Audio Visual Localization of a Short Cartoon Film

In this project, I chose to localize a short cartoon film from an independent director. It tells the story of how a teenage girl was bullied and eventually recovered from it to become a more confident and better self. It used plenty of text to express the feelings, emotions, and their changes of the protagonist, and the way text is displayed and moved is also part of the story. I chose this film as the topic and format were very interesting, and the content is fitting and echoing with Chinese audience.

The project is split into three stages: preparation, localization, and QA and improvement.

Preparation Stage

I first watched the whole video and conducted a localizability check to see if it is suitable for localization. Some works could contain significant cultural differences or target specific groups of people from specific locations, making them hard to localize. This video passed the localizability check and then I started the initial investigation. There are two text categories in the video: embedded in items, like text on the calendar, computer, or phone screen; special effect text that helps develop the plot and show character emotions. The easiest way to add subtitles normally is to add it below the picture, similar to movies. But the text her is part of the picture instead of a simple script, and I wanted to allow the audience to enjoy the entire video instead of having to spread their attention to the bottom of the screen. Therefore, I decided to embed the Chinese translations into the video in a matching style with the original, and try my best to keep the two languages in sync.

After deciding my strategy, the next step is to plan out the details. For text embedded in items, most of them are static. They can be filled in by adding a mask or using the “content aware fill” functionality of After Effects, and replace with translated text. “Content aware fill” has worked well for me in past projects when the picture was relatively clean. For dynamic text on the screen, my initial decision was to use special effects, masking and blending.

Localization Stage

I first sorted out all text in the video and kept them in the media table. Then I used Memsource to translate the entire media table. Using a cat tool helps me keep the translation memory and set the maximum length for some sentences, which is very important for audio visual localization. Since this is an animated piece, I performed transcreation for some sentences instead of going word-by-word.

The next step was to plug translated content back. I used a lot of keyframes to reconstruct the animation effect in the process, and used different approaches for different text.

At the start of the video, there was a laptop image with the video title. Luckily, there was a frame a couple seconds before this point that had the image without the title text, so I used that frame to overlay the Chinese title upon the English one.

Next there was some dynamic text, moving slowly and half transparent. I also created many keyframes in correspondence with the movement of the text, as well as to reach the effect of changing transparency.

A bit later, there was a piece of dynamic text with covered up special effect. In the screen shot below, there is a sentence saying “Who would play with you”, which ran through the hair of the protagonist. Since the text going through the hair disappears, it cannot be reproduced by simple movement. I made it work by adding key frames to the source text and making changes word by word.

Afterwards there was a scene where the protagonist saw another girl having the same birthday date with her. The girl was showing her drawing skills to the protagonist. Praising the girl on the surface, she thought that the girl had no talent internally after seeing the imperfect drawings. The sentence “You have no talent” was delivered through the facial expressions of the protagonist, slowly entering the heart of the girl and hurting her feelings. The text changed shapes here, and my solution for it was to draw the text path with pen tool, and add this path as a mask for the text layer. In the path options of the text layer, I selected the mask I just drew and then adjusted some properties. I also used many key frames, and turned on motion blur to smooth out some motion. Finally, this string entered the girl’s heart and disappeared, which I reproduced by adding keyframes to source text as before.

After plugging all translated text back to the video, I entered the next stage for quality check and perfection.

QA and improvement stage

Here I asked friends who have never seen this video before to take a look. They helped improve the translation quality and make the animation look more natural. This step was simple for my project as I only processed the first half of the entire video, but quality assurance will be more important for larger projects as it would be too hard to go back after the video is finished.

Future plans

Though this project is over, there are more possibilities I can have with this video. For example, I added Chinese subtitles and tried to make them look in sync with the English ones, so all image layers based on Chinese text is still usable. I would love to give the localization scripts of After Effects a try to help import and export text layers on one click, making it easier and faster to localize the video into more languages.