Video Translator’s platform provides a client with the ability to transcribe and translate content to/from a
number of languages, as discussed here. What we are going to do today is transcribe content to a different language, and use a
practical example to demonstrate the technology.
We will be using this video, from the team at NSWHealth. The video is a simple, but effective solution for young mothers, where NSWHealth provide a
useful box of household items aimed at first time mothers.
The idea is to (1)
transcribe from English, (2)
translate to German, (3)
transcribe to German, with the end output being the same video in German. To illustrate using a contrast, we use a
male German voice. That is, we are doing a
English to German video translation. Please
play the video to see the
end result of the process described below.
Please direct your browser to videotranslator.ai, and then click on the
Loginbutton. On successfully logging into the application,
select the template, and using the highlighted button,
View Template: myTemplate.
For a detailed discussion of how to edit the template, have a look at this post. But the template should look similar to below, note we require
audio componentand a
Language Optionshave also been opened up, and the translation possibilities are shown. In this case, we can translate the video into
create an item of type myTemplate, called
upload the video. The voice’s are unsurprisingly,
Australian femalewhich we will switch to
German maleto demonstrate the multiple voices. Below is the original video from NSW Health.
We do a
quick transcription, and then investigate the returned
Captions. The captions are pretty good, but its important to check the result. A good rule of thumb is to check when different people start talking, as the differences in how two individuals sound, can confuse the AI.
In this specific case, the
AI missed the early partof what the gentleman was saying at the
00:30 secondmark. The captions, after some manual editing look like below.
Now that we have the captions, we use the translation AI to produce a German version. Explicitly, we are using the AI to translate from
English to Germanfor the
captions only, hence the
audio component is not selected. Also note that we are creating a
copy of the asset, and
attaching the German translationto this copy of the asset.
We open up the new
de_baby_bundleitem, and now we do two things, (1) we add the
caption into the audio component(copy/paste), and (2)
trigger transcriptionagain, but this time
only on the second component(the audio component).
add the audiointo the original asset as an
overlay, and additionally
mute the original audioto achieve the desired outcome. It should also be noted that we are still using captions all the way through - another alternative would be to remove the captions and put in a text overlay to allow us a finer level of control over the look/feel of the new asset.
Please note that the
German will be incorrect- its just an AI. After translation, it is recommend that a
human subject-matter expert clean upthe results. The Video Translator platform is built on the logic of
AI assisting people, the recommendation is to use the
AI to do the heavy lifting, and use the
human to do the high value tasks.
In this blog post we have covered how to add an audio overlay into your sample content. Please note, Video Translator makes
no representation of a relationship with NSW Health, we are simply using a
publicly available content to do beta testing on our platform prior to opening up our product.
The platform is currently in closed beta, while we work out bugs in the code. If you are interested in trying out our technology, please drop us an email at firstname.lastname@example.org.
As the platform is still in closed beta, (1) the
UI may change in the future, (2) the
pricing numbers shown are indicative only.