Skip to content
Switch branches/tags
Go to file

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Neon Genesis Evangelion - Typing Project E (Shinseiki Evangelion - Typing E Keikaku)
English Translation v2.0

Developed by Derek Pascarella (ateam)

Find me on...
 -> SegaXtreme:
 -> Dreamcast-Talk:
 -> GitHub:
 -> Twitter:
 -> Reddit:

.-----------::[ Patching Instructions ]::-----------
| 1) Copy TOSEC-style GDI files (gdi+raws+bins) to the "gdi" folder.
| 2) Drag the "gdi" folder onto the "extract_gdi.bat" file.  A new folder called
|    "gdi Extracted" will be created.  This folder contains all of the data files
|    from the game.
| 3) Copy all of the files from the "patched_files" folder to the "gdi Extracted"
|    folder, choosing to overwrite existing files when prompted.
| 4) Drag the "gdi Extracted" folder onto the "build_gdi.bat" file.  A small black
|    window will open stating that the GDI build process has begun.  Once it has
|    completed, a success message will appear before the window automatically
|    closes after a few seconds.
| 5) The GDI files (gdi+raws+bins) in the "gdi" folder are now ready for use.

.------------::[ v2.0 Release Notes ]::-------------
| It is with extreme pride that I announce to you that the final, definitive
| version (i.e. version 2) of my English translation patch for "Neon Genesis
| Evangelion - Typing Project E" has arrived!  I have included a lengthy
| explanation of the methods used to bring you this release, but if you have no
| interest in that, just know this: All levels are playable and use localized
| English wordbanks, and every aspect of the game is not only beatable, but
| also playable, and above all, enjoyable!
| There were two major challenges that I set out to tackle for version 2.  The
| first was to swap out any Romanized Japanese wordbanks for localized English
| ones.  For those who have played previous versions of my patch, you will have
| noticed these wordbanks in the following levels/stages:
|   1) Level 2, Stage 2
|   2) Level 2, Stage 3
|   3) Level 3, Stage 2
|   4) Level 3, Stage 3
|   5) Level 4, Stage 1
|   6) Level 4, Stage 2
|   7) Level 4, Stage 3
| The wordbank format expected by these levels is unlike all other levels that
| use English words.  For these levels, the game was designed to take Hiragana
| and then translate it on-the-fly to its Romanized equivalent in order for the
| player to use the Latin alphabet when typing.  This means that the wordbank
| itself contains words comprised entirely of Hiragana.
| The first task was to determine the hex values that correspond to each
| Hiragana character from the game's character map (which, by the way, is
| actually just a giant bitmap image).  Once successfully identifying the hex
| values, my next step was to determine all valid English words that can be
| built using the Romanized spellings of those Hiragana.
| To achieve this, I wrote a series of small programs.  The initial step was to
| generate all permutations/combinations of Hiragana characters, taking into
| account word complexity for the increasing difficulty as the game progresses.
| With all of those permutations/combinations generated, I then used a lookup
| table to substitute the Hiragana with its Romanized spelling (e.g. ら = ra,
| な = na, etc.).  I then compared each potential word against the English
| dictionary, storing separately all valid words.  Before committing these
| words to the game's wordbanks, I curated them by hand to omit uncommon and
| largely unrecognizable words in order to give the game a more natural feeling
| and appeal to a wider audience.
| With this portion of the game done, I was ready to take on the second
| challenge. The holy grail that so many asked about.  Of course, I'm referring
| to making levels 5, 6, and the bonus level 7 not only beatable, but completely
| translated/localized and (enjoyably) playable.  Since I already cracked the
| encoding format for the Japanese wordbanks and successfully reverse-engineered
| a method to populate them with the words of my choice, much of the difficult
| work was already done.  However, like so many things in this game, these last
| three levels are quite different from those that proceed it.
| Recall the level 2, 3, and 4 stages that I listed above.  For those levels,
| Hiragana characters are translated on-the-fly to their Romanized equivalent.
| A similar mechanism is at play for levels 5, 6, and the bonus level 7.
| However, for these levels, instead of taking Hiragana and then Romanizing it
| for the player to input Latin letters, these levels expect a mixture of
| Kanji, Katakana, and Hiragana.  The player is then expected to use Latin
| letter key combinations to build Hiragana characters, the groupings of which
| will build the expected Kanji.  For the Katakana, my assumption is that
| Japanese players use keyboards with the special shift/toggle key (typically
| to the left of the "1" key) to switch to Katakana mode in order to enter
| those characters directly.
| Note that even though these levels allow the player to input Latin alphabet
| characters, it's only for purposes of building Hiragana.  In other words,
| there was no option to find a brand new solution to swap out the wordbank
| with purely English words that have nothing to do with being made of Hiragana
| characters.  I tried, but as soon as the series of Latin alphabet characters
| I inputted were recognized as one that builds a Hiragana character, my text
| was immediately converted.  Without a decent debugger, finding the part of
| the game responsible for the input method proved to be impossible for me.
| Armed with this understanding, the solution hit me like a bolt of lightning!
| Since I already had a large wordbank of English words that can be built with
| the Romanized versions of Hiragana characters, all I needed to do was modify
| the aforementioned character bitmap to replace those Hiragana with the one,
| two, or three Latin alphabet characters used to build them.  You'll notice
| when playing through those last three levels that the spacing of the letters
| in each English word sometimes appears odd, and it's due to me being forced
| to use the same space to fit one, two, or three letters.  I did a lot of
| tweaking to minimize this effect, but eliminating it altogether wasn't a
| possibility.
| In the end, I'm extremely happy with the way my English translation turned
| out.  It's very thorough and makes every effort to make the game not only
| playable and beatable, but also to fully localize it, as well.  The only
| non-English pieces that remain are the spoken dialog heard from time to time,
| like when clearing a stage.  I actually skimmed through my "Neon Genesis
| Evangelion" DVDs with the English dubbing track on them to see if there
| were any clips I could substitute, but unfortunately this proved unfruitful.
| I like to tell myself that this reminds the player that they're still
| enjoying a niche piece of Japanese Dreamcast history.

.------------::[ v1.5 Release Notes ]::-------------
| I am considering this release to be something of a "stop gap" until version 2
| comes along.  With version 1.5, the entire game is now beatable so that
| players can unlock the bonus level and see all content therein.
| As explained in the version 1 release notes, the only parts of the game that
| were left unplayable were level 5, level 6, and the bonus level (i.e. level
| 7).  After hours and and hours of effort, I managed to partially
| reverse-engineer the encoding used for the Japanese word bank data for those
| levels.
| At present, all rounds of these last three levels recycle the same "word" over
| and over again.  That word is "SDFGHJKL", hence why I described my
| reverse-engineering efforts as "partial".  This "word" was chosen for two
| reasons:
|   1) It consists of eight characters and thus serves to help increase the
|      player's total score for each round.
|   2) No combination of those letters correlates to letters-groupings used
|      to generate a Hiragana character.  This is important because the last
|      three levels do not process Roman letters the same way that all previous
|      levels do.  Instead of a 1:1 relationship between what the player types
|      and what appears on-screen, the last three levels convert the Roman
|      letter-groups to Hiragana in order to build the Kanji from each prompt.
| A bug fix was also implemented to eliminate all instances of the player being
| required to type the colon character (:), since the character mapping for the
| standard Western Dreamcast QWERTY keyboard does not align with that of the
| Japanese keyboard for that particular character.  Thus, it was impossible for
| the player to actually input that character.
| Long term, my goal is to take this translation patch to the next and final
| level by using meaningful word to populate the word banks for level 5, 6, and
| the bonus level (i.e. level 7).  While I've made some progress in this
| area, I have not yet arrived at a solution that I am happy with.  That being
| said, I decided to release version 1.5 in order to ensure that all who so
| desire can actually play through and beat the entire game.
| Stay tuned!

.------------::[ v1.0 Release Notes ]::-------------
| I'm very proud to release the very first version of my English translation
| patch for "Neon Genesis Evanglion - Typing Project E" for the Sega
| Dreamcast!  Thanks to a very active community, The Dream™ never dies and I
| want to give my most heartfelt "thank you" to each and every person who
| is passionate about Sega's farewell console.
| In this initial release (version 1), all menus, options, dialog boxes, and
| screens have been fully translated.  Levels 1 through 4 have also been fully
| translated, with the exception of a few short sequences of spoken dialog.
| Levels 5, 6, and the bonus level have only had their menus, dialog boxes and
| screens translated.  However, the levels are not yet playable by a
| non-Japanese speaker.  My goal is to complete a playable version of those
| levels for version 2.  Allow me to elaborate with more detail…
| Part of what made translating this game so attractive to me is the majority
| of the game being QWERTY-keyboard friendly, where an English speaker with
| access to the standard Dreamcast keyboard is able to play and complete each
| level.  Some levels use English words for their word bank, others use the
| Romanji spellings of Japanese words.  In either case, the game is fully
| playable.
| However, the previously mentioned levels (5, 6, and bonus) instead present
| the player with words written in a mixture of Kanji and Katakana/Hiragana.
| The game then expects the player to use the key combinations necessary to
| produce the series of Katana/Hiragana characters that construct one Kanji
| character.  It's obvious how challenging this is from a translation
| perspective.  There are two approaches to solving this:
|   1) Display the Roman alphabet letters that correspond to each Katakana,
|      Hiragana, and Kanji character that the player is asked to type, this way
|      they can easily key in the correct sequence of letters in order to
|      successfully complete the word.  The challenge here is fitting the
|      Roman letters into the space provided on-screen for each character.
|      This is the most straight forward solution as of right now and will
|      likely be polished and implemented in version 2.
|   2) Reverse-engineer the game to either:
|        a. Display as many Roman letters as I'd like for each Katakana,
|           Hiragana, or Kanji character.
|        b. Use entirely custom word banks that would allow for me to avoid the
|           Japanese character problem altogether.
| At this time of this release, I have not been successful in either of these,
| and of course I welcome any and all contributions from others in the
| community with more experience in this area.
| It's also worth mentioning that level 5 allows the player to "cheat" by
| pressing Enter to display the Katakana/Hiragana equivalent of each Kanji word
| to be typed.  As it is much simpler to fit the two or three Roman letters
| into the space provided for each Katakana/Hiragana character, I initially
| believed that level 5 could be solved in this manner.  However, I quickly
| discovered that by cheating, the player is awarded no points for that
| individual word, thus leaving them with no way to progress.
| Lastly, the "WEBPAGE" option on the title screen takes you to a browser that
| displays a website that's stored directly on the disc.  While the browser's
| menu and all of the site's pages are translatable, it's a very low priority
| for me (for obvious reasons).
| In summary, this game is extremely playable.  I have had a lot of fun on this
| project and so you can expect a second and final version to be released once
| I overcome the previously mentioned challenges.  I hope you enjoy the game!

.-----------------::[ Changelog ]::-----------------
| -> 2021-01-24 (v2.0)
|      -Romanized Japanese wordbanks from levels 2, 3, and 4 now fully
|       translated and localized.
|      -Levels 5, 6, and bonus level 7 now fully translated, localized,
|       playable, and enjoyable.
| -> 2021-01-22 (v1.5)
|      -Levels 5, 6, and bonus level 7 now beatable (reduced playability).
|      -Fixed colon (:) bug in level 1.
| -> 2021-01-14 (v1.0)
|      -Initial release.


English translation patch for the Sega Dreamcast game "Neon Genesis Evangelion - Typing Project E".



No packages published