Recording Process For VO
While audio localization has provided a new avenue towards superior localized games maintaining high quality in VO is costly and time-consuming. For example, Alexander O. Smith describes the handling of VO based on his experience in translating VO scripts from Japanese into English for FFXII (2006) (quoted in Jayemanne 2009, n.p.):
Voice scripts pose an additional challenge in the form of timings, and sometimes, matching an actor's take on a character. A single line can take hours of work to get right. Ultimately, practically every line in the FF12 voice script reflects the work of the original writer and editor, one translator's initial take on the line, another translator's crosscheck, the editor's check, the voice director's opinion and the actor's interpretation ....
This description illustrates how input from different specialists has to come together, hence the time-consuming nature of re-voicing. However, VO is a relatively new process in game localization, with poor treatment of voice scripts frequently discussed by gamers, such as problems ranging from the quality of voice acting to lack of synchronization with the image. This is also due to certain work procedures imposed upon translators, such as translating files where all the interventions of one particular character are listed, without indicating who s/he is talking about or at what stage s/he is in the game. Apart from the cases where video captures are made available to voice actors, an added difficulty in re-voicing for games is the need for actors to record isolated strings or even words on their own, sometimes in the absence of sufficient contextual information. Nevertheless, the increasing importance placed on characterization in games seems to be recognizing revoicing as an effective technique to leverage the narrative power of the game in engaging the gamer, drawing more attention to audio localization.
How dubbing takes place for VO scripts in a game can be further illustrated with the Japanese game Catherine (2011), which develops in psycho-drama sequences, featuring conversations between characters. An interview (Ishaan 2011) with the voice director of the North American version of the game shows the extra challenges posed by the way in which voice recording is typically conducted for games. Each voice actor's recording is made individually and according to lines or scenes which are not always in sequence, even though the situational context is clearly the key to providing an appropriate take for each line. Furthermore, in this particular game the voice recording was reportedly conducted concurrently for the original Japanese and the English version, even though the game was not sim-shipped (the Japanese version was released in February 2011, followed by the US version in July 2011). This meant that the common practice of having original takes available at the time of the recording of the English track to provide the feel for a particular scene, was not followed (ibid.). If games include specialized terms with difficult pronunciations particular to the game titles, reference materials such as a pronunciation guide for a VO recording session become necessary. For example, the localization project for Mass Effect 2 (2010) created such a guide to be used across all localized versions, containing a recording of esoteric terms by its original English VO team (Christou et al. 2011, 41). Above all, these examples highlight the way in which in-game VO assets are primarily handled as information objects rather than as a coherent narrative stream in sequence from start to end, unlike in most AVT scenarios.