Is there any trick to scripting audio-based videos where you can't rely on the waveform?

[Messsage]

I’m a fan of audio content, but I have no confidence in actually scripting them. Fap Heros and JOIs are easy enough since you can depend on either a beatbar or a metronome, but for stuff like ASMRs and the like, I don’t really have much to work with. Where what’s being said actually matter. Is there any kind of trick or tools that make scripting them more doable, or is it really just a matter of scripting each part by gut and just hoping it lines up?

[Dock]

Literally the only things I could think of is either manually slowing down the video beforehand (OFS mutes the video if you slow it down too much) and going in it with the recording-scripting method, or maybe using another program to create a visualizer for the asmr video.

Honestly if it’s a straight up lewd asmr audio (like you hear them jacking you off), it may be good to simply rely on people’s innate tendency to do things rythmically, just listen to segments and break them up into the different speeds they make sounds at; you’ll notice pretty quickly that 99% of the time you can make a pattern for x amount of seconds/minutes since irregular, non-rythmic movements are weirdly rare.

Also since ASMR is audio only you have an easier time to get away with making shit up on the spot imo.

[moonlightecli18]

Maybe another waveform in OFS is needed for volume + maybe some filtering.

[Falafel]

If there is a audible consistent beat, Funscript generator would help. It essentially detects audio peaks within a frequenct range and plot a point on each.
https://milovana.com/forum/viewtopic.php?t=24040&hilit=funscript+generator

If said ASMR has an obvious thrusting noise that separates itself from the voice, the tool may also be able to detect those if you play with the built-in EQ.

You can also do it manually by listening and tapping, aka scripting on the fly. If the pace is consistent over a section, equalize these actions afterwards.

---

This could be a stretch, but a spectrogram would be much more helpful than a waveform when scripting audio works. One may possibly process the audio in a way that amplifies these noises to help render them on the waveform, or detected by the software.