Dasha Syrpis

Could you post a short section of your file?

Daniel

Hi Daniel

I'm afraid I can't, because it's highly sensitive material. Basically it's a transcript of an interview where the speakers speak a mixture of English and Kiswahili. So I was wondering if there was a way of automatically detecting the English sections of the interview and copying them across as they are to the target document. It's in Word format. 

Dasha Syrpis

So you have a Word document with a mix of two languages - this is not what is typically called a bilingual document in the world of translation as I know it.

There is no structure to that mix and you want to automatically identify which parts are which language, I hope I got that right.

That is not what Trados Studio is intended to do, but I find your situation interesting. I wrote a "bilingual" text, like a dialogue:

I import it into Studio and let DeepL do a machine translation of everything:

Then I filter for segments where source and target are identical and revert the filter, so that I get all segments where source and target differ:

You will see that segment 4 is German in the source but DeepL translated it into German again, using different words, so it is not identical anymore.

It's not a flawless method, but it might help you with your project.

Hope it helps,

Daniel

Dasha Syrpis

So you have a Word document with a mix of two languages - this is not what is typically called a bilingual document in the world of translation as I know it.

There is no structure to that mix and you want to automatically identify which parts are which language, I hope I got that right.

That is not what Trados Studio is intended to do, but I find your situation interesting. I wrote a "bilingual" text, like a dialogue:

I import it into Studio and let DeepL do a machine translation of everything:

Then I filter for segments where source and target are identical and revert the filter, so that I get all segments where source and target differ:

You will see that segment 4 is German in the source but DeepL translated it into German again, using different words, so it is not identical anymore.

It's not a flawless method, but it might help you with your project.

Hope it helps,

Daniel

Hi Daniel, yes I agree bilingual file, was not the best way to describe it! It is just a mix of two languages one of which does need to be translated. What you've done is very interesting. I'll keep it in mind. Thanks for your help.