, which need not begin at 240 Hz. The student is asked to sing out Sa – Pa – and then the sa of the upper (‘Tara’) octave. From this, the teacher extracts the range of the student’s octave.

After the octave is figured out, it is then ‘captured’ on a ‘drone’ instrument called ‘Sruti box’ or a ‘tanpura’. By ‘capturing’ we mean that these instruments are tuned to produce those three notes at the appropriate frequencies. If the student’s standard octave begins at 260 Hz, the sruti box will play the ‘Sa’ at 260 Hz, Pa at 367 Hz and the upper Sa at 520 Hz. Throughout the session, the ‘drone’ instrument will keep playing these notes. Although this may be quite annoying at the beginning, it is a necessary thing. Sometimes, after a lot of singing, people can ‘lose’ their ‘srutis’ and instead of producing the Sa at 260 Hz (or at whatever happens to be the person’s sruti) they can go completely off their octave and scales. At times like these, you can tune yourself afresh using the drone sounds of the sruti box and come back to your octave. (By the way, for the purpose of writing I am assuming the student to be a male. It can very well be a woman student too. No sexism implied) Not only do beginning students tend to lose their ‘sruti’ – even experts, sometimes in the middle of a big concert can go completely off their sruti and octave and could sound quite miserable.

If you happen to have a sruti box or a tanpura, it is a lot of fun tuning it to your octave or someone else’s. Try it. If you get an opportunity, jump on a concert stage and pluck away at a tanpura, especially if the musician happens to be a big shot. (Just remember not to fall asleep during the concert) The sruti box or the tanpura, by its constant droning also adds an element of harmony to the Karnatic music.

The first bunch of music lessons concern with the production of ‘notes’ – much like Julie Andrews teaching the Sound of Music kids about Do re mi. (there are corresponding Indian movie songs as well about Sa ri ga ma pa) The emphasis here is for the student to stretch his voice to produce different frequencies and at will. A student-friendly scale like the Mayamalavagoulai (Melakarta # 15) is chosen where Keys 1, 2, 5, 6, 8, 9, and 12 are the Sa, Ri, Ga, Ma, Pa, Dha, Ni notes. The upper ‘Sa’ is always added. (Thus we produce EIGHT notes when we sing, although the Ragam has only SEVEN notes. Among other reasons, the eight-ness makes it easy to time this sequence in a binary, Adi Talam which has an eight beat cycle) This particular Ragam is chosen because the notes are ‘spread out’ across the octave, instead of subtly varying, so that a beginning student can reproduce them easily. The notes are drilled into a student in a variety of excruciating exercises, so that at the end of the lessons, when asked, the student can utter a ‘dha’ or a ‘ri’ at the exact frequency. (of course, in his signature octave)

The initial lessons are purely on tone reproduction. Ragams and microtones come afterwards. Initially, the student only learns the solfege notation – a bunch of sa ri ga ma s – he is not taught any text or lyrics. There will be also exercises on keeping Talam.

Then the student graduates to the simple songs called Geetams – the nursery rhymes of Karnatic music. These are often set to mild Ragams – again, the facets of the Ragams are hardly emphasized at this point. The idea is to sing ‘in tune’ following the teacher. Usually the Geetams are only a few lines long and they do not adhere to the Pallavi-Anu Pallavi-Charanam structure always. The Solfege notes (Swarams) are taught and sung out first and then the lyrics (lyrics are called Sahityams in Karnatic music) are sung out afterwards.

The teacher might first go,

‘Ma pa da sa sa ri’
and sing a line of text to fit it.
‘Sreee Gana Naada’
(This can be fitted to a three beat cycle Roopakam, since there are six notes.) This is a Geetam in the Ragam Malahari, a derivative of Mayamalavagoulai and is usually the very first Geetam anyone learns.

If you think about it, there are several ways to express or sing out a ‘tune’ (1) using the lyrics or text (2) humming using syllables like ‘la la’ or ‘aaa’ or ‘whatever’. You will see that this forms the basis for what is called ‘Alap’ or ‘Alapanai’ and lastly (3) you can ‘decompose’ the tune into the corresponding notes and utter just the Solfege notation – this is called Swaram singing. This version of producing the tune is useful especially if you are learning to play an instrument where you don’t care for lyrics anyway.


Once the basics like singing in-tune, keeping Talam etc are mastered, the student is then ready for the more serious compositions. At this point, in addition to learning a particular song, the student is also trained to think in terms of Ragams etc. Even at this point, improvisation is not introduced. Faithful following of the Master is strongly emphasized so that pre-existing songs are reproduced the way they were meant to be – tune, Talam and all. However tempting it may be and however nicer it may actually sound, it is a big no-no for the intermediate student to introduce his own ‘stuff’ and distort the song. You cannot improve on a perfection like a Beethoven Symphony or a Thyagaraja Composition. (What is your reaction to this ?)

The student moves on to the next set of songs called ‘Varnam’s. A Varnam is a song with minimal amount of text. But most syllables will be stretched and twisted quite a bit and a lot of vocal acrobatics will take place. For example, a simple word like ‘Swamy’ might be stretched to several beats (in terms of time) and might go through several notes (in terms of frequency) and might actually sound like:

“Sw a a a a a a m e e e e e e ..”

The idea here is to teach the student a little bit about the Ragam structure, its characteristic phrases, microtonal variations and so on and of course, to be able to train his vocal chords and reproduce such features of the Ragam.

The Varnam is also an exercise in maintaining the Talam.

The Varnam consists of the Pallavi – Anu Pallavi – Charanam structure somewhat. Interspersed with this will be a lot of plain ‘notes’ or ‘swarams’ in the Solfege nonsense syllables. That is, there will be a lot of Sa, ri, ga, ma stuff sung out explicitly throughout the song. The structure of Varnam is approximately,

Anu Pallavi
Initial Swaram (called ‘Muktayi Swaram’)
Charanam 1
Swaram 1 Charanam 1 (These Swarams are called ‘Chittaswarams’)
Swaram 2 Charanam 1
Swaram n Charanam 1

There is only one ‘stanza’ (Charanam) so to speak. Even this stanza is usually only one line long. The basis for the introduction of Swaram in these songs is to teach the student the main phrases of the Ragam without obscuring the tune with lyrics. As the song progresses, from Swaram 1 to Swaram n, the Swarams get longer and more complicated. It takes longer for the student to learn the latter Swarams. All Varnams are pre-composed, including the Swarams.

Another interesting purpose of singing ‘Varnam’ is to practice it in the basic, double and even quadruple speeds to get the hang of the Talam correctly. (and also to push the student to the limit of his learning curve) Varnams are usually very bright and lively pieces. There are also slow paced Varnams called Dance Varnams, essentially used in dance recitals. Instead of being in just one Ragam, some Varnams are composed as Ragamalikai, which means it is a song, which uses different Ragams for its various stanzas. (Usually about three or four Ragams)

Then the student moves on to the mainstream Karnatic songs. These songs are called ‘Kritis’ or ‘Keertanai’s. There are subtle technical differences between the two terms, but we are not going to quibble. The Kritis follow the Pallavi – Anu Pallavi – Charanam structure. Usually, interestingly there will be just one Charanam or one stanza rendered in the Kriti. If ‘Roop tera mastana ..’ were a Karnatic music Kriti, you would only sing one stanza. Of course, there are exceptions. Also, there are songs or Kritis, which do not adhere to the Pallavi – Anu – Charanam format. Kritis are usually ornamental and like movie songs, they are developments of a theme. Even though they are set to a Ragam, their purpose is not bring out every aspect of the Ragam, but simply to produce a musical impact in a listener. So two Kritis set to the same Ragam might sound quite different, even though you might notice many similarities because they are set to the same Ragam. In fact, if you are not musically inclined you won’t even notice that two ‘kritis’ are in the same Ragam.

Sometimes, in a kriti, the first lines will be varied systematically as the singer repeats it over and over. (This happens even in some movie songs). That is, the Pallavi will be sung intentionally in different melodies as it gets repeated. The text of the line will be the same, but the music will be different. Such variations are called ‘Sangatis’. These variations or Sangatis are intended by the original composer (or cleverly introduced by some other musician somewhere along the line) and are parts of the song. If you listen to a Kriti like ‘Vatapi Ganapatim’ in Ragam Hamsadhwani, you will notice that the singer spends a long time mulling over the first line itself. But if you paid close attention to details, the singer is in fact varying the melody of that line over and over – in other words, the singer is producing the various ‘Sangati’s. Sometimes, after all the variations are done, the final version of the Pallavi could sound totally different from what one started out with. In fact, usually, the final version of the Pallavi is what is the official version of the Pallavi from that point on – and this is the version that is sung after each stanza (charanam) subsequently. Even Anu pallavi can have associated Sangatis.

By the time a student starts to learn a lot of Kritis, he or she is already at a reasonable level of expertise. By this time, the student knows how to decompose any musical phrase into its constituent Swarams. They could easily tell you what notes make up ‘Roop Tera Mastana’. At this point, the teacher no longer breaks down the lyrics (the text) into their constituent Swarams. (except when the musical phrases are very tricky) The teacher simply sings out the songs and the student tries to follow without making mistakes. It is no easy task, especially given that there could be subtle variations from Sangati to Sangati. The musical phrases, loaded with microtonal features, are simply reproduced without anyone trying to do too much analysis.


After several Varnams and Kritis are learnt, it is time for the student to learn to improvise – remember the technical term for improvisation in Karnatic music is Manodharma. At this point, it is also apt to call the student a ‘musician’!
In Karnatic music, improvisation is an ‘add-on’ to pre-existing music. In other words, you cannot distort a song like ‘Roop Tera Mastana..’, but you can ADD to it. There are several ways to improvise. Let us examine some of them such as (1) Alapanai (2) Neraval and (3) Kalpana Swaram.

Alapanai is essentially a free format humming. The purpose of the Alapanai is to bring out the total character of a Ragam. Alapanai is sometimes simply called ‘Ragam’ singing or ‘Raga Alapanai’. (It is unfortunate that we tend to use the same word to mean different things in music. If someone says Ragam, it could mean a scale like Shankarabharanam or it could refer to Alapanai. You have to know the context)

The Alapanai is a preamble to a kriti. For example, if the musician is going to perform a song in the Ragam Shankarabharanam, he or she would do an Alapanai just before the song. The musician would try to sing out as many characteristic phrases as possible to enunciate the features of Shankarabharanam. This is a place for him to show his creativity. He would use no texts or words; instead Karnatic musicians use nonsense phrases like ‘Thadhari na’ and so forth to hum out the tune. (Hindustani musicians use simple ‘Aaaaaa’) The Alapanai is not set to any Talam either. So there are no constraints on how long the musical phrases are. The Alapanai can be micro-mini, lasting just a couple of minutes and equally well, it can last nearly a half hour.

Just how exactly does one go about doing the humming and bring out the essentials of a Ragam? There is no clear answer to this question. However there is a method to Alapanai singing, even though there is no prescribed algorithm to perform it. The artiste usually starts out in the lower part of the scale of the Ragam, constructing short melodies in that subset of the octave. Then as the Alapanai proceeds, he would meander into the higher notes (this meandering is technically called ‘Sanchaaram’) – even here he might use just a subset of the scale – and eventually reach even the higher octave. Then he would make up tough melodies, essentially in the very high end of the scale and then would come down in scale slowly and make more complex melodies in the lower part of the scale and would grind slowly to a halt. By the way, this need not be the case in every Alapanai.

Listen to an Alapanai and see if you can identify when the musician is making melodies in the lower part of the scale, when he moves to the higher notes, when he makes complex melodies and finally when he asymptotically comes to a stop. My view is that an Alapanai should be sung before EACH song. It is always nice to elaborate the Ragam to be performed. Also, in a concert situation, an Alapanai lets a musician transition from one song to the next smoothly. Imagine an intense musician performing Shankarabharanam. He cannot abruptly move on to the next song set in perhaps Desh. Here a little Alapanai in Desh could smooth him into the new Ragam.

Even though the Alapanai is an improvisation, very few musicians would walk to the stage cold and start thinking about the Alapanai. The odds are they would have practiced at least some part of the Alapanai leisurely at home or would have rendered them in some other previous occassion. In fact, if you followed a musician from concert to concert, and if he performed an Alapanai in Shankarabharanam in both the concerts, you will see that he actually repeats a lot of phrases. A musician is not graded on how impromptu and extempore his rendition is, but more on what beautiful melodic phrases he comes up with in the Alapanai segment and if some of them are truly ‘out of the world’ phrases.

Neraval is just a technical term given for improvised variation of a line in a song. Typically any one line of the song is chosen and its melody is varied intentionally over and over. Remember, this is exactly the definition for the term Sangati, except that the Sangatis are parts of the original song, introduced by the composer – and they also tend to occur only in the first couple of lines of the song, namely in the Pallavi and Anu Pallavi sections. In ‘neraval’ you typically choose a line from somewhere in the middle and park out there. For example, you can do a Neraval in the song ‘Roop Tera Mastana..’ by choosing a line somewhere in the middle – such as ‘Ankhon se ankhen’ – and endlessly vary the melody to bring out the features of the Ragam. Remember also that since neravals are variations of a line in the song, one must strictly adhere to the Talam structure and the improvised melodies should last appropriate time intervals.

The third avenue for creativity is what is called Kalpana Swaram – or ‘creative Solfege note sequence formation’. Here the musician makes up Swaram sequences (fitted to the Talam, of course) in the Ragam. Sometimes the sequences can be simply mathematical (like Sa ri ga, ri ga ma, ga ma pa…) and some other times very interesting complex patterns.

These ‘Swaram passages’ are made up typically near the end of a kriti. The musician usually launches into “Swaram singing” from a particular point in the song, usually at the Pallavi, after the entire song is sung. At the end of each Swaram passage, he would come back to the Pallavi. The ‘landing’ back to Pallavi should be smooth. For example, if the Swaram sequence ended in the note Ni and the Pallavi started in Ga, there is a perceptible discontinuity and such things are not allowed. The Swaram sequence should more appropriately end in a Ri or a Ma to be able to smoothly go into the Ga of the Pallavi. Also, at the end of each Swaram passage, it is not necessary to repeat the entire Pallavi line. Sometimes just a word or two of the Pallavi is sung.

Of course, it is not necessary choose the Pallavi as the place to ‘home in’. Other lines, especially in the middle of the stanza are often chosen as the launching points for Kalpana Swaram blitz. Even though the Kalpana Swaram sounds very constraining (in terms of sticking to the Talam or finishing the Swaram sequence at the appropriate ‘note’) several performers earn a name for themselves in building up incredibly creative Swaram sequences. There are enough freedoms – for example, the Swarams can be in single speed or in double speed or higher, adding an element of tempo-building to the performance. Sometimes, the musician might make up a ‘structure’ for the Kalpana Swaram. He might start off with a long Swaram passage and as he progresses in the Kalpana Swaram, he might shorten his sequences to smaller and smaller phrases – and might eventually end up in just single notes – at which point, he would conclude his Swaram singing by launching into a long, final, climactic Swaram passage lasting several Talam cycles. Somewhat like a cadence. Usually, after the Kalpana Swaram, the song is concluded.

So, to summarize, let us see the structure of a Karnatic song (‘kriti’), in light of all the Manodharma or improvisation components.

Alapanai (optional and creative component)
Pallavi – Sangati 1
Pallavi – Sangati n
Pallavi – Final Sangati
Anu Pallavi and its Sangatis
Neraval of a line in Charanam (optional and creative component)
Rest of the Charanam, followed by (final Sangati of the Pallavi)
Kalpana Swaram (optional and creative component)