The Drop · August 6, 2026
Vertical music visualizer for TikTok, Reels and Shorts
The app renders one 1920 by 1080 master. Short form feeds are the other way round. Most people find that out after the render, when the thing they built turns into a strip of itself and the words they wanted read are hanging off both sides of the phone.
Photo via Unsplash
The fix is not a different render. It is composing the landscape frame knowing a vertical slice of it is the part anyone will see, so one render feeds a YouTube upload and a Shorts, TikTok and Reels cut. This post is the framing half. The crop mechanics live in the export guide.
Everything here comes from one number
A 9:16 slice out of a frame 1080 pixels tall is 1080 times 9 divided by 16, which is 607.5, rounded to the 608 that the crop command uses. Out of 1920 pixels of width you keep 608. Under a third.
The surviving band runs from x 656 to x 1264. Dead center, narrow, unforgiving. Anything outside it is not compromised or softened, it is gone. So the question worth asking about every setting is whether the thing it draws lands inside those 608 pixels.
Short lines, because the app is not looking at the crop
Text is drawn from the center out, which sounds like it solves the problem, and it does not. The only width rule is that a line gets scaled down once it measures wider than 90 percent of the full frame, so it can run to roughly 1728 pixels before anything shrinks it. That is nearly three times the width you are keeping. A four word line fits the render perfectly and loses its first and last word in the cut.
One or two words per line, three if they are short. Two details matter at this width. The font changes on every word, so a long line is a font pile-up as well as a cropping problem. And each line gets a small random offset when it is drawn, up to 60 pixels sideways, which is nothing across 1920 and a fifth of your half-width at 608. Leave yourself that margin rather than testing it.
The image trick that lands exactly in the band
Each uploaded image has a fit mode. Fill frame zooms until the image covers the whole 16:9 canvas and crops the overhang. Fit whole image scales it to fit inside the canvas and lets the empty space go black.
Now upload a 9:16 image and choose fit whole image. It is taller than it is wide, so the app matches its height to the frame, 1080 pixels, and the width follows the ratio: 607.5, centered. That is the same 607.5 pixels the vertical crop keeps, in the same place. A vertical image dropped into a landscape render reappears in the vertical cut filling the screen edge to edge, with nothing lost.
So for a visualizer for TikTok specifically, prepare the artwork vertical. Square cover art works nearly as well: at fit whole image it draws 1080 wide and the crop keeps the middle 56 percent, which is fine for a centered logo or face. Wide photos are the ones that suffer. If you have to use one, set it to fill frame and check the subject is genuinely in the middle.
One wrinkle: an image swells by about 12 percent on a flash, so one sized exactly to the band pushes past it on every hit. That reads as a punch rather than a mistake, but keep important detail off the outer edge of the artwork.
Effects that survive the crop, and one that thins out
- Zoom punch is the best of them here. It scales the frame up to 1.45 times on a hit, from the center, driving content into the band rather than out of it.
- Shake throws the frame up to 70 pixels in any direction. Across 1920 that is a twitch. Across 608 it is a real move, and anything near the edge of the band walks out of shot for the length of the flash.
- RGB split and glitch slices work on offsets of a few dozen pixels around whatever is already on screen, so they survive the crop unchanged.
- Shapes are drawn in centered coordinates, so checkerboards, rings, stripes and dot grids all still read. They are the safest vertical backdrop because the pattern continues past the crop instead of looking cut.
- Flyers are the one that thins out. They spawn just off the left or right edge of the 16:9 frame and travel across it, so a vertical window only catches each one mid-trip. Expect roughly a third of the traffic the preview shows.
Build for the first second, and do not start at zero
Short form gives you no runway, and the opening of a render is the worst part of it to put in front of someone. That is not a guess, it is how the detection works. Beat detection compares current energy against a rolling average of the last 43 frames, about 0.7 seconds, so the average has not settled at the start. The bass drop detector is blocked outright for the first 1500 milliseconds, deliberately, so a loud intro cannot fake a drop. For a second and a half you are watching the least reactive version of your own visual.
So do not cut from the top of the render. Let the track run a few seconds ahead of the section you want, then take your slice from a moment that already has an event in it: the drop, the beat switch, the point where the hook lands. Fifteen to thirty seconds is the useful length, and it should open on the loudest thing you have rather than build toward it.
If the clip exists to sell something rather than to fill a feed, the type beat post covers the title and description once it is cut.
What the feeds actually ask for
YouTube's own help page puts Shorts at "up to 3 minutes long" with "a maximum resolution of 1080p", so a cropped 1080 by 1920 clip is exactly at spec.
TikTok's in-feed ad specification lists vertical 9:16 as the recommended ratio, and treats the safe zone as a rule rather than a number: it changes with the caption length and whatever else is on the ad, and elements outside it "might be covered or cropped". It ships as downloadable per-format templates, so there is no single margin to design against. What you can rely on is that the interface stacks down the right edge and along the bottom, and the center band keeps you clear of both anyway.
Then export it
The render comes out as a WebM and the vertical version is one crop away. The export post has the ffmpeg line and the per-platform notes. Compose it right and the crop is a formality.
Before you post a flashing clip
Short form autoplays, which means nobody chose to watch your strobe. TikTok says it warns creators when they add effects that could trigger photosensitive epilepsy, and it gives viewers a prompt to "Skip All" similar videos. Treat that as the baseline, not the ceiling: put a flash warning in the caption and keep the settings inside the limits in the strobe video post.
None of this needs a different tool. Open the visualizer, upload a vertical image, keep the lines to a word or two, and picture the middle 608 pixels while you set everything else.
Render it once, post it everywhere
Build a beat-synced 1080p clip in your browser, composed so the vertical cut keeps the good part. Free, no watermark, no signup.
Open the visualizer