re: the sound-cue table, and cutscene dialogue's speaker/portrait/timing
dat/tables.pak holds a 5798-entry SOUNDS record (cue name -> sound id) and a 5135-entry FILES record (.slb bank paths). Cue names are the join key, so a script message id now resolves all the way to the bank that voices it: MSG_VOICE_D_257 -> VOICE_D_257 -> 6945 -> jpn\etc\VOICE_D_257.slb. The prefix rule is MSG_ -> VOICE_, not strip-MSG_. My first rule was the latter; it left 88 names unresolved and I was about to write those families up as text-only announcements, until VOICE_TCAF_592.slb turned up in FILES and refuted it. Corrected rule resolves 1326 of 1338, and SOUNDS and FILES agree on exactly the same 12 absentees. Separately, MSG_DEMO_* is driven by its own IDXD tables in the language packs, which carry speaker, portrait, on-screen seconds and audio cue per page. Field count is 9*PageCount+2 for all 7 distinct PageCounts; 1252/1252 caption-key slots match <ID>_<page>_<line>; the 78 multi-page records equal the 78 counted independently from the caption side; 138 ids close exactly against the caption table both ways. Does not settle the known VOICE_D_452 wrong-recording case -- every cue id is distinct, so bank sharing is not happening at this layer.
This commit is contained in:
Sylpheed RE agent
203
tools/re-capture/sound_cues.py
Normal file
203
tools/re-capture/sound_cues.py
Normal file
@@ -0,0 +1,203 @@
|
||||
#!/usr/bin/env python3
|
||||
"""The sound-cue table in `dat/tables.pak`, and what a message id binds to.
|
||||
|
||||
sound_cues.py summary # table shape and cue families
|
||||
sound_cues.py resolve MSG_VOICE_D_257 # message id -> cue -> id + bank file
|
||||
sound_cues.py demo # the cutscene message table
|
||||
sound_cues.py unbound # script message ids with no recording
|
||||
|
||||
One IDXD object in `dat/tables.pak` carries five records: `SETTINGS`,
|
||||
`BANK_SE`, `FILES` (5 135 `.slb` bank paths), `STAGES` (empty) and `SOUNDS`
|
||||
(5 798 named fields, cue name -> numeric sound id). Cue names are the join key:
|
||||
a script message `MSG_<X>` plays the cue `VOICE_<X>`, and the bank file is
|
||||
`jpn\\Voice\\VOICE_<X>.slb`. See docs/re/structures/sound-cue-table.md.
|
||||
"""
|
||||
import collections
|
||||
import os
|
||||
import re
|
||||
import struct
|
||||
import sys
|
||||
import zlib
|
||||
|
||||
TABLES = '/work/sylph_extract/dat/tables.pak'
|
||||
|
||||
|
||||
def be32(b, o):
|
||||
return struct.unpack_from('>I', b, o)[0]
|
||||
|
||||
|
||||
def cstr(b, pool, off):
|
||||
base = pool + off
|
||||
return b[base:b.index(b'\0', base)].decode('latin1')
|
||||
|
||||
|
||||
def pak_entries(path):
|
||||
d = open(path, 'rb').read()
|
||||
segs, base, i = b'', path[:-4], 0
|
||||
while os.path.exists('%s.p%02d' % (base, i)):
|
||||
segs += open('%s.p%02d' % (base, i), 'rb').read()
|
||||
i += 1
|
||||
out = {}
|
||||
for k in range(be32(d, 4)):
|
||||
h, off, cs = struct.unpack_from('>III', d, 0x10 + 12 * k)
|
||||
raw = segs[off:off + cs]
|
||||
if raw[:2] == b'Z1':
|
||||
try:
|
||||
raw = zlib.decompress(raw[10:])
|
||||
except Exception:
|
||||
pass
|
||||
out[h] = raw
|
||||
return out
|
||||
|
||||
|
||||
def records(b):
|
||||
"""IDXD -> {record name: [(field name or None, value)]}."""
|
||||
n = be32(b, 4)
|
||||
o = 8 + 16 * n
|
||||
recs = [struct.unpack_from('>IIII', b, 8 + 16 * i) for i in range(n)]
|
||||
m = be32(b, o)
|
||||
o += 4
|
||||
fl = [struct.unpack_from('>III', b, o + 12 * i) for i in range(m)]
|
||||
o += 12 * m
|
||||
pool = o + 4
|
||||
out = {}
|
||||
for (_nh, no, fb, fe) in recs:
|
||||
name = cstr(b, pool, no) if no != 0xFFFFFFFF else None
|
||||
out[name] = [(cstr(b, pool, fno) if fno != 0xFFFFFFFF else None,
|
||||
cstr(b, pool, fvo)) for (_k, fno, fvo) in fl[fb:fe]]
|
||||
return out
|
||||
|
||||
|
||||
def sound_table():
|
||||
obj = [v for v in pak_entries(TABLES).values()
|
||||
if v[:4] == b'IDXD' and b'DEMO_017' in v][0]
|
||||
r = records(obj)
|
||||
sounds = {k: v for k, v in r['SOUNDS']}
|
||||
files = [v for _, v in r['FILES']]
|
||||
return sounds, files, r
|
||||
|
||||
|
||||
def cue_for(message_id):
|
||||
"""`MSG_VOICE_D_257` -> `VOICE_D_257`; `MSG_TCAF_592` -> `VOICE_TCAF_592`.
|
||||
|
||||
The `MSG_` prefix becomes `VOICE_`, except that a name already carrying
|
||||
`VOICE_` keeps the one it has rather than doubling it.
|
||||
"""
|
||||
x = message_id[4:] if message_id.startswith('MSG_') else message_id
|
||||
return x if x.startswith('VOICE_') else 'VOICE_' + x
|
||||
|
||||
|
||||
def wstr(b, pool, off):
|
||||
base = pool + 2 * off
|
||||
e = base
|
||||
while b[e:e + 2] != b'\0\0':
|
||||
e += 2
|
||||
return ''.join(chr(c) for c in
|
||||
struct.unpack('>%dH' % ((e - base) // 2), b[base:e]))
|
||||
|
||||
|
||||
def captions(lang='E'):
|
||||
"""`MSG_*_<page>_<line>` -> text, from the pack's IXUD field tables."""
|
||||
pak = '/work/sylph_extract/dat/GP_MAIN_GAME_%s.pak' % lang
|
||||
out = {}
|
||||
for b in pak_entries(pak).values():
|
||||
if b[:4] != b'IXUD':
|
||||
continue
|
||||
n = be32(b, 4)
|
||||
o = 8 + 16 * n
|
||||
m = be32(b, o)
|
||||
o += 4
|
||||
fl = [struct.unpack_from('>III', b, o + 12 * i) for i in range(m)]
|
||||
o += 12 * m
|
||||
pool = o + 4
|
||||
for (_k, no, vo) in fl:
|
||||
if no == 0xFFFFFFFF:
|
||||
continue
|
||||
try:
|
||||
key, val = wstr(b, pool, no), wstr(b, pool, vo)
|
||||
except Exception:
|
||||
continue
|
||||
if key.startswith('MSG_') and val.strip():
|
||||
out.setdefault(key, val)
|
||||
return out
|
||||
|
||||
|
||||
def demo_messages(lang='E'):
|
||||
"""The cutscene message table: id -> per-page (speaker, face, dur, cue)."""
|
||||
pak = '/work/sylph_extract/dat/GP_MAIN_GAME_%s.pak' % lang
|
||||
out = {}
|
||||
for b in pak_entries(pak).values():
|
||||
if b[:4] != b'IDXD' or b'MSG_DEMO' not in b:
|
||||
continue
|
||||
for name, fs in records(b).items():
|
||||
if not name or not name.startswith('Message_'):
|
||||
continue
|
||||
named = {k: v for k, v in fs if k}
|
||||
pos = [v for k, v in fs if k is None]
|
||||
pc = int(named['PageCount'])
|
||||
out.setdefault(named['ID'], []).append(
|
||||
[tuple(pos[9 * p:9 * p + 9]) for p in range(pc)])
|
||||
return out
|
||||
|
||||
|
||||
def main():
|
||||
what = sys.argv[1] if len(sys.argv) > 1 else 'summary'
|
||||
sounds, files, recs = sound_table()
|
||||
stems = {os.path.basename(f.replace('\\', '/'))[:-4]
|
||||
for f in files if f.lower().endswith('.slb')}
|
||||
|
||||
if what == 'resolve':
|
||||
mid = sys.argv[2]
|
||||
c = cue_for(mid)
|
||||
print('message %s' % mid)
|
||||
print(' cue %s' % c)
|
||||
print(' sound id %s' % sounds.get(c, '<not in SOUNDS>'))
|
||||
hit = [f for f in files
|
||||
if os.path.basename(f.replace('\\', '/'))[:-4] == c]
|
||||
print(' bank file %s' % (hit[0] if hit else '<no bank>'))
|
||||
return
|
||||
|
||||
if what == 'demo':
|
||||
cap = captions()
|
||||
print('# Cutscene message table — dat/GP_MAIN_GAME_E.pak, IDXD objects')
|
||||
print('#')
|
||||
print('# One line per page. A page is one subtitle box; successive pages are')
|
||||
print('# successive utterances, which is why the speaker changes mid-message.')
|
||||
print('# Columns: id, page, speaker, portrait, on-screen seconds, audio cue,')
|
||||
print('# then the caption text.')
|
||||
print()
|
||||
for mid, variants in sorted(demo_messages().items()):
|
||||
for pages in variants:
|
||||
for i, g in enumerate(pages):
|
||||
text = ' '.join(cap.get('%s_%03d_%02d' % (mid, i, l), '')
|
||||
for l in range(4)).strip()
|
||||
print('%-16s p%d %-14s %-18s %5ss %-11s %s'
|
||||
% (mid if i == 0 else '', i,
|
||||
g[0].replace('Character', ''),
|
||||
g[1].replace('Face', ''), g[3], g[4], text))
|
||||
return
|
||||
|
||||
if what == 'unbound':
|
||||
for c in sorted(set(sounds) - stems):
|
||||
print('%-24s id=%s (cue with no bank file)' % (c, sounds[c]))
|
||||
return
|
||||
|
||||
print('records : %s' % ', '.join(str(k) for k in recs))
|
||||
print('SOUNDS cues : %d' % len(sounds))
|
||||
print('FILES bank paths : %d (distinct stems %d)' % (len(files), len(stems)))
|
||||
print('cues with no file : %d' % len(set(sounds) - stems))
|
||||
print('files with no cue : %d' % len(stems - set(sounds)))
|
||||
fam = collections.Counter(re.match(r'[A-Za-z]+', k).group(0) for k in sounds)
|
||||
print('\ncue families and their sound-id ranges:')
|
||||
rng = collections.defaultdict(list)
|
||||
for k, v in sounds.items():
|
||||
if v.lstrip('-').isdigit():
|
||||
rng[re.match(r'[A-Za-z]+', k).group(0)].append(int(v))
|
||||
for p, _ in fam.most_common():
|
||||
v = rng.get(p, [])
|
||||
if v:
|
||||
print(' %-8s n=%-5d %6d .. %6d' % (p, len(v), min(v), max(v)))
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
main()
|
||||
Reference in New Issue
Block a user