Game-Translator
The Witcher 2 Subtitle
logo
LOCALIZATION MOD
RELEASED vExperimental-1 Austronesian Lang

The Witcher 2 Subtitle The Witcher 2 Subtitle

Bahasa Indonesia, Melayu, Filipino

Memuat data interpretasi naratif secara real-time...

Product Narrative

The Full Story

Geralt si White Wolf lagi kejepit di tengah konspirasi pembunuhan raja-raja, dan dia butuh bantuan lo buat bersihin namanya. The Witcher 2 itu game brutal, penuh intrik, dan dialognya tajem-tajem. Sayang banget kan kalau lo main tapi nggak nangkep nuansa omongannya yang kasar, puitis, atau lucu?


Tenang, gue udah buatin solusinya: mod lokalisasi paling niat se-Nusantara! Gue proses 364,582 kata pake neural pipeline 8 tahap biar hasilnya 87 persen pas sama konteks aslinya. Dari Roche yang omongannya kasar khas serdadu sampe Iorveth yang puitis tapi rasis, semuanya gue sesuaikan pake slang native kita biar nggak kaku kayak buku PPKN. Gas download sekarang, tunjukin ke dunia kalau Witcher bisa ngomong lebih lokal dari abang nasi goreng depan komplek!

Current Milestone

Available Now

Author's Notes

=== Audit Teknis & Semantik Lokalisasi THE WITCHER 2 ASSASSIN'S OF KING ===

1. SKALA LINGUISTIK & CAKUPAN

- Skala Proyek: Sekitar 364,582 kata diproses melalui alur neural 8-tahap.

- Cakupan Bahasa: Dukungan trilingual penuh untuk pasar Indonesia, Malaysia, dan Filipina.


2. VALIDASI NEURAL & AKURASI


- Skor Keselarasan Semantik (Platt Score): Indonesia: 87%, Malay: 85%, Filipino: 83%

(Skor ini mengukur seberapa akurat terjemahan mempertahankan makna asli dari teks sumber.)

- Gaya Bahasa Karakter: Penyesuaian gaya (gaul, formal, santai) telah diterapkan pada 109 karakter unik.

- Pemulihan Struktur Otomatis (Tag Repair): 29 tag kode game telah dipulihkan secara presisi.


3. KAPABILITAS ENGINE

- Pipeline: Austronesian Localization System (Neural LoRA-Adaptive Architecture).

- Pengenalan Entitas: Ekstraksi penuh untuk terminologi spesifik game dan konstanta lore.

Linguistic Analysis Report

Stylometric Register Analysis

Discourse analysis using Gemma embeddings. Classifies rhetorical register across the corpus to ensure tonal consistency with source narrative assets.

Casual
53.1%
Standard
18.7%
Formal
28.1%
Emotional Spectrum

Emotional tone mapped via dot-product similarity between extracted dialog embeddings and predefined sentiment anchors using zero-shot semantic alignment.

Neutral/Functional
28.2%
Stoic/Restrained
26.5%
Positive/Warm
17.4%
Negative/Intense
15.7%
Complex/Ambivalent
12.2%
Archetypes
30 detected
Geralt Of Rivia
11.9%
Vernon Roche
9.1%
Cecil Burdon
8.5%
Troll
8.2%
Ui
7.5%
Npc
6.7%
Ves
6.4%
Zoltan Chivay
5.9%
Cynthia
3.9%
Saesenthessis (saskia)
3.8%
System Ui / Prompts
2.9%
Letho (kingslayer)
2.5%
Sabrina Glevissig
2.1%
Bernard Loredo
1.9%
Emhyr Var Emreis
1.9%
Jan Natalis
1.9%
Iorveth
1.7%
Henselt
1.7%
Triss Merigold
1.4%
Dandelion
1.3%
Sheala De Tancarville
1.2%
Radovid V
1.0%
Foltest
1.0%
Detmold
0.9%
Kimbolt
0.8%
Stennis
0.8%
Aryan La Valette
0.7%
Maria Louisa La Valette
0.6%
Shilard Fitz-oesterlen
0.6%
System
0.5%

DISCLOSURE: Profiling data generated algorithmically via zero-shot inference and semantic vector alignment. Represents AI interpretation of the dataset corpus, not explicit ground-truth statistics from the underlying game engine or internal metrics. Use as a heuristic guide for context mapping.

Cross-Lingual Quality Matrix

Semantic alignment quantified via Multilingual E5 Large Instruct (RoBERTa based) bitext mining. NER entities preserved using GLiNER heuristic extraction protocols to maintain terminological invariance.

ID
Indonesian
28,246 / 28,860 lines
98%
Semantic Sim.
87 %
Lex. Density
71.2 %
src
63.1%
Lex. Diversity
5.3 %
src
4.4%
MS
Malay
28,432 / 28,860 lines
99%
Semantic Sim.
85 %
Lex. Density
71.9 %
src
63.1%
Lex. Diversity
4.0 %
src
4.4%
TL
Tagalog
27,742 / 28,860 lines
96%
Semantic Sim.
83 %
Lex. Density
57.7 %
src
63.1%
Lex. Diversity
4.9 %
src
4.4%

* Sim = Cosine Similarity (Vector Space) · Density = Content/Total Tokens · Diversity = TTR (Type-Token Ratio) · "src" = Source Baseline · Named Entities enforced via GLiNER mining.

Corpus Volume & Metrics
83,676 Token Lines
Src Density
63.1%
Src Diversity
4.4%
Syntactic Error Report

Heuristic markup verification utilizing multi-pass validation and correction to ensure syntactical integrity of control codes and visual tags.

29
Mismatch
29
Fixed
0
Partial

Name

Label
Retrieving Portrait...
Narrative Profile

Associated Entities
Semantic Archetypes

NLP Pipeline Intelligence

Featured Preview Auto-Detected

Line Identity 0
Source (English)
Loading...
Indonesian (ID)
Loading...
Malay (MS)
Loading...
Tagalog (TL)
Loading...

Pipeline Receipts

Merger (S7) 2026-03-26 11:48
Tagger (S1) 2026-03-26 10:38
Tag Repair (S6) 2026-03-26 10:17
Validator (S5) 2026-03-26 09:16
Re-Import (S4) 2026-03-26 08:52
Corrector (S3) 2026-03-26 08:37
Translator (S2) 2026-03-26 04:31
Splitter (S0) 2026-03-26 00:19

Released Archive

Austronesian Showcase

Location
Image
Video