How to translate 5000 words in 3 hours

Yes, this is indeed possible.

I remember those times 13 years back when translating 5000 words per DAY would be a major feat…. and I would feel completely exhausted, needing at least 2 days to recover from such a marathon.

Today, thanks to modern machine translation and voice dictation technologies, it is very well possible to translate 5000 words in just 3 hours…. and not to mentally collapse afterwards.

How to do it… in some reasonable quality

This is how it’s done in short: use machine translation and voice recognition.

Prerequisities

1. The text needs to be relatively easy to translate. It should have long fluent sentences with not too many specialized terms that would require any substantial research. A typical example would be a coffee-machine manual.

2. You need to have some experience as a translator so that you can quickly translate yourself those parts that do not come correct from the machine-translation engine. Also, you will need to understand the source language very well to quickly identify possible mistakes in the machine translation.

3. You need to be familiar with CAT tools. Ideally with SDL Trados or memoQ that have the actual potential to enhance translation productivity.

Or, sure, you can translate with one Word-document on the left, the other on the right. But this is very exhausting, as you have to steadily look with your eyes for the source text in your document. On the other hand, CAT tools have the advantage to neatly arrange individual sentences next to each other and to highlight the sentence (segment) currently being translated.

4. You will need some special technologies to achieve this goal. Especially the machine translation feature. Google Translate and DeepL provide very good results. You will to plug-in this into your CAT project, e.g. in SDL Studio 2011, this would be in Project Settings / Translation Memories and Automated Translation:

Alternatively, you can start Bohemicus and leave it running in the background. It will automatically deliver machine translation when you press the Ctrl+Space keyboard shortcut.

Step by step instructions

1. Prepare the project in SDL Studio (Trados) or memoQ. You can see this video to see how a project is prepared e.g. in memoQ (this video is on translating a webpage, but the preparation part is useful in this case too):

2. Connect the Google Translate / DeepL engine, see above. Or run Bohemicus.

3. Run your voice recognition software. You will need it when you translate segments (sentences) that come out as total nonsense from the machine translation engine. That’s because it’s usually faster to just dictate a sentence than to type it manually. At least for me  – I can type at some 200 chars/min… so it’s still faster to just dictate the text, instead of typing it. Also much more comfortable. You can also use e.g. the Bohemicus voice recognition software that works with 100+ languages (https://bohemicus-software.cz).

Summary

You will be basically reviewing the machine translated text, not actually translating. You can alternatively also use Bohemicus to machine translate and dictate your text (along with many other features that will make your translation process much easier). Bohemicus will enable you to use machine translation and voice dictation in just any CAT tool or any other Windows-based application.

Translating 5000 words in just 3 hours can be done. I know it. Because I have done it.

A final tip: Doing 5000 words in 3 hours can be exhausting, because you still need to process all those words with your brain. I suggest you eat well before you start working, to give your brain enough energy.

Please watch the video below to see this in action:

This entry was posted in Uncategorized and tagged , , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *