How to Generate Subtitles for a Video

Direct answer

To generate subtitles, start with the final audio or video and open FreeAudioTrim Audio & Video Transcription. Choose the spoken language and an available Whisper model, then review the timed transcript while listening to the source. Correct the text before downloading SRT for YouTube and most editors or VTT for web players. Supported media stays on your device during transcription.

When to use this workflow

This workflow is best when you need subtitles that are useful beyond a quick social caption. It works for YouTube videos, online courses, interviews, podcasts turned into clips, client edits, accessibility captions, translated subtitle preparation, and videos that will move through Premiere Pro, DaVinci Resolve, or a web player.

It is also a good fit when the media is private. Browser-based processing helps avoid upload queues and unnecessary copies of client footage, interviews, lectures, or internal recordings. Browser support, device power, file size, and format compatibility can still affect performance, so long files may take longer on older machines.

How to make a subtitle file

1. Start with clean, final audio

Subtitles follow the audio. If you trim the video after generating captions, the timing can drift. Before you create subtitles, remove unwanted sections with the audio cutter, reduce long pauses with remove silence from audio, or even out speech volume with normalize audio volume.

2. Generate the transcript and subtitle draft

Turn your final audio or video into editable text, SRT, or VTT. Select the spoken language instead of asking the model to guess. The tool lists 100 language options and offers up to three model sizes depending on the device. A live preview shows provisional text while the uploaded file is processed.

3. Review the words before fixing the timing

Play the media in the transcript player. The current segment highlights as the audio moves, and clicking a segment jumps to that part of the recording. Select Edit transcript to correct names, brand terms, technical vocabulary, punctuation, and numbers. Automatic subtitles are a draft, not final proof.

Edits flow into TXT, SRT, VTT, and the translation workflow. The editor changes the words inside each segment but does not provide manual timecode editing or automatic speaker labels.

4. Choose TXT, SRT, or VTT

Choose the export based on where the subtitles will go next:

• Choose TXT for plain transcripts, notes, copy editing, summaries, translation preparation, or sending text to a reviewer. TXT does not preserve subtitle timing.
• Choose SRT for YouTube, social platforms, client review, Premiere Pro, DaVinci Resolve, and most editor workflows. It includes numbered captions, start and end times, and subtitle text.
• Choose VTT for websites, HTML5 video, learning platforms, and web players. It can also support extra cue settings in compatible players.

5. Sync and shape each subtitle cue

FreeAudioTrim builds timed cues from the model's segments and protects the full-file timeline when long recordings are processed in overlapping windows. You can hide timestamps in the readable transcript without deleting the underlying times. SRT and VTT downloads remain timed because those formats require cue timing.

Check the exported file in the final platform. Break long cues on complete phrases rather than in the middle of a name or thought. If every cue is late or early by the same amount, shift the whole file in your subtitle editor. If the offset grows over time, check whether the video changed after transcription.

6. Import, check, and deliver

For YouTube, upload the SRT in the subtitles area and preview it against the published video. In Premiere Pro or DaVinci Resolve, import the SRT, place it on the caption or subtitle track, then check the first cue, scene changes, long lines, and the final cue. Deliver a sidecar file when the platform or client needs switchable captions, or burn captions into the picture only when the delivery spec calls for visible text.

YouTube, Premiere Pro, DaVinci Resolve, and web video

YouTube subtitles: Use SRT for most uploads. Review the captions inside YouTube before publishing because platform auto-sync and line wrapping can change how viewers experience the text.

Premiere Pro captions: SRT is the simplest handoff format. Import, review the caption track against the final timeline, fix line breaks, and export captions as a sidecar file or burn them into the video if the client wants open captions.

DaVinci Resolve subtitles: SRT is also the practical choice. Import the subtitle file, check the subtitle track, adjust any drift, and confirm styling or burn-in settings before final render.

Website captions: Use VTT when you need captions for an HTML5 video player or a web-based learning platform. Keep a TXT transcript as a backup for accessibility, indexing, editorial review, and translation.

Preparing subtitles for translation

Correct the source transcript before translating. FreeAudioTrim supports 20 translation languages and sends the edited segments to a temporary Google Translate view. Only transcript lines are marked for translation, while interface labels and timestamps stay untouched.

The translation page can download TXT, SRT, or VTT. It rebuilds timed cues around the translated sentences, but you still need to check reading speed and line breaks. For English-to-Arabic or Arabic-to-English subtitles, check punctuation, names, numbers, and mixed-language phrases in the final player.

You can also open Refine with ChatGPT. The button opens a dedicated custom GPT but does not send the transcript automatically. Copy only the text you want to share.

How the translation handoff works

The available source and target choices are English, Modern Standard Arabic, Spanish, French, German, Hindi, Urdu, Turkish, Simplified Chinese, Portuguese, Russian, Japanese, Korean, Italian, Dutch, Persian, Bengali, Indonesian, Ukrainian, and Vietnamese. The source and target must be different.

When you open translation, FreeAudioTrim copies the corrected segment text and original cue times into a temporary local-storage session. A clean page opens in a new tab and loads the Google Translate website element. The session expires after 12 hours, so an old translation link cannot remain open indefinitely.

Google Translate changes only elements marked as transcript text. Cue numbers, timecodes, buttons, and page labels are protected. After translated text appears, FreeAudioTrim breaks it into sentences and rebalances those sentences across the original time span. That retiming is a useful starting point, but it cannot judge reading comfort as well as a person watching the finished video.

The active format button controls the download. TXT contains the translated text, SRT uses numbered cues and comma-based milliseconds, and VTT starts with WEBVTT and uses period-based milliseconds. The translated filenames end in _translated.txt, _translated.srt, or _translated.vtt.

What the ChatGPT button does

Refine with ChatGPT opens the FreeAudioTrim Subtitle Translator custom GPT in another tab. It can help with grammar, punctuation, readability, translation wording, and terminology when you choose to provide the text.

The button does not copy the transcript, attach the media, put transcript text in the URL, send anything automatically, or return ChatGPT edits to FreeAudioTrim. You decide what to share and move any accepted changes back yourself.

How subtitles move through a real project

Subtitles often travel through more than one tool: a transcript editor, a translator, YouTube Studio, Premiere Pro, DaVinci Resolve, or a client review link. A clean source transcript makes every handoff easier.

Privacy note: if the video includes confidential client audio, unreleased footage, or internal training material, use a no-upload workflow where supported and keep the reviewed TXT, SRT, or VTT files organized with the project.

Practical tip: finish the edit first, generate subtitles from the final audio, export SRT for editor and YouTube handoff, then prepare translation from the corrected source text.

Arabic and mixed Arabic-English subtitles

Arabic subtitles need a careful right-to-left review. Check punctuation at the end of lines, numbers, English product names, hashtags, URLs, and mixed Arabic-English phrases. Some video editors and players handle bidirectional text differently, so test the subtitle file in the final platform before delivery.

If you work with Gulf Arabic, dialect speech, or code-switching between Arabic and English, expect to do more manual review. The transcript may capture the general meaning but still miss names, local terms, or where one language switches into another.

Arabic users can also start from the Arabic transcription and subtitle tool for a workflow written around Arabic-language needs.

Common subtitle mistakes

• Generating subtitles before the edit is locked: Every cut after transcription can push captions out of sync.
• Exporting only TXT when you need timed captions: TXT is useful, but it will not upload as timed subtitles on most video platforms.
• Trusting automatic text without review: Always check names, numbers, industry terms, and unclear speech.
• Ignoring reading speed: Captions that disappear too quickly are technically present but hard to use.
• Using VTT everywhere: VTT is great for web players, but SRT is usually easier for YouTube and editor handoff.
• Forgetting accessibility context: Captions should include important spoken content and useful sound cues when they matter to understanding the video.

Limitations to know

Automatic subtitles are a draft. Noisy audio, fast speech, overlapping speakers, dialects, technical names, and mixed Arabic-English phrases can all produce errors that only show up when you watch the captions against the final video.

The tool can miss words, misspell names, repeat a phrase, or produce text during silence. It does not know which speaker is talking. Right-to-left display also depends on the destination. Test Arabic SRT or VTT files in the actual player before sending subtitles to a client or publishing a translated video.

What the tool does not do

The current tool does not capture a microphone, transcribe a meeting continuously, or provide live microphone dictation. Its live feature is a changing preview while an uploaded file is processed.

It also does not provide speaker diarization, automatic speaker names, or manual timestamp-value editing. The built-in Enhance audio option performs peak normalization. It does not remove noise, music, room echo, clipping, or overlapping voices.

Google translation is not local, and ChatGPT refinement is not a one-click round trip. Codec support is never guaranteed by the extension alone. Accuracy also varies by model, language, dialect, and recording quality, so there is no fixed WER that applies to every file.

Right-to-left text and export behavior

Arabic, Urdu, Persian, Hebrew, and other right-to-left languages are displayed with the appropriate text direction. Exports add direction marks, and Latin words, URLs, or numbers inside RTL text can receive bidirectional isolation to reduce scrambled display order.

That browser-side handling cannot control every video editor or player. Test the exported subtitle file in its final destination, especially when a cue mixes Arabic and English.

A practical FreeAudioTrim workflow

For the cleanest subtitle workflow, use FreeAudioTrim tools in this order:

1. Need audio from a video first? Use Extract Audio from Video.
2. Need to cut dead space or wrong takes? Use Audio Cutter Online.
3. Need more consistent voice volume? Use Normalize Audio Volume.
4. Need the subtitle file? Use Audio & Video Transcription Online and export TXT, SRT, or VTT.

For more detail, read the guides on how to transcribe audio to text, how to transcribe video to text, and how to extract audio from video.

FAQ

Can I generate subtitles without uploading my video?

Yes, for supported browser-based workflows. FreeAudioTrim is designed so supported media can be processed locally in your browser, which means your file stays on your device instead of being uploaded to a server.

What is the difference between subtitles and captions?

People often use the words interchangeably. In practice, subtitles usually focus on spoken dialogue, while captions may also include meaningful sound cues, speaker identification, and other audio context for accessibility.

What is the difference between TXT, SRT, and VTT?

TXT is plain text with no timing. SRT is the most common timed subtitle format for YouTube, social platforms, and editing software. VTT is a timed caption format commonly used by websites and HTML5 video players.

Which subtitle format should I use for YouTube?

Use SRT for most YouTube subtitle uploads. It is widely supported, easy to review, and works well as a handoff format between subtitle tools and publishing platforms.

Which subtitle format should I use for Premiere Pro or DaVinci Resolve?

Use SRT unless your delivery spec asks for something else. Import the SRT, check it against the final timeline, fix any line breaks or drift, then export a sidecar subtitle file or burn captions into the video.

How do I fix subtitles that are out of sync?

If all captions are late or early, shift the full subtitle file by the same offset. If timing drift changes throughout the video, check whether the edit changed after transcription, whether silence was removed, or whether the video frame rate changed. In that case, regenerate from the final audio or fix the affected sections manually.

Can I generate Arabic subtitles?

Yes, but Arabic subtitles need careful review. Check right-to-left display, punctuation, dialect words, names, and mixed English terms before publishing.

Can I translate subtitles after generating them?

Yes. Start with a corrected TXT transcript, translate the text, then rebuild or review the timed subtitle file. Translation can change sentence length, so expect to adjust line breaks and timing.

How accurate are automatic subtitles?

Automatic subtitle accuracy depends on audio quality, noise, accents, overlapping speech, microphone placement, language, and specialized vocabulary. Use it as a strong draft, then review anything important before publishing.

How do subtitles help accessibility?

Captions help viewers who are deaf or hard of hearing, people watching without sound, and anyone who needs support following speech. For accessibility, include important spoken words, speaker changes where needed, and meaningful sound cues that affect understanding.