How to Make an SRT File for a Song

Subtitling a song is not the same job as subtitling speech. Vocals stretch across notes, choruses repeat with slightly different phrasing, and long instrumental sections leave gaps that automatic tools like to fill with nothing at all. This guide covers how to make an SRT file for a song, what the format can and cannot do with music, and when a different format is the better choice.

What an SRT file actually contains

An SRT file is plain text. Every entry has three parts: a sequence number, a timestamp range, and the text to display.

1
00:00:12,340 --> 00:00:16,180
I heard there was a secret chord

2
00:00:16,420 --> 00:00:20,900
That David played and it pleased the Lord

Timestamps use hours:minutes:seconds,milliseconds, with a comma before the milliseconds rather than a period. That comma trips up a lot of hand written files. A blank line separates each entry, and the sequence numbers run in order from one.

Because it is plain text, you can open an SRT in any editor and fix a typo without special software. That simplicity is exactly why the format is supported nearly everywhere, and also why it cannot do some of the things music needs.

Why songs are harder to subtitle than speech

Most subtitle guides assume you are working with a podcast or an interview. Music breaks several of those assumptions at once, and knowing where it breaks tells you where to check the output.

Sustained notes stretch one line across many seconds. A vocalist holding a word for six seconds gives automatic timing very little to anchor to. The line often ends early, so the subtitle vanishes while the singer is still on the word.

Repeated choruses drift. The same words return three or four times, and automatic alignment sometimes matches a chorus to the wrong occurrence. The result reads fine in isolation but sits several seconds out of place.

Instrumental gaps confuse pacing. An eight bar solo between verses is silence as far as the lyrics are concerned. Tools that expect continuous speech can stretch the previous line across the gap rather than ending it.

Layered vocals compete. Backing vocals, harmonies, and ad libs overlap the lead. Transcription may capture a background phrase and place it in the middle of the lead line. If your song has distinct lead and backing parts, separating them before you start makes the timing noticeably cleaner.

None of this makes automatic timing useless. It just means the review pass should be targeted: check the long notes, the repeated sections, and the instrumental breaks, and leave the straightforward verses alone.

How to make an SRT file for a song step by step

Step 1: Add the song

Open the SRT File Generator and point it at your track. There are three ways in, and they suit different starting points.

Upload a file from your computer if you already have the audio. MP3, WAV, MP4, MOV, WebM, FLAC, and M4A all work, and if you upload a video the audio is extracted automatically, so there is no need to convert anything first.

Import from Google Drive or Dropbox if the track lives in cloud storage, which saves downloading it first only to upload it again.

Paste a YouTube link if the song is already on YouTube. The audio is pulled from the video for you, so this is the fastest route when you are subtitling something you did not record yourself, such as a cover you are studying or a track you are preparing for a karaoke night.

Step 2: Transcribe or paste the lyrics

You have two routes here, and they suit different situations.

Let AI transcribe the vocals when you do not have the lyrics to hand, or when the track is your own unreleased song that no lyrics site will have. Paste the lyrics yourself when the wording matters and you already have it, since printed lyrics get spelling, names, and stylised phrasing right where transcription can mishear a heavily produced vocal.

Step 3: Check the timing at the hard spots

Play the track back against the subtitles and look specifically at the four trouble spots from the previous section. Sustained notes and instrumental gaps are where drift shows up first, and they are quick to nudge once you know to look.

Every line sits on the timeline as its own block, and you fix the timing by dragging rather than by typing timestamps. Drag the middle of a block to slide the whole line earlier or later while its duration stays the same, which is what you want when a line is correct but simply lands late. Drag the left edge to change only when the line appears, and the right edge to change only when it disappears, which is the fix for a sustained note that vanishes while the singer is still holding it.

Neighbouring lines act as limits, so a block will not slide past the one before or after it and subtitles cannot end up overlapping. That means you can drag quickly through a rough patch without breaking the sequence, and fixing one drifted section on its own is usually faster than retiming the whole song.

Step 4: Download the .srt file, or burn it into the video

You have two ways to finish, and which one you want depends on where the result is going.

Download the .srt on its own when the subtitles need to stay as a separate file. This is what you want for YouTube, where the caption file is uploaded alongside the video and stays editable and machine readable, and for a video editor where you want to restyle the captions later without touching the timing.

Export the video with the subtitles burned in when the words need to survive anywhere the video travels. Burned in text is part of the picture, so it plays the same on every platform and cannot be switched off. That matters most on social feeds, where a separate caption file is often ignored and much of the audience watches muted.

The two are not mutually exclusive, and for a YouTube upload the strongest combination is usually both: burn the lyrics in so the video reads well on its own, and upload the .srt as well so YouTube still has machine readable text to index and translate.

When SRT is the wrong format for music

This is the part most SRT tutorials skip, because most of them are not written with music in mind.

SRT stores line-level timing only. A line appears at its start timestamp and disappears at its end timestamp, as one block. There is no way to record when each individual word arrives, which means SRT cannot produce word by word karaoke highlighting. If you have seen a lyric fill in colour syllable by syllable and wondered why your SRT will not do it, this is why. The format has nowhere to put that information.

For word level timing you need a format built for it. Stay with SRT when the file is going to YouTube or a video editor, since nothing beats it for universal support and it gets indexed as text. Switch to LRC when the file is going to a music player or a karaoke app, because those expect a lyrics file and LRC can carry enhanced word timing. Reach for ASS when you want styled karaoke that fills in word by word, since it stores per syllable timing alongside colour and position.

If you are deciding between these, word level vs line level timing explains the difference in depth.

A practical rule: SRT is for video, LRC is for music players, ASS is for styled karaoke. If your file is heading into a video timeline or a captions uploader, SRT is right. If it is heading into a karaoke experience, it is not, and you want the Karaoke Editor instead.

Frequently asked questions

Can an SRT file highlight lyrics word by word?

No. SRT only stores line-level timing, so a line appears and disappears as a block. For word by word highlighting you need LRC with enhanced timing, or ASS.

Should I use SRT or LRC for a song?

Use SRT when the file is going to a video platform or a video editor, such as YouTube captions or a Premiere Pro timeline. Use LRC when the file is going to a music player or a karaoke app that expects a lyrics file.

Can I make an SRT file from a YouTube video?

Yes. Paste the YouTube link into the SRT File Generator and the audio is pulled from the video automatically, with no need to download it yourself first. From there the process is the same as an uploaded file.

Does YouTube accept SRT files for music videos?

Yes. YouTube Studio accepts .srt directly under the Subtitles tab. Uploaded captions are indexed as text, unlike lyrics that are burned into the video image.

Do I need the lyrics before I start?

No. AI can transcribe the vocals directly from the audio. Pasting the official lyrics still helps, because printed lyrics get the exact wording right where transcription can mishear a stylised or heavily produced vocal.

Conclusion

Making an SRT file for a song is quick once the timing is handled, and the format's universal support is hard to beat for anything heading into a video platform or an editing timeline. The one thing to keep in mind is its limit: line-level timing only, which rules out word by word karaoke effects no matter which tool you use.

Start with the SRT File Generator if you need captions for video. If what you actually want is lyrics filling in word by word as the song plays, head to the Karaoke Editor instead.