• Options
• Login
• Desktop Style

Templatic Voynich generator

• Index

  • The Voynich Ninja
  • Voynich Research
  • Analysis of the text
  • Templatic Voynich generator
• Templatic Voynich generator

• Templatic Voynich generator

  Labyrinthinesecurity > 04-06-2026, 10:21 PM

  Dear all,
  
  I have made a template driven voynich text generator which passes the statistic tests, including my own 4 signature tests which are quite demanding. Is it something that would be worth sharing, or template generators are already well described and known? I couldnt find much litterature except the groundbreaking work from Stolfi, but I dont know if he made an actual generator. Thanks for your inputs

• RE: Templatic Voynich generator

  bi3mw > 04-06-2026, 10:45 PM

  I just tried out a text generator.
  You are not allowed to view links. Register or Login to view.
  Why don't you just give us a detailed overview of your project ?

• RE: Templatic Voynich generator

  Labyrinthinesecurity > 05-06-2026, 10:10 AM

  Ok here is an overview:
  
  1. Grapheme normalization
  Each Voynich word is first normalized:

  • Replace
    ch → C
    
    
  • Replace gallows letters (t, k, p, f) →G
    
    

  Then map characters into a reduced alphabet:

  • C,G stay unchanged
    
  • y → Y
    
  • q →Q
    
  • vowels → V
    
  • all other letters → X
    
    

  This produces a skeleton string encoding coarse word shape.
  
  Example for word "chekcheor":
  Step-by-step:

  • ch → C
    
  • k → G
    
  • vowels → V
    

  Resulting skeleton: chekcheor → CVGCVVX
  
  2. Frame extraction from skeletons
  Each skeleton is split into:

  • ONSET: substring before the first vowel (V)
    
  • CODA: substring after the last vowel
    
  • The central vowel region is discarded
    

  This yields a frame:    frame=(ONSET,CODA)
  
  Example using skeleton CVGCVVX:
  
  Structure:

  • first V at position 1
    
  • last V at position 5
    

  So:
  ONSET = C
  CODA  = X
  
  Frame = (C, X)
  
  3. Frame induction from corpus
  All observed (ONSET, CODA) pairs are extracted from the corpus:

  • frames are defined as distinct observed pairs
    
  • frames are ranked by frequency
    
  • the most frequent frames form the model’s state space
    

  
  4. Template association with frames
  Each corpus word has:

  • a skeleton (full structural pattern)
    
  • a derived frame (via ONSET/CODA split)
    

  Thus:

  • each frame is associated with a set of observed skeleton templates
    
  • this association is induced statistically from the corpus
    

  
  Example for frame (C, X)
  
  Observed templates:
  CV
  CVX
  CVGCVVX (chekcheor)
  ...
  
  5. Markov model over frames
  A Hidden Markov Model is trained where:

  • hidden states = frames
    
  • transitions = empirical frame-to-frame transitions observed in lines
    
  • additional biases exist for:
    • line start states
      
    • line end states
      
    • line length distribution
      

  
  6. Frame sequence generation
  At generation time:

  • a frame is sampled from the Markov chain
    
  • subsequent frames are generated via transition probabilities
    

  
  Example generated path: (C, X) → (V, X) → (V, X) → (C, X)
  Each step is chosen by Markov probabilities.
  
  7. Template emission conditioned on position
  Each frame emits a template according to position-dependent empirical distributions:

  • initial position distribution
    
  • middle position distribution
    
  • final position distribution
    

  These distributions are learned from observed corpus frequencies of templates within each frame context.
  The emission selects existing templates only.
  
  8. Surface word realization (lexical sampling)
  Each emitted template indexes a bucket of attested corpus words:

  • the final output word is sampled from this set
    
  • no new words are constructed at this stage
    

  
  The generator is trained separately on Currier A and Currier B, producing different texts for these two languages. 
  
  Example (Currier A): 
  lchdy qokol olo chekcheor chol
  tchor chol qoaiin qoteol cho dy
  tchory
  chopchal chody tos kcheey ainy chaiin
  ydar cho cholkol cheeykeem
  kcho dair dar daiin chok chy daiin
  alam
  toleechal daiin chol saiin
  oldal cheeky chol chotchy keol chan okar cholfy
  daiin chory daiiin ches aiin ykeol
  choly dor choo chody aiin olchy qoty
  ykeey al otoldy saiin choek chos chor
  ol
  dar daiin damo chol alaiinom okoldy daiin
  daiin okaraiin ols ldy ykoaiin dain cheokeey
  okeeor cheos qokchey qokod chetchy oeeeb
  chkor tchor chod otchol chaiin chain cho
  otcheey pykchy okoldg ytchom otoldy ol
  lchal choky dar cheky chor ykaiin dal
  qotchy qokeol qotomody chear chey
  kochor olkor chol chor chor chaiin kchy
  schey qopchy kchol olchor olaiin chokchol oty cheodar
  dchokchy chotchy daiin dal

• RE: Templatic Voynich generator

  Koen G > 05-06-2026, 10:11 AM

  This isn't News. Thread moved.

• RE: Templatic Voynich generator

  nablator > 05-06-2026, 11:49 AM

  (05-06-2026, 10:10 AM)Labyrinthinesecurity Wrote: You are not allowed to view links. Register or Login to view.Example (Currier A): 

  
  No sh? Did you convert both sh and ch to C for training, then convert C back to ch?

• RE: Templatic Voynich generator

  Labyrinthinesecurity > 05-06-2026, 12:16 PM

  (05-06-2026, 11:49 AM)nablator Wrote: You are not allowed to view links. Register or Login to view.

  (05-06-2026, 10:10 AM)Labyrinthinesecurity Wrote: You are not allowed to view links. Register or Login to view.Example (Currier A): 

  
  No sh? Did you convert both sh and ch to C for training, then convert C back to ch?

  
  You have sharp eyes sh is considered CH, and cth is tch (wame.for.f.p k) for reasons to be.explained in a dedicated.info theory thread.

• RE: Templatic Voynich generator

  Mauro > 05-06-2026, 12:30 PM

  I made a pseudo-Voynich generator some time ago. No idea of how it compares to yours, but it reproduced character distribution and bigrams distributions; it reproduced character, bigrams and word entropies, words lengths distribution over the text (Zipf's law) and the binomial words length distribution over the dictionary. It produced a very similar dictionary, with just slightly less hapax legomena (and with some words which I found offensive, the most frequent being 'po'). It was just a Markov chain generator, about 1200 states and ~8000 transitions. In the end I decided it was a not-that-interesting approach.
  
  Sample text:
  

  Quote:shey cheol otey kir okeedy tail shyy olshedy cheeodam oror al oty eeey odaiin loiiin qo okshey lchedy oty chol oteeody chory chol shedefy oteas cheodl cthochs shey qokeeody teedy oain sheey odchedy ykeeo okeo otal opcheold sheckhy shedy qokeedy atam shedy oteol lodaiin cheykches ykeedy or okeedy chol qol cheodaiin choldain otedy okeotos okchor rarod ykaiin rol osheeo dal otal qokaiin dar otoky oychey ctholechey olkaiidy okaiin oteeey cheey lsheo opchedy ar okar cheey chol cthol qokeeo chal okar otshedy y lkchokchol chcfhor chy otal chor chcthy okeeody kchos kaldy rchos qoky char qokeey oekeo chekary ykaiin al okeodaiin lkaiin otain chy qokaiin qod chey dain yair qokain qopchdy dal chcth orearary kchodaiin olky ctharoly shedy kos osary okeedy otaiin psho chokchey qoky chedy cheor sal dalsheody tol ail ol ykeeody otar chykchy otar chedy fchedy dar lshedy chedaiin qokaldy qoor poldaiin dy alo qockhey chedy okaiir cheocphey qokshdy qoteeedy r chqeos o chocfhor aky chean otain sar okeeody shey or darar ctheol chcthy qolkal otalor okeaiin cheodaiin or daiin shey aiin chol om shedydy shedy dolkeeody qokchedy daiin cheey sodaiin aiin oeeshy qokeey kolram shokocfhy saiir shar s cthy cheey ol opchd tol ol arorchy qocheoty dchor or lolkar ol sar qokeedy okedy qotdaiin okedy daiin otey cholpchedy y daiir dykeey sheealy chor chey pchocthy otykechy y or aiiin qofol shody daiin otaiin qotedy chytchy daiin daiir qotaiin qokeey chor dar qokaiin lkain dlshedy ykeed okeeey shedy checkhy aiin ar chedaiin keched chol roldalol kol rom chan lkeedy qokain aiin okeey oteedy qokeeey otair otody otaiin odal yteol okeod ar okedy qokeedy okeody doltol otaiin otolchey shokey qokeedy yky ctholkary qokal okeokeokeody qokar

• RE: Templatic Voynich generator

  oshfdk > 05-06-2026, 12:47 PM

  (05-06-2026, 10:10 AM)Labyrinthinesecurity Wrote: You are not allowed to view links. Register or Login to view.lchdy qokol olo chekcheor chol
  tchor chol qoaiin qoteol cho dy
  tchory
  chopchal chody tos kcheey ainy chaiin
  ydar cho cholkol cheeykeem
  kcho dair dar daiin chok chy daiin
  alam
  toleechal daiin chol saiin
  oldal cheeky chol chotchy keol chan okar cholfy
  daiin chory daiiin ches aiin ykeol
  choly dor choo chody aiin olchy qoty
  ykeey al otoldy saiin choek chos chor
  ol
  dar daiin damo chol alaiinom okoldy daiin
  daiin okaraiin ols ldy ykoaiin dain cheokeey
  okeeor cheos qokchey qokod chetchy oeeeb
  chkor tchor chod otchol chaiin chain cho
  otcheey pykchy okoldg ytchom otoldy ol
  lchal choky dar cheky chor ykaiin dal
  qotchy qokeol qotomody chear chey
  kochor olkor chol chor chor chaiin kchy
  schey qopchy kchol olchor olaiin chokchol oty cheodar
  dchokchy chotchy daiin dal

  
  I think the number of rare or non attested glyph combinations here is a bit high:
  pykchy (I don't think 'pykc' exists in the MS)
  cheeykeem (I don't think ykeem is in the MS)
  toleechal (appears once as toleeshal in the MS)
  oeeeb (appears once)
  qotomody (appears once)
  
  I don't know what the idea behind this generator is, if it was supposed to produce believable Voynichese, it needs some tweaking.

• RE: Templatic Voynich generator

  Labyrinthinesecurity > 05-06-2026, 02:29 PM

  (05-06-2026, 12:30 PM)Mauro Wrote: You are not allowed to view links. Register or Login to view.I made a pseudo-Voynich generator some time ago. No idea of how it compares to yours, but it reproduced character distribution and bigrams distributions; it reproduced character, bigrams and word entropies, words lengths distribution over the text (Zipf's law) and the binomial words length distribution over the dictionary. It produced a very similar dictionary, with just slightly less hapax legomena (and with some words which I found offensive, the most frequent being 'po'). It was just a Markov chain generator, about 1200 states and ~8000 transitions. In the end I decided it was a not-that-interesting approach.
  
  Sample text:
  

  Quote:shey cheol otey kir okeedy tail shyy olshedy cheeodam oror al oty eeey odaiin loiiin qo okshey lchedy oty chol oteeody chory chol shedefy oteas cheodl cthochs shey qokeeody teedy oain sheey odchedy ykeeo okeo otal opcheold sheckhy shedy qokeedy atam shedy oteol lodaiin cheykches ykeedy or okeedy chol qol cheodaiin choldain otedy okeotos okchor rarod ykaiin rol osheeo dal otal qokaiin dar otoky oychey ctholechey olkaiidy okaiin oteeey cheey lsheo opchedy ar okar cheey chol cthol qokeeo chal okar otshedy y lkchokchol chcfhor chy otal chor chcthy okeeody kchos kaldy rchos qoky char qokeey oekeo chekary ykaiin al okeodaiin lkaiin otain chy qokaiin qod chey dain yair qokain qopchdy dal chcth orearary kchodaiin olky ctharoly shedy kos osary okeedy otaiin psho chokchey qoky chedy cheor sal dalsheody tol ail ol ykeeody otar chykchy otar chedy fchedy dar lshedy chedaiin qokaldy qoor poldaiin dy alo qockhey chedy okaiir cheocphey qokshdy qoteeedy r chqeos o chocfhor aky chean otain sar okeeody shey or darar ctheol chcthy qolkal otalor okeaiin cheodaiin or daiin shey aiin chol om shedydy shedy dolkeeody qokchedy daiin cheey sodaiin aiin oeeshy qokeey kolram shokocfhy saiir shar s cthy cheey ol opchd tol ol arorchy qocheoty dchor or lolkar ol sar qokeedy okedy qotdaiin okedy daiin otey cholpchedy y daiir dykeey sheealy chor chey pchocthy otykechy y or aiiin qofol shody daiin otaiin qotedy chytchy daiin daiir qotaiin qokeey chor dar qokaiin lkain dlshedy ykeed okeeey shedy checkhy aiin ar chedaiin keched chol roldalol kol rom chan lkeedy qokain aiin okeey oteedy qokeeey otair otody otaiin odal yteol okeod ar okedy qokeedy okeody doltol otaiin otolchey shokey qokeedy yky ctholkary qokal okeokeokeody qokar

  
  Interesting! Was it Currier A or B?  In any case, it fails 2 of the 4 signatures: 
  
  FOUR-SIGNATURE EVALUATION
    ───────────────────────────────────────────────────────
  
    Sig1  E->S:      66.2%  [66.2, 66.2]
          VMS ref:  64.3%  [59.1, 68.6]
          Criterion: range [50, 79]%
  
    Sig2  Bilateral (full corpus):
          Generator: Se=0, Ee=0  → NO
          VMS ref:  Se=3, Ee=7  → YES
          Criterion: must have bilateral (ref has it)
          (chunks:  gen=0%  ref=90%)
  
    Sig3  MI:        2.0376  [2.0376, 2.0376]
          VMS ref:  0.4980  [0.4759, 0.5241]
          Criterion: range [0.249, 0.996]
          Quality:  overshoot (4.1x) (4.09x reference)
  
    Sig4  Shape:    Zipfian  R²=0.918  CV=0.97
          VMS ref:  Intermediate  R²=0.805  CV=1.53
          Criterion: non-Plateau (ref=Intermediate)
  
    ───────────────────────────────────────────────────────
    VERDICT (calibrated to Currier B):
    ───────────────────────────────────────────────────────
      E->S        : ✓ PASS      [range [50, 79]%]
      Bilat      : ✗ FAIL      [must have bilateral (ref has it)]
      MI          : ✗ FAIL      [range [0.249, 0.996]]
      Shape      : ✓ PASS      [non-Plateau (ref=Intermediate)]
    ───────────────────────────────────────────────────────
      Joint (adaptive):  2/4
  
    Reference values:
      VMS Currier B: E->S=64.3%  MI=0.4980  Bilateral=YES  Intermediate
  
    ★ 2/4 signatures passed

• RE: Templatic Voynich generator

  dashstofsk > 05-06-2026, 02:44 PM

  Would it be possible for your generator to output its text in IVTFF format? And if it could generate 20 pages and about 7000 words total I would then be better able to compare it with quire 13.
• Next Oldest Next Newest