From dc67d294c2e61eed6870ca022a431d097dab32a0 Mon Sep 17 00:00:00 2001 From: "Reece H. Dunn" Date: Sat, 17 Mar 2018 17:27:44 +0000 Subject: [PATCH] docs: describe the Kirshenbaum (ASCII-IPA) transcription scheme. --- docs/phonemes.md | 24 ++++- docs/phonemes/kirshenbaum.md | 196 +++++++++++++++++++++++++++++++++++ 2 files changed, 216 insertions(+), 4 deletions(-) create mode 100644 docs/phonemes/kirshenbaum.md diff --git a/docs/phonemes.md b/docs/phonemes.md index ed67daa0..728a3132 100644 --- a/docs/phonemes.md +++ b/docs/phonemes.md @@ -1,5 +1,6 @@ # Phonemes +- [Phoneme Transcription Schemes](#phoneme-transcription-schemes) - [Consonants](#consonants) - [Other Symbols](#other-symbols) - [Gemination](#gemination) @@ -54,6 +55,20 @@ Instead, it is designed to be a transcription guide for authors of espeak-ng languages and voices on how to specify phonemes so that the IPA and feature transcriptions are consistent. +## Phoneme Transcription Schemes + +| BCP47 Variant | Abbreviation | Transcription Scheme | Encoding | +|---------------------------------------|--------------|------------------------------------------------------|----------| +| `foncxs` | CXS | Conlang X-SAMPA | ASCII | +| `fonipa` | IPA | International Phonetic Alphabet | Unicode | +| [`fonkirsh`](phonemes/kirshenbaum.md) | | Kirshenbaum (ASCII-IPA) | ASCII | +| `fonxsampa` | X-SAMPA | Extended Speech Assessment Methods Phonetic Alphabet | ASCII | +| `fonzsampa` | Z-SAMPA | Zompist Bulletin Board (ZBB) SAMPA | ASCII | + +1. `foncxs`, `fonkirsh`, and `fonzsampa` are private use extensions defined in + the [bcp47-extensions](https://raw.githubusercontent.com/espeak-ng/bcp47-data/master/bcp47-extensions) + file. + ## Consonants | | `blb` | | `lbd` | | `dnt` | | `alv` | | `pla` | | `rfx` | | `alp` | | `pal` | | `vel` | | `uvl` | | `phr` | | `glt` | | @@ -189,10 +204,6 @@ consonants, but Wikipedia does not. This model uses the Wikipedia descriptions. | `sml` | æ | | ɐ | | | | | `low` | a | ɶ | | | ɑ | ɒ | -__NOTE:__ The `smh` vowels are more `cnt` than the other vowels. However, this -distinction is not needed to classify these vowels, so is not included in the -above table. - ### Other Symbols | Symbol | Features | @@ -220,6 +231,11 @@ above table. | `cnt` | center | `kirshenbaum` | | `bck` | back | `kirshenbaum` | +__NOTE:__ The `smh` vowels are more `cnt` than the other vowels. However, this +distinction is not needed to classify these vowels so there are no features +for front-central and back-central. The `fnt` and `bck` features are used +instead. + ### Rounding | Feature | Name | Origin | diff --git a/docs/phonemes/kirshenbaum.md b/docs/phonemes/kirshenbaum.md new file mode 100644 index 00000000..5a66e7f6 --- /dev/null +++ b/docs/phonemes/kirshenbaum.md @@ -0,0 +1,196 @@ +# Kirshenbaum (ASCII-IPA) Transcription Scheme + +- [Phoneme Transcription Schemes](#phoneme-transcription-schemes) +- [Consonants](#consonants) + - [Other Symbols](#other-symbols) +- [Vowels](#vowels) + - [Other Symbols](#other-symbols-1) +- [Diacritics](#diacritics) +- [References](#references) + +---------- + +The Kirshenbaum (ASCII-IPA) transcription scheme was created by Evan +Kirshenbaum and others between 1991 and 1993 for use in Usenet and e-mail. +Since then, the support for Unicode has greatly improved. An ASCII encoding +of IPA is still useful as it is often easier to type. + +The Kirshenbaum transcription scheme defines a set of phoneme features that are +described in eSpeak NG in the [phonemes](phonemes.md) document. The `cainteoir` +and `espeak-ng` phoneme features defined in that document are not useable in +transcriptions using the Kirshenbaum transcription scheme. They are used here +where relevant to define the Kirshenbaum phonemes in a consistent way across +the different transcription schemes. + +Broad (phonemic) transcriptions are enclosed in `/` characters (`/p/`), while +narrow (phonetic) transcriptions are wrapped in `[` and `]` characters (`[p]`). + +A phoneme can be transcribed as a named phoneme (`p`), or as a set of features +wrapped in `{` and `}` characters (`{vls,blb,stp}`). A feature (typically a +diacritic) of a phoneme can be added or modified by wrapping the feature in `<` +and `>` characters after the base phoneme (`p`). + +## Phoneme Transcription Schemes + +| BCP47 Variant | Abbreviation | Transcription Scheme | Encoding | +|----------------------------|--------------|------------------------------------------------------|----------| +| `foncxs` | CXS | Conlang X-SAMPA | ASCII | +| [`fonipa`](../phonemes.md) | IPA | International Phonetic Alphabet | Unicode | +| `fonkirsh` | | Kirshenbaum (ASCII-IPA) | ASCII | +| `fonxsampa` | X-SAMPA | Extended Speech Assessment Methods Phonetic Alphabet | ASCII | +| `fonzsampa` | Z-SAMPA | Zompist Bulletin Board (ZBB) SAMPA | ASCII | + +1. `foncxs`, `fonkirsh`, and `fonzsampa` are private use extensions defined in + the [bcp47-extensions](https://raw.githubusercontent.com/espeak-ng/bcp47-data/master/bcp47-extensions) + file. + +## Consonants + +| | `blb` | | `lbd` | | `dnt` | | `alv` | | `pla` | | `rfx` | | `alp` | | `pal` | | `vel` | | `uvl` | | `phr` | | `glt` | | +|-----------------------------|----------|----------|-------|----------|-----------|-----------|----------|----------|-------|-------|-------|-------|-------|-------|----------|----------|----------|----------|----------|----------|-------|----------|-------|--------| +| | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | +| `nas` | | `m` | | `M` | | `n[` | | `n` | | | | `n.` | | | | `n^` | | `N` | | `n"` | | | | | +| `stp` | `p` | `b` | | | `t[` | `d[` | `t` | `d` | | | `t.` | `d.` | | | `c` | `J` | `k` | `g` | `q` | `G` | | | `?` | | +| `sib` `afr` | | | | | | | | | | | | | | | | | | | | | | | | | +| `afr` | | | | | | | | | | | | | | | | | | | | | | | | | +| `lat` `afr` | | | | | | | | | | | | | | | | | | | | | | | | | +| `sib` `frc` | | | | | | | `s` | `z` | `S` | `Z` | `s.` | `z.` | | | | | | | | | | | | | +| `frc` | `P` | `B` | `f` | `v` | `T` | `D` | | | | | | | | | `C` | `C` | `x` | `Q` | `X` | `g"` | `H` | `H` | `h` | `h` | +| `lat` `frc` | | | | | | | `s` | `z` | | | | | | | | | | | | | | | | | +| `apr` | | | | `r` | | `r[` | | `r` | | | | `r.` | | | | `j` | | `j` | | `g"` | | | | | +| `lat` `apr` | | | | | | `l[` | | `l` | | | | `l.` | | | | `l^` | | `L` | | | | | | | +| `flp` | | | | | | | | `*` | | | | `*.` | | | | | | | | | | | | | +| `lat` `flp` | | | | | | | | `*` | | | | | | | | | | | | | | | | | +| `trl` | | `b` | | | | | | `r` | | | | | | | | | | | | `r"` | | | | | +| `clk` | `p!` | | | | `t!` | | `c!` | | | | | | | | `c!` | | `k!` | | | | | | | | +| `lat` `clk` | | | | | | | `l!` | | | | | | | | | | | | | | | | | | +| `imp` | | `` b` `` | | | | `` d` `` | | `` d` `` | | | | | | | | `` J` `` | | `` g` `` | | `` G` `` | | | | | +| `ejc` | `` p` `` | | | | `` t[` `` | | `` t` `` | | | | | | | | `` c` `` | | `` k` `` | | `` q` `` | | | | | | +| `ejc` `frc` | | | | | | | | | | | | | | | | | | | | | | | | | +| `lat` `ejc` `frc` | | | | | | | | | | | | | | | | | | | | | | | | | + +1. `h` is transcribed as `{glt,apr}` in ASCII-IPA. It is transcribed as + `{glt,frc}` in this document to match the assignment used in the + Wikipedia IPA table, and the placement in other documents in eSpeak NG. + The IPA chart places them as being `{frc}` or `{apr}`, as does the main + Wikipedia page for the `[h]` phoneme. + +1. The `` q` `` phoneme is incorrectly placed in the segment table (Appendix D) + of the ASCII-IPA document. According to the diacritics list (Appendix C), + the `` ` `` character denotes an `{ejc}` phoneme for a `{vls}` base + phoneme, not `{imp}` which is used for `{vcd}` base phonemes. The table + above uses the correct assignment. + +1. The `L` phoneme has an alternate meaning in ASCII-IPA. This document only + describes the variant preferred by Evan Kirshenbaum. + +### Other Symbols + +| | `bld` | | `alv` | | `pla` | | `pal` | | `lbv` | | `vel` | | +|-----------------------------|-------|-------|-------|-------|-------|-------|-------|-------|----------|----------|----------|-------| +| | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | `vls` | `vcd` | +| `nas` | | | | | | | | | | `n` | | | +| `stp` | | | | | | | | | `t` | `d` | | | +| `afr` | | | | | | | | | | | | | +| `vzd` `frc` | | | | | | | | | | | | | +| `ptr` `apr` | | | | | | | | | | | `w` | `w` | +| `fzd` `lat` `apr` | | | | | | | | | | | | | + +1. The `w` phoneme is transcribed as being `{vcd,lbv,frc}` or `{lbv,apr}`, and + `w` as `{vls,lbv,frc}` in ASCII-IPA. They are transcribed as + `{ptr,vel,apr}` in this document to match their definition as labialized + velar approximants in the IPA chart. + +## Vowels + +| | `fnt` | | `cnt` | | `bck` | | +|--------|-------|-------|----------|-------|-------|-------| +| | `unr` | `rnd` | `unr` | `rnd` | `unr` | `rnd` | +| `hgh` | `i` | `y` | `i"` | `u"` | `u-` | `u` | +| `smh` | `I` | `I.` | | | | `U` | +| `umd` | `e` | `Y` | `@` | | `o-` | `o` | +| `mid` | | | `@` | `@.` | | | +| `lmd` | `E` | `W` | `V"` | `O"` | `V` | `O` | +| `sml` | `&` | | | | | | +| `low` | `a` | `a.` | | | `A` | `A.` | + +1. The `&` phoneme is transcribed as being `{low,fnt,unr,vwl}`, and `a` as + `{low,cnt,unr,vwl}` in ASCII-IPA. This does not match their placement in + the IPA chart according to the Unicode characters those ASCII-IPA + phonemes map to. As such, the table above reflects their IPA chart + positions. + +1. The `&.` phoneme is defined in the segment table (Appendix D) of the + ASCII-IPA definition. According to the IPA charts, this is not a valid + phoneme. As such, it is not listed in the table above. + +1. The `W` and `Y` phonemes have an alternate meaning in ASCII-IPA. This + document only describes the variant preferred by Evan Kirshenbaum. + +### Other Symbols + +| Symbol | Features | +|----------|-------------------------------| +| `R` | `unr` `mid` `cnt` `rzd` `vwl` | +| `R` | `unr` `lmd` `cnt` `rzd` `vwl` | + +1. The `R` phoneme is transcribed as being `{umd,cnt,unr,rzd,vwl}` in + ASCII-IPA. However, the IPA phoneme it is representing is based on the + `{lmd}` vowel. The table above keeps the `` modifying feature to be + compatible with ASCII-IPA, but assigns it to `{lmd}` in this case to be + compatible with the IPA chart. + +1. The `R` phoneme has an alternate meaning in ASCII-IPA. This document only + describes the variant preferred by Evan Kirshenbaum. + +## Diacritics + +| Symbol | Feature | +|--------|---------| +| `:` | `{lng}` | +| `!` | `{clk}` | +| `[` | `{dnt}` | +| `;` | `{pzd}` | +| `^` | `{pal}` | +| `` | `{fzd}` | +| `` | `{asp}` | +| `` | `{unx}` | +| `` | `{rzd}` | +| `` | `{ptr}` | +| `` | `{brv}` | + +1. The `` diacritic has an alternate meaning in ASCII-IPA. This document + only describes the variant preferred by Evan Kirshenbaum. + +1. The `{ptr}` (protruded) feature is specified as `{lzd}` (labialized) in + ASCII-IPA. + +1. The `{brv}` (breathy voice) feature is specified as `{mrm}` (murmured) + in ASCII-IPA. + +The ASCII-IPA transcription scheme defines several diacritics that vary in +meaning depending on whether the base phoneme is a vowel or consonant. + +| Symbol | Vowel | Consonant | +|--------|---------|-----------| +| `~` | `{nzd}` | `{vzd}` | +| `-` | `{unr}` | `{syl}` | +| `.` | `{rnd}` | `{rfx}` | +| `"` | `{cnt}` | `{uvl}` | + +The ASCII-IPA transcription scheme defines several diacritics that vary in +meaning depending on the voicing of the base phoneme. + +| Symbol | Voiceless | Voiced | +|---------|-----------|---------| +| `` ` `` | `{ejc}` | `{imp}` | + +## References + +1. Kirshenbaum, Evan, + [Representing IPA phonetics in ASCII](https://web.archive.org/web/20160419125856/http://www.kirshenbaum.net/IPA/faq.html) (HTML). 04-01-1993. + Updated 30-04-2003. + +2. Kirshenbaum, Evan, + [Representing IPA phonetics in ASCII](https://web.archive.org/web/20160419125856/http://www.kirshenbaum.net/IPA/ascii-ipa.pdf) (PDF). 22-02-2001. + Updated 06-09-2011.