AnuLaya Composition Authoring Guide

A practical guide to writing tabla compositions — from simple theka to a full kaida with paltas and chakradhar tihai.

Type in lowercase English; the app always renders the canonical Devanagari alongside. Every example below shows the raw input and the grid as it appears in the app.

---

Table of contents

Part 1 — Basics (start here)

1. What is a composition?
2. Designed for phone keyboards
3. Quick start
4. Composition metadata
5. Bols and Devanagari
6. Basic notation
7. Phrases per beat (concatenation)
8. Comments and subtitles

Part 2 — Advanced

9. Dot subdivision (unequal beats)
10. Parentheses (nested grouping)
11. Composite shortcuts
12. Repeats
13. Speed markers (layi)
14. Tihai with ^
15. Variables
16. Taal grid and vibhags

Part 3 — Recipes and reference

17. Recipe: full kaida (Teentaal)
18. Recipe: tukda and tihai (Rupak)
19. Recipe: song accompaniment (Keherwa)
20. Common pitfalls
21. Quick reference

---

Part 1 — Basics

Everything in Part 1 uses Dadra (6 beats, vibhags 3+3) so you can focus on notation without worrying about taal structure.

---

1. What is a composition?

A composition has:

• Metadata — title, taal, tempo, category.
• Bols — the notation text (what this guide is mostly about).

The app parses your bols text and lays it out on a taal-aligned grid. Playback walks the grid at your chosen BPM. Everything else is just ways to pack more than one stroke into a beat, or to repeat/reuse phrases.

---

2. Designed for phone keyboards

Most compositions are typed on phones, one-thumb. The notation stays on the primary keyboard layer wherever possible — no shift, minimal symbol-layer trips:

• Lowercase — no shift key, autocorrect-safe. Case is ignored anyway.
• Space between beats, - for rest — both on the main letter layer.
• Dot for subdivision (dha.ge) — reachable without switching layers. Concatenation (dhage) works too.
• x N for repeat — plain letter, not × or *.
• [ ], $, #, ##, @Nx, ^ — all single-tap ASCII, familiar from Markdown.
• Composites (tirkit → ti ra ki te) — four taps for a common 4-bol phrase.
• Forgiving parser — autocorrect capitals, extra spaces, and double-tap-space-periods don't break anything.

Roman is offered because it's fast to type; the app always renders the canonical Devanagari alongside.

---

3. Quick start

Open the app, go to Compositions → New, and fill in:

Field	Value
Title	Dadra Theka
Taal	Dadra
Tempo	120
Category	theka
Bols	dha dhin na dha tin na

Input:

dha dhin na dha tin na

Grid (Dadra):

1	2	3		4	5	6
dha	dhin	na		dha	tin	na
धा	धिं	ना		धा	तिं	ना

Hit play — you hear the 6-beat Dadra cycle. Each space-separated token is one beat. The gap marks the vibhag boundary between beats 3 and 4.

---

4. Composition metadata

Field	Required	Notes
title	yes	Display name.
taal	yes	Dadra, Rupak, Keherwa, Bhajani, Dhumali, Jhaptaal, Ektaal, Chautaal, Deepchandi, Dhamar, Teentaal, Tilwada.
tempo	yes	BPM (20–300). Base tempo; speed markers and global speed multiply on top.
bols	yes	The notation text.
speed	no	Global layi (1 = baraabar, 2 = dugun, 3 = tigun, 4 = chaugun). Default 1.
category	no	theka, kaida, rela, tukda, tihai, song, other.
description	no	Short prose — who taught it, style, what to watch for.
author	no	Composer / guru credit.
tags	no	e.g. ["kayda", "intermediate", "benares"].
visibility	no	private or published. Default private.
phrases	no	Space-separated bol phrases for grid highlighting (see §4.1).

4.1 Phrases — coloring patterns across beats

Tabla phrases often cut across beat boundaries. For example, in a Teentaal kaida the phrase dha ti dha ge starts mid-beat and runs through the next — the grid shows it split across cells, making it hard to see where one phrase ends and the next begins.

The phrases field solves this. Enter your recurring phrases as space-separated concatenated bols:

DhaTiDhaGe NaDhaTiRaKiTe TinNaKeNa

The app matches these patterns in the grid and color-codes each phrase with a distinct color (blue, green, purple, orange, etc.). Bols belonging to the same phrase get the same color, even when they span multiple beats.

Example — Teentaal kaida theme with phrases:

Field	Value
Bols	dha ti dha ge na dha tira kite dha ti dha ge tin na ke na
Phrases	DhaTiDhaGe NaDhaTiRaKiTe TinNaKeNa

Grid (Teentaal, beats 1–8):

1	2	3	4		5	6	7	8
dha	ti	dha	ge		na	dha	ti ra	ki te
धा	ति	धा	गे		ना	धा	ति र	कि ट

Grid (Teentaal, beats 9–16):

9	10	11	12		13	14	15	16
dha	ti	dha	ge	tin		na	ke	na
धा	ति	धा	गे	तिं		ना	के	ना

• DhaTiDhaGe — appears twice (beats 1–4 and 8–11)
• NaDhaTiRaKiTe — beats 5–7
• TinNaKeNa — beats 12–15

Without phrase colors, beats look like four unrelated bols. With colors, you instantly see it's the same DhaTiDhaGe phrase.

Phrases are optional. The app can auto-detect common phrases, or you can enter them manually for precise control.

---

5. Bols and Devanagari

Tabla bols are traditionally written in Devanagari, but it is easier to type in Roman. The Roman spellings here are a canonical one-to-one mapping chosen so round-trip is lossless — some Roman forms are deliberate (e.g. te for ट, not ta) to avoid collisions.

5.1 Canonical mapping

Devanagari	Roman	Input		Devanagari	Roman	Input
धा	Dha	dha		ता	Ta	ta
धे	Dhe	dhe		तिं	Tin	tin
धें	Dhen	dhen		तुं	Tun	tun
धि	Dhi	dhi		त	T	t
धिं	Dhin	dhin		ति	Ti	ti
धुं	Dhun	dhun		ट	Te	te
दि	Di	di		त्र	Tra	tra
दिं	Din	din		ना	Na	na
द	Da	da		ने	Ne	ne
ड़	De	de		ग	Ga	ga
घे	Ghe	ghe		गे	Ge	ge
घें	Ghen	ghen		घि	Ghi	ghi
क	Ka	k, ka		के	Ke	ke
कि	Ki	ki		कत्	Kt	kt
क्र	Kra	kra		क्ड	Kda	kda
र	Ra	r, ra		-	Rest	-, s

Placeholder: ~ — layout-only (advanced; use - for ordinary silence).

5.2 Key disambiguations

Pairs that sound close in English but are distinct strokes in Devanagari:

• Te (ट) vs Ta (ता) — ट is retroflex; in casual romanization both get written "ta". Here ट is te so there's no collision. Example: तिरकिट → tirakite, not tirakita.
• T (त) vs Ta (ता) — त is the bare consonant (short, dry); ता has the आ vowel (fuller). Phrases like tkta = त क ता.
• Ge (गे) vs Ghe (घे) — same stroke, different Devanagari. Add the h for घे. Example: घे ना गे ना → ghe na ge na.
• Di / Dhi / Ti — don't drop the h.
• Na (ना) vs Ne (ने) — separate bols; mind the vowel.
• De (ड़) vs Dhe (धे) vs Da (द) — unrelated strokes; use the exact letters.

5.3 Single-character shortcuts

Input	Roman	Devanagari
t	T	त
r	Ra	र
k	Ka	क

Example: tkta = T Ka Ta = त क ता.

---

6. Basic notation

6.1 One bol per beat

Input:

dha dhin na dha tin na

Grid (Dadra):

1	2	3		4	5	6
dha	dhin	na		dha	tin	na
धा	धिं	ना		धा	तिं	ना

Each space-separated token is one beat.

6.2 Rests

Use - (hyphen) to silence a beat.

Input:

dha dhin - dha - -

Grid (Dadra):

1	2	3		4	5	6
dha	dhin	-		dha	-	-
धा	धिं	-		धा	-	-

s and S also mean rest.

6.3 Multi-line rows

Newline = new row. Rows are visual only — playback is seamless. Use them to make long compositions readable.

Input:

dha dhin na
dha tin na

Grid (Dadra):

1	2	3		4	5	6
dha	dhin	na		dha	tin	na
धा	धिं	ना		धा	तिं	ना

Two input rows, same single-cycle grid. The line break is visual in the editor; the grid lays it out by taal.

6.4 Multiple cycles

A composition can span any number of taal cycles. Each row in the input doesn't have to be one cycle — the grid wraps by taal automatically.

Input:

dha ti dha dha tin na
ta ti dha dha dhin na

Grid (Dadra, cycle 1):

1	2	3		4	5	6
dha	ti	dha		dha	tin	na
धा	ति	धा		धा	तिं	ना

Grid (Dadra, cycle 2):

1	2	3		4	5	6
ta	ti	dha		dha	dhin	na
ता	ति	धा		धा	धिं	ना

Two lines of 6 bols each = two Dadra cycles. The second cycle starts on sam (beat 1) of the next round. Notice the bhari→khali shift: dha in cycle 1 becomes ta in cycle 2.

6.5 A simple tihai (spanning cycles)

A tihai repeats a phrase three times to land on sam. Even with basic notation (no […]x3 syntax yet), you can write one out longhand:

Input:

dha - k t 
dha - k t 
dha - k t

Grid (Dadra, cycle 1):

1	2	3		4	5	6
dha	-	k		t	dha	-
धा	-	क		त	धा	-

Grid (Dadra, cycle 2):

1	2	3		4	5	6
k	t	dha		-	k	t
क	त	धा		-	क	त

The phrase dha - k t repeats 3 times across 12 beats (2 Dadra cycles). The rests (-) between each dha create the damdaar tihai. (In §14 you'll learn [dha - k t]x3 as a shorthand for this.)

6.6 Partial cycles are padded with silence

If your bols don't fill a complete taal cycle, the remaining beats are filled with silence.

Input:

dha - - dha

Grid (Dadra):

1	2	3		4	5	6
dha	-	-		dha	-	-
धा	-	-		धा	-	-

Only 4 bols entered, but Dadra needs 6 beats per cycle — beats 5 and 6 are automatically silent.

6.7 Pipes as visual separators

| is ignored by the parser — use it for readability, typically at vibhag boundaries.

Input:

dha dhin na | dha tin na

Same grid as §6.1. The pipe helps you eyeball the vibhag split while editing.

6.8 Case is ignored

All three produce the same grid:

dha dhin na dha tin na        ← recommended (no shift key)
Dha Dhin Na Dha Tin Na
DHA DHIN NA DHA TIN NA

Use lowercase — it's fastest. Mixed case like DhaGeTiTe can help readability for long concatenations (§7).

---

7. Phrases per beat (concatenation)

Multiple bols touching each other (no spaces) share a single beat. This is the simplest way to pack more than one stroke into one beat.

7.1 Two bols in a beat

Input:

dha dhin na dhadha tin na

Grid (Dadra):

1	2	3		4	5	6
dha	dhin	na		dha dha	tin	na
धा	धिं	ना		धा धा	तिं	ना

Beat 4 holds two bols, each taking half the beat.

7.2 Four bols in a beat

Input:

dha dhin na tirakite tin na

Grid (Dadra):

1	2	3		4	5	6
dha	dhin	na		ti ra ki te	tin	na
धा	धिं	ना		ति र कि ट	तिं	ना

Beat 4 has four bols (parser segments longest-match-first). TiRaKiTe parses identically — camelCase aids readability for long tokens.

7.3 Rests concatenate too

Input:

dha dhin na dhadha-- tin na

Grid (Dadra):

1	2	3		4	5	6
dha	dhin	na		dha dha - -	tin	na
धा	धिं	ना		धा धा - -	तिं	ना

Beat 4 has four slots — dha, dha, rest, rest — each 1/4 of the beat.

7.4 Mixed jati (misra) — different speeds per beat (Keherwa)

Concatenation naturally creates misra jati (mixed groupings) when adjacent beats have different numbers of bols. No special syntax needed — just write each beat with the bols you want.

Input:

dhagena dhitedhite dhagena tinakena
takena titetite dhagena dhinnagena

Grid (Keherwa, cycle 1):

1	2	3	4		5	6	7	8
dha ge na	dhi te dhi te	dha ge na	ti na ke na		ta ke na	ti te dhi te	dha ge na	dhin na ge na
धा गे ना	धि ट धि ट	धा गे ना	तिं ना के ना		ता के ना	ति ट ति ट	धा गे ना	धिं ना गे ना

Beat 1 has 3 bols (each 1/3 of the beat), beat 2 has 4 bols (each 1/4 of the beat). Played back-to-back, this creates the 3+4 misra jati feel — the grouping shifts naturally without any special notation.

Beat 1 (tisra — 3)	Beat 2 (chatusra — 4)
dha · ge · na	dhi · te · dhi · te
1/3 · 1/3 · 1/3	1/4 · 1/4 · 1/4 · 1/4

---

8. Comments and subtitles

8.1 # comments — hidden

Lines starting with # are ignored at playback. Use them to organize your source: group variable definitions, mark sections, leave reminders.

Input:

# Basic Dadra theka
dha dhin na dha tin na

# Variation with doubled 4th beat
dha dhin na dhadha tin na

Comments don't appear on the grid. Think of them like headings in a notebook — they make a 100-line composition skimmable when you return to edit it. Grids are the same as §6.1 and §7.1.

8.2 ## subtitles — shown on the grid

Two hashes = a section header the viewer sees.

Input:

## Theka
dha dhin na dha tin na

## Variation
dha dhin na dhadha tin na

## Theka and ## Variation render as headers above their respective rows in the grid. Use # for editor-only notes; use ## for structure you want displayed.

Each subtitle section must end on a complete taal cycle. If the bols don't fill the cycle, the remaining beats are padded with silence — just like §6.6.

Input:

## Phrase
dha - k t

## Theka
dha dhin na dha tin na

Grid (Dadra):

Phrase

1	2	3		4	5	6
dha	-	k		t	-	-
धा	-	क		त	-	-

Theka

1	2	3		4	5	6
dha	dhin	na		dha	tin	na
धा	धिं	ना		धा	तिं	ना

The "Phrase" section has only 4 bols — beats 5 and 6 are silent. The "Theka" section starts fresh on its own sam.

---