How to prepare a video for dubbing: a checklist
Part of the work of a good dub happens before the upload. These are the checks that make a lesson or a screen recording easier to dub well and easier to review.

The checklist
To prepare a video for dubbing, start with a file you hold the rights to, in a common format, with speech that is as clean as you can make it. Keep one lesson per file, know how many people speak, and decide whether the music stays. You do not need a script.
- Hold the rights to the video, and upload the file itself, not a link.
- Keep the speech clean: music low under the voice, little room noise.
- Put one lesson in each file, within your plan's video length.
- Know how many people speak.
- Decide whether to keep the music and background sound.
- Skip the script: the transcript is written for you, and you can edit it.
- List the product names, menu items and numbers you will check in review.
- Make sure every voice in the file is yours, or that its owner gave you written permission.
The rest of this page goes through each point and the reason behind it.
Hold the rights and upload the file itself
A dub is a new version of your video, so start with a video that is yours to change: your own recording, or one you have the rights to adapt. If a course belongs to a client or a school, check your agreement before you translate it.
Upload the file, not a link to where it is published. DubForge takes MP4, MOV, M4A or WAV and does not accept YouTube links, because a link does not show that you hold the rights to the video. The file's real type, length and size are checked before anything is charged, and the file goes to private storage.
Use the best copy you have. Export it from your editing project, or use the original recording, rather than a copy that was downloaded and compressed again on the way.
Clean speech makes a better dub
Everything in a dub starts from the speech: it is transcribed, translated and spoken again. In DubForge the voice is also taken from your file, so what your microphone captured is what the dub has to work with.
- Music low under the voice. A music bed that competes with the speech makes words harder to tell apart. If you are still editing, lower it under the parts where you talk.
- Little room noise. Fans, keyboard clatter and echo stay in the recording. They are easier to deal with before you record than after.
- One voice at a time. People talking over each other are hard to follow in any language, and hard to separate.
You do not need a studio. A screen recording made with a decent headset microphone in a quiet room is a good starting point.
Music and background sound do not have to be removed first. DubForge separates them from the speech and mixes them back under the dub.
One lesson per file
Upload each lesson as its own file rather than a whole course in one export. A five-minute lesson is something you can review in one sitting, and a problem in lesson 4 is easier to find when lesson 4 is its own project. If you are dubbing a full course, dubbing online courses covers how to plan it.
Each plan also sets the longest video it takes:
| Plan | Longest video |
|---|---|
| Credits without a plan | 20 minutes |
| Creator | 20 minutes |
| Pro | 1 hour |
| Studio | 3 hours |
Credits are counted by length: one credit dubs one second of video in one language, rounded up to the next second, and each target language is its own dub. If the recording has minutes before the lesson starts or after it ends, trim them before you upload. Keep the pauses inside the lesson, though: they are part of how the lesson reads. The full arithmetic is in how much it costs to dub a video.
Speakers, music and background
When you create the project, you say how many people speak: one, two, three or more, or not sure. Count everyone whose voice is in the file, including a student who asks one question or a co-host who speaks for ten seconds. Each person is then dubbed in their own voice and speaks when they spoke.
You also choose whether to keep the music and background. Keep them when they are part of the lesson, such as a sound the software makes or a short sting between chapters.
If you would rather remix the audio in your own editor, the downloads include the voice alone and the music and background alone, each as a WAV file, next to the full mix and the dubbed video.
No script needed, but list your names
You do not need a script or subtitles to start. Speech recognition writes the transcript from the video, and in the editor you can correct it next to the translation, segment by segment.
What helps is a short list of the words that must come out right. Write it before you upload, while the lesson is fresh in your mind:
- Product and software names, spelled the way the company spells them.
- Menu items, buttons and settings, as they appear on screen.
- Numbers that matter: values typed into fields, versions, keyboard shortcuts.
- Names of people, files or projects you mention.
This is a habit, not a setting. Keep the list next to you when you review the dub and tick each word as it comes up.
Fix what you can in the file. Noisy speech, a whole course in one export or a voice you forgot was there are all harder to deal with after the dub.
Check whose voices you are dubbing
Dub your own voice, or voices whose owners gave you written permission. DubForge asks you to confirm this in each project, for that exact file, and a public figure's voice needs their written permission too.
Recorded classes are where this catches people out: a student's question, a guest in a webinar, a colleague who explains one step. Each is someone else's voice. Ask them in writing before you upload, or cut that part. What a permission should say is in voice consent for AI dubbing.
To see how one of your own recordings comes out before you plan a whole course, the free trial dubs 30 seconds once, for an account with a confirmed email and no card. It has every editor feature, and the trial dub is watermarked. Plans are on the pricing page.
Questions
Can I dub a YouTube video from its link?
Not in DubForge. Upload the file itself, as MP4, MOV, M4A or WAV, because a link does not show that you hold the rights to the video. Export it from your editing project or use the original recording.
Should I remove background music before dubbing a video?
You do not have to. DubForge separates music and background sound from the speech, mixes them back under the dub, and lets you choose whether to keep them. Music kept low under the voice still helps.
How long can a video be for AI dubbing?
On DubForge it depends on the plan: up to 20 minutes on Creator or with credits bought without a plan, up to 1 hour on Pro and up to 3 hours on Studio. Long courses are easier to review split into one file per lesson.
Does background noise affect an AI dub?
It can. Noise and echo make speech harder to transcribe, and the dub's voice is taken from your recording, so a quieter room and a closer microphone give it a better start.