Video Translation 101: 3 Editors’ Tips to Lower On-screen Title Costs

March 23, 2017
new-articulate-review-features-for-e-learning-translation-qa
Two New Articulate Review Features for e-Learning Translation QA
March 20, 2017
are-timed-synchronized-voice-over-audio-different-yes
Are Timed & Synchronized Voice-Over Audio Different? (Yes!)
March 27, 2017

On-screen titles are often a source of frustration during otherwise straight-forward video translation projects. Why? It’s simple – they require a large amount of labor to replace and re-format during localization, and in video editing software all this work has to be done by hand, without the aid of translation exports. Fortunately, there are three steps that video editors can take to make their timelines much more localization-friendly – and much more cost-effective.

This blog will list three tips for editors that will significantly lower on-screen title replacement costs.

[Average read time: 4 minutes]

Why are on-screen titles so tricky to localize?

On-screen titles replacement (also known as OST replacement) within a video editing timeline can be very complex, especially if the titles are animated. On top of that, different video editing and title compositing programs – Adobe Premiere, Adobe After Effects, Final Cut Pro, and Avid Media Composer – have pretty different workflows when it comes to title creation. This means that some programs are easier to localize in than others. To add to this confusion, once a video is exported from the source file into a format like MP4, MOV, WMV, and FLV; it’s almost impossible to tell in which program it was created. That said, After Effects allows for more sophisticated animation and title sequences, so that it’s usually a safe bet for more complex videos.

Second, as we’ve mentioned above, OST in the programs above have to be replaced manually. That means that a localization video editor has to copy and paste each title translation into the editing timeline – one-by-one. Because of that, any complexity that gets added to this process – anything from heavy rendering to preview animation effects, to animation that slows down the copying-and-pasting – can add a lot of labor and become a serious cost factor. In fact, for many videos, the OST replacement can be even more labor-intensive than the voice-over recording, even though that requires a director, native voice-over talent, and a QA reviewer.

For more information on OST replacement itself, as well as standard methods, check out our previous blog post, Video Translation 101: Replacing On-Screen Titles Cost-Effectively.

The Editors’ Tips

With that out of the way, here are the three editors’ tips. Keep in mind that they specifically address how labor-intensive the final OST replacement is – the idea is to make it as quick (and cost-effective) as possible.

1. Make sure linguistic units are in one text box

A “linguistic unit” is basically just localization-speak for a full sentence or phrase. In general, keeping sentences and phrases together makes translation much faster and more accurate. The opposite is also true – separating them in the text, whether by adding line breaks, hard returns, extra tabs or spaces, or by splitting them up in different text boxes, can make translation more difficult and more costly.

In the following screen shot, the on-screen title on the left is in one text box, whereas the one on the right is split into two.

forced-format-on-screen-title-video-translation-localization.png

Note that they look identical in their respective viewers – once the video is exported, it’d be impossible to tell that the text was split into two titles. The title set on the right would take at least twice as long to replace as the one on the left. Not only does keeping full sentences and phrases in the same text reduce the number of titles you have to copy and paste by hand, but it also avoids translation errors that may require additional changes or corrections to the OSTs.

Additionally, editors should make sure to avoid forced formats inside the text boxes – the following blog post goes deep into that. It focuses on captions in e-Learning courses, but all of the tips apply to video editing timelines.

2. Avoid complex animations – especially if they split up words or phrases.

Complex animations on text take a while to render, which can make any editing process take longer. But some kinds of animations really add to the OST replacement labor, including:

  • Animations that convert words to outlines: Often, reproducing these animations requires outlining the localized words (trying to match the original English formatting), and then manually re-implementing the animations. Very time-consuming, since the source animation timings can’t be used directly.
  • Animations separate letters in a word one-by-one: Commonly used to create a typewriter effect, or to really emphasize a word on the screen. These animations usually take a long time to produce in the English source video – and then take just as long to produce in the foreign-language video. An in-house editor spending 2-3 hours to impress his or her clients will end up adding that same amount of time to a localization project – per language.
  • Animations that create multiple instances of the same OST: Commonly-seen in Final Cut Pro and After Effects. While these animations may not add more re-work time by themselves, they do create multiple instances that require manual replacement. If you have to replace the same OST ten times, that increases the work for that OST ten-fold.

These are the most common ones, but there are many other kinds of animations that add a significant amount of work during localization.

3. Make sure each title is one instance with multiple references.

Some animation effects can create multiple instances for an OST, as discussed above. But there are many other reasons that editors may create multiple instances of the same title. For example, a video may cut from title to B-roll footage and then back to the title. Ideally, the editor would create one instance of this title, and then layer the B-roll on top of it on the timeline. However, the editor may be tempted to create a second instance of the title, and put it on the same layer as the B-roll footage, right after it. The following screen capture shows those two options – on the left, there is one instance of the title, with the other footage above it on the timeline; while on the right, there are two instances of the title with the B-roll footage cut in between them:

split-on-screen-titles-for-video-translation-localization.png

Needless to say, the timeline on the right has twice as much work as the one on the left. There are many other cases where editors may create multiple instances – for example, if they want to reveal a bulleted list one item at a time. In this case, creating one title and adding masks that disappear to reveal the bulleted items would be more localization-friendly. That said, “instancing” may be unavoidable in some cases, as when a video revisits a title much later on in a timeline. But in general, keeping the instances of each title down can save a large amount of time and money during translation.

Remember – manual work multiplies by the number of languages in a project

This is crucial when assessing whether or not it’s really worth making a change to an on-screen title to optimize it for localization. If replacing it is going to take only 10 minutes, it’s tempting to just roll that cost into the budget. However, for projects that have many languages – at JBI, we commonly see multimedia localization projects in 24 to 28 languages – this can be a real liability. Ten minutes in 24 languages will add four hours to a project. If there are ten titles that require that same amount of work, in all those languages, that effectively adds a week’s worth of work – 40 hours. It’s crucial to keep this in mind when developing the English-language original video – small tweaks that may not take very long in the original English will get multiplied by your localization project’s language set.