The Source Layer setting determines which layer includes labels whose encoding will be converted.
This setting determines what encoding the source and destionation layers are assumed to have, for transcoding purposes.
For example, if the Source Layer contains pronunciations encoded with
CELEX DISC labels, and you want the Destination Layer to contain those
pronunciations encoded using ARPAbet labels, select the DISC → ARPAbet
as the Translation.
The following table presents some common encodings and equivalences or near-equivalences between phonemes. 1
| Example | IPA | SAM-PA | DISC2 | CPA3 | Kirshenbaum4 | ARPAbet | CMU Dict |
|---|---|---|---|---|---|---|---|
| Vowels | |||||||
| kit | ɪ | I | I | I | I | IH | IH |
| dress | ɛ | E | E | E | E | EH | EH |
| trap | æ | { | { | ^/ | & | AE | AE |
| strut | ʌ | V | V | ^ | V | AH | AH |
| foot | ʊ | U | U | U | U | UH | UH |
| another | ǝ | @ | @ | @ | @ | AX | |
| fleece | iː | i: | i | i: | i: | IY | IY |
| bath | ɑː | A: | # | A: | A: | AA | AA |
| lot | ɒ | Q | Q | Q | A. | AO | AO |
| thought | ɔː | O: | $ | O: | O: | ||
| goose | uː | u: | u | u: | u: | UW | UW |
| nurse | ɜː | 3ː | 3 | @: | V” | ER | ER |
| face | eɪ | eI | 1 | e/ | eI | EY | EY |
| price | aɪ | aI | 2 | a/ | aI | AY | AY |
| choice | ɔɪ | OI | 4 | o/ | OI | OY | OY |
| goat | ǝʊ | @U | 5 | O/ | @U | OW | OW |
| mouth | aʊ | aU | 6 | A/ | aU | AW | AW |
| near | ɪǝ | I@ | 7 | I/ | I@ | IY R | IY R |
| square | ɛǝ | E@ | 8 | E/ | E@ | EH R | EH R |
| cure | ʊǝ | U@ | 9 | U/ | U@ | UH R | UH R |
| timbre | æ | {~ | c | ^/~ | &~ | ||
| détente | ɑ̃ː | A~: | q | A~: | A~: | ||
| lingerie | æ̃ː | {~: | 0 | ^/~: | &~: | ||
| bouillon | ɒ̃ː | O~: | ~ | O~: | A.~: | ||
| Consonants | |||||||
| pat | p | p | p | p | p | P | P |
| bad | b | b | b | b | b | B | B |
| tack | t | t | t | t | t | T | T |
| dad | d | d | d | d | d | D | D |
| cad | k | k | k | k | k | K | K |
| game | g | g | g | g | g | G | G |
| bang | ŋ | N | N | N | N | NG | NG |
| mad | m | m | m | m | m | M | M |
| nat | n | n | n | n | n | N | N |
| lad | l | l | l | l | l | L | L |
| rat | r | r | r | r | r | R | R |
| fat | f | f | f | f | f | F | F |
| vat | v | v | v | v | v | V | V |
| thin | Ɵ | T | T | T | T | TH | TH |
| then | ð | D | D | D | D | DH | DH |
| sap | s | s | s | s | s | S | S |
| zap | z | z | z | z | z | Z | Z |
| sheep | ʃ | S | S | S | S | SH | SH |
| measure | Ʒ | Z | Z | Z | Z | ZH | ZH |
| yank | j | j | j | j | j | Y | Y |
| had | h | h | h | h | h | HH | HH |
| wet | w | w | w | w | w | W | W |
| cheap | ʧ | tS | J | T/ | tS | CH | CH |
| jeep | ʤ | dZ | _ | J/ | dZ | JH | JH |
| loch | x | x | x | x | x | ||
| bacon | ŋ̩ | N, | C | N, | N- | ||
| idealism | m̩ | m, | F | m, | m- | ||
| burden | n̩ | n, | H | n, | n- | ||
| dangle | l̩ | l, | P | l, | l- | ||
| car alarm | * | r* | R | r* | |||
| uh-oh | ʔ | ? | ? | Q | |||
| father | ɚ | AXR | |||||
| wetter | ɾ | DX | |||||
1 In the table, some phoneme representations are highlighted with a bold typeface; this highlighting is intended to indicate representations that are unpredictable in some way, either because they're substantially different from IPA or from English orthographical convention, or they're different from the corresponding representation in an otherwise-similar set of representations. Others are highlighted with an italic typeface; these are examples of representations that actually use a combination of two phonemes, where in other sets only one phoneme is used.
2 SAM-PA and DISC phonemes taken from CELEX English Guide (1995) § 2.4.1 pp. 31-32, Tables 3 & 4.
3 The Computer Phonetic Alphabet (CPA) was developed for seven European languages, based on the IPA - Kugler-Kruse (1987)
If you select the option Custom Translation
, then you can specify an arbitrary
list of Mappings from the Source Layer label characters to
the Destination Layer characters.
If you select the Custom Translation
option, you can specify a list
of Mappings - a list of characters to match against, which starts off
empty.
To add a mapping, click the + button on the right hand side.
To remove a mapping, select it in the list and click the - button on the right hand side.
The Source Characters column contains the characters that will be applied to the source layer annotation labels. If one of the rows matches a character or sequence of characters on the source layer, then the corresponding value from the Destination Characters column is copied to the resulting label on the destination layer.
For example, if you add a mapping ng → ŋ
, then all instances
of ng
on the source layer will be translated as ŋ
on the
destination layer.
The characters are checked in the order you specify. You can move a pattern up or down in the order by selecting it and using the ↑ and ↓ buttons on the right.
For characters in the source layer that don't match any value in the Source Characters column, the Characters with no mapping setting determines whether they are copied as-is, or ignored.
This is the layer that new annotations will be added to. You can pick an existing layer, or add a new layer.
The Language setting determines which language to target for annotation. Leave this blank to ignore language, or enter an ISO language/locale code, e.g. en-NZ , to target varieties of a particular language. This setting is treated as a regular expression, so es.* will target all varietys of Spanish, etc.
If you specify a language, then only utterances in the language specified will be annotated. There are three factors that determine the language of an utterance: