I timed a pile of Beluga-style uploads with a stopwatch before building Dingcut. Not the text, the gaps. What I found is that the good ones follow rules the creators probably never wrote down, and the bad ones break those rules in the same handful of ways.
This post is those rules.
The base rhythm
On the channels that consistently get views, a normal message runs about 1.4 seconds from the previous one. That splits roughly as 0.3 seconds of empty screen, then the popup with the ding, then 1.1 seconds of hold.
That 0.3 second gap is not nothing. Cut it to zero and the video feels like a slideshow. Stretch it past half a second on every line and it feels slow, even if each line is funny.
This is the Beluga preset in Dingcut. Slow burn is 0.5 and 1.6. Machine gun is 0.1 and 0.5.
The ding is on the beat, not on the message
The Discord notification ding plays as the message appears. Not slightly before, not slightly after. Viewers have heard that sound tens of thousands of times and they will notice a 100 millisecond drift even if they cannot say what is wrong.
If you are doing this by hand in a timeline editor, this is where the forty minutes go. Line up sound to card, nudge, play, nudge again.
The hold is where the comedy is
The single strongest pattern: the message before the punchline holds longer than everything else.
Normal line, normal line, normal line, then a question that sits on screen for three seconds. Then the answer, one word, with a boom.
The long hold works because the viewer's brain fills the silence. They start predicting the answer. When it arrives short and flat, the prediction collapses and that is the laugh.
Beginners do the opposite. They rush to the punchline because they are nervous the viewer will scroll. The viewer will not scroll during a well-set-up pause. They scroll during a rushed one.
One boom
The vine boom is the punctuation for the punchline. It goes on one line per video, maybe two if the video has a second act.
Channels that put a boom on every third message train the viewer to ignore it. By the time the real punchline arrives the boom means nothing.
Same logic for the zoom punch. Save it.
Reactions land late
When a message gets emoji reactions in a real Discord server, they arrive a second or two after the message. Beluga videos copy this. The skulls pile onto "rip channel" after the boom, not with it.
Dingcut does this automatically. Reactions are on a slight delay after the card appears, and they stagger in rather than all at once.
Machine gun sections
Fast sections work when they are a contrast, not a default. Three or four messages at 0.6 seconds each after a run of normal-pace lines feels like someone losing their mind in the chat. A whole video at that pace is unwatchable.
The best use is an escalation: "now faster" followed by "ok / ok / ok / STOP" with an air horn on the stop.
Leaving the server
"X left the server" with the leave sound is the second most common ending after a boom. It works because it is the format's version of walking out of the room. Use it after a punchline, not as one.
Where the numbers came from
The presets in Dingcut are the typical values from those uploads. If your channel has a house style, set your own defaults in Custom and every new project starts there. The point of presets is that you should spend your time on the two lines that matter and let the other ten sit on the grid.