Romanization Challenge Thread v2.0

[Darren]

/p (pʲ) t ts c k (kʷ)/ <p pi t č ć k ku>
/s ʂ ɕ/ <s š ś>
/m (mʲ) n ȵ (ŋ)/ <m mi n ń ng>
/β~w l r ʎ j/ <u l r ł i>

With the alt analysis because yes
/a ə/ <a o>

/ma tprəjn sən pəwɲcn sənsarjs karjpəc sasprtww alək wrəm nəʂ kəws nəj alaswnəj || kjəltəj nəjʂ wrəsəɕɕ snəj wawləjʂw snəj psl klw ɕwatsj ʂəjʂ kəlpə-wrkʂntwən arwər pəpʲtkʷnt wsalw jəjtwəjntw βəʂlən ʂəjɲcm alasap klw krəwplwnəja kəlpə-wrkʂntw nəkntm kəppaɲ pakr takərm snəj wawləjʂw snəj psl klw nəktm ɕawən wləjsəntjəw psləɕɕl ʂjək pakr takm cəmʲ alaswnəjjs nw tsrəʂʂwnəj prətjpəkʂ nantsw tmʲəw tsrəʂʂwnəj ɲ arkjʂəwʂjən pkʷən prwccəməw plskən/

Ma tproin son pouńćn sonsaris karipoć sasprtuu alok urom noš kous noi alasunoi. Kioltoi noiš urosośś snoi uauloišu snoi psl klu śwači šoiš kolpo-urkšntuon aruor popitkunt usalu ioituointu uošlon šoińćm alasap klu krouplunoia kolpo-urkšntu nokntm koppań pakr takorm snoi uauloišu snoi psl klu noktm śauon uloisontiou pslośśl šiok pakr takm ćomi alasunoiis nu čroššunoi protipokš nanču tmiou čroššunoi ń arkišoušion pkuon pruććomou plskon

[bradrn]

Tocharian A

/p (pʲ) t ts c k (kʷ)/ ⟨p py t c ć k kw⟩
/s ʂ ɕ/ ⟨s š ś⟩
/m (mʲ) n ȵ (ŋ)/ ⟨m my n ń ng⟩
/β~w l r ʎ j/ ⟨w l r ł y⟩

/a e ə o i ɨ u/ ⟨a e ə o i ɨ u⟩

/ma tɨpren sən poɲcɨn sənsaris karipəc saspɨrtβu alək βrəm nəʂ kos ne alasune || kjəlte neʂ βrəsəɕɕi sne βaβleʂu sne psɨl klu ɕβatsi ʂeʂ kəlpəβrɨkʂɨntβən arβər pəpʲɨtkunt βsalu jetβejntu βəʂlən ʂeɲcɨm alasap klu kropluneja kəlpəβrɨkʂɨntu nəkɨntɨm kəppaɲ pakɨr takərɨm sne βaβleʂu sne psɨl klu nəktɨm ɕaβən βlesəntjo psɨləɕɕɨl ʂjək pakɨr takɨm cəmi alasunejis nu tsrəʂʂune prətipəkʂ nantsu tɨmʲo tsrəʂʂune ɲi arkiʂoʂjən pkʷən pruccəmo pɨlskən/

Ma tɨpren sən pońćin sənsaris karpiəć saspɨrtwu alək wrəm nəš kos ne alasune. Kyəlte neš wrəsəśśi sne wawlešu sne psɨl klu śwaci šeš kəlpəwrɨkšɨntwən arwər pəpyɨtkunt wsalu yetweyntu wəšlən šeńćɨm alasap klu kropluneya kəlpəwrɨkšɨntu nəkɨntɨm kəppań pakɨr takərɨm sne wawlešu sne psɨl klu nəktɨm śawən wlesəntyo psɨləśśɨl šyək pakɨr takɨm ćəmi alasuneyis nu crəššune prətipəkš nancu tɨmyo crəššune ńi arkišošyən pkwən pruććəmo pɨlskən.

Or, alternately:

/p (pʲ) t ts c k (kʷ)/ ⟨p py t c ć k kw⟩
/s ʂ ɕ/ ⟨s š ś⟩
/m (mʲ) n ȵ (ŋ)/ ⟨m my n ń ng⟩
/β~w l r ʎ j/ ⟨v l r ł j⟩

/a e ə o i ɨ u/ ⟨a e y o i w u⟩

Ma twpren syn pońćin synsaris karpiyć saspwrtvu alyk vrym nyš kos ne alasune. Kjylte neš vrysyśśi sne vavlešu sne pswl klu śvaci šeš kylpyvrwkšwntvyn arvyr prpjwtkunt vsalu jetvejntu vyšlyn šeńćwm alasap klu kropluneja kylpyvrwkšwntu nykwntwm kyppań pakwr takyrwm sne vavlešu sne pswl klu nyktwm śavyn vlesyntjo pswlyśśwl šjyk pakwr takwm ćymi alasunejis nu cryššune prytipykš nancu twmjo cryššune ńi arkišošjyn pkwyn pruććymo pwlskyn.

[Nortaneous]

/p (pʲ) t ts c k (kʷ)/ <p p t c č k kv>
/s ʂ ɕ/ <s š sʼ>
/m (mʲ) n ȵ (ŋ)/ <m mʼ n nʼ ŋ>
/β~w l r ʎ j/ <v l r lʼ j>

/a e ə o i ɨ u/ <a e y o i 0 u>

/ma tɨpren sən poɲcɨn sənsaris karipəc saspɨrtβu alək βrəm nəʂ kos ne alasune || kjəlte neʂ βrəsəɕɕi sne βaβleʂu sne psɨl klu ɕβatsi ʂeʂ kəlpəβrɨkʂɨntβən arβər pəpʲɨtkunt βsalu jetβejntu βəʂlən ʂeɲcɨm alasap klu kropluneja kəlpəβrɨkʂɨntu nəkɨntɨm kəppaɲ pakɨr takərɨm sne βaβleʂu sne psɨl klu nəktɨm ɕaβən βlesəntjo psɨləɕɕɨl ʂjək pakɨr takɨm cəmi alasunejis nu tsrəʂʂune prətipəkʂ nantsu tɨmʲo tsrəʂʂune ɲi arkiʂoʂjən pkʷən pruccəmo pɨlskən/
Ma tpren syn ponʼčn synsaris karipyc sasprtvu alyk vrym nyš kos ne alasune. Kjylte neš vrysyssʼi sne vavlešu sne psl klu śvaci šeš kylpyvrkšntvyn arvyr pypʼtkunt vsalu jetvejntu vyšlyn šenʼčm alasap klu kropluneja kylpyvrkšntu nykntm kyppanʼ pakr takyrm sne vavlešu sne psl klu nyktm śavn vlesyntjo pslyssʼl šjyk pakr takm čymi alasunejis nu cryššune prytipykš nancu tmʼo cryššune nʼi arkišošjyn pkvyn pruččymo plskyn.

[Rounin Ryuuji]

Tocharian A

/p (pʲ) t ts c k (kʷ)/
/s ʂ ɕ/ s sz sh
/m (mʲ) n ȵ (ŋ)/
/β~w l r ʎ j/

/a e ə o i ɨ u/

/ma tɨpren sən poɲcɨn sənsaris karipəc saspɨrtβu alək βrəm nəʂ kos ne alasune || kjəlte neʂ βrəsəɕɕi sne βaβleʂu sne psɨl klu ɕβatsi ʂeʂ kəlpəβrɨkʂɨntβən arβər pəpʲɨtkunt βsalu jetβejntu βəʂlən ʂeɲcɨm alasap klu kropluneja kəlpəβrɨkʂɨntu nəkɨntɨm kəppaɲ pakɨr takərɨm sne βaβleʂu sne psɨl klu nəktɨm ɕaβən βlesəntjo psɨləɕɕɨl ʂjək pakɨr takɨm cəmi alasunejis nu tsrəʂʂune prətipəkʂ nantsu tɨmʲo tsrəʂʂune ɲi arkiʂoʂjən pkʷən pruccəmo pɨlskən/

/p (pʲ) t ts c k (kʷ)/ p (py) t ts ch k kw
/s ʂ ɕ/ s sz sh
/m (mʲ) n ȵ (ŋ)/ m (mh) n nh (ng)
/β~w l r ʎ j/ v l r lh y
/a e ə o i ɨ u/ a i eo o i w u

Ma twpren seon ponhchin seonsaris karipeoch saspwrtvu aleok vreom kos ne alasune. Kyeolte nes vreoseosshi sne vavleszu sne pswl klu shvasti szesz keolpeovrwkszwntveon arveor peopywtkunt vsalu yetveyntu veoszleon szenhchwm alasap klu kropluneya keolpeovrwkszwntu neokwntwm keoppanh pakwr takeorwm sne vavleszu sne pswl klu neoktwm shaveon vleseontyo pswleosshwl szyeok pakwr takwm cheomi alasuneyis nu tsreosszune preotipeoksz nantsu timho tsreosszune nhi arkiszoszyeon pkweon pruccheomo.

[axolotl]

So this isn't exactly a romanization challenge, it's more of a reconstruction and romanization challenge, but I thought here would be a good place to post it, just as a bit of an experiment to see if people can figure out what has happened here. If you get the reconstruction right (it's not a lot of reconstruction, just a few sound changes), you should be able to use historical spelling to write this language using only the 26-letter Latin alphabet, with no phonetic ambiguity.

Consonants:

p t k
b d g
s z
m n l r
mˀ nˀ lˀ rˀ
j w

Notes:
Current syllable structure is (C)V(C). This was not always the case - previously, it was (C)V(C)(C).
b d g z only occur after m n l r, but they are fully phonemic - p t k s also occur after m n l r, though the voiced consonants occur more frequently than the voiceless consonants in this post-liquid position.
mˀ nˀ lˀ rˀ only occur wordfinally. These sounds come from more than one source.
Word-final p t k are often unreleased or have some level of glottalization.
j w may only occur word-initially or between vowels.

Vowels:
ɑ æ e ø o i y u

The vowels æ ø y arose by i-umlaut and u-umlaut. The umlaut was blocked by certain consonantal elements between the two vowels. It is possible to write this vowel inventory without ambiguity using only the letters a e i o u, if the historical characteristics of the consonants are depicted correctly.

Here's something that might help with that.
/yndi undi unti/ and similar sequences -> all can occur
/ynti/ and similar sequences -> does not occur in native words
/ɑni/ /æni/ /ɑwi/ /æwi/ and similar sequences -> all can occur

[bradrn]

I haven’t quite figured this one out yet, but in the meantime here’s a phonetic romanization:

/p t k/ ⟨p t k⟩
/b d g/ ⟨b d g⟩
/s z/ ⟨s z⟩
/m n l r/ ⟨m n l r⟩
/mˀ nˀ lˀ rˀ/ ⟨mh nh lh rh⟩
/j w/ ⟨y w⟩

/ɑ æ e ø o i y u/ ⟨a ä e ö o i ü u⟩

[Killerwhale]

Modern Suraic

m n n̪ ɲ ŋ
p pʰ b ɓ t̪ t̪ʰ d̪ t tʰ tʼ d ɗ k kʰ kʼ g
tɕ tɕʰ tɕʼ dʑ ʈʂ ʈʂʰ tʂʼ ɖʐ
s z ɕ ʂ x
ɾ l j w

a e o i u ɨ

(C)V(R) syllable structure, where R is either a liquid or a glide. Diachronically, consonant clusters have been mercilessly reduced to the point where in the modern language they are found in loanwords only. Vowels can carry one of two lexical tones (high and low), which are complicated by pitch accent. Hiatus is broken up by a glottal stop, which is not found anywhere else.

Middle Suraic

m n
p b t̪ d̪ t d k g ʔ
tɕ dʑ
f v s z ɕ ʂ ʐ h
ɾ l j w

a e ɛ o ɔ u ɨ

(C)V(W)(C) syllable structure, where W is a glide or a liquid. Consonant clusters are extremely common, but all the plosives before other plosives have already debuccalised to a glottal stop and all the fricatives before plosives have debuccalised to /h/.

[bradrn]

Modern Suraic:

/m n n̪ ɲ ŋ/ ⟨m n nh ny ng⟩
/p pʰ t̪ t̪ʰ t tʰ tʼ k kʰ kʼ/ ⟨p ph x xh t th tt k kh kk⟩
/b ɓ d̪ d ɗ ɡ/ ⟨b bb dh d dd g⟩
/tɕ tɕʰ tɕʼ dʑ ʈʂ ʈʂʰ tʂʼ ɖʐ/ ⟨cy cyh ccy jy c ch cc j⟩
/s z ɕ ʂ x/ ⟨s z sy x h⟩
/ɾ l j w/ ⟨r l y w⟩

/a e o i u ɨ/ ⟨a e o i u ɨ⟩

Middle Suraic:

/m n/ ⟨m n⟩
/p b t̪ d̪ t d k g ʔ/ ⟨p b ṱ ḓ t d k g q⟩
/tɕ dʑ/ ⟨c j⟩
/f v s z ɕ ʂ ʐ h/ ⟨f v s z x š ž h⟩
/ɾ l j w/ ⟨r l y w⟩

/a e ɛ o ɔ u ɨ/ ⟨a e ɛ o ɔ u ɨ⟩

Could we have a sample text please? I’d like to see how these look in practise.

(Also, is that you, Birdlang? If not, it’s an amazing coincidence that you both have conlangs called Suraic with nearly identical inventories…)

[Karch]

/m n n̪ ɲ ŋ/ <m n nh ny ng>
/p pʰ b ɓ t̪ t̪ʰ d̪ t tʰ tʼ d ɗ k kʰ kʼ g/ <p ṗ b ḅ th ṭh dh t ṭ tʼ d ḍ k ḳ kʼ g>
/tɕ tɕʰ tɕʼ dʑ ʈʂ ʈʂʰ tʂʼ ɖʐ/ <ch c̣h chʼ j tr ṭr trʼ dr>
/s z ɕ ʂ x/ <s z sh sr x>
/ɾ l j w/ <r l y w>

/a e o i u ɨ/ <a e o i u ü>
/á à/ <á a> <û>

[dɮ the phoneme]

sinitic quebec french or something

/pʰ tʰ kʰ ʔ/
/b̥ d̥ g̊/
/f s ʂ x ʁ/
/m n ɲ ŋ/
/ r j/
/i y u/
/ ɪ ʏ ʊ/
/e ø o/
/ ɛ ɵ ~ ə ɔ/
/æ a/
/i˞ ɨ˞/

/e˞ o˞/
/ ɵ˞/
/æ˞ a˞/

Syllable structure: (C)CV(ʔ, ɴ), where ɴ is a moraic nasal.

Stress: left-aligned iambs

major allophones:

tʰ d̥ kʰ g̊ ʂ x > tsʰ dz̥ c ɟ̊ ɕ ç / _{i ɪ y ʏ j}
m > w / V_V
b d > ɓ ɗ / word-initially
ʁ > χ / [-voice]_
epenthetic ɨ̆ breaks up certain clusters
ɴ assimilates in place to following consonant, becomes nasalization on the vowel in utterance-final position. Becomes a homorganic nasalized sonorant before a sonorant.
V > Vː in stressed syllables
k g > q ɢ / _{a ʁ}

/b̥ d̥ g̊/ are lenis and may be realized with or without voicing depending on context.

/ʔb̥aɴ jøraʔ b̥ɔstɪ tki˞ɴne ske kʁʏʔ pəxtu xʂɪɴ jætnæ˞ʔry g̊u saɴ ʁɲɛ | sɪɴfraɴ ʂja g̊ʂaʔpkɵ˞ toɴxe/

[Killerwhale]

Alright, here's something unusual for you guys: a language that's super-heavy on prosody, to the point where each phoneme has multiple variants depending on which type of prosody applies to the word (palatalised, labialised and/or nasalised).

Consonants are shown here with palatalised, nasalised and then palatalised+nasalised forms following their default ones, labialisation on everything is just a Cʷ at most and is usually audible only on vowels anyway.
p > pʲ > m > mʲ
t > t͡ʃ > n > ɲ
k > kʲ > ŋ > ŋ
f > fʲ > mv > mvʲ
s > ʃ > nz > ɲʒ
x > ʃ > ŋg > ŋgʲ
ɾ > ɾ > ndɾ > ɲd͡ʒ
l > lʲ > nd > ɲd͡ʒ
j > j > ɲd͡ʒ > ɲd͡ʒ
w > ɥ > mb > mbʲ

Vowels are shown as plain, palatalised, labialised, then palatalised+labialised. Nasalisation on everything is just a Ṽ that doesn't change the quality of the actual vowel.
ɯ>i>u>y
ə>e>o>œ
a>a>a>a

Syllable structure is a very simple (C)V, with vowel sequences being broken by epenthenic glottal stops.

[bradrn]

Ooh, interesting! Reminds me of Marshallese. In this case, I suspect the straightforward thing to do would be to mark nasalisation on consonants only and palatalisation/labialisation on vowels only:

p > pʲ > m > mʲ ⟨p p m m⟩
t > t͡ʃ > n > ɲ ⟨t t n n⟩
k > kʲ > ŋ > ŋ ⟨k k g g⟩
f > fʲ > mv > mvʲ ⟨f f v v⟩
s > ʃ > nz > ɲʒ ⟨s s z z⟩
x > ʃ > ŋg > ŋgʲ ⟨x x ng ng⟩
ɾ > ɾ > ndɾ > ɲd͡ʒ ⟨r r nr nr⟩
l > lʲ > nd > ɲd͡ʒ ⟨l l d d⟩
j > j > ɲd͡ʒ > ɲd͡ʒ ⟨y y j j⟩
w > ɥ > mb > mbʲ ⟨w w b b⟩

ɯ>i>u>y ⟨ï i u ü⟩
ə>e>o>œ ⟨ë e o ö⟩
a>a>a>a ⟨a (i)a a (i)a⟩
And I suppose you could do ⟨an en ën etc.⟩ for words consisting of one nasalised vowel without consonants.

I’m not 100% sure how well this would work in practise, though, so a sample text would be appreciated!

[Killerwhale]

[ɲd͡ʒỹɲœ̃ ɯ̃ndrãndə̃nzə̃, aʃ ʃyrajyʃy ʃit͡ʃe tosu]
/jɯtə̃ʲʷ ɯraləsə̃, asʲ sɯrajɯsɯʲʷ xɯtəʲ təsɯʷ/

Not sure how to indicate the prosody, so I just marked it on the last vowel of each word.

[bradrn]

One change: I think I’ll represent /aʲ aʲʷ/ as ⟨(y)a⟩ rather than ⟨(i)a⟩, since that’s ambiguous with [iʔa].

Sample text:

[ɲd͡ʒỹɲœ̃ ɯ̃ndrãndə̃nzə̃, aʃ ʃyrajyʃy ʃit͡ʃe tosu]
/jɯtə̃ʲʷ ɯraləsə̃, asʲ sɯrajɯsɯʲʷ xɯtəʲ təsɯʷ/

Jünö ïnradëzë, asy sürayüsü xite tosu.

[aʃ] confuses me though… wasn’t the syllable structure supposed to be (C)V? If it was [ʃa] that would straightforwardly be ⟨sya⟩, but I’m not sure how to represent [aʃ].

[Killerwhale]

Ah, sorry, should be /asaʲ/ instead.

[bradrn]

sinitic quebec french or something

Looks like I somehow never saw this when it was posted, so let me do it now:

/pʰ tʰ kʰ ʔ/ ⟨p t k q⟩
/b̥ d̥ g̊/ ⟨b d g⟩
/f s ʂ x ʁ/ ⟨f s x h r⟩
/m n ɲ ŋ ɴ/ ⟨m n ny ŋ m~n⟩
/ r j/ ⟨r y⟩

/i y u/ ⟨i ü u⟩
/ ɪ ʏ ʊ/ ⟨ì ỳ ù⟩
/e ø o/ ⟨e ö o⟩
/ ɛ ɵ ~ ə ɔ/ ⟨è ə ò⟩
/æ a/ ⟨ä a⟩

/i˞ ɨ˞/ ⟨ir ur⟩
/e˞ o˞/ ⟨er or⟩
/ ɵ˞/ ⟨ər⟩
/æ˞ a˞/ ⟨är ar⟩

Allophony is not represented (except for /ɴ/).

/ʔb̥aɴ jøraʔ b̥ɔstɪ tki˞ɴne ske kʁʏʔ pəxtu xʂɪɴ jætnæ˞ʔry g̊u saɴ ʁɲɛ | sɪɴfraɴ ʂja g̊ʂaʔpkɵ˞ toɴxe/
Qban yöraq bòstì tkirnne ske krỳq pəhtu hxìn yätnärqrü gu san rnye. Simfran xya gxaqpkər tonhe.

[Killerwhale]

Another take on my last phonology, except this time it's less about prosody and more about a tiny and weird inventory with a lot of allophony.

p t k s h j

That's it for the standard dialect, folks! Southern dialects also add /ts/ to the mix as a distinct phoneme, and a few dialects have /r/ as a separate phoneme as well.

Vowels are, on the phonemic level, pretty simple as well:

i y ɯ u a

All the vowels come in long, nasalised and high tone forms, or any combination of the above.

Syllable structure is the CV, with null onsets being filled with an epenthetic glottal stop.

Now, for the really interesting part, which is allophony and alternative analyses.

You see, the vowels all come in either long or short varieties, and these varieties generally have different quality, so it's possible to also analyse the vowel system as the following:

i: ɪ y: ø ɯ: ʌ u: o a: ɛ

Alternatively, it's possible to analyse the syllable structure as C(w)V instead and add a /w/ to the phonemic inventory, the /w/ never appears on the surface level, but it transforms the bizarre and weird /i y ɯ u a/ vowel system into a saner /i a ɯ/ one by turning the rounded vowels into underlying /wV/ sequences. Or you can give each consonant a labialised counterpart that turns the following non-low vowel rounded.

Now, for the consonants. The consonants show great variance in realisation depending on the vowel following them.

/p/ becomes /w/ before a long vowel, or /m/ before a nasal one. Long nasal vowels still make for /m/
/t/ is /l/ before a long vowel word-initially, and /r/ between any two oral vowels. Nasal vowels of any length and quality turn it into /n/, a short front vowel that is not nasalised following a word-initial /t/ results in /t͡ʃ/
/j/ becomes /ɲ/ before a nasal vowel
/k/ becomes /ŋ/ before a nasal vowel
/s/ is /ʃ/ before a front vowel of any quality or nasality

So really, if you were to say that all or some of the allophones are different consonants, you could end up with a larger consonant inventory.

[Nortaneous]

fake hluic

/p b mb t d nd k g ŋg q ɴɢ ʔ/
/s ʃ h/
/m n ŋ/
/l r ʁ j w/
/a ɛ e o i ɨ u/ + /ã ẽ ə̃ õ/

/tj dj ndj/ [tʃ dʒ ndʒ], but these can't occur in codas, unlike /ʃ/

maʃ bjida hu o ndin ʔimɨrtjaʁi do ʁod ekjub madoʁbɨʁ, leʔɨʃtuʁ kɨrsɛ ʁamij ejeʔ. ʔa djãw ndjɨj ʁamʁun ɨʁɨrne, a ŋgu emjan ɨnakɨm, mjan yɨr ɨkãtuŋ naʁɨrleŋ, i ʁom qãʁamʁun lõ maʁɨnaduʁ.

[Starbeam]

Hope i'm doing this right.
---
fake Hluic

/m n ŋ/ m n n
/mb nd (ndʒ) ŋg ɴɢ/ mb nd (ndi) ňg ňq
/p b t d (tʃ) (dʒ) k g q ʔ/ p b t d (ti) (di) k g q ´/`
/s ʃ h/ s x h
/l r ʁ j w/ l ř r i u
/a ɛ e o i ɨ u/ a æ e o i y u
/ã ẽ ə̃ õ/ ã ẽ ỹ õ

vowels receive an acute or grave accent depending on whether the glottal stop sound is preceding or following them, respectively.
<ä ň ňg ňq ř š>= <ae nh ng nq rr sh>
the tilde is replaceable by the vowel following <nn>
the acute accent is replaceable by the vowel preceding or following <`>
apostrophes can be used to separate multigraphs
<j w> can replace semivowel <i u>

/maʃ bjida hu o ndin ʔimɨrtjaʁi do ʁod ekjub madoʁbɨʁ, leʔɨʃtuʁ kɨrsɛ ʁamij ejeʔ. ʔa djãw ndjɨj ʁamʁun ɨʁɨrne, a ŋgu emjan ɨnakɨm, mjan jɨr ɨkãtuŋ naʁɨrleŋ, i ʁom qãʁamʁun lõ maʁɨnaduʁ./

Maš biida hu o ndin ímyřtari do rod ekiub madorbyr, leýštur kyřsæ ramii eyè. Á diãw ndiyi ramrun yryřne, a ňgu emian ynakym, mian iyr ykãtuň naryřleň, i rom qãramrun lõ marynadur. (with diacritics)

Mash biida hu o ndin 'imyrrtari do rod ekiub madorbyr, le'yshtur kyrshae ramii eye'. 'A diannw ndiyi ramrun yryrrne, a ngu emian ynakym, mian iyr ykanntunh naryrrlenh, i rom qannraamrun lonn marynadur (without diacritics)

[Karch]

/p b mb t d nd k g ŋg q ɴɢ ʔ/ <p b mb t d dd k g gg kh gh '>
/s ʃ h/ <s sh h>
/m n ŋ/ <m n ng>
/l r ʁ j w/ <l r q y w>
/a ɛ e o i ɨ u/ <a ai ei o i e u>
/ã ẽ ə̃ õ/ <an in en un>
/an/ <ann>

/tj dj ndj/ [tʃ dʒ ndʒ], but these can't occur in codas, unlike /ʃ/

maʃ bjida hu o ndin ʔimɨrtjaʁi do ʁod ekjub madoʁbɨʁ, leʔɨʃtuʁ kɨrsɛ ʁamij ejeʔ. ʔa djãw ndjɨj ʁamʁun ɨʁɨrne, a ŋgu emjan ɨnakɨm, mjan yɨr ɨkãtuŋ naʁɨrleŋ, i ʁom qãʁamʁun lõ maʁɨnaduʁ.
Mash byida hu o ddinn 'imertyaqi do qod eikyub madoqbeq, lei'ishtuq kersai qamiy eiyei'. 'A dyanw ddyey qamqunn eqernei, a ggu eimyann enakem, myann yer ekantung naqerleing, i qom khanqamqunn lun maqenaduq.

or:
/ŋg q ɴɢ ŋ ʁ e ɨ ã ẽ ə̃ õ/ <ŋg q ŋq ŋ gh e ɨ añ eñ ɨñ oñ>
[tʃ dʒ ndʒ] <c j nj>

Mash byida hu o ndin 'imɨrcaghi do ghod ekyub madoghbɨgh, le'ɨshtugh kɨrsai ghamiy eye'. 'A jañw njɨy ghamghun ɨghɨrne, a ŋgu emyan ɨnakɨm, myan yɨr ɨkañtuŋ naghɨrleŋ, i ghom qañghamghun loñ maghɨnadugh.