Pitch Analysis - A fun tool to investigate Shruti in CM

Suji Ram · Post by **Suji Ram** » 24 Jan 2008, 08:00

cmlover wrote:Very nice Suji!
I literally had a treat listening to the audio gamaka with the visual.
The GPGP sounds so sweet when one can see it happening. I will run it on the software and play the selection to follow closely. (Now I am playing them separately!). The subtle GMGM which lasts less than 0.5 secs can be relished using the visual. These fine gamakas will be missed when you listen to the music fast but the beauty of CM is in these fine prayogas. Thanks to you. Just noticed that the climb between 12 to 13 shown in the graph is not noise but 'nOkku' for the Pancamam to follow is very elegant. This is a great educational software.
I would love to see 'pitchappa' execute the same phrase vocally!

Thanks CML,
The climb from 12 to 13 is actually D- from P
It would be nice to take up a familiar varnam with gamakas and see the pattern.

cmlover · Post by **cmlover** » 24 Jan 2008, 08:59

I agree! But the audio has to be very good to microanalyze the gamakams. Noisy ones will mislead us, like the MS 'ga' until Pitchappa gave us a noise free sample!
It will be nice to build a glossary of different gamakas in different ragas with visual representations.

vasanthakokilam · Post by **vasanthakokilam** » 24 Jan 2008, 11:27

Maybe I should wait for Uday to recreate the sample visually with audio!

Yup, let us wait for Uday.

vasanthakokilam · Post by **vasanthakokilam** » 24 Jan 2008, 13:52

Good work Suji. As CML said, good reference material for getting acquainted with the tool.

I also learnt that 'ma' of raghurama as 'pa ma pa ma' though yours sounds nice as well. I looked at Hyd Bros version of this song ( from MIO) and their swings are even wider.

uday_shankar · Post by **uday_shankar** » 24 Jan 2008, 17:43

vasanthakokilam wrote:'pa ma pa ma'

No. The gamaka is P-G3 P-G3. That's the kampita version of M1. In fact, on the flute it is played identical to the G-P-G jaaru of mohanam that I described earlier in the thread

. Such is the weirdness of CM.

I'm skipping town for a few days. Don't know if I can record anything before I leave tonite. If not, then monday. Sorry.

cmlover · Post by **cmlover** » 24 Jan 2008, 21:11

Uday
Can we all collaborate on the project of producing a visual glossary/dictionary of important gamakas. You provide the clean sample and we will do the visuals. No need to hurry, we have all the time in the world (except perhaps I have much less

Each gamakam should be fullydescribed in context and annotated. Variations also may be documented. If you provide the Flute version (as well as the chitravina version) Suji will folloowup with Violin version and our Pitchappa willl attempt the vocal version. If Ananya also can produce a natural abhinaya interpretation then we are in business! Let 2008 be a banner year for CM!
Folks! Think about it! More ideas welcome!

cmlover · Post by **cmlover** » 25 Jan 2008, 01:24

OK! Here is a gamakam quiz. Not that I know the answer!
I can tell you the flute is at 1 kaTTai (normalized by me). Try to figure out the raga. (I have the answer). Here is the audio.
http://www.sendspace.com/file/qu9ggi

You will hear the phrase executed by Flute followed by the violinist as close as possible. Try to figure out the phrase (I do not have the answer!). Now look at the visual picture

Can we get the gamakas on the notes ?
To help in the calculations here is the shruti chart.

We can call on Pitchappa later

arunk · Post by **arunk** » 25 Jan 2008, 02:02

cmlover - i think you may have got the shruthi incorrect here - based on what I identify the raga here. I believe the flat note at beginning and end of each sections is neither pa nor sa.

Arun

cmlover · Post by **cmlover** » 25 Jan 2008, 02:40

No Arun! I heard the whole song and got the shruti from that using your algorithms. The actual shruti was 2 kattai for the flute. Using the computer I transformed the shruti to 1 for ease of calculation and identification. Perhaps that has introduced some artefacts but the gamakas will not change! Could you guess the raga?

The nyaasam need not always be sa or pa, is it not?

arunk · Post by **arunk** » 25 Jan 2008, 02:59

I may be wrong but it sounds like bahudAri, something like g ... (violin) (flute) g m p d n~(s) n p(d)(p) m g... (and then violin)

Per this, the starting long note is 435-440 range (i had to use the pitch detector - wasnt clear enough from your graph) and in your table that is da2, but it is ga3. That is why I thought your shruthi here seemed off.

Btw, the flute ga3, but at 880, however it ends around 440-450. Atleast as per pitch detector.

Arun

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 04:08

I also doubt it is 1 kattai. Even going by the overall tone of the flute sound, It does not sound like 1 kattai. Normally, unless the artist used that big long flute ( the so called sindhu bhairavi flute

), the flute sruthi ( before your normalization ) would be 4.0 to 6.0

cmlover · Post by **cmlover** » 25 Jan 2008, 05:46

Arun
You are right, it is bahudari (I expected Suji to beat you

Here is the original before I pitch reduced
http://www.sendspace.com/file/ygwhb1
You can see that it is at 2 kaTTai relative to the earlier one. I may be off on the shruti calculation based on listening to the whole. But not as bad as what VK is thinking

Still I am quite weak in that area

I will check on your deciphering of the phrase....

arunk · Post by **arunk** » 25 Jan 2008, 06:17

cmlover - It looks like you have raised the pitch and not lowered it by 1 whole tone (i.e. 2 semitones or 2 sthanams or 2 increments of kattais). This is what I perceive aurally when I listen to both one after another.

The first one (i.e. the one you changed) is at pitch F I think Ga3 (the most obvious swara where the tune rests on in many places) is A, and so Ga2 is G#, R2 is G, R1 is F# and Sa is F i.e. 4.

So original was probably D# i.e. 2.5 kattai (F is 4, E is 3, D# is 2.5, D is 2, C# is 1.5 and C is 1). That I believe is common for flute (?)

Arun

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 06:27

If the starting and ending note of the flute is GA, then that G# makes the kattai 2.5. So CML, you are right, you were not that far off, just half a kattai difference. It sounded too high a pitch to me ( on an absolute basis ) compared to a normal low kattai long flute. Anyway, I will let the experts comment on this.

CML, this is actually interesting psychoacoustical experiment, the clip being short, without any external reference to establish the Sa. I automatically kept going to the perception that the nyasa swara is either a 'Ga' or 'Ma'. Absent any background drone, why is that? It could have been 'Sa'. But even in isolation, that initial violin sound did not seem right if I mentally construed that to be the 'Sa'. Do the rest of the notes ( which I did hear before hearing that initial portion in isolation ) leave some residue to make the brain perceive it as Ga and not 'Sa'?

arunk · Post by **arunk** » 25 Jan 2008, 06:30

G# => kattai 2.5 (??)

arunk · Post by **arunk** » 25 Jan 2008, 06:32

Btw ga3 is an important swaram in bahudAri

Arun

cmlover · Post by **cmlover** » 25 Jan 2008, 06:38

Arun/VK
My profound apologies. You are absolutely right; the original was 2.5 kaTTai (D#) (as I have now checked carefully!) and due to my oversight (?stupidity

) I had raised by two semi tones. In other words the shruti had become 3.5 (which does not exist

. Hence your guess and calculations fall in place. I was indeed wondering about the nonadherence of the pitchplot with Bahudari (would have even jumped to the silly conclusion of shruti lapse

(God forbid!)
Congratulations to your acumen and thanks for reslving the mystery for me!

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 06:57

arunk wrote:G# => kattai 2.5 (??)

Well, I really meant to say, if G ( and not G#, that was my mistake, pl. see below ) is Ga, then it will be 2.5 kattai. Does that make sense?

BTW, in the graph, the nyasa swara actually maps closer to G# than G for the flute portion and closer to G on the violin portion Did I read that right? This is more to do with the tool than the music. The piano keys are laid out in such a way that the white keys are wider than the black keys. That gives the impression that there is a wider range for the swaras for the white keys than the black keys. That does not sound right but that is how the tool displays the Y axis.

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 06:59

I had raised by two semi tones. In other words the shruti had become 3.5 (which does not exist

In Kattai Arithmetic, 2.5 + 1/2 + 1/2 = 4

cmlover · Post by **cmlover** » 25 Jan 2008, 07:09

Perhaps this revised table will now help!

My apologies (since shruti 3.5 does not exist, I just interpolated!)

cmlover · Post by **cmlover** » 25 Jan 2008, 07:14

vasanthakokilam......
In Kattai Arithmetic, 2.5 + 1/2 + 1/2 = 4

The new math

musically!

arunk · Post by **arunk** » 25 Jan 2008, 08:01

vasanthakokilam wrote:
arunk wrote:G# => kattai 2.5 (??)
Well, I really meant to say, if G ( and not G#, that was my mistake, pl. see below ) is Ga, then it will be 2.5 kattai. Does that make sense?

Yes. However, in this case Ga was actually closer to A (i.e. the cmlover modified sample). -It was around 435 Hz (Remember A above middle C is 440) Analyze cmlover's first sample in sonicvisualizer and see the individual points of the pitch detector to see that the values are closer to A. The graph that cmlover posted is a bit misleading - it doesnt show enough resolution, and the y-axis markers are a bit ambiguous.

arunk · Post by **arunk** » 25 Jan 2008, 08:02

cmlover - no problem. No need to apologize and all

We all make mistakes!

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 08:03

CML, I am confused. No interpolation is needed. When you take a 2.5 kattai song and raise it by two semi tones ( for 4 kattai ), these are the frequencies.

175 S
185 R small
196 R big
208 G small
220 G big
233 M small
247 M big
262 P
278 D small
294 D big
312 N small
330 N big
350 S

Multiply by 2 for other octaves.

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 08:09

arunk wrote:
vasanthakokilam wrote:
arunk wrote:G# => kattai 2.5 (??)
Well, I really meant to say, if G ( and not G#, that was my mistake, pl. see below ) is Ga, then it will be 2.5 kattai. Does that make sense?
Yes. However, in this case Ga was actually closer to A (i.e. the cmlover modified sample). -It was around 435 Hz (Remember A above middle C is 440) Analyze cmlover's first sample in sonicvisualizer and see the individual points of the pitch detector to see that the values are closer to A. The graph that cmlover posted is a bit misleading - it doesnt show enough resolution, and the y-axis markers are a bit ambiguous.

Yes, perfectly. When I wrote that, I was looking at the unmodified original which was 2.5 kattai ( hence Ga is G ) and you are referring to the modified one which was 2 semi tones higher and hence Ga is A. Sorry for switching bits on you!! I think it all reconciles.. true?

arunk · Post by **arunk** » 25 Jan 2008, 08:11

cml: If sa is 170, its harmonics would be 340, 510, 680, 850. You were showing 170,340,680 and thus listing only even harmonics (1, 2, 4).

For 175, it would be 175, 350, 525, 700 etc.

nth partial = pitch_freq * n

Arun

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 08:13

Arun, I think CML is showing octaves and not partials.

cmlover · Post by **cmlover** » 25 Jan 2008, 08:25

VK
As I understand 2.5 kaTTai +2 semitines = 2.5 + 1 whole tone which will be still 3.5 kaTTai only which does not exist since there is no kaTTai between E and F (which is one whole tone. However software can interpolate and hence I interpolated. Am I wrong?
Or do I understand that 3.5 is already embedded in F and so we can use F as the aadhaara Shadjam?

Arun
I am going only by octaves!

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 08:29

CML, The fact that in our kattai convention 3.5 does not exist does not mean anything. It is just a numbering convention so it maps nicely to the keyboard key layout. Each successive key is still a semi-tone apart. So the interval between 3 kattai and 4 kattai is only a semitone.

I offer below the full table I just did in Excel. This is for the first octave. Multiply by two for other octaves successively. And as you can see, I am using the rather unsophisticated, small-big naming convention to keep things simple. The frequencies are rounded.

arunk · Post by **arunk** » 25 Jan 2008, 08:31

ah - octaves ok.

cmlover - no no. Between E and F is still only 1 semitone.

C C# D D# E F F# G G# A A# B C

So you see B-C and E-F are like semitone apart with NO sharps in between. That is just the way the western notation is, our kattai terminology simply maps it to a numeric scheme.

1 1.5 2 2.5 3 4 4.5 5 5.5 6 6.5 7 and 1 (i.e. 8)

So just like there is no E#, there is no 3.5
Just like there is no B#, there is no 7.5

cmlover · Post by **cmlover** » 25 Jan 2008, 08:34

VK
3 kaTTai to 4 kaTTai is one whole tone (according to WM). There is no semitone in between!

cmlover · Post by **cmlover** » 25 Jan 2008, 08:36

OK!
I stand corrected!

cmlover · Post by **cmlover** » 25 Jan 2008, 08:46

Thanks VK! Now I understand perfectly!

vasanthakokilam · Post by **vasanthakokilam** » 25 Jan 2008, 08:56

Arun, I think CML is showing octaves and not partials.

Just a tangential observation:

This difference of partials being simple multipliers and octaves being power of 2 multipliers is the nature's gift to us to make music even possible, isn't it? ( generating partials in the first place and making our aural perception work with octaves/logarithmic etc. ) I may be over stating something here but partials are how people originally stumbled upon swarasthanas( swayambu swaras ), right? And which led to the various tuning techniques........

cmlover · Post by **cmlover** » 25 Jan 2008, 20:29

VK
The octave phenomenon is the natural physiological consequence as characterized by Fechner's Law. On the otherhand the partials and octave is purely a mathematical construct. One could choose to approximate wave forms using non-harmonic functions just as well though it will be highly cumbrous and non-intuitive! Actually Haar functions are more elegant than sinusoids. Further the 'linearity' approximations that we use out of convenience is pseudo since the world is essentially nonlinear.

cmlover · Post by **cmlover** » 25 Jan 2008, 20:34

Arun
Just to pick your brain, let us know how you discovered that it was bahudari. Was it a lucky accident from the context or logically arrived at or simply intuitive. Once you knew it was bahudari then everything fell in place by knowing the importance of Ga as the nyasa svaram which also revealed my 'goofing'

I would like to know how we can guess a raga from just some odd enunciated phrases just as we are able to know from hearing just a few words, who is talking....

arunk · Post by **arunk** » 25 Jan 2008, 21:18

I would say it was just intuitive. As soon as I listened to it, I knew it was bahudari, and knew the nyasa was G3 . How? I am not sure but I am sure it came about in the gmpdn~snpdpmg...part as that reflects the scale. I mean to say when you listen to initial long note, there is no way to tell it is G3. It probably became bahudari in the phrase that followed and thus I knew the first was G3.

But if Uday were here he probably would say, assuming the swaras wer (approxly) : gmpdn~snpdmg...., , based on our conditioning of scales, the long note must be G3.

Why?

A possible explanation here:

G3 M1 P D2 N2~ S' N2 P /D2 P M1 G3

Note that there are gamakas on the notes/swaras which are labelled above as N2, and P/D2P

Let us ask why didnt my brain consider the possibility that the long note was Sa, pa or ma i..e. other "major" ratios which are good candidates for long rests - well atleast sa and pa are. I am just throwing ma in as it is a simple major ratio.

Now, if you take G3 to be Sa it would become: S R1 G2 M1 M2~ D1 M2 G2 /M1 G2 R1 S . Isnt that odd, with M1 and M2 and used like M1 M2~ D1? Now we do have ragas with that but will they employ M1 and M2 but would a phrase like be "common" or even found, and particularly with this kind of gamaka on M2? Theoretically maybe, but it possible based on conditioning our brain never considers it.

Now if you take G3 as P, it would be become: p d1 n2 S R1~ G2 R1 n2 /s n2 d1 p... Now this does make sense (or let's say it doesnt look outlandish immediaty). But I am not sure our R1 would take gamakas like the one on N2 here, nor would the contour of p-d-p can appear in n2-s-n2. I have no idea, it just never occured to me. But also remember that the combination of S R1 G2 and D1 all generally point to melancholy and sombre, and this tune doesnt. So that could play a part in our subconscious thinking in eliminating this possibility.

If you take G3 as M1, you have M1 M2 D1 N2 N3~ R1 ...., I wont complete - it neither has Sa nor pa, and so doesnt make a whole lot of sense.

May maybe given the nature of carnatc music (and even many other forms), scale wise bahudari (i.e G3 for the first long note) perhaps fits naturally. You combine that with the kind of gamakas on specific notes AND the sub-conscious familiarity of the listener to mentally (subconsciously) recognize them, perhaps led me to the result quickly.

Of course, this is all just conjecture - but does provide some explanation as to why even vk felt that the long note didnt provide "closure" and why he was looking for sa.

Arun

Suji Ram · Post by **Suji Ram** » 25 Jan 2008, 21:24

Irrespective of the pitch it was a clear bahudAri.. as Arun explained
I was listening to bahudAri a lot recently. At first I thought my mind was overinfluenced by this raga.
And the shruti was close to what I set for violin..

cmlover · Post by **cmlover** » 25 Jan 2008, 22:02

Nice analysis! Will come handy at other times too provided the mind computes fast enough

vasanthakokilam · Post by **vasanthakokilam** » 26 Jan 2008, 01:11

cmlover wrote:VK
The octave phenomenon is the natural physiological consequence as characterized by Fechner's Law. On the otherhand the partials and octave is purely a mathematical construct. One could choose to approximate wave forms using non-harmonic functions just as well though it will be highly cumbrous and non-intuitive! Actually Haar functions are more elegant than sinusoids. Further the 'linearity' approximations that we use out of convenience is pseudo since the world is essentially nonlinear.

My understanding is:

1) Musicality and Octaves have something to do with how we perceive music, in logarithmic fashion. Meaning, 100-200 combination sounds the same as 900-1800 to our brains ( pitch differences notwithstanding ). So we have evolved that way. That is nature.

2) Any natural sound involves harmonics. That contributes to the frequency of 140Hz of violin sound different from 140Hz of flute. And that is nature as well. And there are no approximations there.

So, Both of these natural phenomena played a part in the very early ancient perception of music and how the various notes/swarasthanas were identified. The math behind harmonics may be theoretical but it is commonly understood that the swayambu swaras that emanate because of harmonics gave the early humans the idea that there are sweet sounding swaras in between the octave.

(I do not understand why you consider sinusoidal analysis as approximate. I thought It is a mathematical reality that any complex wave form can be decomposed into various sine waves ( Fourier ). Please set me right if I am wrong)

cmlover · Post by **cmlover** » 26 Jan 2008, 07:23

VK
your item 1) is correct physiologically. Nature loves the logarithmic law wherein the rate of change of stimulus is proportional to its size. The exponential growth function however fails when the size is too large. The classic example is the law of population growth or of even economy.

In regard to 2) it is only a matter of convenience to approximate functions using a harmonic (sinusoidal) base. You might as well approximate using a polynomial base. We know very well there are many functions (eg., piecewise discontinuous functions) whch cannot be well approximated by sinusoids. For example a shriek or explosion cannot be easily represented harmonically. You may call it non-musical but it is only a matter of mental conditioning. CM is only a small subset of myriads of musical forms. Take the sound of the vedic chant which is highly musical which cannot be represented within CM system (I know some call it a form of Kharaharapriya, but that is a crude approximation since it is made of only three notes udaatta, anudaatta and svarita. One has to be conditioned and educated to appreciate CM and the ability is not inborn!There is no need to claim any divinity to the notes employed in CM. Onc can choose a non-harmonic shruti base and approximate and construct music with them. Will it be liked is yet another question

The sound of wind blowing or that of the waves in the ocean are indeed musical but cannot be accommodated within CM. Let us admit our limitations and learn to appreciate all that appeals to our senses intuitively without any prejudice.

arunk · Post by **arunk** » 26 Jan 2008, 07:40

cmlover - I think vk has a point. If harmonics are NOT real, then how come ma1 and pa and ga3 are major intervals, and also related to sa (or each other) by simple ratios? That the timbre many of the instruments (including voice) is harmonic is neither a mathematical model nor some sort of an approximation. They are indeed harmonic in nature although their harmonicity may vary.

But we also pay a bit too much attention to harmonics and resonance to determine what is musical. But I think it is safe to say that hamonics play a significant role in almost all that humans find musical. There are probably exceptions of course.

Arun

vasanthakokilam · Post by **vasanthakokilam** » 26 Jan 2008, 08:00

Arun, my points exactly. Please also comment on the perception of octave. Meaning, why power of 2 separated notes have significance? What fundamentally is in us that makes those logarithmic ratios instantly attractive? Is it in our ear sensing mechanisms or in the brain? I remember reading somewhere that the ear structure has something to do with how we perceive the various harmonics and how the brain puts together the perception of timbre. But I am not sure if it is known where the Octave perception is implemented in us. All this just to convince myself that octave perception and musicality is not conditioned through reasoning or repeated listening but fundamental to how we are put together.

CML, I need to think about the general purpose nature's law you speak about: Rate of change is proportional to size... Can you give me some examples to guide my thinking. Also relate it to the progression of musical frequency: What is the rateof change here and what is the size?

On the other front, you conveniently subsituted CM with music and the same time taking the opposite position that not all music is covered in CM

But the key thing is, any continous sound can be represented harmonically, CM or otherwise.

Let me break it down...

1) Vedic chant and CM. Agreed that they are not the same, but Vedic chants can be expressed harmonically.

2) Yes, it is understood that any complex waveform can be expressed as a combined waveform of arbitrary types, not just sine waves. But that is tangential. If our ears can take apart a complex wave form into its constituent sine waves and send to our brain in some coded form, that is all that matters. It is not an approximation, that is just one way of decomposing a wave form. I do not think there is an entropy involved in such mathematical decomposition since you can mathematically put it back and get back the original signal. Mathematically... is the key word.

3) The way our brain perceives and puts together the sensory input is approximate since it may discard higher partials with faint amplittudes. Our sensory organs may approximate any sensory input by throwing away edge cases since speed is important for survival. It is doing it in near real time ( wire speed as we call it

).

4) So there is some entropy involved in that conversion. If you somehow capture all the signals from the ear to the brain and put it back together, it may not be same as the original signal that hit the ear. That is most probably true ( I really do not know but it is an educated guess )

5) You said "a shriek or explosion cannot be easily represented harmonically". Is that really true? yes, I am challenging you on that even though I do not know one way or the other

6) You said "We know very well there are many functions (eg., piecewise discontinuous functions) whch cannot be well approximated by sinusoids.". That is true. But a shreik is not that.

arunk · Post by **arunk** » 26 Jan 2008, 08:14

vk - FFT says any waveform can be reproduced by combining sine waveforms. I do not think this implies that the original waveform is indeed in reality formed that way. You may be making a jump in assuming that it is how out ear+brain processes it.

Consider this. I construct an audio file with two *pure tones*: 200Hz and 400Hz. Being pure tones neither have any partials. Now do you think humans will perceive them to have the "same note" - i.e s and S? Repeat say with have four tones, 200, 260, 400, 520.

Arun

vasanthakokilam · Post by **vasanthakokilam** » 26 Jan 2008, 08:38

Arun: I really do not know the details but I remember reading that our ear can break down a complex wave form into its harmonics but I can not find that article now. In any case, take it with a pinch of salt I may not have understood the article properly.

Regarding your experiment, see if this article sheds any light: http://www.sciencedirect.com/science?_o ... c2bfa14b17

If you can not surf there due to link issues, here is the abstract ( only the abstract is free anyway ):

Hiroyuki Uno1, *, Masao Maekawa and Hiromichi Kaneko
Department of Physiology, Dokkyo University School of Medicine, Mibu, Tochigi 321-02, Japan
Received 30 September 1996; revised 28 March 1997; accepted 2 April 1997. Available online 11 February 1998.

Abstract
We trained zebra finches on a go/no-go operant conditioning procedure to discriminate between two harmonic structures that differed only by the presence or absence of the second harmonic. Test sessions revealed that the birds responded to probe stimuli based on the presence or absence of the second harmonic regardless of the other frequency components. Some of the birds also tended to use the fundamental frequency as a discriminative cue when the number of harmonics was few. These results indicate that birds use multiple strategies for this type of harmonic structure discrimination.

cmlover · Post by **cmlover** » 26 Jan 2008, 08:39

VK
As far as I know there is nothing structurally in our hearing apparatus that analyzes sound harmonically which is put together by the brain. Brain is holistic in recognizing sound. Per ception of music is acquired; note that animals can sense sound but not enjoy music! You agree periodic functions are not necessarily sinusoidal. Though the brain can handle complexity, we intrinsically prefer simplicity and consider it elegant. Even visually we appreciate simplicity and symmetry. The musical notes being simple ratios are appealing and are easily comprehended by the brain . That makes us happy and hence the joy coming out of those sounds

But again there is an intellectual joy coming out of understanding the innards which may not be associated with the melody or rhythm. The blaring of the autobus horn may be raucous but it is music to the ears of the driver. Why do we says 'kuzhalinidu, yaazh inidu tham makkaL mazhalaiSol kELaadavar

arunk · Post by **arunk** » 26 Jan 2008, 09:01

Ok. Here is a sample for analysis. I have (yet again) been mesmerized by this raga. Now I find the usage da in pa-da-ni-da (note that aro/avaro of this raga doesnt cover everything that is allowed) to be special. Some say it is trisruthi, i.e. lower than D2. I will admit that it does seem lower - but then I think maybe not, it is just delivered softer i.e. a different touch. Then I think maybe it is a combination of both.

Here is pitchappa trying his amateur best: http://www.sendspace.com/file/o69eww

He has tried the standard aro/avaro, then the phrase, then *tried* standard D2 in P D2 N2 S, then again P D2 N2 D2 of our raga, and then finally a P D1 N2~ S. Questions

1. Is this a fair sample even though amateurish?
2. How do the 3 das compare - what do your ears say? What do your eyes (i.e. graphs) say?
3. Has Pitchappa overstepped his limits and thus we should pass a resolution banning him from these kind of "pseudo-scholarly" attempts

Note: Pitchappa says he had to sing softly as to not wake his little one up. Also his arrival back to sa, is let us say shaky

Arun

vasanthakokilam · Post by **vasanthakokilam** » 26 Jan 2008, 09:01

CML, Arun, I found the article where I had read about this first.

http://dissertations.ub.rug.nl/FILES/fa ... ost/c5.pdf

Here is the relevant excerpt:

In the inner ear, the different frequency components of a complex sound excite
different parts along the basilar membrane (von BÃ©kÃ©sy, 1960). High frequencies excite the
basal part of the basilar membrane, while low frequencies travel through the cochlea and have
their maximal response at the apical part of the basilar membrane. Due to the approximately
logarithmic cochlear frequency map and the limited frequency resolving power of the cochlea,
the relatively widely spaced lower harmonics of a complex sound are resolved, while the
higher harmonics, which are spaced more closely, remain unresolved.
A resolved harmonic causes the basilar membrane to vibrate nearly sinusoidally at the
place of resonance.

vasanthakokilam · Post by **vasanthakokilam** » 26 Jan 2008, 09:04

Arun, you seemed to have posted the delete link. I did not click on it yet, but noticed the word delete in the link...

arunk · Post by **arunk** » 26 Jan 2008, 09:06

oops! Fixed that (hopefully before someone clicked on it!)