翻译技术速递Machine Subtitling and Voiceover
机器字幕和画外音
This report was co-authored by Yulia Akhulkova and Belén Agulló García.
Now is an exciting time for media localization, with technologies improving across the board making automated subtitling and dubbing an actual possibility.
With machine dubbing, we're still far from having automated voices that sound like real human actors expressing emotions and delivering high-quality acting performances. The current automated technology is mainly used for voiceover recordings (for example, documentaries, e-learning, news, corporate videos, and audio description). However, interesting advances are being made in the field of voice synthesis, as we’ll discuss further.
Automated subtitling and voiceover: closer than putting a man back on the moon?
When it comes to subtitling, there are two main ways software can automatically generate subtitles:
You might have seen examples of this technology on YouTube. The platform has an automatic subtitle generation function. The output quality varies depending on the quality of the audio and the pronunciation of the speaker, as well as the language of the video. That is how automatic subtitles are already replacing humans. Although automatic captions including transcription and synchronization are not always available on YouTube currently, YouTube Studio allows you to provide a transcription for your videos. Their solution will automatically synchronize the video with the transcription using the first of the aforementioned methods.
Source: YouTube
There are many software solutions that offer automatic transcription, including Dragon Naturally Speaking, Webcaptioner, and Microsoft Translate. As Mara Campbell from True Subtitles pointed out in her presentation at Media4All, even the cheapest, worst speech recognition software offers greater productivity than doing it manually. We have spotted an increase in services offering automatic transcription and synchronization for subtitles such as Happyscribe, Limecraft, Trint, and Rev.
But what about automatically translating subtitles into different languages?
本报道由尤利娅·阿库尔科娃和贝伦·阿古洛加西亚联合报道。
技术的全面进步让自动字幕和配音成为可能，对媒体本地化而言，这是令人兴奋的时代。
虽然有了机器配音，但我们还远不能产生自动化声音，听起来就像真正的人类演员用声音表达情感，进行高质量的表演。当前的自动化技术主要用于画外音录制（例如，纪录片，电子学习，新闻，企业视频和音频描述）。然而，正如我们将进一步讨论的那样，在语音合成领域正在取得有趣的进展。
自动字幕和画外音: 比把人送上月球容易？
说到字幕，软件自动生成字幕的方式主要有两种:
你可能在YouTube上看到过类似的例子。该平台具有字幕自动生成功能。输出质量取决于音频的质量和说话人发音优劣，以及视频的语言。这就是自动字幕已经取代人类的方式。虽然目前YouTube上并不总是提供自动字幕（包括转录和同步），但YouTube Studio允许您为您的视频提供转录。他们的解决方案是使用前述方法中的第一种，自动同步视频和转录。
图片来源:YouTube
有许多软件解决方案提供自动转录，包括Dragon Naturally Speaking，Webcaptioner和Microsoft Translate。正如来自True Subtitles的玛拉坎贝尔在Media4All的演讲中指出的那样，即使是最便宜，最差的语音识别软件也能提供比人工更高的工作效率。我们发现提供自动转录和同步字幕的服务有所增加，例如Happyscribe，Limecraft，Trint和Rev。
但若是字幕自动翻译成多种语言呢？
以上中文文本为机器翻译，存在不同程度偏差和错误，请理解并参考英文原文阅读。
阅读原文