You Wanna Get Personal Speak To Me! Localization of Audio in Apps
|Dawni Jacobs in Programming Monday, August 11, 2014|
Human voices “get to us” like little else. Just ask the execs working at BMW in the 1990s, when the car company had to recall a line of BMWs because its male drivers were not willing to take directions from a female GPS voice.
A trustworthy voice can immediately create a strong interpersonal connection - even with a machine, as anyone can attest who watched Joaquin Phoenix interacting with Scarlett Johansson's disembodied operating system voice in Hollywood's recent release, Her.
But a highly computerized or unnatural voice jeopardizes that same connection. By angering or amusing us, the voice makes us more aware of its quality than its utility and function.
App developers understand this and work hard to match their apps with voices that suit their purpose and help users feel a connection with the app. Clearly, this is also an essential consideration for the app's localizers. The localized Chinese, Arabic, or Hindi apps should speak as naturally to their new users as the original app did to its original audience. Obviously, this can't be achieved by leaving the audio portions in the original language. That would be about as effective as not releasing the app in the new markets at all.
So what are the strategies for localizing the voice audio portions of your app? How can you help this process?
You may already know that voice audio exists in one of two formats: as standalone audio files (typically .mp3 or .wav) or integrated into a video file (mostly .mp4, .avi, .fla, or .wmv).
In an ideal world, the voice portion of any format would be re-recorded in the new language. If your app contains only audio files, localizing them will usually mean re-recording them in the other target languages. If you are very cost-conscious and at the same time localizing into a larger number of languages - say, more than six - it might make sense to budget some upfront engineering cost for the option of an on-screen display of translated text. If your budget allows for the re-recording of the target languages, however, that's what you should do no matter the number of languages - and your users will thank you for it.
The considerations for audio localization of videos are slightly different, though the cost and the number of languages have a similar correlation.
First of all, the typical video file is not properly internationalized. In this context, a properly internationalized file is designed so that every occurrence of spoken text has a corresponding text-based file that places text in the correct location on-screen (for instance, close to the head of the speaker if applicable). In the original voiced version of the app the on-screen text is not displayed, but in the localized versions it can be filled with the appropriately translated text and displayed in the correct position on the screen. (Yes, the localization team might have to struggle a bit to fit the translated text in the allotted space - but they'll manage.)
Since the cost and effort to post-internationalize your video files might seem prohibitive if you're localizing into only one language, it often ends up being more cost-effective to simply re-record the voice portion of the video. The more languages you're localizing into, the more economic sense the text-only option might make, even with the additional post-internationalizing development cost (a one-time expense).
If you're localizing your app into more than two languages, our experience shows that you'll save about 33% by going with the text-only option. Your localization team can help you locate that financial sweet spot.
What should you pay attention to as you're developing and preparing your app for localization?
- First things first: if you're developing a voice-heavy app that you know you want to localize at some point, be sure to build in correctly timed options for onscreen-text display. This will allow you to opt for either onscreen or voiceover tracks down the road. (You'll be our very first client to do so -- so don't worry if you're already past this step.)
- To streamline the localization effort and reduce cost, maintain a written script. This will eliminate the need to have the audio transcribed.
- Separate noise and voice tracks. During localization this will ensure that the noise track remains untouched.
- Have an idea of what type of voice suits your app. Your localization team will listen to the original voices and select voice talent in the target languages accordingly, but before they select a portfolio of voices they will likely ask for your input (age, gender, mood, dialect, "like Julia Roberts," etc.). Once they give you that portfolio to choose from, it's going to be your call.
- Trust your app localization team to produce a high-quality recording. The recording will take place in a professional recording studio to eliminate external noises and guarantee high-quality output. Any recording parameters will be documented by the sound recording team, so if additional recordings are necessary the same voice talent can be re-engaged and all the other variables will stay the same as well. (After a microbrew of two, your audio localizer might share some of the less obvious pitfalls of this process: tracking down the voice talent after she's moved away or dealing with a change in her voice because of allergy season.)
- Attempt to provide your localization team with the final content before the recording takes place. We know that life moves fast, and sometimes it's necessary to add content even after the recording is done. But anything you can do to minimize post-recording changes will be appreciated by the localization team. The same is true for last-minute changes. The voice talents prepare intensely for the recording sessions, including practicing the text and marking up the script, and this naturally has to be redone for any late changes.
There's nothing more satisfying for app users than to have an app "speak" in their own language. And there's nothing more rewarding for us at PTIGlobal than to help app developers make their babies speak many languages, just as effectively and naturally as they did in their original app. Drop in to our Portland office for a local microbrew and a chat with an amiable audio localizer about how we can help. Or shoot us an email and we’ll be glad to give you a call back. Or simply call us yourself at 503-297-2165.
Read more: http://www.ptiglobal.com/
This content is made possible by a guest author, or sponsor; it is not written by and does not necessarily reflect the views of App Developer Magazine's editorial staff.