This is a continuation of the Ragas and Scales post covering why a raga based melody sounds so different from a western melody even if both can be said to be based on the same scale. I am comparing the raga Sankarabharanam with its western counterpart – the major scale, and covered a few obvious differences in the previous post. Now I will cover a couple more here. First one again is perhaps obvious, but the second is a very interesting and important difference.

Discrete vs. Continuous
In carnatic music, the gamakas not only adorn/embellish swaras with pitch modulations, but also perform the role as “links” between previous/succeeding swaras in the melody. When going from one swara to another swara that typically takes gamakas, it is common for this transition to be started with a slide as the initial part of the gamaka. Let us go back to the sample that played the Sankarabharanam swaras just straight up and down in Carnatic style:

If you listen carefully, you will find that the modulation of ri (the second swara being struck) in this case included a lead-in from the previous swara sa. The same applies for ma (4th swara), and for dha (6th swara). This sort of a movement where a “slide” is implicitly part of a swara that has gamakas is quite common although not mandatory. In my opinion, it helps in maintaining melodic continuity as you transition swaras – something which seems important in Carnatic music. Even when you have jumps like sa-pa (D-A), sa-ma (D-G), and smaller jumps, it is quite common for it to include a slide from the first swara to the next.

Looking at the flip side, in western music notes are held flat more often, a lot lot more often. While slides are indeed part of the music, they are not used as often, and not “implicitly” (as explained above) as in Carnatic music. So it is typical for a Carnatic music listener to perceive a western melody to have a “discrete flow”, and a Carnatic melody to have a much more “continuous flow”.

Skippin’ n Jumpin swaras/notes
The last difference I cover seems like an important key to the puzzle. I learned about it first in a DVD by Guitar Prasanna called Ragamorphism, where he deals with this same topic – how is a raga from a scale? He says that in a western melody, there will be lots of jumps between notes. By a “jump”, we mean e.g. a transition from D to F# in the major scale on D, and thus skipping the E note which is indeed part of the scale (note: for the majors scale on D, F is absent). In a Carnatic melody, a jump would e.g. be a transition from sa to ga for a raga that includes ri in the ascent. In a scale, you are free to jump from one note to any other note with aesthetics as being the only criteria. On the other hand, in a Carnatic melody, you have to honor the ordering of the swaras in the underlying raga structure. Prasanna mentions that this does not mean that jumps are disallowed in Carnatic melodies – just that in general, order must be followed. Prasanna then demonstrates by first playing something that is based on the major scale, and then something in Sankarabharanam.

That section from that DVD was a very good clue to an answer I was searching for a while. It wasn’t a complete answer as Prasanna does not explain much beyond that i.e. what is meant “general order must be followed”. He just let his guitar do the talking. I delved into this a bit more and I think I am now able to understand it better.

Western music does have more jumps compared Carnatic music even though the latter Carnatic music does permit jumps – sa-pa, sa-ma, ga-da, ri-pa, ni-ri, da-ri etc. all figure regularly. So what does “Western music have more jumps that Carnatic” mean? This can be summed up by two points:

  1. Two consecutive jumps are quite rare in Carnatic music. An example of this is D-A-E in major scale, where D-A is a jump followed by A-E which a jump. In Carnatic, this would e.g. be ni ri n or sa ga pa say in a raga which includes all swaras in ascent and descent. I will refer to these as Double Jumps from now on.
  2. The # of jumps itself is not that high in carnatic music. Hence, the probability of a swara being either preceded or followed by one of its neighbouring swara in the raga structure is very high. So for example, if we take ri in Sankarabharanam, the chances of it occurring as one of sa-ri, ri-sa, ri-ga, ga-ri is very high. In other words, “general order of the swaras in the raga structure” is followed.

Note that by ‘jump’ here we also include jump from one note/swara to the same note/swara up or down an an octave.

But how rare is rare and how high is high? How does it compare to western music? We of course need some metrics that allows us to see whether there is any truth to the above assertions. So here it is: I analyzed the melody lines of a few Western and Carnatic pieces, tabulating how many note/swara transitions there are, how many of them are double jumps, and how many are single jumps, and thus how often is the next note a neighbor (as per scale/raga structure). Here are the results:

Name Transitions1 Double Jumps Single Jumps Next / Previous
is neighbour
Pastorale – Domenico Zipelli 119 53 (44.54%) 7 (5.8%) 59 (49.58%)
Minuet in G – Mozart2 92 44 (47.83%) 6 (6.52%) 42 (45.65%)
Minuet in G – Bach3 111 7 (6.32%) 13 (11.71%) 91 (81.97%)
Gavotte – Arcango Corelli 69 20 (28.98%) 4 (5.8%) 45 (65.22%)
Carnatic – Sankarabharanam raga based songs
Sami ninne – Adi tala varnam 331 4 (1.21%) 50 (15.1%) 277 (83.69%)
Chalamela – Ata tala varnam 499 11 (4.01%) 64 (12.83%) 424 (84.96%)
Carnatic – other ragas
ninnukOri – mohana raga varnam 271 3 (1.11%) 17 (6.27%) 251 (92.62%)
evvari bodhana – Abhogi raga varnam 290 3 (1.03%) 17 (5.86%) 296 (93.11%)
vanajaskhiro- kalyani raga varnam 345 11 (3.19%) 38 (11.01%) 296 (85.80%)
era napai – todi raga varnam4 458 44 (9.61%) 61 (13.32%) 353 (77.07%)

1 – Does not include transitions involving accidental notes in western pieces (very few if any for those shown).
2 – Minuet in G by Mozart involves key shifts from G-Major scale to C-Major Scale. The pieces also contain a few accidentals.
3 – Minuet in G by Bach involves key shifts from G-Major scale to C-Major Scale. The pieces also contain a few accidentals.

In the above table,

  • Transitions is the total number of note/swara transitions or change of note/swara including jumps to the same note in a different octave in the song/melody (a repeat of same note is not a transition: B C C D D or S R R G G would contain 3 transitions). In case of western pieces, this also does not include transitions involving accidental notes. The pieces shown above have very few accidentals (if any).
    • Also, it should be noted that only the melody of the treble clef was considered. This does distort the picture a bit as the bass line melody could be “compensating” for some jumps. However, my guess is that this is not a significant distortion
  • Double jumps is the # of transitions that are part of two consecutive jumps as explained above.
  • Single Jumps is the # of occurrences of a transition from note/swara X to note/swara Y that skips over one or more intervening notes/swaras as defined by the scale/raga.

I should note that the western pieces picked are not exactly a representative set. They are just the ones my colleague gave me “as a start” when I asked her for some pieces which are in the major scale, did not use accidentals much and stuck to one key. Judging by the number of transitions between the western pieces and the carnatic pieces, we can see that perhaps more complex ones from western should have been picked. However, it is also possible that in more complex western pieces, there would be more accidentals, more key/scale changes, and also polyphony, making a comparison less meaningful.

What the data means
As you can see from the table, the varnams involve a lot of transitions, which I think is not that unique to varnams, and would be typical for all carnatic melodies. In spite of the large number of transitions, these melodies involve very few double jumps. Looking at the last column, the percentage of time that a swara is preceded/followed by the neigbouring swara in the raga structure is quite high – 80% or even higher. This I believe is a key characteristic of how a Carnatic raga uses its underlying swara structure. This seems to apply to all ragas although it is easier to quantify for ragas that have a symmetrical structure.

You may notice that even among the Carnatic set, the mohanam and the abhogi varnams have very very few jumps, with the last column showing a figure > 90%. This could be because these ragas have pentatonic structure – this is just my guess. Also you may notice that the todi varnam has a much higher percentage of double-jumps compared to other songs. A large portion of the varnam has the pa (fifth) is completely absent. Skipping pa (but done judiciously) is not uncommon for that raga. Perhaps symmetrically, it is also not uncommon for the sa (tonic) to be skipped. All this allows for many M-D-M or D-M-D, and N-R-N occurrences, and thus perhaps a high jump %, and a lower (<80%) last column figure.

In the western side, Bach’s Minuet does have “Carnatic like” usage of the scale implying that even this is possible in western music. However, all others involve significant # of double jumps, which I suspect is more common.

However, an analysis of a bigger sample set with more representative pieces may be needed before we can confirm this.


So there you go. There are several factors that make a carnatic melody sound very different from a western melody based on the equivalent scale. You have

  • Gamakas
  • Continuity between swaras
  • Fundamental difference in how the basic structure is used in terms of avoiding double jumps, not too many jumps, and thus honoring order determined by the raga structure.

Have you ever wondered why sometimes a film song that is supposed to based on raga has only a faint hint of that raga? We all know film songs can be like that, but how/why are they different? Why can’t they resemble Carnatic songs in that raga more? Well – it could the absence of one or more of the above. Besides it may employ accidentals or foreign notes and also change keys change like western pieces. So film songs that are loosely based on a raga are sort of like western pieces. However, they are typically much closer – they may include one of the above.

That’s it folks. Hope you found the Ragas and Scales comparison interesting.

The end