🧱Song Structure
These are controls to specify the "tone" of the song structurally and indicate where lyrics start and finish.
The controls in this section are used to control structural elements. There's two main controls here: Clip Start and Lyric Timing:
Clip Start
The Clip Start option essentially controls where the AI "sees" your clip as beginning contextually.
It goes without saying songs tend to have a structure, and that structure follows conventions: introductions tend to set the tone and build up energy/tension, mid sections tend to be where the "exciting/impactful" bits of a song are, penultimate sections tend to offer breaks from the previous and outros conclude the song appropriately. All of this is accomplished with things like tempo, vocal delivery and instrumental choices.
This slider allows you to "tell" Udio where overall in terms of song context; 0 meaning "the beginning" and 100 meaning "the end", your clip "is" within a song. This will cause the AI to treat the contents of the clip differently. You could have two prompts prepared in exactly the same manner, same settings and lyrics, but with one selected as "Clip Start 0%" and another "Clip Start 40%" they will sound different to each other. The first will involve building up and tone setting and the second will probably be more dynamic and impactful.
You don't have to obey organisation strictly; you can use this slider to manipulate your song in precise manners. For example halfway through your song, if you want to return to a much gentler and earlier tone in the song, setting the clip start to 10-20% will help that, even if the clip itself is roughly 60% into the song.
See below for some general indicators:
0% - Start of a song, sets tone
10-30% - Early song, builds themes, building escalation towards end of clip
35-50% - Mid section, dynamic, impactful, catchy and "big" parts.
50-70% - Post-mid, usually returns to themes earlier in song but with more energy and build in them, or provides a difference.
70-85% - Setup for outro, provides breaks and tone setting. Instrumental solos tend to work well in this area.
86%+ - Concluding song.
The above is not intended as a manual telling you how you should build your song! Rather it's a guide for if you place the slider in the above areas; the AI will "build" the prompt you give it in a direction according to the above general rules. So if you want a catchy, powerful section anywhere in your song, 35-50% clip start could be an option to help you get this.
Lyric Timing
This slider allows you control over how the lyrics within a section will stop and start in terms of being vocalised. When activated the slider has two "ends":
The first slider "end" indicates when the lyrics will start. The white box identifies this in seconds into the clip (and this is the case for both 32 second and 130 second clips). Slide it along to decide when your lyrics will "start" to be sung. The opposite can be said for the other slider, use this to control when the lyrics "end".
Note: This slider is best thought of as a strong recommendation to the AI, not a direct command. It will sometimes ignore your recommendation because perhaps with the instrumental pattern the clip generated, it does not "work" to have your vocalist start to sing when you specified. You may have to regenerate a few times to get an idea of whether your specification is doable, or whether it's clearly not (where the AI consistently ignores your recommendation).
If you have minimal lyrics in a section and use these sliders, you can use them to have fewer lyrics delivered over a longer period. This can result in some very nice vocalisations and delivery, particularly if you indicate such in your lyrical prompting. See the Knowledgebase on this for details.
Last updated