Discuss traditional rhythms, singing etc
By chin
I'm quite a "newbie" in understanding of different families of ternary rhythms...
Only until recently I had been able to play the shuffle bell pattern, including the offbeat... (eg. Soli, Mendiani, Soboninkun etc.)
I'm still struggling with Dununbas, although I am really really interested in them. (In fact I had stopped for a while... since I couldn't even play the shuffle offbeat before, now that I can do that maybe it's time to try again)

Since I've been studying those ternary rhythms more in depth, I started to find that there are "vague" differences among them. (maybe vague to me only...) Then I started to geek out here and found there had been discussions (heated) here concerning the families. But to be honest, I had been trying to understand by reading the same 2 threads for days but am not really sure if I really understand. So I would appreciate if someone could help this newbie out! :) However due to very limited exposure to ternary rhythms at my area, my repetoire is really limited...

I noticed:
1) Dununbas (and other rhythms like Mendiani)
2) Dya (and rhythms like soko)
3) Binary

My questions:
a) I noticed that the signal of 1) and 2) are different, is it always 1) brugudu gudu gudu while 2) gudugudu gudu gudu gu?
b) Chauffe: I read in another thread that the chauffe of dununba in 1) is always upbeat (to recall, during my only chance to study with Famoudou, it's the case too) and 2) is always "towards" the beat, am I correct too?
c) Bell pattern of kenkeni of Dya family is usually **.**.**.**. am I right? How about Dununba family, is there a usual pattern?
d) I read about something about "Dya bell pattern" (or something similar), but I cannot really understand this... is it the bell pattern of Sangban? That there could be "one stroke" bell?
e) Very importantly... when I hear/learn a ternary rhythm, actually how do I distinguish which family it goes? What is the salient feature? Or all the above? Would there be exceptions? Then if I already am able to distinguish to which family, and if I am on dununba is it that I can automatic play the respective chauffe to the family?
f) I already read that soli could be regarded as another small family, why? The version I learned is from Mamady, I am not very sure of other version... say Hamanah version (which I'd be very very interested... I seem to be gravitated towards Hamanah music)? And maybe that is why it shall be regarded separately? (if so, is it possible for me to find notation/clip somewhere?)

Thanks a lot in advance for kind reply! :)
By bubudi
hi chin, what a first post! please introduce yourself in the introduction section and let us know how long you have been playing, where you live and who your teachers are, and anything else you wish to share.
chin wrote:I started to geek out here and found there had been discussions (heated) here concerning the families. But to be honest, I had been trying to understand by reading the same 2 threads for days but am not really sure if I really understand.
this is to be expected. this music has very few absolutes and occassionally 2 masters will even disagree. with regards to the families, this is made more complicated by the fact there are many ways to classify rhythm families. as discussed on the other thread, you can group them by function, or feel, for example. so you will hear people talk about 'dununba family' or 'kassa family' classified on their function. in the other thread we mostly talked about form and feel. i tend to look at it more from the point of view of rhythms evolving from several core rhythms while others may take a different approach.
I noticed:
1) Dununbas (and other rhythms like Mendiani)
2) Dya (and rhythms like soko)
3) Binary
for hamana this may be true, but with some debate over whether soli fits into the first category or an additional one. also, there are 2 types of binary feel at least. the rhythms in the categories depend on the geographical region. for instance, sega sidibe considers suku (soli), denba (maraka) and dansa to be the core rhythm families in mande music.
a) I noticed that the signal of 1) and 2) are different, is it always 1) brugudu gudu gudu while 2) gudugudu gudu gudu gu?
quite often, but not always. also, in many villages they will not use many calls, if at all.

i'll do my best to answer yout other questions a little later.
By chin
:oops: :oops: :oops:

I'm really sorry and I didn't mean to be so rude... but I guess I was too overwhelmed and excited and forgot to even say hi to all of you... Please forgive my weirdness... :oops: :oops:

Thanks a lot bubudi, I understand what you said on classification can be on functions or evolvement from core rhythms. What I have been struggling to understand is, what is the significant difference between the 1) and 2), put aside those "vague", exceptions and debatable rhythms.

I hope this doesn't sound dumb... but when I was reading another thread about dununbas, it occurred to me that I understand every single word in the thread but somehow I fail to follow how the logic falls... maybe I know too little on the music...

Anyway, binary aside, is it possible to deduce into a generic saying of "if I hear x feature and y feature, I can know that it probably is 1... so it is logical that I play the chauffe of dununba as xx"??

Thanks again bubudi! :rasta:
By chin
I found that I was not very clear before... right now I'm trying to understand the categorisation musically instead of functionally/socially.

By bubudi
dununba chauffs vary depending on the particular dununba rhythm but yes they tend to be offbeat (while the djembe chauff keeps it anchored to the beat). i'm not sure exactly what you mean by 'towards the beat'?

correct about most dja family kenkeni bell patterns. dununba rhythm kenkeni patterns are almost always '-xx -xx -xx -xx', with some exceptions which are all, to the best of my knowledge, played outside hamana. another interesting kenkeni bell pattern for dununba rhythms, which is good to use if for some reason you don't have a strong sangban bell, is:
x-x x-x x-x x-x
--o -oo --o -oo

this is good for hand independence too, as you're required to play the kenkeni strokes and bell strokes at different times. i will stress, though, that it's not traditionally found in hamana.

i think the reference to the one stroke bell is having single strokes rather than double strokes. for example, the three single strokes in the dja sangban pattern:
x-x -x- xx- x-x

more later!
User avatar
By African Rythm
There are 4 families of rhythm in mandingo music. 2 families of binary rhythms, 2 of ternary rhythm. I try to explain every properties of each family, on my blog. But I wrote it in french, and I'm not confident in my english...

Could anybody help me to translate my post in english, so that I can show you how I conceptualize the west african music ?

Here are my first post about the 4 families :

2 posts about the 2 ternary families :

And 2 other posts for a further understanding of one ternary family that I called "shuffle ternary"
By djembeweaver
michi wrote:Thanks for posting this!

Google Translate does a fairly decent job of converting it to English, but it would be great if someone who speaks both French and English could do a proper translation!


Hmmnn...it's written in a very formal academic french. I translated the first half of the first page but it took quite a while as I had to look quite a few phrases up. If someone better than me doesn't do it I might get round to it at some point. Of course, I could paraphase it but a full translation would be better.

By djembeweaver
OK it doesn't look like anyone with better french than me is going to do this so here goes. This is only the first article. I will translate the others when I get round to it. I don't claim that it is all perfect as it pushed my french skills to their limit and beyond. It's kind of like writing in classical score...I can do it but its a ball-ache!

Certain phrases (most notably 'debit') I can't find a good translation for even though I understand the sense. In the end I decided to translate 'debit' as 'spacing'. Please correct me if you have a better translation...

The 4 Rhythmic Families of West Africa

Here is the theory (and its vocabulary) which supports my understanding and teaching of the rhythms of West African music. Careful, it's only a model and therefore revisable...

West African rhythms are divisible into four families. To those who are familiar with the concepts of binary / ternary: two families are binary and two are ternary.

Hereafter I will call the binaries Binary Classic (BC) - the most common - and Binary Inverse (BI) - more rare (they are used mostly in Mali, like the rhythm Sandia).

For the ternaries I will use the nomenclature already used in Jazz: Ternary Iambic (TI) and Ternary Shuffle (TS). The latter are very specific to West Africa (the dununbas from Haut Guinea, for example, are part of that). The former are very close to ternary rhythms of other cultures (like the maloya de la Reunion or the rumba of Cuba for example).

For each family there is a corresponding spacing and a rhythmic key.

By focusing on the concept of 'spacing' I do not mean to imply that the divisions in time (of the quarter notes and the dotted quarter notes) are never regular. Thus, the subdivisions of time are defined by an irregular spacing of notes. In this model, each spacing defines a family.

The rhythmic keys are the rhythmic phrases of reference on which the rhythms are built. A bit like Cuban claves, which structure music like rumba and salsa.

Within a family, a rhythm will be defined by the overlapping of several rhythmic phrases within the same spacing and played with the same rhythmic key.

Each family has equally noteworthy properties, like the systematic accentuation of syncopations (notes displaced from the pulse or time)

As we shall see later, it is possible to switch from one family to another by 'forcing' the interpretation of spacing, or by referring to finer subdivisions of time (including sixteenth notes and sixteenth note triplets for example). Thus, a rhythm can, depending on the speed of the pulse, pass from one family to another.

Of course this theory has its limitations. Primarily because certain rhythms don't fit into any of the four families, as I will explain in the following articles. Luckily, the model using spacing and keys allows us to understand the functioning of these atypical rhythms, for which we can create new families. I leave you to discover the other limitations of this model over the course of these articles. But above all I will begin by showing you why it works in most cases.

To summarize: There are four families of rhythms, two binary and two ternary. Each one is defined by its (irregular) spacing of notes. Each family corresponds to rhythmic phrases of reference: the keys.

If any part of this article is not clear, or if you do not understand certain technical words, pose the question in the comments. It will be my pleasure to answer them.

And now, to enter into the detail of each family, and therefore of the following theory and the beginning of the practical exercises, I invite you to read the following articles concerning each of the families....


Phew.....more to come when I have the energy!

By djembeweaver
And here's the next one:

The Spacing of Iambic Ternary Rhythms

Ternary iambic music is present in all styles of music in Mali, Guinea, Burkina Faso and Ivory Coast. Whilst theoretically simple to understand, this ternary music is sometimes very difficult to hear correctly. That is to say to listen to the beat of the true pulse.

A little audio example will illustrate my point. Here is a piece of music for hunters, the donsos by the Malian Yoro Sidibe (photo above)

In this example we hear two instruments: the ngoni, a lute-harp (that Yoro Sidibe is playing in the above photo), and the karignan, a metal percussion instrument. It is actually the karignan player who fools us by playing an accented beat on a syncopated beat rather than on a pulse beat.

I will try to explain how iambic ternary music functions in theory.

The principle characteristic of these rhythms is that the first sub-pulse lies further from the preceding pulse.

Look at the diagram below.

The three spaces are therefore different. The interval between the pulse and the first sub-pulse is the longest (6/15 of a dotted crotchet or quarter note, being a bit longer than a quaver or 8th note); the interval between the two sub-pulses is the shortest (4/15 of a dotted crotchet or quarter note, a little less than a quaver or an 8th note); and the last interval, between the second sub-pulse and the pulse, equates to a quaver or 8th note.

This distortion of spacing results from an important property: The first sub-pulse (pink in the diagram) is the most accented note in this ternary music. It is where the karignan plays in Yoro Sidibe's song.

The first sub-pulse is not systematically played stronger than the pulse. Where it is, it can lead us to mis-hear the pulse.

Above all these themes are often apparent in the first sub-pulse.

Listen to this, where the pulse note is accented:

And here are some pieces that use iambic ternary rhythms:

The first is a sample of Malinke flute, played by the Guinean Mamady Manaré. The second, played by Kanazoé and his group from Burkina Faso, is a piece from the Senoufo people. Finally the last is a soninké rhythm very popular in Mali called maraka, interpreted by the late Karim Tounkara on djembe and on dunun khassonke by Dramané ("left hand") Diabate.

Did you find the pulse in these examples?

Do you have examples of iambic ternary rhythms to share?

Questions? Comments?

In the next article of the iambic ternary category I will propose practical exercises for internalizing this timing, then I will seek to describe the principle rhythmic figures of reference, the iambic ternary keys.

OK my head hurts now.

User avatar
By michi
Hi Jon,

thanks heaps for the heroic effort!

That's really interesting, and it lines up with what's described in "Séga Kan Do".

Sega refers to the swing where the first micro-pulse after the beat is delayed as the "sentiment of Demba Foly." (In the book, that's the only rhythm that is quoted as an example of this micro-timing. What others fall into that family?)

Sega also gives examples of the opposite swing, where the first micro-pulse after the beat comes a little early: Suku, Garangedon, and Mendiani. (This swing is particularly evident in the classical Mendiani call.)

A good exercise is to play tsstsstsstss echauffements without swing (easy), with the Suku swing (harder for me), and the Mendiani swing (hardest for me).

Would be interesting to find out whether that difficulty is the same for most people, or whether some people have an easier time with the Mendiani swing than the Demba Foli one…

Jon, if you can spare the time, I'd love to read the remainder of the articles! Thanks again for doing all this work!


By djembeweaver
No problem Michi (actually that's not strictly true!)

This is exactly how I categorize rhythms. In my opinion all rhythms with the soko-style signal are Iambic Ternary. Hence Soko, Sorsornet, Wassolonka (Toulounbeh) etc etc. I find this type of swing much easier to play than the shuffle (most noticeable for me in Mendiani but I think the swing gets more pronounced the closer to Mali you get).

Anyway, here is the third article:

Translators note: there is one sentence that I am struggling to translate. It is the second sentence below the flow diagram. The sentence reads: Entre un temps et le suivant nous avons la durée de la noire (solfège). This translates as, "Between one pulse and the next we have the duration of a noire (solfege)" Now then, 'noire' in this context (music theory) means a crotchet or quarter note. The problem is that in 6/8 (or 12/8) the length of a pulse beat is a dotted crotchet or quarter note. In french this would be 'noire pointée', as the author points out in the following brackets where he shows what fraction of a dotted crotchet each sub-pulse represents. 'Solfege' normally means the orthography of the music (music theory if you like). This might be a mistake on the part of the author but I'm just not sure so I've left this phrase in the original french (yes a cop out I know)...

Spacing of Ternary Shuffle Rhythms

The ternary shuffles are particularly characteristic of Mandeng Africa. This family covers nearly half of the ternary rhythms used in Mandeng music. To date I have not found their equivalents in the music of other cultures, unlike the iambic ternaries which are often used in other african music and further afield (La Reunion, Cuba)

To begin this series of articles on the ternary shuffles, I will try to explain their spacing. That's to say the way in which the irregular subdivisions of the beat (of the dotted quarter notes) are put together in all the rhyhtms of this ternary family. It's precisely this spacing that allows me to define this family.

Let's begin with a little musical extract; here is Soli, a rhythm from Haut-Guinea, played by a master of this genre, Famoudou Konate...

Like all ternary rhythms the pulse is divided into 3, thus there is one note on the pulse and 2 sub-pulses. The defining feature of TS is that the first sub-pulse is consistently very close to the pulse (Remember: the divisions of time are irregular!)

A diagram to illustrate:

In this diagram I have represented the average spacing of TS along a temporal line (thank you Matthieu Lamarche for the maths!) Between one pulse and the next we have the duration of a noire (solfege). The gap between the pulse and the first sub-pulse is therefore shorter (4/15 of a dotted quarter note, a little less than an 8th note) than the gap between the two sub-pulses (5/15 being a third of a dotted quarter note, i.e. an 8th note or quaver) or the gap between the second sub-pulse and the next pulse beat (6/15 of a dotted crotchet or quarter note, a little more than an 8th or quaver)

Here is the structure in isolation, at several tempos (again, thank you Matthieu):

We can hear this in the extract at the top of the page, where Famoudou Konate, the soloist, plays all the notes in the structure (we call this an echauffement...I have pointed out the first few instances of this in the extract). I leave it to you to listen to this.

A consequence of this deformation is that the second sub-pulse, in pink in the diagram, is the most accented note in these rhythms. Sometimes it is more accented than the pulse itself! For example in the well-known rhythm Dununba...

This is why in the first extract Famoudou Konate plays a "solo accompaniment" based on the second sub-pulse (listen between 1:40 & 1:50)

Once the structure has been selected, all the accompaniment instruments and all the voices respect exactly the same spacing. The result is the groove particular to ternary shuffles. These pieces very often have a pronounced accent on the second sub-pulse.

And now I suggest you listen to how these rhythmic properties are applied in melodic phrases:

The first is a piece of balafon music of toussian ethnicity, Fama, interpreted by Daouda Diabate from Burkina (careful...this extract is based on a measure of 6 time; a time signature of 18/8). The second is the magnificent song Ko Benna Toumana Le, by the guinean artist Djen Doumbia. Finally the last is a guinean piece, Mamaya, interpreted here by the voice and kora of Djeli Moussa Diawara.

User avatar
By michi
Thanks heaps for that Jon!

Here two sound clips that illustrate the lambic swing. The first one is from "Rhythmen der Malinke", Mendiani:
Mendiani excerpt, Famoudou Konaté
(940.07 KiB) Downloaded 366 times
You can clearly hear the "bent" spacing during the echauffement.

This one is from Fadouba's CD, Soko. It's the call at the beginning.
Soko excerpt, Fadouba Oulare
(311.9 KiB) Downloaded 338 times
This also indicates that the standard ternary calls (bri-bidi-bidi-bidi and bribidibi-dibi-dibi-di) don't necessarily indicate the swing. Soli uses the first and soko the second, but both are lambic.

But what also seems to be true that the following accompaniment is always associated with bri-bidi-bidi-bidi, and never with bribidibi-dibi-dibi-di:
Code: Select all
Conversely, the following accompaniment is always associated with bribidibi-dibi-dibi-di, and never with bri-bidi-bidi-bidi:
Code: Select all
Whereas the standard ka-dika is used with either call and with either swing.

Another observation: lambic swing emphasizes the third micro-pulse. For example, think about Garangedon and Soli. Many of the solo phrases place a slap on the third micro-pulse just before the down-beat. Conversely, rhythms that use the other swing, such as Maraka, emphasize the second (delayed) micro-pulse. Many of the solo phrases and accompaniments put a slap on the second micro-pulse.

It seems to be true that, at least generally, bri-bidi-bidi-bidi is associated with lambic swing, and bribidibi-dibi-dibi-di with the other swing, but there are exceptions, such as Soko. (I can't think of an example for the opposite. Anyone know of such a rhythm?)

User avatar
By djembefeeling
hey jon,

thank you so much for this effort! After reading these articles I wonder why the one ternary family is called iambic and the other shuffle. In a binary shuffle the second pulse (or in the terms of the article: the first sub-pulse) is delayed while in the ternary shuffle it is brought forward. That's confusing. Does anybody know why it is still called ternary shuffle?

For me, TS is easier to play than TI. I think this is because I started to practise deflected microtiming with TS and never really practised systematically the microtiming of TI after that. It's easier to deflect after you get used to some way, I guess...

As for Soko, I don't think that Fadouba's intro is a paradigm for TS in Soko. It is just an intro, and there are tons of examples of bands using e.g. binary intros for ternary rhythms and vice versa. But I know of soloists who like to "swing" on TI once in a while in solophrases. I guess this is due to the fact that they use to impose a binary structure on TI, and the first sub-pulse of three sounds like a strong swing?! In Mali, the typical Wassulunka microtiming also integrates a fourth note into the big space between the pulse and the first sub-pulse. this can be confused with (a very strong) swing, as well. So there are several examples of "swing" in TI.