An amazing voice or presentation will not make up for poor content so preparing the script is very important. Before the audio can be right, the words have to be right. Starting with a more generic script without source language colloquialisms and idiomatic expressions content helps.
When recording audio in your target language, pick a gender and dynamic quality of voice that is appropriate for your topic. A localization expert can help you pick out a native speaker who fits your specifications with good enunciation and voice intonation that comes across clearly and concisely.
Before you start recording your audio, take your last chance to review the translated script one more time. It will be more expensive and time consuming to go back and re-record because the script was wrong.
Once the script is completed, communicate your expectations for the narration and engineering specifics. For the narration, let your project manager know the speed, emphasis, and dynamic you are hoping to create. If the text needs to be recorded within a certain time-frame, communicate this as well. Also, specify engineering details, such as the wave range, hard limit, normalize, and the file format (.wav or .mp3 for example) necessary for the final product.
If the previous steps are well executed, the actual audio recording should go quickly and smoothly. Keep in mind, voice talents have a time limit on how much they can work in a day. Typically, 3 hours is the maximum they can work in a day before the quality of voice begins to suffer. Plan your recording sessions accordingly for maximum efficiency.
Once the audio is edited according to your specifications for engineering, a native speaker should review the recordings to ensure it was recorded correctly and pauses were inserted in the appropriate places.
Adding translated, localized audio to a multimedia presentation may mean taking a few extra steps but taking your message from so-so to creating a memorable experience for your target audience is worth the extra time and effort.