Text-to-Speech for Localization of Learning Multimedia
Localizing multimedia materials for online learning and courseware content requires careful planning. In this blog, Senior Multimedia Localization Engineer at Welocalize, Michael Anderson, introduces text-to-speech and outlines why it offers learning solution providers and companies with online training programs a cost-effective and timely solution for localizing audio and video voice-over work.
The main style of voice-over (VO) recording, commonly used by global learning companies, is voice-over studio recording and on occasion, at-home VO recording. Using studios to record multilingual versions of learning videos and courses is a significant investment in studio hires, voice talent and subsequent editing services. There are also additional costs associated to changes made to the source material, which may require additional budgeting for studio time and talent.
Welocalize is helping companies reduce some of these associated costs with the newly added multimedia localization solution, text-to-speech (TTS). This Welocalize service offering provides clients with a cost-effective alternative, as well as a potential faster time-to-market in comparison with the use of recording studios and voice talents for voice-over related projects. This capability can be used for learning courses, instructor-led training content, training videos, as well as online help features on websites or in software and mobile applications.
What is Text-to-Speech?
Instead of voice talents sitting in a studio, recording the localized versions, text documents (or scripts) are loaded into synthetic voice software, turning written word into speech. The technology has been available for some time; however, significant recent improvements have been made within the last couple of years which means the speech output is less robotic and more intelligent.
A key benefit of using this approach for localizing and translation of content is that the more you use the software engine, the more it remembers and retains common terms in the translation memory (TM), which in turn continues to reduce cost and further speeds-up the translation process. The process includes adding the content to the software and based on the quality of the voice output, simply tweaking the word document to gain the exact translation and voice.
Welocalize has piloted the text-to-speech solution with several learning clients and similar to machine translation (MT), you have to continue to train and educate the software for each client’s needs. We are also able to provide in-depth cost analysis for TTS and compare it to historical VO costs, so the client can see where the costs in terms of time and money are being made and if there is an opportunity to streamline the process.
The market potential for multimedia localization of eLearning programs and learning content is huge. There is growing use of voice in the learning space and also for technology, devices and communication is growing rapidly. Welocalize is excited to offer an alternative multimedia localization solution that fits our client’s needs for delivering learning content in multiple languages. We are continually looking for innovative ways to improve and customize the multimedia localization process to ensure our client’s communications reach their global audience in their local language and local dialect.
Based in California, Michael Anderson is Senior Multimedia Localization Engineer at Welocalize. He is also a well-known film-maker, producing documentaries in his spare time.
TAKE OUR VOICE CHALLENGE! You can meet senior Welocalize language experts at the Learning Solutions Conference and Expo taking place in Orlando, Florida, March 25-27, 2015. Welocalize will also be running demonstrations of Text-to-Speech versus Voice-Over at booth #607. These demonstrations will enable attendees to compare the costs, turnaround time and quality of TTS vs VO. If you are attending the event and would like to schedule a demonstration, please visit: http://www.welocalize.com/voicechallenge/