Finding sruthi...

vikram_venkat · Post by **vikram_venkat** » 06 Jan 2006, 17:17

Hi,
On listening to a song, or just a piece of melody, we are able to sing the 'sa' pa' 'sa' for that song..(sruthi)(assuming there is no tambura or chords playing in the background...) Can anybody define how we do that?? or, in other words, how can we find the sruthi(or find 'sa'), given the western notes of a song??(like some....F#GCD......)

-Vikram

vnbharadwaj · Post by **vnbharadwaj** » 08 Apr 2009, 03:59

Personally, I do not think "we" referred to by you, includes everybody. Only people who have inbuilt music in them can do that

vasanthakokilam · Post by **vasanthakokilam** » 08 Apr 2009, 05:37

Interesting topic. We discussed that before in some other thread but I did not get a good hang of it. Hope those in the 'know' will educate us and give us practical tips so we can do it ourselves. ( possibly with some music samples - non commercial ).

mohan · Post by **mohan** » 08 Apr 2009, 06:05

Sruti is a musical reference point. In Carnatic music we usually set the sruti then phrase the music around it.
The reverse is also possible - for example, when a decent musician sings we hear the raga played and use the various notes as reference point, then locate the Sa (and sometimes Pa).

Since our Carnatic trained ears are tuned to hearing the sruti as a reference point, when the sruti is lacking some confusion can arise. In some cases we may be listening to a Mohanam but it is really a Hindolam because different listeners have latched on to the incorrect reference point (sruti-bedham effect).

Identifying the actual sruti or Western tonic note (C, F# etc) is a totally different exercise and only extremely gifted people with so called 'perfect pitch' will be able to identify the actual frequency of the base sruti.

VK RAMAN · Post by **VK RAMAN** » 08 Apr 2009, 06:10

Those who can understand when some one is not in syc with the sruthi and those who can set their own sruthi should be able to identify the sruthi of others singing too. It is a learnt technique with the exception of some who have in built music in them.

srikant1987 · Post by **srikant1987** » 08 Apr 2009, 06:16

mohan wrote:In some cases we may be listening to a Mohanam but it is really a Hindolam because different listeners have latched on to the incorrect reference point (sruti-bedham effect).

This is very wild indeed! I would expect it to be confused with shuddha saveri, if at all. Hindolam has a vastly different mood.

vasanthakokilam · Post by **vasanthakokilam** » 08 Apr 2009, 07:15

I think the original poster's question is not about detecting sruthi in a western music piece but given a CM melody line without any tonic reference (sruthi box sound), what are the clues people use to latch on to the tonic. I agree that sometimes people latch on to the wrong one. Even in such wrong cases, people do not latch on to all possible wrong tonics. So there seem to be some clues that we intutively use to latch on to the sruthi. What are those?

Vijay once mentioned about 'intrinsic tonic'. I think this is what he is also referring to. How do we perceive that intrinsic tonic?

srikant1987 · Post by **srikant1987** » 08 Apr 2009, 07:23

Oh yes intrinsic tonic.

Sometimes as I do something else, I might just burst into a hum. Upon repeated observation I find there's SOMETHING, most of the times a fan, producing the tonic in which I'm humming.

arunk · Post by **arunk** » 08 Apr 2009, 07:53

In a cm piece, if you get the raga (and assuming you got it right), you have subconsciously determined the tonic i.e. sa - right?

Arun

mohan · Post by **mohan** » 08 Apr 2009, 07:57

arunk wrote:In a cm piece, if you get the raga (and assuming you got it right), you have subconsciously determined the tonic i.e. sa - right?

That is right but if you got the wrong raga you have may have gotten the tonic wrong.

srikant1987 · Post by **srikant1987** » 08 Apr 2009, 08:03

The great proportion of gamakas in CM at least these days ensure that if we get a raga wrong, it's most commonly due to some other reason than a shruti bhedam we unknowingly perform on it.

bilahari · Post by **bilahari** » 08 Apr 2009, 09:08

Srikant raises an interesting point in post 8. When all of us hum music, what tonic do we use? Personally, I find myself humming always in D/D# which is the shruti I've been using on violin for 2 years now and it's latched on in my mind. How a tonic can get internalised is a mystery to me...

vasanthakokilam · Post by **vasanthakokilam** » 08 Apr 2009, 10:40

Mohan, in the wrong case, they have latched on to the wrong tonic. Yes, but there is still something intrinsic in it that they latched on, albeit wrongly. I think there is something valuable in looking into both right and wrong latch. ( We are still talking about cases where there is no external sruthi reference ). Is it the melodic contour that through pattern matching gets us to the raga directly ( right or wrong one ) which then gets us to the tonic?

mohan · Post by **mohan** » 08 Apr 2009, 10:42

In one of the quizzes that DRS had set long a go there was an example where many members got one raga and the others got a sruti-bedham equivalent.

srikant1987 · Post by **srikant1987** » 08 Apr 2009, 13:18

There's one classic case I can think of: hindolam and shuddha dhanyasi, merged (?) together in mArgazhi pUvE of May maadham.

The thing is, if you make hindolam's madhyamam (which is an IMPORTANT note for it, and FLAT note most of the times, indeed afaik) and make it your shadjamam, you'll get shuddha dhanyasi. The panchamam of this shuddha dhanyasi is the shadjamam of hindolam (both unshaken notes).

And we've also STANDARDISED gamakas for sadharna gandharam and kaishiki nishadam. (This was talked about in a lecture I attended.)

So we just have to look for a flat madhyamam of shuddha dhanyasi as against the (std-) gamaka nishadam of hindolam.

bilahari · Post by **bilahari** » 08 Apr 2009, 13:20

VK, for me, personally, it's the melodic contour. For instance in Uday's first quiz, even though there was no shruti reference, the initial phrase of SuruTTi was totally unambiguous and that's what I latched on to. Of course, once I knew where the Ni was by the end of the phrase, orienting myself became easier thereafter.

uday_shankar · Post by **uday_shankar** » 08 Apr 2009, 14:52

mohan wrote:
arunk wrote:In a cm piece, if you get the raga (and assuming you got it right), you have subconsciously determined the tonic i.e. sa - right?
That is right but if you got the wrong raga you have may have gotten the tonic wrong.

Both Arun and mohan are profoundly right.

So why does somebody get the wrong rAga ? There are several possibilities:

1. Got the tonic/shruti right but made a trivial mistake in identification due to insufficient knowledge of raga nuances or insufficient distincitive phrasings in the piece listened to. For example got confused that something was sriranjani instead of ritigowlai. or simhendra madhyamam vs shammukhapriya.

2. Got the tonic wrong but identified the "correct" rAga in the wrongly perceived tonic/shruti ! For example, as mohan said Hindolam vs mohanam vs shuddha saveri vs suddha dhanyasi vs madhyamavati.

In any alapana, it is the responsibility of the performer to bring out the rAga swaroopa ASAP and only then embark on swara based virtuosity display (if at all). So both the mistakes stated above can often be laid squarely at the doorstep of the performer. Usually each rAga has distinctive phrases, kampita gamakams, jaarus, kaarvais, etc.. which cannot be associated with the shruti bhedam cousin, what bilahari calls "melodic contour". For example, see what happens if we played a simple phrase G, P G ~ ,, R~, , S, , in mohanam. The ~ character is supposed to indicate kampita gamakam on the note. So here's what would be the result in the cousins:

G P G~, R~, S,, - Mohanam (Nice mohanam phrase)
S G S~, N~, D,, - Hindolam (Huh! Ever heard of a kampita gamakam on S in any raga or N in hindolam ?! )
D S D~, P~, M,, - Shuddha Saveri (Huh! Ever heard of a kampita gamakam on P in any raga or kaarvai in M in shuddha saveri ?)

I guess you get the jist...feeling too lazy to type out the rest. So that simple phrase can only be mohanam.

Performers who have migrated to the far end of scale-based rAga dileneation such a L Subramaniam may be particularly guilty as charged

. For example of If LS plays dharrmavathi it could well be sarasangi. Check this out sometime when you get a chance to listen to an LS alapana.

bilahari,
BTW, my suratti example did have the tonic loud and clear. You just didn't notice it consciously (but must have registered it subconsciously) but simply latched on to the disctinctive suratti-only beginnig phrase RMPN~, N~. However, it wasn't suratti enough to several others (they were fresh off thinking it was GPDS(nS),S(nS), in mohana kalyani

) and hence a guitar redo. Lesson learned.

Suji Ram · Post by **Suji Ram** » 09 Apr 2009, 03:52

bilahari wrote:Srikant raises an interesting point in post 8. When all of us hum music, what tonic do we use? Personally, I find myself humming always in D/D# which is the shruti I've been using on violin for 2 years now and it's latched on in my mind. How a tonic can get internalised is a mystery to me...

how true, I latch on to E for humming (my violin setting) That Sruti keeps ringing in my ears...

vijay · Post by **vijay** » 10 Apr 2009, 14:05

I am not sure there is a great difference between identying ragams through notes and finding sruthi. First you find the base sruthi of the piece and compare it with your own reference sruthi (which you generally hum songs in - this tends to be quite fixed for me although I have never tried to figure out what exactly this is - perhaps B or B. as per Radel tambura). So the only challenge is to map the base srtuhi of the piece with your own inetrnal reference point.

Of course this is a circuitous way of identifying pitch (as opposed to gifted/perfect pitch cases) but over time I am sure the skill can be acquired, if one persists - don't think you need to be extraoridnarily gifted although an ear for notes is a must - not that I can do it myself though..

On intrinsic reference points, I have earlier made the observation (as VK recalls) that it is not always possible in western/quasi western music. In fact some scaled used by ewstern composers exceed the octave! However, the distance between a specific notes can be mapped on to one's reference point by the trained ear to estimate the picth of a particular note. This can be then used to extrapolate the pitches of othre notes in the phrase/melody.

Of course all of the above refer to western music. The process in CM/HM is much simpler and one can easily use characteristic phrases/gamakas to get the raga as Uday has pointed out above.

vasanthakokilam · Post by **vasanthakokilam** » 10 Apr 2009, 19:43

Let me summarize and see if this is correct.

R - Set of all ragas

A sequence of notes are sounded in the carnatic idiom. The artist has a raga in mind and his/her best in bringing out the raga in that sequence of notes. No external sruthi reference is provided. How does a CM rasika latch on to the tonic. Here the rasika is just an avid knowledgeable CM listener but not well versed in swara identification etc.

P - A subset of Ragas that are possible for this sequence for all possible tonics in the octave purely based on scale.

P1 - Many ragas from P that may not be in vogue. P1 represents a subset of P that are ready for melodic pattern match in people's minds.

T - When people apply the melodic pattern match in the CM context to identify the raga for the above piece, the built in knowledge of distinctive phrases, kampita gamakams, jaarus, kaarvais etc. come in to picture. T is a subset of P1 based on those considerations. ( as per Uday's post above ). The size of set T can be greater than 1 as this thread and the thread on sruthi bedham amply demonstrate

S - The listener intutively collapses the set T to a singleton S ( due to a variety of personal reasons including mood, personal experience etc.. ) . Once that is gotten, the tonic comes out as a side result of this collapse to single raga ( i.e raga identification ).

If this is all correct, then tonic identification seems to be a derived property in the raga identification process rather than the primary thing. We get to the tonic through the ragas rather than the other way around. So in way, we non-note identifying rasikas are living in the grama-murchana times though the underlying musical foundations of the practioners have changed!!

Is this a reasonable summary of what has been discussed above?

arunk · Post by **arunk** » 10 Apr 2009, 20:09

vasanthakokilam wrote: S - The listener intutively collapses the set T to a singleton S ( due to a variety of personal reasons including mood, personal experience etc.. ) . Once that is gotten, the tonic comes out as a side result of this collapse to single raga ( i.e raga identification ).

How would you eliminate the possibility that this is not based on intuitive (subsconscious) idea of where the tonic is?

Just like phrase identification (without knowing what the swaras) is intuition based, one could say tonic is also intuition based - both are in the mix.

Let us say a listener who has no swara knowledge was able correctly identify a single swara (intonated with gamaka) as kAmbhOji and Abhogi - there is no tampura. If that is too hard to believe, say it was two swaras. He finds that this happens to be the da (say ma-da in AbhOgi, and pa-da... - both on the way up).

Cant one then claim that he has subconsciously, automatically placed the swaras relative to a tonic, which he didn't hear? Certainly, his explicit awareness of the tonic may come later on it being explained by someone, but even in the process of identifying this swara correctly, cant one argue that in his sub-conscious, he has awareness of the tonic? Remember that the user's knowledge-base is from listening to these ragas by various artists i.e.. various tonics. So his pattern matching is not a strict absolute match (i.e. a straight-forward memory recall). There is some behinds the scenes intelligent processing. That could involve taking into account a tonic (at a deeper level).

Arun

vasanthakokilam · Post by **vasanthakokilam** » 11 Apr 2009, 12:58

Arun:

>Remember that the user's knowledge-base is from listening to these ragas by various artists i.e.. various tonics.
>So his pattern matching is not a strict absolute match (i.e. a straight-forward memory recall).

That is a good point and it does complicate matters! But this point has been very helpful. After considerable debate within myself ( how much fun is that!! ) I have a possible solution/resolution that is based on this subconscious tonic idea. Or rather a hypothesis on what that subconscious process is.

Let me indulge you all in my tortuous thought process before presenting the hypothesis.

At time T1 if the melodic contour using frequency set F1 constructed by the characteristic note sequence, gamakams, jaarus, kaarvais etc. can only uniquey belong to one raga, then raga reveals itself directly and tonic becomes a derived property. Someone who cares to find it may do so by extending the melody along the raga lines towards the tonic. At time T2, the same melodic contour but this time using frequency set F2 still matches the same raga without any reference to the tonic. So the melodic contour similarity is what someone is latching on to.

But if two different ragas can match that melodic contour, then the tonic comes into play as a tie breaker. But I would think that except with rare individuals, both ragas are not going to be simultaneously perceivable without some conscious effort. Consider two cases: At time T1, person A intuitively perceives raga A and person B intutive perceives raga B. Second case is person A at time T1 perceives raga A and at an entirely different time T2 perceives raga B. This all seems to point to some intelligent pattern matching to directly latch on to the melodic contour. A memory recall usually is a dictionary look up ( IN: sounds Out: A classified shape ) plus some semantic reasoning. Our brains have evolved for quick conclusions, whether it is right or not, consistent or not, but it needs an answer. The "semantic" part of semantic reasoning is highly infuenced by the need for that decision/conclusion and it can vary based on the environment, internal states etc. ( I do not think this is all well understood ). No matter what, what it needs to do is to classify the input to a normalized structure and somehow fit in with past experiences. Shape similarity is something we do with other inputs ( like visual ) all the time. Two shapes offset by a constant are classified to belong to the same class if the need at that time ( emotional or physical, fight or flight etc.) for that classification is not the actual magnitude but just a match against a previousy learnt shape. ( different sized tigers with different stripes all mean the same.. Just run for your life! )

If we take the visual example and apply to auditory input, then a case can be made for the 'subconscious tonic'. If we look at a picture without much contrast or background, brain struggles with it a bit and decides on a shape. How does it do that? It has learnt that shape before. But while learning that shape, it had enough contrast/background. But when presented with a shape without the same contrast/background, in its search of memory, it has to substitute various backgrounds against this background-less input and pull out a learnt shape. If we play tricks with the brain by providing ambigous foreground and background, brain will detect different shapes and figures depending on the currently perceived background. It can actually switch fairly rapidly. ( the old woman / young woman picture trick is played through these kinds of ambigous foregound/background thing ).

Applying it to the auditory stimulus, specifically in raga learning and recall, the tonic plays the role of the contrast/background. The shapes ( ragas ) are learnt and internalized when the tonic was present. But when presented with an auditory stimulus without the background ( tonic ), it tries for a match by overlaying it with various backgrounds ( tonics ) and finds the first one that matches.
So this subconsious tonic is that search struggle it goes through by taking the input sounds and overlaying it with different tonics to find that shape it had learnt before. This will also explain why different people's brains might come up with a different answer for the same input since their past listening histories are different. This is an idealized description/hypothesis of what may be happening but something very complex should be going on in there since it has to normalize on quite a few parameters to find that match by using various background tonics.