Text Data Collection Guideline

LDC-IL collected text corpus from different sources. They are mainly books, magazines, and newspapers. LDC-IL has different Sampling approach over while extracting text from these three sources.

Sampling Approach for Books

The books were identified so that the representation of different domains can be catered. After identifying the books, the next step is to extract typically 10 pages of text from it. LDC-IL follows a sampling method to collect the pages from a book. For example, if the book has 200+ pages we collect every 20th page of the book.

Other generic principles that have been normally followed in the sampling tasks across languages are as follows:

• Contents containing obnoxious or vulgar texts should be avoided.
• Prefer books published after 1990.
• Text extracts containing poems and formulae should be avoided.
• Pages containing diagrams, tables or figures should be avoided.
• Books containing less than 50 pages are not part of sampling.
• If the text contains content other than the intended language, those texts should be avoided if the other language content is longer than one sentence.

Sampling Approach for Magazines

In the case of magazine texts, which are typically brief and span various domains, the entire magazine should be included in the corpus, excluding advertisements, image captions, tables, and similar non-text elements. A magazine corpus generally encompasses a variety of text types, such as cookery, health, cinema, stories, contemporary articles, and more.

Sampling Approach for Newspaper

The newspaper corpus is contemporary text in nature. The text may contain political news, editorials, sports news etc. The newspaper has separate subcategories which cover all the newspaper domains. Classifieds, very small news snippets were avoided.

Proof Reading

Once the text is in digital form, it is proofread to eliminate any typographical errors. For proof reading the following steps should be taken:

• Removing any poetic text or poetic structures that appear within the running text.
• Eliminating incomplete sentences, particularly those at the end of paragraphs.
• Ensuring the correct use of the visargaha symbol and the colon ‘:’ symbol, verifying that each is used appropriately.
• Unnecessary space near the punctuation mark should be removed.
• The digitised text should be true to the hard copy.
• Ensuring that the Title, Author, and Headline fields are written in Roman script using the LDC-IL transliteration scheme.

Data Encoding

The collected data should be encoded in a machine-readable format for further analysis. When storing the data, it is important to follow to certain standards to ensure ease of storage and overlong retrieval. The LDC-IL Text corpus uses Unicode encoding and is stored in XML format. Large-scale language resources depend on metadata to ensure the authenticity, which serves as a compulsory fragment of any corpus.

LDC-IL XML Markup Standard ( link it here https://dev.ldcil.org/standardsTextXML.aspx)

Speech Data Creation

The LDC-IL speech corpus was precisely accumulated to meet the diverse needs of the research and development community for numerous types of speech-based linguistic analysis. The availability of speech technology in Indian languages has been minimal because speech data accumulation is highly challenging in Indian languages. To provide language support for Indian languages in various speech aided applications LDC-IL started crating speech data.

The data was collected for a wide range of tasks such as Automatic Speech Recognition (ASR), Speech-to-Text (STT) systems, linguistic analysis, and speech therapy. To achieve this, various types of content were recorded. Thus the datasets were designed in such a way that to cover all the phonemes and allophones of the language in all possible environments. To ensure the real world usage continuous speech was recorded in natural environments. This careful planning and expert input have resulted in a versatile and comprehensive speech corpus that supports a broad range of linguistic and technological applications.

Speech Data Collection Guideline

Speech Data Collection

• Data should be taken from native speaker
• The data should be collected from all possible dialect.
• Collect metadata such as speaker ID, age, gender, language, dialect, recording environment, and date.
• Obtain signed consent forms from each participant.
• Data should not be taken from too noisy environment.
• Allow the participant to be as comfortable sit that reduce the stress levels so to get natural speech.
• Place the microphone at an optimal distance and angle to capture clear audio.
• Record audio in wav format
• Ensure the recording equipment functioning correctly and data is stored securely.
• Carry spare batteries, memory cards etc.
• Regularly back up recordings to prevent data loss.

Organizing the data

• After the field work is completed, the data has to be stored in a server as soon as possible for safe keeping.
• Keep a backup of the saved data.
• After the data is stored, it is segmented and mapped with its corresponding text and metadata.

Data Verification

• Metadata needs to be verified.
• Audio against text mapping needs to be checked.
• The data duplication needs to be checked.
• Naming Conventions need to be verified.

Speech Data Types

• Contemporary Text (News)
• Creative Text
• Sentence
• Date
• Command and Control Words
• Place Name
• Person Name
• Most Frequent Word-Part
• Most Frequent Word-Full Set
• Phonetically Balanced-Full Set
• Form and Function Word-Full Set

Speech Annotation

A dataset containing language-specific information is called an annotated corpus. It can be used to train machine learning algorithms to capture the computational attributes of language structure. It simplify the system to easily identify patterns.

An annotated speech corpus provides a wide range of linguistic information, which is particularly useful for analyzing the phonetic aspects of a language. Speech data is annotated at various levels, including phone, phoneme, syllable, word, and sentence. Annotating data from the morphosyntactic to the pragmatic level requires careful attention to the structure of speech and its attributes. Annotated data can be used for speech recognition, speech synthesis, and many other language-related technologies.

LDC-IL Guideline for Speech Annotation-1.0

Guidelines for Phonetically Normalized Speech Annotation

In phonetically normalized speech annotation layer should be carried out as per the pronunciation of the speaker in the audio. The wrong pronunciation, that is, deviation from the text should be transcribed accordingly.

Following are the instances which should be marked in the annotation.

1. Use of ‘#’

‘#’ symbol is used to mark excessively noisy or unnecessary parts of the audio – generally these portions also contain human speech but are not legible. The audio is post-processed to remove such portions from the audio files that are being released publicly. # should be used within the sentence.

2. Use of ‘0’

‘0’ is used to mark silences at the beginning and end of the audio files as well as long silences or non-speech sounds in the intermediate portions of the audio files - these are marked only for those portions which do not have any kind of human speech. These portions are marked so that it can be removed in the post-processing step.

3. Silences are removed

Any silence longer than 50 ms should be marked.

4. Cut-off speech and intended speech should be marked.

Example: [mini]*ster —shows that the speaker intended to speak minister but spoke mini in an unclear fashion and ster clearly.

5. Annotation of speech disfluency

Restarts/false starts should be marked. For example, if the speaker intends to speak “Keralam” but speaks “Ke Keralam”, this should be marked as Ke-Keralam.

6. Numbers

All number sequences should be spelled out. Years should be transcribed in spoken format.

7. Mispronunciations

If a speaker mispronounces a word and the mispronunciation is not an actual word, transcription should be done as the word is spoken.

8. Utterances longer than 30 seconds should be further split into multiple parts - this split is made at the point of a long silence of around 500 ms.

Guidelines for Orthographically Normalized Speech Annotation

Orthographically normalized textual layer is prepared over the phonetically normalized text by using the guidelines given below

1. Small portion of a word like a grammatical element or a letter is missed then it is corrected as per the correct writing pattern. For example, if the speaker speaks ‘avan viitti poyi’ ‘He went home’ in an informal way then it is annotated as ‘avan viittil poyi’ in the proper standard Malayalam sentence. There is no valid word ‘viitti’ in the language, so it is annotated as a proper word according to the context.
2. Any deviation in the phonetically annotated text is corrected according to standard writing form. For example, if the audio is ‘Maiyam Engineering’ it must be corrected as ‘Marine Engineering’.
3. Restarts/false starts are removed. For example, the speaker intends to speak ‘keralam’ but speaks ‘kekeralam’. If the word ‘ke’ is a valid word morpheme, then it has been kept, otherwise ‘ke’ is not marked.
4. Cut-off speech is written as standard form and remove [ ]* symbol
5. Incomplete sentence is standardized to the extent of available audio.

LDC-IL Guideline for Speech Annotation-2.0

1. Use of ‘#’

‘#’ symbol is used to mark excessively noisy or unnecessary parts of the audio – generally these portions also contain human speech but are not legible. The audio is post-processed to remove such portions from the audio files that are being released publicly. # should be used within the sentence.

2. Use of ‘0’

‘0’ is used to mark silences at the beginning and end of the audio files as well as long silences or non-speech sounds in the intermediate portions of the audio files - these are marked only for those portions which do not have any kind of human speech. These portions are marked so that it can be removed in the post-processing step.

3. Removal of Silence

A period of relative quietness occurs in the audio files where speaker stops to think or hesitates before saying a word. The expression relative quietness is used here because there is no actual silence in the speech signal due to line and environmental noise. Any silence longer than 50 ms to be marked using ‘#’ or ‘0’ based on the context.

4. Cut-off speech

In case a small portion of word like grammatical element or a letter is missed then it should be corrected as per the correct writing pattern.

For example if the speaker speaks ‘avan viitti poyi’ ‘He went home’ in an informal way then it is annotated as ‘avan viittil poyi’ in the proper standard Malayalam sentence. There is no valid word ‘viitti’ in the language so it should be annotated as a proper word according to the context.

5. Dis-fluency

For any deviation in the utterance from the standard form of pronunciation, transcription should to be kept in the standard writing form

For example if the audio is “Maiyen Engineering”, it has to be corrected as “Marine Engineering”

6. Mispronunciations

If a speaker mispronounces a word and the mispronunciation is an actual word in the language, transcription to be done as the word is spoken.

For example if the audio is ‘avaɭ’ ‘she’ instead of ‘avan’ ‘he’ then transcribe according to the wave.

7. Unnecessary repetition

Eliminate unnecessary repetition of words by marking as ‘0’ or ‘#’ as the case may be.

8. Restart/False start

Eliminated starts by marking as ‘0’ or ‘#’ as the case may be.

For example ‘kaɽɳaːʈakajuʈe t̪alast̪ʰaːnamaːɳ beŋgaɭuːru’ ‘Bangalore is the capital of Karnataka’ is the sentence with Restart and false starts words then following way it should be annotated.

For example “bengaluru is the capital of karnatka”

In Audio: ma mandya is a district of gar karnataka

Annotation: | 0 | mandya is the district of | # | karnataka

9. Utterances should be no longer than 30secs. So the annotator should find a silence and split the sentence appropriately or if the utterance is too long then break the sentence at phrase endings by marking boundary.
10. Incomplete sentence should be annotated to the extent of available audio only.
11. Numbers: All number sequences are spelled out
12. No punctuations should be used in the annotations.

Parts of Speech Tagging

It is the process of assigning a word in a text as corresponding to a particular part of speech on the basis of its definition and its occurrence in a given context. The process is basically to design or provide help in creation of appropriate language technology. Since each PoS tag is attached to a single word, preprocessing mechanisms such as splitting, tokenization, etc. have already been performed to filter out typesetting based-raw corpus. This is in response to meet the requirement of standardization amongst the Indian languages that exhibit a very rich system of morphology where words appear long with complex morpho-phonemic and morpho-syntactic changes at the junctures.

PoS Tagging Guidelines

In order to develop various TagSets for individual languages, the LDC-IL has undertaken certain linguistic modus operandi as laid down below:

• Defining the traditional parts of speech along with the examples
• Understanding the concept of Form and Function (Pronouns, Demonstratives, Numerals, etc.)
• Recognizing the fuzzy boundaries between the grammatical classes, i.e., a lexical item may function as a specific category and the same may function as different category in different context (Gerunds vs. Infinitive/Participle etc).
• Working out the syntactic relation between the modifier-modified (Adj-Noun; Participle-Noun).
• Realizing the morpho-syntactic features a particular lexical item carries in a given syntactic configuration. (Person-Number-Gender/Case; Tense-Aspect-Mood/Mod).

LDC-IL POS Tag Set

The priority is to cover all the Scheduled languages and then take up other non-scheduled languages. LDC-IL planned to make PoS tagged data on 22 Scheduled Languages such as Assamese, Bengali, Bodo, Dogri, Gujarati, Hindi, Kannada, Kashmiri, Konkani, Maithili, Malayalam, Manipuri, Marathi, Nepali, Odia, Punjabi, Sanskrit, Santali, Sindhi, Tamil, Telugu, Urdu. 15 PoS tag set were prepared and they are listed below:

1. Assamese
2. Bengali
3. Bodo
4. Dogri
5. Gujarati
6. Hindi
7. Kannada
8. Kashmiri
9. Malayalam
10. Manipuri
11. Nepali
12. Odia
13. Punjabi
14. Tamil
15. Urdu

Chunking

The process of annotating tagged tokens with structures in a non-hierarchical and non-recursive way is Chunking. It is acknowledged that segmentation and labeling are the most common operations in language processing. Chunking is a popular representative of a segmentation process aiming to segment the tagged tokens into meaningful structures. In the meantime, chunks generally do not try to analyze entire sentences, but only try to build “chunks” of words. In this line of view, the rule system of chunks is relatively simple, robust, and efficient.

Chunking Guidelines

The scheme has adopted certain set of linguistic norms which should be followed by the Resource Persons working on chunking. The chunking of linguistic expression is purely based on specific categorical label and hence the following linguistic guides are being introduced for the ease of annotators.

• Identifying different chunk levels along with the typical examples.
• Keeping in mind that minimal recursive phrases (nominal or verbal) should be captured.
• Understanding the idea that chunking operates on the minimal non-recursive phrases and within such minimal construction, there is no nested structure.
• Make sure that nested non-recursive clusters are identified with their heads. (Possessive Constructions, Spatial Relational Nouns, Nested Modifier inside noun phrases).
• Having the knowledge as well as hands-on experience of linguistic phenomena such as scrambling of the lexical items, dislocated element, spelling out of boundary elements realized as case markers or tense, mood, aspects etc, between two expressions that operate on the data of the language concerned.

Parallel Corpora

A parallel corpus contains translations of the same document in two or more languages, aligned at least at the sentence level. Parallel corpora are generally data sets of translated sentence pairs. It can be used to train and test machine translation models. Parallel corpora play a vital role in translation studies and contrastive linguistics. Moreover, it is significantly aiding the exploration of inter-linguistic phenomena and serves as a valuable resource for language teaching.

LDC-IL is creating parallel corpora in 270 Indian languages. The data is creating by outsourcing all over India including the remote part of the country through an online platform called Trankit.

Guideline for Translator

• The translation should be as natural as the target language.
• Translate according to the cultural nuances of the target language and try to convey the intended meaning of the source language.
• Source language style should match with target language.
• Keep consistency in the terminology.
• Maintain readability.
• Keep coherence in the sentence structure that helps to give natural flow of language.
• Find an equivalent expression of idiomatic expression in the target language.
• Follow the grammar and syntax rules of the target language.
• Ensure technical terminology in the target language, especially in scientific, medical, legal, mathematical, and technological domains.
• Careful about the ethical issues
• When the source text is intentionally ambiguous, the translator must decide whether to maintain that ambiguity or clarify it in the translation. [if you can make the same ambiguity in the target language then keep it otherwise explain it]

Guidelines for the Reviewers

• The sentence which has correction/error should be commented and rectified. The edits you make on the portal can be viewed by all concerned individuals (including the user, project manager and the reviewer). Reviewers should be careful in not penalizing the translator without any substantial reason and the reason must be specified for the edits in the comments given to it.
• Machine translation is not acceptable if it violates the target language structure and flow. However, translators should not be penalized just because a translation is similar to what any MT system gives.
• The translation should be as natural as per the flow of target language.
• Translate according to the cultural nuances of the target language and try to convey the intended meaning of the source language.
• Source language style should match with target language, wherever possible. Adaptation is not intended.
• Maintain readability i.e. the target text should sound natural in the language.
• Keep coherence in the sentence structure that helps to give natural flow of language.
• Find an equivalent expression of idiomatic expression in the target language, wherever possible.
• Follow the grammar and syntax rules of the target language.
• Ensure technical terminology in the target language, especially in scientific, medical, legal, mathematical, and technological domains.
• When the source text is intentionally ambiguous, the translator must decide whether to maintain that ambiguity or clarify it in the translation. [if you can make the same ambiguity in the target language then keep it otherwise explain it]
• In case of any doubt, please use the "Raise Issue" button to mark the issue to the translator and the project manager.

Trankit Review Procedure for Parallel Corpus

As the reviewers keep on reviewing, they should make a note of their review comments, if required, for each of the segments. This helps in preparing a score card automatically that can be shared with the translator as well.

Preparing the Score Card as you review

The scorecard is the means of providing feedback primarily to the translator, but ultimately to any stakeholders. The scorecard provides insight into the quality of the translation, the errors that were found (if any) and general assessment.

During the review, the reviewer is supposed to record any errors found into the scorecard, including the location of the error (i.e. the segment or the task), source and target text, corrected target text, type of error and its severity.

The final error score is calculated in ampersand based on number of errors (with severity weighted in) with respect of the volume.

How to generate the Review Report

You do not need to anything separately to generate the Review Report. As you go on reviewing the segments of any given task, you can keep on making comments and giving your scores for each of the segments, wherever necessary.

If a segment has any error, you can just edit the text, give a comment and hit the score button to give an error score to the segment. The E button pop's out a window where you can select the error type and its severity. You can also use this button to give kudos to the translator for coming up with an extra-ordinary translation for the given segment.

Do not forget to fill in both error type and severity for all errors. However, the exceptions are Kudos, Preferential Change and Repeat - these do not have any severities, so no point in filling it in (the severity field is greyed out when these three error types are selected.

Error Definitions and Severities

1. Naturalness

When the translation is not fluent, well-written in the eyes of a native speaker, but is not necessarily due to wrong meaning, wrong grammar, punctuation or spelling. Definition in the Oxford English dictionary: The ease with which a text may be scanned or read; the quality in a book, etc., of being easy to understand and enjoyable to read.

EXAMPLE:

Literal and word-to-word translations sound awkward and inauthentic, (no effort to paraphrase).

Native speakers would not have used unnatural sentence structure and word order or word combination that. No effort to paraphrase and adjust to the norms of your language.

Literal translations that make it difficult to identify the main topic.

2. Compliance

When the translator fails to follow requirements specified in the Style guide and/or project instructions.

EXAMPLE:

When a segment is left un-translated (missing translation), while it shouldn't be, e.g. there is no instruction telling the translators to leave anything un-translated.

Identical sentences were translated inconsistently when they should be consistent (maybe because a big file was translated by several translators).

Grammar and punctuation errors are described in the style guide (if available).

3. Grammar

Grammatical error, word order replacement, gender, singular/plural, or any other violation of language rules need to be checked.

The translation does not adhere to the target language-specific rules with regard to grammar or syntax.

4. Meaning

Translation does not convey the meaning or nuance intended in the original source.

Omissions or additions that improve the language flow or makes translation more natural should not be a meaning error.

EXAMPLE:

"Error: Duplicate site or pages" translated as though it means "Error: please duplicate these sites or pages" when what is meant is "an error occurred where some sites were duplicated".

5. Punctuation/Spelling

Punctuation: General punctuation error and capitalization e.g. wrong use of commas, colons, dashes, etc. Spelling: Typos, use of outdated/uncommon spelling (or character), etc.

EXAMPLE:

Leaving in the period inside the quotation mark in [Click "OK," then exit.] when it should have been moved outside the quotation mark in your language.

Using double quotation marks (") when you should use single quotation marks (') in your languages.

6. Terminology

Terms not following the translation in glossary/TD or specification in any other forms of communication including style guide and feedback from previous projects.

7. Preferential Change

If there is a mistake in the source, reasonable misunderstanding of the source due to lack of information then don’t consider as the fault of a translator even it is inappropriate translation.

When the translation needs to be changed but the translation is not wrong in any way. This category has no weight in the final score.

EXAMPLE:

"Home" can be either the "homepage" or a "home" label for phone number. When there is no description or screenshot and the previous/next string is "office", translating that into "home" phone number would make sense, even though the right translation might be "homepage".

8. Repeat

The Repeat code is used to mark repetitive instances of an error that was already penalized. It has no weight in the final score.

If the reviewer has already penalized the translator for an error once, the translator cannot be penalized again for the same.

EXAMPLE:

The same grammatical, spelling, or standard punctuation error repeated throughout a project.

9. Kudos

The Reviewer can assign Kudos to mark excellent translations, but Kudos doesn’t influence the scoring formulas in any way. This is because particularly successful translations do not mitigate the quality impact of any concrete error.

Error category notes

1. An error should be objectively classified as one of the error categories according to the guidelines above. If an error falls into more than one error category, it will be categorized based on the most severe error category.
2. Changing the writing style according to personal preferences does not play a role in a translation review, unless specifically required by a project (which is unlikely here). Only indisputable language errors (Errors that can be verified against a grammar/usage book, approved TMs and glossaries, language specific style guides (if available and provided) and translation guidelines for the relevant language.) are classified as errors and included in the error count. There are cases in which a change would greatly improve the intelligibility level of the text, but the text does not contain an objective language error. The reviewer uses Compliance category in such cases.

Severity Levels

• A CRITICAL ERROR is any error:

That causes an application to crash. (such as when the string has a placeholder which should not have been translated, check if your project tasks have placeholders in it).

That modifies or misinterprets the source text in a way that makes the end-user to be disoriented and that would cause the reader to seek un-called for clarifications.

That results in a potentially offensive statement (morally, politically, regarding religion or culture) or can cause embarrassment to the client/project owner.

That is in a highly visible part of documentation or software (e.g. cover page, menu command, site web page, site opening page, etc.).

For which the localizer/translator has repeatedly ignored previous reviewer feedback without any particular reason for not implementing it. The reviewer should be informed beforehand of any correction from previous reviews not implemented.

For which the localizer/translator has failed to implement query answers regarding key information.

• A MAJOR ERROR is any error:

That appears in an important or visible location (header, TOC, chapter title, help topic title, web page section title, etc.) regardless of error category.

That result in a significant change in the meaning, which causes that the user is very likely to be misled (severe Accuracy or context-dependent Terminology errors).

In grammar or syntax that is a gross violation of generally accepted language conventions.

For which the localizer/translator has ignored previous language QA feedback without any particular reason for not implementing it. The reviewer/customer should be informed beforehand of any correction from previous QA not implemented.

For which the localizer/translator has failed to implement query answers that do not influence key product features or information.

• A MINOR ERROR is any error:

That results in a slight change in the meaning.

That would not confuse or mislead a user but could be noticed.

In the Formal Category that does not result in misinterpretation of the source or in the wrong functionality of the product (e.g. page layout, spacing, tables and numbered/bulleted lists, fonts, font style, etc.).

If the use of punctuation or capitalization not resulting in a loss of the meaning.

In style that does not result in misinterpretation of the source.

In grammar or syntax that is a minor violation of generally accepted language conventions.

In the Style Guide/Product Guidelines category that does not result in misinterpretation of the source or in the wrong functionality of the product.

Giving Overall Feedback

Please provide your overall impression, point out repetitive issues, what the localisation team should focus on to improve the quality, what the good points are and what the main problem is.

To provide constructive and actionable feedback to the translator, please answer these questions:

• How do you rate the overall linguistic experience?
• What did the translator do well?
• What are the most significant or recurrent errors found?
• Can you guess the root cause of the most significant errors?
• What is the impact of the errors for the end-user?
• Would you recommend any special training?

Transliteration

Unicode has become the standard for encoding text from different languages and scripts. However, legacy systems may have different approach and it struggle data processing and communication. The Linguistic Data Consortium for Indian Languages (LDC-IL) has addressed this by developing a transliteration schema that converts Indian scripts into Roman script using the ASCII range.

By transliterating Indian scripts into ASCII characters, LDC-IL ensure that these systems can process, store, and transmit text data without encountering errors or data corruption. Transliteration allows text data in Indian scripts to be easily incorporated into ASCII-based Roman script to ensure the text documents can be exchanged without loss of information or integrity. This transliteration schema plays a crucial role in accessibility and usability for users who are not familiar with Indian scripts and maintaining the continuity and functionality of various digital systems.

LDC-IL Transliterator Application converts 9 Indic Scripts to Roman and Vice-versa. They are Assamese/Bengali, Devanagari, Gujarati, Gurumukhi, Kannada, Malayalam, Odia, Tamil and Telugu.

LDC-IL Transliteration Scheme

Character	Unicode Value in Decimal	Roman Transliteration Schema	Character Description
ঁ	2433	m'	ASSAMESE/BENGALI SIGN CANDRABINDU
ং	2434	M	ASSAMESE/BENGALI SIGN ANUSVARA
ঃ	2435	H	ASSAMESE/BENGALI SIGN VISARGA
অ	2437	a	ASSAMESE/BENGALI LETTER A
আ	2438	A	ASSAMESE/BENGALI LETTER AA
ই	2439	i	ASSAMESE/BENGALI LETTER I
ঈ	2440	I	ASSAMESE/BENGALI LETTER II
উ	2441	u	ASSAMESE/BENGALI LETTER U
ঊ	2442	U	ASSAMESE/BENGALI LETTER UU
ঋ	2443	x	ASSAMESE/BENGALI LETTER VOCALIC R
ঌ	2444	q	ASSAMESE/BENGALI LETTER VOCALIC L
এ	2447	E	ASSAMESE/BENGALI LETTER E
ঐ	2448	ai	ASSAMESE/BENGALI LETTER AI
ও	2451	O	ASSAMESE/BENGALI LETTER O
ঔ	2452	au	ASSAMESE/BENGALI LETTER AU
ক	2453	ka	ASSAMESE/BENGALI LETTER KA
খ	2454	kha	ASSAMESE/BENGALI LETTER KHA
গ	2455	ga	ASSAMESE/BENGALI LETTER GA
ঘ	2456	gha	ASSAMESE/BENGALI LETTER GHA
ঙ	2457	ng'a	ASSAMESE/BENGALI LETTER NGA
চ	2458	ca	ASSAMESE/BENGALI LETTER CA
ছ	2459	cha	ASSAMESE/BENGALI LETTER CHA
জ	2460	ja	ASSAMESE/BENGALI LETTER JA
ঝ	2461	jha	ASSAMESE/BENGALI LETTER JHA
ঞ	2462	nj'a	ASSAMESE/BENGALI LETTER NYA
ট	2463	Ta	ASSAMESE/BENGALI LETTER TTA
ঠ	2464	Tha	ASSAMESE/BENGALI LETTER TTHA
ড	2465	Da	ASSAMESE/BENGALI LETTER DDA
ঢ	2466	Dha	ASSAMESE/BENGALI LETTER DDHA
ণ	2467	Na	ASSAMESE/BENGALI LETTER NNA
ত	2468	ta	ASSAMESE/BENGALI LETTER TA
থ	2469	tha	ASSAMESE/BENGALI LETTER THA
দ	2470	da	ASSAMESE/BENGALI LETTER DA
ধ	2471	dha	ASSAMESE/BENGALI LETTER DHA
ন	2472	na	ASSAMESE/BENGALI LETTER NA
প	2474	pa	ASSAMESE/BENGALI LETTER PA
ফ	2475	pha	ASSAMESE/BENGALI LETTER PHA
ব	2476	ba	ASSAMESE/BENGALI LETTER BA
ভ	2477	bha	ASSAMESE/BENGALI LETTER BHA
ম	2478	ma	ASSAMESE/BENGALI LETTER MA
য	2479	ya	ASSAMESE/BENGALI LETTER YA
র	2480	ra	BENGALI LETTER RA
ল	2482	la	ASSAMESE/BENGALI LETTER LA
শ	2486	sha	ASSAMESE/BENGALI LETTER SHA
ষ	2487	Sa	ASSAMESE/BENGALI LETTER SSA
স	2488	sa	ASSAMESE/BENGALI LETTER SA
হ	2489	ha	ASSAMESE/BENGALI LETTER HA
়	2492	'	ASSAMESE/BENGALI SIGN NUKTA
ঽ	2493	ঽ	ASSAMESE/BENGALI SIGN AVAGRAHA
া	2494	A	ASSAMESE/BENGALI VOWEL SIGN AA
ি	2495	i	ASSAMESE/BENGALI VOWEL SIGN I
ী	2496	I	ASSAMESE/BENGALI VOWEL SIGN II
ু	2497	u	ASSAMESE/BENGALI VOWEL SIGN U
ূ	2498	U	ASSAMESE/BENGALI VOWEL SIGN UU
ৃ	2499	x	ASSAMESE/BENGALI VOWEL SIGN VOCALIC R
ৄ	2500	X	ASSAMESE/BENGALI VOWEL SIGN VOCALIC RR
ে	2503	E	ASSAMESE/BENGALI VOWEL SIGN E
ৈ	2504	ai	ASSAMESE/BENGALI VOWEL SIGN AI
ো	2507	O	ASSAMESE/BENGALI VOWEL SIGN O
ৌ	2508	au	ASSAMESE/BENGALI VOWEL SIGN AU
ৎ	2510	t	ASSAMESE/BENGALI LETTER KHANDA TA
ড়	2524	D'a	ASSAMESE/BENGALI LETTER RRA
ঢ়	2525	Dh'a	ASSAMESE/BENGALI LETTER RHA
য়	2527	Ya	ASSAMESE/BENGALI LETTER YYA
ৠ	2528	X	ASSAMESE/BENGALI LETTER VOCALIC RR
ৡ	2529	Q	ASSAMESE/BENGALI LETTER VOCALIC LL
ৢ	2530	q	ASSAMESE/BENGALI VOWEL SIGN VOCALIC L
ৣ	2531	Q	ASSAMESE/BENGALI VOWEL SIGN VOCALIC LL
0	2534	0	ASSAMESE/BENGALI DIGIT ZERO
1	2535	1	ASSAMESE/BENGALI DIGIT ONE
2	2536	2	ASSAMESE/BENGALI DIGIT TWO
3	2537	3	ASSAMESE/BENGALI DIGIT THREE
4	2538	4	ASSAMESE/BENGALI DIGIT FOUR
5	2539	5	ASSAMESE/BENGALI DIGIT FIVE
6	2540	6	ASSAMESE/BENGALI DIGIT SIX
7	2541	7	ASSAMESE/BENGALI DIGIT SEVEN
8	2542	8	ASSAMESE/BENGALI DIGIT EIGHT
9	2543	9	ASSAMESE/BENGALI DIGIT NINE
ৰ	2544	ra	ASSAMESE LETTER RA
ৱ	2545	wa	ASSAMESE LETTER WA
ँ	2305	m'	DEVANAGARI SIGN CANDRABINDU
ं	2306	M	DEVANAGARI SIGN ANUSVARA
ः	2307	H	DEVANAGARI SIGN VISARGA
अ	2309	a	DEVANAGARI LETTER A
आ	2310	A	DEVANAGARI LETTER AA
इ	2311	i	DEVANAGARI LETTER I
ई	2312	I	DEVANAGARI LETTER II
उ	2313	u	DEVANAGARI LETTER U
ऊ	2314	U	DEVANAGARI LETTER UU
ऋ	2315	x	DEVANAGARI LETTER VOCALIC R
ऌ	2316	q	DEVANAGARI LETTER VOCALIC L
ए	2319	E	DEVANAGARI LETTER E
ऐ	2320	ai	DEVANAGARI LETTER AI
ऑ	2321	ao	DEVANAGARI LETTER CANDRA O
ओ	2323	O	DEVANAGARI LETTER O
औ	2324	au	DEVANAGARI LETTER AU
क	2325	ka	DEVANAGARI LETTER KA
ख	2326	kha	DEVANAGARI LETTER KHA
ग	2327	ga	DEVANAGARI LETTER GA
घ	2328	gha	DEVANAGARI LETTER GHA
ङ	2329	ng'a	DEVANAGARI LETTER NGA
च	2330	ca	DEVANAGARI LETTER CA
छ	2331	cha	DEVANAGARI LETTER CHA
ज	2332	ja	DEVANAGARI LETTER JA
झ	2333	jha	DEVANAGARI LETTER JHA
ञ	2334	nj'a	DEVANAGARI LETTER NYA
ट	2335	Ta	DEVANAGARI LETTER TTA
ठ	2336	Tha	DEVANAGARI LETTER TTHA
ड	2337	Da	DEVANAGARI LETTER DDA
ढ	2338	Dha	DEVANAGARI LETTER DDHA
ण	2339	Na	DEVANAGARI LETTER NNA
त	2340	ta	DEVANAGARI LETTER TA
थ	2341	tha	DEVANAGARI LETTER THA
द	2342	da	DEVANAGARI LETTER DA
ध	2343	dha	DEVANAGARI LETTER DHA
न	2344	na	DEVANAGARI LETTER NA
प	2346	pa	DEVANAGARI LETTER PA
फ	2347	pha	DEVANAGARI LETTER PHA
ब	2348	ba	DEVANAGARI LETTER BA
भ	2349	bha	DEVANAGARI LETTER BHA
म	2350	ma	DEVANAGARI LETTER MA
य	2351	ya	DEVANAGARI LETTER YA
र	2352	ra	DEVANAGARI LETTER RA
ऱ	2353	Ra	DEVANAGARI LETTER RRA
ल	2354	la	DEVANAGARI LETTER LA
ळ	2355	La	DEVANAGARI LETTER LLA
ऴ	2356	Za	DEVANAGARI LETTER LLLA
व	2357	va	DEVANAGARI LETTER VA
श	2358	sha	DEVANAGARI LETTER SHA
ष	2359	Sa	DEVANAGARI LETTER SSA
स	2360	sa	DEVANAGARI LETTER SA
ह	2361	ha	DEVANAGARI LETTER HA
़	2364	'	DEVANAGARI SIGN NUKTA
ा	2366	A	DEVANAGARI VOWEL SIGN AA
ि	2367	i	DEVANAGARI VOWEL SIGN I
ी	2368	I	DEVANAGARI VOWEL SIGN II
ु	2369	u	DEVANAGARI VOWEL SIGN U
ू	2370	U	DEVANAGARI VOWEL SIGN UU
ृ	2371	x	DEVANAGARI VOWEL SIGN VOCALIC R
ॄ	2372	X	DEVANAGARI VOWEL SIGN VOCALIC RR
े	2375	E	DEVANAGARI VOWEL SIGN E
ै	2376	ai	DEVANAGARI VOWEL SIGN AI
ॉ	2377	ao	DEVANAGARI VOWEL SIGN CANDRA O
ो	2379	O	DEVANAGARI VOWEL SIGN O
ौ	2380	au	DEVANAGARI VOWEL SIGN AU
ॐ	2384	@M	DEVANAGARI OM
क़	2392	k'a	DEVANAGARI LETTER QA
ख़	2393	kh'a	DEVANAGARI LETTER KHHA
ग़	2394	g'a	DEVANAGARI LETTER GHHA
ज़	2395	j'a	DEVANAGARI LETTER ZA
ड़	2396	D'a	DEVANAGARI LETTER DDDHA
ढ़	2397	Dh'a	DEVANAGARI LETTER RHA
फ़	2398	ph'a	DEVANAGARI LETTER FA
य़	2399	Ya	DEVANAGARI LETTER YYA
ॠ	2400	X	DEVANAGARI LETTER VOCALIC RR
ॡ	2401	Q	DEVANAGARI LETTER VOCALIC LL
ॢ	2402	q	DEVANAGARI VOWEL SIGN VOCALIC L
ॣ	2403	Q	DEVANAGARI VOWEL SIGN VOCALIC LL
।	2404	.	DEVANAGARI DANDA
॥	2405	..	DEVANAGARI DOUBLE DANDA
0	2406	0	DEVANAGARI DIGIT ZERO
1	2407	1	DEVANAGARI DIGIT ONE
2	2408	2	DEVANAGARI DIGIT TWO
3	2409	3	DEVANAGARI DIGIT THREE
4	2410	4	DEVANAGARI DIGIT FOUR
5	2411	5	DEVANAGARI DIGIT FIVE
6	2412	6	DEVANAGARI DIGIT SIX
7	2413	7	DEVANAGARI DIGIT SEVEN
8	2414	8	DEVANAGARI DIGIT EIGHT
9	2415	9	DEVANAGARI DIGIT NINE
ઁ	2689	m'	GUJARATI SIGN CANDRABINDU
ં	2690	M	GUJARATI SIGN ANUSVARA
ઃ	2691	H	GUJARATI SIGN VISARGA
અ	2693	a	GUJARATI LETTER A
આ	2694	A	GUJARATI LETTER AA
ઇ	2695	i	GUJARATI LETTER I
ઈ	2696	I	GUJARATI LETTER II
ઉ	2697	u	GUJARATI LETTER U
ઊ	2698	U	GUJARATI LETTER UU
ઋ	2699	x	GUJARATI LETTER VOCALIC R
ઌ	2700	q	GUJARATI LETTER VOCALIC L
ઍ	2701	ae	GUJARATI VOWEL CANDRA E
એ	2703	E	GUJARATI LETTER E
ઐ	2704	ai	GUJARATI LETTER AI
ઑ	2705	ao	GUJARATI VOWEL CANDRA O
ઓ	2707	O	GUJARATI LETTER O
ઔ	2708	au	GUJARATI LETTER AU
ક	2709	ka	GUJARATI LETTER KA
ખ	2710	kha	GUJARATI LETTER KHA
ગ	2711	ga	GUJARATI LETTER GA
ઘ	2712	gha	GUJARATI LETTER GHA
ઙ	2713	ng'a	GUJARATI LETTER NGA
ચ	2714	ca	GUJARATI LETTER CA
છ	2715	cha	GUJARATI LETTER CHA
જ	2716	ja	GUJARATI LETTER JA
ઝ	2717	jha	GUJARATI LETTER JHA
ઞ	2718	nj'a	GUJARATI LETTER NYA
ટ	2719	Ta	GUJARATI LETTER TTA
ઠ	2720	Tha	GUJARATI LETTER TTHA
ડ	2721	Da	GUJARATI LETTER DDA
ઢ	2722	Dha	GUJARATI LETTER DDHA
ણ	2723	Na	GUJARATI LETTER NNA
ત	2724	ta	GUJARATI LETTER TA
થ	2725	tha	GUJARATI LETTER THA
દ	2726	da	GUJARATI LETTER DA
ધ	2727	dha	GUJARATI LETTER DHA
ન	2728	na	GUJARATI LETTER NA
પ	2730	pa	GUJARATI LETTER PA
ફ	2731	pha	GUJARATI LETTER PHA
બ	2732	ba	GUJARATI LETTER BA
ભ	2733	bha	GUJARATI LETTER BHA
મ	2734	ma	GUJARATI LETTER MA
ય	2735	ya	GUJARATI LETTER YA
ર	2736	ra	GUJARATI LETTER RA
લ	2738	la	GUJARATI LETTER LA
ળ	2739	La	GUJARATI LETTER LLA
વ	2741	va	GUJARATI LETTER VA
શ	2742	sha	GUJARATI LETTER SHA
ષ	2743	Sa	GUJARATI LETTER SSA
સ	2744	sa	GUJARATI LETTER SA
હ	2745	ha	GUJARATI LETTER HA
઼	2748	'	GUJARATI SIGN NUKTA
ા	2750	A	GUJARATI VOWEL SIGN AA
િ	2751	i	GUJARATI VOWEL SIGN I
ી	2752	I	GUJARATI VOWEL SIGN II
ુ	2753	u	GUJARATI VOWEL SIGN U
ૂ	2754	U	GUJARATI VOWEL SIGN UU
ૃ	2755	x	GUJARATI VOWEL SIGN VOCALIC R
ૄ	2756	X	GUJARATI VOWEL SIGN VOCALIC RR
ૅ	2757	ae	GUJARATI VOWEL SIGN CANDRA E
ે	2759	E	GUJARATI VOWEL SIGN E
ૈ	2760	ai	GUJARATI VOWEL SIGN AI
ૉ	2761	ao	GUJARATI VOWEL SIGN CANDRA O
ો	2763	O	GUJARATI VOWEL SIGN O
ૌ	2764	au	GUJARATI VOWEL SIGN AU
ૠ	2784	X	GUJARATI LETTER VOCALIC RR
ૡ	2785	Q	GUJARATI LETTER VOCALIC LL
ૢ	2786	q	GUJARATI VOWEL SIGN VOCALIC L
ૣ	2787	Q	GUJARATI VOWEL SIGN VOCALIC LL
0	2790	0	GUJARATI DIGIT ZERO
1	2791	1	GUJARATI DIGIT ONE
2	2792	2	GUJARATI DIGIT TWO
3	2793	3	GUJARATI DIGIT THREE
4	2794	4	GUJARATI DIGIT FOUR
5	2795	5	GUJARATI DIGIT FIVE
6	2796	6	GUJARATI DIGIT SIX
7	2797	7	GUJARATI DIGIT SEVEN
8	2798	8	GUJARATI DIGIT EIGHT
9	2799	9	GUJARATI DIGIT NINE
ਁ	2561	M'	GURMUKHI SIGN ADAK BINDI
ਂ	2562	M	GURMUKHI SIGN BINDI
ਃ	2563	H	GURMUKHI SIGN VISARGA
ਅ	2565	a	GURMUKHI LETTER A
ਆ	2566	A	GURMUKHI LETTER AA
ਇ	2567	i	GURMUKHI LETTER I
ਈ	2568	I	GURMUKHI LETTER II
ਉ	2569	u	GURMUKHI LETTER U
ਊ	2570	U	GURMUKHI LETTER UU
ਏ	2575	E	GURMUKHI LETTER EE
ਐ	2576	ai	GURMUKHI LETTER AI
ਓ	2579	O	GURMUKHI LETTER OO
ਔ	2580	au	GURMUKHI LETTER AU
ਕ	2581	ka	GURMUKHI LETTER KA
ਖ	2582	kha	GURMUKHI LETTER KHA
ਗ	2583	ga	GURMUKHI LETTER GA
ਘ	2584	gha	GURMUKHI LETTER GHA
ਙ	2585	ng'a	GURMUKHI LETTER NGA
ਚ	2586	ca	GURMUKHI LETTER CA
ਛ	2587	cha	GURMUKHI LETTER CHA
ਜ	2588	ja	GURMUKHI LETTER JA
ਝ	2589	jha	GURMUKHI LETTER JHA
ਞ	2590	nj'a	GURMUKHI LETTER NYA
ਟ	2591	Ta	GURMUKHI LETTER TTA
ਠ	2592	Tha	GURMUKHI LETTER TTHA
ਡ	2593	Da	GURMUKHI LETTER DDA
ਢ	2594	Dha	GURMUKHI LETTER DDHA
ਣ	2595	Na	GURMUKHI LETTER NNA
ਤ	2596	ta	GURMUKHI LETTER TA
ਥ	2597	tha	GURMUKHI LETTER THA
ਦ	2598	da	GURMUKHI LETTER DA
ਧ	2599	dha	GURMUKHI LETTER DHA
ਨ	2600	na	GURMUKHI LETTER NA
ਪ	2602	pa	GURMUKHI LETTER PA
ਫ	2603	pha	GURMUKHI LETTER PHA
ਬ	2604	ba	GURMUKHI LETTER BA
ਭ	2605	bha	GURMUKHI LETTER BHA
ਮ	2606	ma	GURMUKHI LETTER MA
ਯ	2607	ya	GURMUKHI LETTER YA
ਰ	2608	ra	GURMUKHI LETTER RA
ਲ	2610	la	GURMUKHI LETTER LA
ਲ਼	2611	La	GURMUKHI LETTER LLA
ਵ	2613	va	GURMUKHI LETTER VA
ਸ਼	2614	sha	GURMUKHI LETTER SHA
ਸ	2616	sa	GURMUKHI LETTER SA
ਹ	2617	ha	GURMUKHI LETTER HA
਼	2620	'	GURMUKHI SIGN NUKTA
ਾ	2622	A	GURMUKHI VOWEL SIGN AA
ਿ	2623	i	GURMUKHI VOWEL SIGN I
ੀ	2624	I	GURMUKHI VOWEL SIGN II
ੁ	2625	u	GURMUKHI VOWEL SIGN U
ੂ	2626	U	GURMUKHI VOWEL SIGN UU
ੇ	2631	E	GURMUKHI VOWEL SIGN EE
ੈ	2632	ai	GURMUKHI VOWEL SIGN AI
ੋ	2635	O	GURMUKHI VOWEL SIGN OO
ੌ	2636	au	GURMUKHI VOWEL SIGN AU
ਖ਼	2649	Kh'a	GURMUKHI LETTER KHHA
ਗ਼	2650	g'a	GURMUKHI LETTER GHHA
ਜ਼	2651	j'a	GURMUKHI LETTER ZA
ੜ	2652	Ra	GURMUKHI LETTER RRA
ਫ਼	2654	ph'a	GURMUKHI LETTER FA
0	2662	0	GURMUKHI DIGIT ZERO
1	2663	1	GURMUKHI DIGIT ONE
2	2664	2	GURMUKHI DIGIT TWO
3	2665	3	GURMUKHI DIGIT THREE
4	2666	4	GURMUKHI DIGIT FOUR
5	2667	5	GURMUKHI DIGIT FIVE
6	2668	6	GURMUKHI DIGIT SIX
7	2669	7	GURMUKHI DIGIT SEVEN
8	2670	8	GURMUKHI DIGIT EIGHT
9	2671	9	GURMUKHI DIGIT NINE
ੰ	2672	m'	GURMUKHI TIPPI
ಁ	3201	m'	KANNADA SIGN CANDRABINDU
ಂ	3202	M	KANNADA SIGN ANUSVARA
ಃ	3203	H	KANNADA SIGN VISARGA
ಅ	3205	a	KANNADA LETTER A
ಆ	3206	A	KANNADA LETTER AA
ಇ	3207	i	KANNADA LETTER I
ಈ	3208	I	KANNADA LETTER II
ಉ	3209	u	KANNADA LETTER U
ಊ	3210	U	KANNADA LETTER UU
ಋ	3211	x	KANNADA LETTER VOCALIC R
ಌ	3212	q	KANNADA LETTER VOCALIC L
ಎ	3214	e	KANNADA LETTER E
ಏ	3215	E	KANNADA LETTER EE
ಐ	3216	ai	KANNADA LETTER AI
ಒ	3218	o	KANNADA LETTER O
ಓ	3219	O	KANNADA LETTER OO
ಔ	3220	au	KANNADA LETTER AU
ಕ	3221	ka	KANNADA LETTER KA
ಖ	3222	kha	KANNADA LETTER KHA
ಗ	3223	ga	KANNADA LETTER GA
ಘ	3224	gha	KANNADA LETTER GHA
ಙ	3225	ng'a	KANNADA LETTER NGA
ಚ	3226	ca	KANNADA LETTER CA
ಛ	3227	cha	KANNADA LETTER CHA
ಜ	3228	ja	KANNADA LETTER JA
ಝ	3229	jha	KANNADA LETTER JHA
ಞ	3230	nj'a	KANNADA LETTER NYA
ಟ	3231	Ta	KANNADA LETTER TTA
ಠ	3232	Tha	KANNADA LETTER TTHA
ಡ	3233	Da	KANNADA LETTER DDA
ಢ	3234	Dha	KANNADA LETTER DDHA
ಣ	3235	Na	KANNADA LETTER NNA
ತ	3236	ta	KANNADA LETTER TA
ಥ	3237	tha	KANNADA LETTER THA
ದ	3238	da	KANNADA LETTER DA
ಧ	3239	dha	KANNADA LETTER DHA
ನ	3240	na	KANNADA LETTER NA
ಪ	3242	pa	KANNADA LETTER PA
ಫ	3243	pha	KANNADA LETTER PHA
ಬ	3244	ba	KANNADA LETTER BA
ಭ	3245	bha	KANNADA LETTER BHA
ಮ	3246	ma	KANNADA LETTER MA
ಯ	3247	ya	KANNADA LETTER YA
ರ	3248	ra	KANNADA LETTER RA
ಱ	3249	Ra	KANNADA LETTER RRA
ಲ	3250	la	KANNADA LETTER LA
ಳ	3251	La	KANNADA LETTER LLA
ವ	3253	va	KANNADA LETTER VA
ಶ	3254	sha	KANNADA LETTER SHA
ಷ	3255	Sa	KANNADA LETTER SSA
ಸ	3256	sa	KANNADA LETTER SA
ಹ	3257	ha	KANNADA LETTER HA
಼	3260	'	KANNADA SIGN NUKTA
ಾ	3262	A	KANNADA VOWEL SIGN AA
ಿ	3263	i	KANNADA VOWEL SIGN I
ೀ	3264	I	KANNADA VOWEL SIGN II
ು	3265	u	KANNADA VOWEL SIGN U
ೂ	3266	U	KANNADA VOWEL SIGN UU
ೃ	3267	x	KANNADA VOWEL SIGN VOCALIC R
ೄ	3268	X	KANNADA VOWEL SIGN VOCALIC RR
ೆ	3270	e	KANNADA VOWEL SIGN E
ೇ	3271	E	KANNADA VOWEL SIGN EE
ೈ	3272	ai	KANNADA VOWEL SIGN AI
ೊ	3274	o	KANNADA VOWEL SIGN O
ೋ	3275	O	KANNADA VOWEL SIGN OO
ೌ	3276	au	KANNADA VOWEL SIGN AU
ೞ	3294	Za	KANNADA LETTER FA
ೠ	3296	X	KANNADA LETTER VOCALIC RR
ೡ	3297	Q	KANNADA LETTER VOCALIC LL
ೢ	3298	q	KANNADA VOWEL SIGN VOCALIC L
ೣ	3299	Q	KANNADA VOWEL SIGN VOCALIC LL
0	3302	0	KANNADA DIGIT ZERO
1	3303	1	KANNADA DIGIT ONE
2	3304	2	KANNADA DIGIT TWO
3	3305	3	KANNADA DIGIT THREE
4	3306	4	KANNADA DIGIT FOUR
5	3307	5	KANNADA DIGIT FIVE
6	3308	6	KANNADA DIGIT SIX
7	3309	7	KANNADA DIGIT SEVEN
8	3310	8	KANNADA DIGIT EIGHT
9	3311	9	KANNADA DIGIT NINE
ം	3330	M	MALAYALAM SIGN ANUSVARA
ഃ	3331	H	MALAYALAM SIGN VISARGA
അ	3333	a	MALAYALAM LETTER A
ആ	3334	A	MALAYALAM LETTER AA
ഇ	3335	i	MALAYALAM LETTER I
ഈ	3336	I	MALAYALAM LETTER II
ഉ	3337	u	MALAYALAM LETTER U
ഊ	3338	U	MALAYALAM LETTER UU
ഋ	3339	x	MALAYALAM LETTER VOCALIC R
ഌ	3340	q	MALAYALAM LETTER VOCALIC L
എ	3342	e	MALAYALAM LETTER E
ഏ	3343	E	MALAYALAM LETTER EE
ഐ	3344	ai	MALAYALAM LETTER AI
ഒ	3346	o	MALAYALAM LETTER O
ഓ	3347	O	MALAYALAM LETTER OO
ഔ	3348	au	MALAYALAM LETTER AU
ക	3349	ka	MALAYALAM LETTER KA
ഖ	3350	kha	MALAYALAM LETTER KHA
ഗ	3351	ga	MALAYALAM LETTER GA
ഘ	3352	gha	MALAYALAM LETTER GHA
ങ	3353	ng'a	MALAYALAM LETTER NGA
ച	3354	ca	MALAYALAM LETTER CA
ഛ	3355	cha	MALAYALAM LETTER CHA
ജ	3356	ja	MALAYALAM LETTER JA
ഝ	3357	jha	MALAYALAM LETTER JHA
ഞ	3358	nj'a	MALAYALAM LETTER NYA
ട	3359	Ta	MALAYALAM LETTER TTA
ഠ	3360	Tha	MALAYALAM LETTER TTHA
ഡ	3361	Da	MALAYALAM LETTER DDA
ഢ	3362	Dha	MALAYALAM LETTER DDHA
ണ	3363	Na	MALAYALAM LETTER NNA
ത	3364	ta	MALAYALAM LETTER TA
ഥ	3365	tha	MALAYALAM LETTER THA
ദ	3366	da	MALAYALAM LETTER DA
ധ	3367	dha	MALAYALAM LETTER DHA
ന	3368	na	MALAYALAM LETTER NA
പ	3370	pa	MALAYALAM LETTER PA
ഫ	3371	pha	MALAYALAM LETTER PHA
ബ	3372	ba	MALAYALAM LETTER BA
ഭ	3373	bha	MALAYALAM LETTER BHA
മ	3374	ma	MALAYALAM LETTER MA
യ	3375	ya	MALAYALAM LETTER YA
ര	3376	ra	MALAYALAM LETTER RA
റ	3377	Ra	MALAYALAM LETTER RRA
ല	3378	la	MALAYALAM LETTER LA
ള	3379	La	MALAYALAM LETTER LLA
ഴ	3380	Za	MALAYALAM LETTER LLLA
വ	3381	va	MALAYALAM LETTER VA
ശ	3382	sha	MALAYALAM LETTER SHA
ഷ	3383	Sa	MALAYALAM LETTER SSA
സ	3384	sa	MALAYALAM LETTER SA
ഹ	3385	ha	MALAYALAM LETTER HA
ാ	3390	A	MALAYALAM VOWEL SIGN AA
ി	3391	i	MALAYALAM VOWEL SIGN I
ീ	3392	I	MALAYALAM VOWEL SIGN II
ു	3393	u	MALAYALAM VOWEL SIGN U
ൂ	3394	U	MALAYALAM VOWEL SIGN UU
ൃ	3395	x	MALAYALAM VOWEL SIGN VOCALIC R
ൄ	3396	X	MALAYALAM VOWEL SIGN VOCALIC RR
െ	3398	e	MALAYALAM VOWEL SIGN E
േ	3399	E	MALAYALAM VOWEL SIGN EE
ൈ	3400	ai	MALAYALAM VOWEL SIGN AI
ൊ	3402	o	MALAYALAM VOWEL SIGN O
ോ	3403	O	MALAYALAM VOWEL SIGN OO
ൗ	3415	au	MALAYALAM AU LENGTH MARK
ൠ	3424	X	MALAYALAM LETTER VOCALIC RR
ൡ	3425	Q	MALAYALAM LETTER VOCALIC LL
ൢ	3426	q	MALAYALAM VOWEL SIGN VOCALIC L
ൣ	3427	Q	MALAYALAM VOWEL SIGN VOCALIC LL
0	3430	0	MALAYALAM DIGIT ZERO
1	3431	1	MALAYALAM DIGIT ONE
2	3432	2	MALAYALAM DIGIT TWO
3	3433	3	MALAYALAM DIGIT THREE
4	3434	4	MALAYALAM DIGIT FOUR
5	3435	5	MALAYALAM DIGIT FIVE
6	3436	6	MALAYALAM DIGIT SIX
7	3437	7	MALAYALAM DIGIT SEVEN
8	3438	8	MALAYALAM DIGIT EIGHT
9	3439	9	MALAYALAM DIGIT NINE
ൺ	3450	N'	MALAYALAM LETTER CHILLU NN
ൻ	3451	n'	MALAYALAM LETTER CHILLU N
ർ	3452	Ra'	MALAYALAM LETTER CHILLU RR
ൽ	3453	la'	MALAYALAM LETTER CHILLU L
ൾ	3454	La'	MALAYALAM LETTER CHILLU LL
ൿ	3455	k'	MALAYALAM LETTER CHILLU K
ଁ	2817	m'	ODIA SIGN CANDRABINDU
ଂ	2818	M	ODIA SIGN ANUSVARA
ଃ	2819	H	ODIA SIGN VISARGA
ଅ	2821	a	ODIA LETTER A
ଆ	2822	A	ODIA LETTER AA
ଇ	2823	i	ODIA LETTER I
ଈ	2824	I	ODIA LETTER II
ଉ	2825	u	ODIA LETTER U
ଊ	2826	U	ODIA LETTER UU
ଋ	2827	x	ODIA LETTER VOCALIC R
ଌ	2828	q	ODIA LETTER VOCALIC L
ଏ	2831	E	ODIA LETTER E
ଐ	2832	ai	ODIA LETTER AI
ଓ	2835	O	ODIA LETTER O
ଔ	2836	au	ODIA LETTER AU
କ	2837	ka	ODIA LETTER KA
ଖ	2838	kha	ODIA LETTER KHA
ଗ	2839	ga	ODIA LETTER GA
ଘ	2840	gha	ODIA LETTER GHA
ଙ	2841	ng'a	ODIA LETTER NGA
ଚ	2842	ca	ODIA LETTER CA
ଛ	2843	cha	ODIA LETTER CHA
ଜ	2844	ja	ODIA LETTER JA
ଝ	2845	jha	ODIA LETTER JHA
ଞ	2846	nj'a	ODIA LETTER NYA
ଟ	2847	Ta	ODIA LETTER TTA
ଠ	2848	Tha	ODIA LETTER TTHA
ଡ	2849	Da	ODIA LETTER DDA
ଢ	2850	Dha	ODIA LETTER DDHA
ଣ	2851	Na	ODIA LETTER NNA
ତ	2852	ta	ODIA LETTER TA
ଥ	2853	tha	ODIA LETTER THA
ଦ	2854	da	ODIA LETTER DA
ଧ	2855	dha	ODIA LETTER DHA
ନ	2856	na	ODIA LETTER NA
ପ	2858	pa	ODIA LETTER PA
ଫ	2859	pha	ODIA LETTER PHA
ବ	2860	ba	ODIA LETTER BA
ଭ	2861	bha	ODIA LETTER BHA
ମ	2862	ma	ODIA LETTER MA
ଯ	2863	ya	ODIA LETTER YA
ର	2864	ra	ODIA LETTER RA
ଲ	2866	la	ODIA LETTER LA
ଳ	2867	La	ODIA LETTER LLA
ଵ	2869	ଵ	ODIA LETTER VA
ଶ	2870	sha	ODIA LETTER SHA
ଷ	2871	Sa	ODIA LETTER SSA
ସ	2872	sa	ODIA LETTER SA
ହ	2873	ha	ODIA LETTER HA
଼	2876	'	ODIA SIGN NUKTA
ା	2878	A	ODIA VOWEL SIGN AA
ି	2879	i	ODIA VOWEL SIGN I
ୀ	2880	I	ODIA VOWEL SIGN II
ୁ	2881	u	ODIA VOWEL SIGN U
ୂ	2882	U	ODIA VOWEL SIGN UU
ୃ	2883	x	ODIA VOWEL SIGN VOCALIC R
ୄ	2884	X	ODIA VOWEL SIGN VOCALIC RR
େ	2887	E	ODIA VOWEL SIGN E
ୈ	2888	ai	ODIA VOWEL SIGN AI
ୋ	2891	O	ODIA VOWEL SIGN O
ୌ	2892	au	ODIA VOWEL SIGN AU
ୟ	2911	Ya	ODIA LETTER YYA
ୠ	2912	X	ODIA LETTER VOCALIC RR
ୡ	2913	Q	ODIA LETTER VOCALIC LL
0	2918	0	ODIA DIGIT ZERO
1	2919	1	ODIA DIGIT ONE
2	2920	2	ODIA DIGIT TWO
3	2921	3	ODIA DIGIT THREE
4	2922	4	ODIA DIGIT FOUR
5	2923	5	ODIA DIGIT FIVE
6	2924	6	ODIA DIGIT SIX
7	2925	7	ODIA DIGIT SEVEN
8	2926	8	ODIA DIGIT EIGHT
9	2927	9	ODIA DIGIT NINE
ୱ	2929	wa	ODIA LETTER WA
ஂ	2946	M	TAMIL SIGN ANUSVARA
ஃ	2947	H	TAMIL SIGN VISARGA
அ	2949	a	TAMIL LETTER A
ஆ	2950	A	TAMIL LETTER AA
இ	2951	i	TAMIL LETTER I
ஈ	2952	I	TAMIL LETTER II
உ	2953	u	TAMIL LETTER U
ஊ	2954	U	TAMIL LETTER UU
எ	2958	e	TAMIL LETTER E
ஏ	2959	E	TAMIL LETTER EE
ஐ	2960	ai	TAMIL LETTER AI
ஒ	2962	o	TAMIL LETTER O
ஓ	2963	O	TAMIL LETTER OO
ஔ	2964	au	TAMIL LETTER AU
க	2965	ka	TAMIL LETTER KA
ங	2969	ng'a	TAMIL LETTER NGA
ச	2970	ca	TAMIL LETTER CA
ஜ	2972	ja	TAMIL LETTER JA
ஞ	2974	nj'a	TAMIL LETTER NYA
ட	2975	Ta	TAMIL LETTER TTA
ண	2979	Na	TAMIL LETTER NNA
த	2980	ta	TAMIL LETTER TA
ந	2984	na	TAMIL LETTER NA
ன	2985	n'a	TAMIL LETTER NNNA
ப	2986	pa	TAMIL LETTER PA
ம	2990	ma	TAMIL LETTER MA
ய	2991	ya	TAMIL LETTER YA
ர	2992	ra	TAMIL LETTER RA
ற	2993	Ra	TAMIL LETTER RRA
ல	2994	la	TAMIL LETTER LA
ள	2995	La	TAMIL LETTER LLA
ழ	2996	Za	TAMIL LETTER LLLA
வ	2997	va	TAMIL LETTER VA
ஶ	2998	ஶ	TAMIL LETTER SHA
ஷ	2999	sha	TAMIL LETTER SSA
ஸ	3000	sa	TAMIL LETTER SA
ஹ	3001	ha	TAMIL LETTER HA
ா	3006	A	TAMIL VOWEL SIGN AA
ி	3007	i	TAMIL VOWEL SIGN I
ீ	3008	I	TAMIL VOWEL SIGN II
ு	3009	u	TAMIL VOWEL SIGN U
ூ	3010	U	TAMIL VOWEL SIGN UU
ெ	3014	e	TAMIL VOWEL SIGN E
ே	3015	E	TAMIL VOWEL SIGN EE
ை	3016	ai	TAMIL VOWEL SIGN AI
ொ	3018	o	TAMIL VOWEL SIGN O
ோ	3019	O	TAMIL VOWEL SIGN OO
ௌ	3020	au	TAMIL VOWEL SIGN AU
ௗ	3031	La	TAMIL AU LENGTH MARK
0	3046	0	TAMIL DIGIT ZERO
1	3047	1	TAMIL DIGIT ONE
2	3048	2	TAMIL DIGIT TWO
3	3049	3	TAMIL DIGIT THREE
4	3050	4	TAMIL DIGIT FOUR
5	3051	5	TAMIL DIGIT FIVE
6	3052	6	TAMIL DIGIT SIX
7	3053	7	TAMIL DIGIT SEVEN
8	3054	8	TAMIL DIGIT EIGHT
9	3055	9	TAMIL DIGIT NINE
ఁ	3073	m'	TELUGU SIGN CANDRABINDU
ం	3074	M	TELUGU SIGN ANUSVARA
ః	3075	H	TELUGU SIGN VISARGA
అ	3077	a	TELUGU LETTER A
ఆ	3078	A	TELUGU LETTER AA
ఇ	3079	i	TELUGU LETTER I
ఈ	3080	I	TELUGU LETTER II
ఉ	3081	u	TELUGU LETTER U
ఊ	3082	U	TELUGU LETTER UU
ఋ	3083	x	TELUGU LETTER VOCALIC R
ఌ	3084	q	TELUGU LETTER VOCALIC L
ఎ	3086	e	TELUGU LETTER E
ఏ	3087	E	TELUGU LETTER EE
ఐ	3088	ai	TELUGU LETTER AI
ఒ	3090	o	TELUGU LETTER O
ఓ	3091	O	TELUGU LETTER OO
ఔ	3092	au	TELUGU LETTER AU
క	3093	ka	TELUGU LETTER KA
ఖ	3094	kha	TELUGU LETTER KHA
గ	3095	ga	TELUGU LETTER GA
ఘ	3096	gha	TELUGU LETTER GHA
ఙ	3097	ng'a	TELUGU LETTER NGA
చ	3098	ca	TELUGU LETTER CA
ఛ	3099	cha	TELUGU LETTER CHA
జ	3100	ja	TELUGU LETTER JA
ఝ	3101	jha	TELUGU LETTER JHA
ఞ	3102	nj'a	TELUGU LETTER NYA
ట	3103	Ta	TELUGU LETTER TTA
ఠ	3104	Tha	TELUGU LETTER TTHA
డ	3105	Da	TELUGU LETTER DDA
ఢ	3106	Dha	TELUGU LETTER DDHA
ణ	3107	Na	TELUGU LETTER NNA
త	3108	ta	TELUGU LETTER TA
థ	3109	tha	TELUGU LETTER THA
ద	3110	da	TELUGU LETTER DA
ధ	3111	dha	TELUGU LETTER DHA
న	3112	na	TELUGU LETTER NA
ప	3114	pa	TELUGU LETTER PA
ఫ	3115	pha	TELUGU LETTER PHA
బ	3116	ba	TELUGU LETTER BA
భ	3117	bha	TELUGU LETTER BHA
మ	3118	ma	TELUGU LETTER MA
య	3119	ya	TELUGU LETTER YA
ర	3120	ra	TELUGU LETTER RA
ఱ	3121	Ra	TELUGU LETTER RRA
ల	3122	la	TELUGU LETTER LA
ళ	3123	La	TELUGU LETTER LLA
ఴ	3124	Za	TELUGU LETTER LLLA
వ	3125	va	TELUGU LETTER VA
శ	3126	sha	TELUGU LETTER SHA
ష	3127	Sa	TELUGU LETTER SSA
స	3128	sa	TELUGU LETTER SA
హ	3129	ha	TELUGU LETTER HA
ా	3134	A	TELUGU VOWEL SIGN AA
ి	3135	i	TELUGU VOWEL SIGN I
ీ	3136	I	TELUGU VOWEL SIGN II
ు	3137	u	TELUGU VOWEL SIGN U
ూ	3138	U	TELUGU VOWEL SIGN UU
ృ	3139	x	TELUGU VOWEL SIGN VOCALIC R
ౄ	3140	X	TELUGU VOWEL SIGN VOCALIC RR
ె	3142	e	TELUGU VOWEL SIGN E
ే	3143	E	TELUGU VOWEL SIGN EE
ై	3144	ai	TELUGU VOWEL SIGN AI
ొ	3146	o	TELUGU VOWEL SIGN O
ో	3147	O	TELUGU VOWEL SIGN OO
ౌ	3148	u	TELUGU VOWEL SIGN AU
ౠ	3168	X	TELUGU LETTER VOCALIC RR
ౡ	3169	Q	TELUGU LETTER VOCALIC LL
ౢ	3170	q	TELUGU VOWEL SIGN VOCALIC L
ౣ	3171	Q	TELUGU VOWEL SIGN VOCALIC LL
0	3174	0	TELUGU DIGIT ZERO
1	3175	1	TELUGU DIGIT ONE
2	3176	2	TELUGU DIGIT TWO
3	3177	3	TELUGU DIGIT THREE
4	3178	4	TELUGU DIGIT FOUR
5	3179	5	TELUGU DIGIT FIVE
6	3180	6	TELUGU DIGIT SIX
7	3181	7	TELUGU DIGIT SEVEN
8	3182	8	TELUGU DIGIT EIGHT
9	3183	9	TELUGU DIGIT NINE