Difference between revisions of "LexiBot"

From CCRMA Wiki
Jump to: navigation, search
Line 32: Line 32:
 
=== Lyrics ===
 
=== Lyrics ===
  
* Program chooses everything (do nothing)
+
* Program chooses everything to do with lyrics (do nothing)
  
 
* Subject area by keyword
 
* Subject area by keyword
Line 45: Line 45:
 
** Input a sentence given the vocabulary, which will then be mimicked
 
** Input a sentence given the vocabulary, which will then be mimicked
 
** Choose a sentence type
 
** Choose a sentence type
 +
 +
* Overall narrative structure
 +
** Telling a story
 +
** Verse/chorus structure
 +
** Other types of lyrics (repetitive, nonsensical)
  
 
* Write own set of grammar rules with new vocabulary
 
* Write own set of grammar rules with new vocabulary
 +
 +
==== Issues ====
 +
* Generating syllable recordings for new words
 +
* Ensuring that program's word-selection criteria match those intended by user
 +
* RHYMING
  
 
=== Music (Genre and Instrumentation) ===
 
=== Music (Genre and Instrumentation) ===
 +
* Program uses all music
 +
 +
* Choose music genre and instrumentation
 +
** Choose genre (ex. techno)
 +
** Choose band (profile) whose style you'd like to emulate (ex. Daft Punk, Elliott Smith)
 +
 +
* Choose instrumentation, determine own harmonies/accompinament
 +
** Preset harmonic progression
 +
** "Play" accompinament on ascii keyboard, deciding chord progression as you go
 +
 +
* Create own instrumentations using ChucK, do own keyboard mappings
 
=== Music (Melody) ===
 
=== Music (Melody) ===
 +
* Predetermined melody
 +
 +
* Preset melodic types (dependent on harmonies?)
 +
 +
* Creating one's own melodies
 +
** Preset the melody
 +
** "Play" the melody on the ascii keyboard
 +
 +
==== Issues ====
 +
* Matching a played melody to the number of (randomly generated) syllables
 +
* Matching a generated melody to syllable count
 +
* Matching the melody to the harmonic structure
 +
 
=== The Synthesized Voice ===
 
=== The Synthesized Voice ===
 +
* Currently using doctalk.
 +
** http://ccrma.stanford.edu/~jeffcoop/sound/doctalk.zip
 +
** Can set voice, frequency, inflection
 +
*** 220a proj used robotoid at 110Hz, flat intonation, for example
 +
 +
=== Pitch Shifting ===
 +
* Currently changing pitch by changing sample rate by ratio of desired pitch/given pitch
 +
** Only works within small range

Revision as of 15:34, 15 January 2008

LexiBot Stuff

Blair and Ge

Links/References

Modules (Editable or Selectable Features)

Lyrics

  • Program chooses everything to do with lyrics (do nothing)
  • Subject area by keyword
    • Keyword entry
      • Use something like WordNet to determine associated words
    • Keyword selection (from available keyword areas)
    • General theme selection
  • Sentence/String structure
    • Input a sentence given the vocabulary, which will then be mimicked
    • Choose a sentence type
  • Overall narrative structure
    • Telling a story
    • Verse/chorus structure
    • Other types of lyrics (repetitive, nonsensical)
  • Write own set of grammar rules with new vocabulary

Issues

  • Generating syllable recordings for new words
  • Ensuring that program's word-selection criteria match those intended by user
  • RHYMING

Music (Genre and Instrumentation)

  • Program uses all music
  • Choose music genre and instrumentation
    • Choose genre (ex. techno)
    • Choose band (profile) whose style you'd like to emulate (ex. Daft Punk, Elliott Smith)
  • Choose instrumentation, determine own harmonies/accompinament
    • Preset harmonic progression
    • "Play" accompinament on ascii keyboard, deciding chord progression as you go
  • Create own instrumentations using ChucK, do own keyboard mappings

Music (Melody)

  • Predetermined melody
  • Preset melodic types (dependent on harmonies?)
  • Creating one's own melodies
    • Preset the melody
    • "Play" the melody on the ascii keyboard

Issues

  • Matching a played melody to the number of (randomly generated) syllables
  • Matching a generated melody to syllable count
  • Matching the melody to the harmonic structure

The Synthesized Voice

Pitch Shifting

  • Currently changing pitch by changing sample rate by ratio of desired pitch/given pitch
    • Only works within small range