Can you make a lyric video before you have the lyrics?

Pause any lyric video at a random second

What is on screen is one or two lines of text, held for exactly as long as it takes to sing them, sitting on top of something that barely moves. That is the whole form. The picture does maybe a fifth of the work. The other four fifths are four decisions repeated a few dozen times: when a line arrives, where it breaks, how long it stays, and how often the screen changes.

Which makes the question in the title uncomfortable rather than clever. Yes, you can start before the lyrics exist. You can choose the background, the typeface, the color, the aspect ratio, the loop that runs under everything. No, you cannot finish, and you cannot lay out a single frame that will survive, because every layout decision is an answer to a specific line of a specific length landing at a specific moment. Build the frames first and you rebuild all of them the day the words arrive.

What this site does, before you invest any more reading

We write lyrics. There is one free field for your theme, capped at 300 characters, plus closed lists for the rest: point of view, song language, audience age, structure, hook direction, vocal, musical style, and one to three emotions. Out comes a title and a finished lyric. It costs 20 credits, which is $0.20 at the site rate of 100 credits per dollar. The numbers live on the pricing page and the tool itself is on the generators page.

We do not produce audio. No vocal, no instrumental, no mp3, not a single second of sound. We do not render video either: no timeline, no karaoke sync, no export button. If you searched for a lyrics video generator hoping one link would turn an idea into a finished file, we are one link of a three link chain, and a cheap one. Suno belongs to a different company that we have no relationship with, do not resell, and do not speak for. Sing the words there or anywhere else, that part is yours.

And if the lyrics are already written, sitting in a note on your phone, then you do not need us at all. Skip to the sections below on breaking lines and timing them, because that is where your actual problem is.

The four decisions, only one of which is visual

Sync. The moment each line appears and disappears against the recording. Not "roughly the right bar". A specific frame.

Line breaking. Where a long line splits into two cards, or two screen rows, or none.

Readability. Contrast, weight, size, and how much text you allow in the frame at once. This is the visual one, and it is the one everybody solves first because it is the easy one.

Rhythm of appearance. How often the screen changes. Not how fast the song is: how often the eye is asked to do work.

Get the fourth one wrong and the video feels broken even when every line is technically on time. That is the part nobody warns you about.

Timing is a reading instruction, not a technical detail

Here is the thing that separates a lyric video people sing along to from one people scroll past: a line that appears at the same instant it is sung is already late.

Reading is not instant. The viewer has to see the line, parse it, and get to the end of it. If the text lands on the first syllable, the eye chases the voice for the whole line and never catches up, and the viewer settles into being a spectator. Push the same line about half a second earlier and something else happens: the viewer reads ahead, arrives at the last word just before the singer does, and says it with them. That half second is the difference between watching a video and being inside a song.

So the sync question is not "when is this line sung". It is "when does this line need to be readable so the viewer reaches the end of it on the beat". Those are different questions with different answers, and the gap between them grows with line length. A three word line barely needs a head start. A fourteen word line needs one badly, or it needs to stop being a fourteen word line.

The reverse error is just as visible. A line that stays after the next one has started singing reads as a stuck frame. Cheap sync tools tend to hold each line until the next cue, which means a phrase sung two seconds before a long instrumental sits there for eight. Nothing is technically wrong and it still looks wrong. Fade it out and leave the frame empty, because empty is a legitimate state here. A lyric video with no text during an instrumental is doing exactly its job.

The line length that breaks the video, in both directions

Take a real line:

I keep the porch light on for somebody who is never coming back

Sixty three characters. On a phone, at a size anyone can read while half watching, that is two full rows with nothing to spare, and three the moment you pick a size that reads across a room. Three rows held for four seconds is not a lyric video, it is a paragraph with music behind it, because the eye does not scan a block the way it scans a line.

Now the other failure. Suppose you write in short punchy fragments because someone told you short is strong:

Gone. All of it. Gone again.

Time that honestly against the recording and the screen changes roughly every six tenths of a second. Three cuts where the ear hears one phrase. It flickers, and on a small screen with any motion behind it the flicker reads as a glitch: viewers do not think "interesting pacing", they think the file is broken.

The working range sits between those. Pick a character ceiling before you write anything and hold to it: around forty characters per screen row is a sane starting point, with a hard limit of two rows. Then measure your actual lines against it instead of trusting your eye, because a line always looks shorter in a text editor than it does at 1080 by 1920 with a safe margin.

The line above splits cleanly:

I keep the porch light on for somebody who is never coming back

Twenty five characters, then thirty seven. Both fit. Both are complete units of sense. Neither needs the other one on screen to mean something.

Which brings up the one thing you can actually do in advance, at the writing stage, and it is the whole practical argument of this article. Write lines that survive being cut in half. When you draft a verse, glance at each line and ask where you would break it if you had to. If the answer is "there is no good place", that line is going to fight you later, in the editor, at eleven at night, when changing it means recutting a vocal. Fix it now, while it is still just text. Our songwriting tips and prompts go deeper on shaping lines, and song structure templates covers the section level version of the same problem.

Why automatic splitting cuts in the wrong place

Every automatic splitter works on one of two rules. Either it breaks at punctuation and pauses, or it fills a character budget and breaks at the last word that fits. Neither rule knows what a phrase is.

Run our line through a thirty five character budget and the machine gives you:

I keep the porch light on for somebody who is never coming back

Twenty nine characters, then thirty three. Perfectly legal. Also wrong, because the first card now ends on "for", a word that promises something and delivers nothing. The viewer reads a card that hangs open, waits, and collects the payoff on the next one. Do that forty times across a song and the whole video stutters, even though every line is on time and every card fits.

Punctuation splitters fail differently. Lyrics are often written with no punctuation at all, so the splitter finds nothing to grab and dumps the entire line into one card. Or the writer uses commas for breath rather than grammar, the splitter cuts on every breath, and you get the flicker problem from the previous section.

The fix is not a smarter splitter. The fix is to decide the breaks yourself, once, in the lyric document, before anything touches a timeline. Write your sheet with the breaks already in it: one screen card per line of the file. Then whatever tool you use is transcribing your decisions instead of inventing its own.

There is a second reason to do this by hand, and its name is repeats. A chorus that appears four times is rarely four identical screens. The third one picks up an ad lib, the last one drops a word or holds one longer. Copy the timings from chorus one into chorus four and the one that drifts is the final chorus, which is the emotional peak of the song. Type each repeat separately. Five minutes of boring work saves the ending.

A lyric video is not subtitles, and the difference is not size

People reach for subtitle tools because the shape looks the same: text, timed, over video. The jobs are opposite.

Subtitles serve somebody else's speech. They exist so you can follow what you cannot hear, and they succeed when you stop noticing them. Hence the bottom of the frame, the small neutral type, and the licence to be a little loose. A subtitle two tenths of a second late is invisible, a subtitle that hangs through a pause bothers nobody, and no one in the room checks whether the caption left on the exact frame.

A lyric video inverts all of that, because the text is the image. There is nothing else to look at. A line two tenths late against a sung syllable is visible to everyone, including the people who could not tell you why the video felt cheap. The text carries the composition, so its position, weight and timing are directing decisions rather than accessibility settings.

What follows from that in practice:

• Subtitle habits pull you toward small text at the bottom. Lyric text usually wants the middle third and a size you can read across a room.
• Subtitle tools are tuned to a comfortable reading speed for prose. Song lines are not prose, and their pace is set by the melody, not by the eye.
• Subtitle tools like to hold the previous caption until the next one starts. Lyric videos need empty frames.
• A subtitle file is a fine delivery format. It is a bad place to make decisions.

If your editor only imports subtitle files, use one, but write the breaks yourself first and treat the file as packaging rather than as a design tool.

You have music but no words. Or words but no music

Two different dead ends, two different orders of work.

A track and no lyrics. The melody already exists, so the syllable count per line is already decided, and that is a hard constraint you did not choose. Count the syllables of the melody line by line before you write a word, then write to that count. If our generator is the tool you use, put the constraint into the theme field: it holds 300 characters, so spend them on what actually decides the outcome. Something like "verse lines of 8 syllables, chorus lines of 6, one word hook at the end of each chorus" costs about 85 characters and buys a lyric that fits the melody you already have. Pick the structure, style and emotions from the closed lists after that. The whole thing is twenty cents on the generators page, which is cheap enough to run it twice with different emotions and keep the better one.

A lyric and no track. Easier problem, more dangerous temptation. The temptation is to start building the video now, because the text was the part you were waiting for. Do not. Timing is measured against a recording, the recording does not exist, so anything you time now gets retimed later. What is useful without audio: finalize the line breaks, count the characters per card, mark which cards repeat, and decide where the empty frames go. That work survives.

Neither one. Words first, always. The lyric is the cheapest thing to change, and it shapes everything downstream. A rewrite at the lyric stage costs a retype. The same rewrite after a vocal exists costs a re-record. After the video is cut it costs the video. The middle step is laid out in turning lyrics into a song, and the first one in detail in the song lyrics generator guide.

Where this advice stops working

Rap at speed. Dense verses break the whole model. A bar packed into a second and a half cannot be read in a second and a half, so line by line display turns into a blur. Fast rap videos usually run word by word or in short karaoke style chunks, or they hold a full bar and accept that the viewer is reading rather than singing along. Both are legitimate. What does not work is pretending pop timing rules apply.

Screamed or heavily distorted vocals. Here the video does the vocal's job, because nobody is decoding the words by ear anyway. The timing can be looser and the type can be bigger, since comprehension has moved entirely to the screen.

Instrumental passages longer than a couple of seconds. Empty frame. Resist the urge to fill it with the song title or a handle, unless you want the viewer's eye trained to ignore text.

Languages with different line widths. A line that fits forty characters in English may not fit at all in a language that spells the same meaning longer. Set the ceiling per language, in the real font, at the real size, instead of inheriting a number from an English tutorial. If you are working across two languages, the closest thing we have to a map is writing songs in another language.

Ad libs and backing vocals. Decide early whether they appear on screen at all. Putting every "oh" and "yeah" up there doubles the card count and halves the readability. Most videos are better without them, and the ones that include them usually set them smaller and off center so the main line stays the main line.

Quick answers

Can an AI lyrics video generator do all of this for me? Tools in that category handle the mechanical parts: they take audio plus a text file, guess the alignment, and render frames. The guess is decent on clear pop vocals and poor on dense, layered or distorted ones. The line breaks are still yours, the empty frames are still yours, and the half second of anticipation is still yours. Those are the parts that make it watchable.

Do I need the final master before timing? No, but you need the final vocal take. Timing against a rough mix is fine. Timing against a scratch vocal that will be re-sung is wasted work.

How long does the timing pass actually take? Longer than you expect on the first song and much faster after that. What cuts the time is the lyric sheet with the breaks already decided, not the tool.

What if the lyric will not fit the frame no matter how I break it? Then the lyric is too wordy for the form, and you are looking at a rewrite rather than a layout problem. Cut the line at the writing stage. It is the one place in this chain where changing your mind is free.