Three Generation Modes in Studio#
lyrics_type=user
Uses user-provided lyrics (specified in the lyrics
field) to generate music. In this mode, the prompt
acts as tags for defining the music style. This corresponds to Suno's custom mode.
lyrics_type=instrumental
Creates instrumental music, where the prompt
is used for relevant keywords or styles.
lyrics_type=generate
Automatically generates lyrics. The prompt
field describes the type of song you want. This corresponds to Suno's descriptive mode.
Key Fields#
Field | Type | Description |
---|
prompt | string | Keywords or descriptions for the song or style,is required |
lyrics | string | Lyrics to be used in the music, when lyrics_type is user ,is required |
lyrics_type | string | Lyrics type: user (custom), generate (descriptive), instrumental (pure instrumental), is required |
model_type | string | Model type. The studio130-v1.5 model consumes twice the credits of studio32-v1.5 . studio130-v1.5 will generate 130 seconds length of music, and studio32-v1.5 will generate 32 seconds length of music, is required |
negative_prompt | string | Negative keywords to avoid in the output, is optional |
config.mode | string | Extension mode: precede (extend the start of the song), continuation (extend the end), or regular (generate complete music), is required |
config.source.song_id | string | The ID of the song to extend, when use extend mode, is required |
Additional Parameters#
song_section_start/end
#
Defines the start and end positions for generating a section of the song:0%
(0) corresponds to the beginning of the song.
50%
(0.5) corresponds to the middle of the song.
100%
(1) corresponds to the end of the song.
This parameter is primarily used in song extension modes.lyrics_placement_start/end
#
Specifies where to place the lyrics in the song:8%
places lyrics at the start of the song.
50%
places lyrics in the middle.
90%
places lyrics at the end of the song.
This parameter is particularly useful in song extension modes.prompt_strength
#
Controls how strongly the prompt
influences the generated content:Higher values increase adherence to the prompt
but may make the audio sound less natural.
lyrics_strength
#
Controls how strongly the lyrics influence the generated content:Lower values may produce more natural audio but might neglect parts of the lyrics.
clarity_strength
#
Adjusts the priority between audio clarity and instrument separation:Higher values produce clearer audio but may reduce the naturalness of the sound.
generation_quality
#
An experimental feature that balances generation quality and speed.context_length
#
The context length determines how much of the track the model considers when generating output. Longer context improves coherence but limits style changes, while shorter context is more flexible but less coherent.
Other Field Descriptions#
Field | Type | Description |
---|
song_section_start | number | The start position for extension (0-1 decimal), is optional |
song_section_end | number | The end position for extension (0-1 decimal), is optional |
lyrics_placement_start | number | The start position for lyrics placement (0-1 decimal), is optional |
lyrics_placement_end | number | The end position for lyrics placement (0-1 decimal), is optional |
prompt_strength | number | Strength of the prompt (0-1 decimal), default is 0.5, is required |
clarity_strength | number | Clarity strength (0-1 decimal), default is 0.25, is required |
lyrics_strength | number | Lyrics strength (0-1 decimal), default is 0.5, is required |
generation_quality | number | Options: 0.25 , 0.5 , 0.75 , 1 . Default is 0.75. Higher values improve quality but reduce speed, is required |
seed | number | Random seed. Default -1 means random generation, is optional |
bypass_prompt_optimization | boolean | Whether to skip prompt optimization. Default is false , is optional |
Modified at 2024-12-31 09:49:52