Fix "sprecher" data & Migrate array-based "sprecher" to object-based "sprechrollen"

[YourMJK]

Fix "sprecher" data

Actors (~360 changes)

Using data from hoerspielforscher.de, Wikipedia and Google search:

• fixed typos in names
• fixed wrong attributions
• updated names to current preference
• deobfuscated pseudonyms by adding real name in brackets
• one canonical name per person

Roles (~480 changes)

• fixed typos and punctuation in names
• updated names to most recent/common usage
• one canonical name per role

Migrate array-based "sprecher" to object-based "sprechrollen"

Changed JSON structure by replacing all previous "sprecher", e.g.:

"sprecher" : [
  [
    "Role name 1",
    "Actor name 1"
  ],
  [
    "Role name 2",
    "Pseudonym [Actor name 2]"
  ],
  ...
]

with new "sprechrollen", e.g.:

"sprechrollen" : [
  {
    "rolle" : "Role name 1",
    "sprecher" : "Actor name 1"
  },
  {
    "rolle" : "Role name 2",
    "sprecher" : "Actor name 2",
    "pseudonym" : "Pseudonym"
  },
  ...
]