How Well Does MTL Work?
-
@someoldguy I'm afraid not. It does, however, have a Windows app that you can easily push text to through a keyboard shortcut (going via the clipboard).
Edit: it should have a MacOS app as well.
-
@waterdweller I saw that. Google has the advantage of being plugged into Chrome, so you just need to browse the page and (usually) get comic relief. A plug-in like the "reader view" would be very helpful.
I use auto-translate a lot on a forum that gets posts advertising apartment cleaning and dating services in various Russian cities, so I can tell who needs to be banned and for what reason. :)
Sometimes you can't copy/paste text very easily. Some pages even prevent you from highlighting text to get it into the clipboard, meaning you have to type things in. If I understood Japanese well enough to be able to type it in, I probably wouldn't need the translation.
-
Sometimes, Deep L works surprisingly well. But it's a lot of effort to learn the plot about a book without really "reading" it since you need to edit out the Japanese punctuation characters to avoid some bad Deep L glitches.
Someone more tech saavy could probably make a find / replace macro to replace these with their English equivalents.
。」』「『、…─
-
Something I'm really disappointed about Google translate is that they failed to build a community around it. A couple years ago, I was curious so signed up to be Google Translate Contributor.
https://translate.google.com/about/contribute/
You register the language you are good at, and then you go through a few different "tasks". One is that it will gives you bunch of short phrase or words one after another, and you enter its translation. Another task is that it will give you a word/phrase and it has multiple different translation of it, and then you pick the ones you think are good translation.
It is like a game and you go though bunch of them in a batch of a few dozen or so per session. You make more contribution of them and you can earn "batch" of different colors. But that's it! There is no list of contributors, no rankings of contributors, no place for contributors to discuss with each other. You are nobody.
It is a very lonely process and not fun at all, and didn't motivate me to keep on going.
-
This post is deleted! -
Since people posted here about DeepL, I gave it a try.
My impression is that DeepL gives a adequate translation about 85% of the time, makes bad inferences 10% of the time, and the remaining 5% it just makes stuff up that has nothing at all to do with the original text.
Here's an example of it making stuff up, from Town Building Game web novel.
JP:
常に余裕がある表情を心掛けて、笑顔を絶やさぬように気を張っていたので表情筋が固まってしまった。
がさっ、という物音と視線を感じたので足下に向けると、ディスティニーが俺を真似して自分の頬をモミモミしている。
……くっ、ちょっとかわいいと思ってしまった。
DeepL:
I was always trying to keep a relaxed expression on my face, and I was so careful to keep smiling that my facial muscles froze.
This is a great way to make sure you are getting the most out of your time with your family and friends.
I thought it was a little cute. ......
Me:
I'd been trying so hard to smile like I had everything under control that the muscles on my face had hardened.
I felt something at my feet, and when I looked down, Destiny was imitating me, giving himself a cheek massage.
I thought that was a little cute.I think my verdict is that, as with other MTL, use DeepL with caution and as a last resort. You'll need to know the source language anyway so you can fill in the gaps when the MTL makes mistakes, and if the MTL makes only small mistakes, you won't notice.
-
@rsog412 Odd, since this is what I get:
-
@myskaros said in How Well Does MTL Work?:
@rsog412 Odd, since this is what I get:
Interesting! Mine happened when I had it translate the entire chapter, so maybe it does better when the sentence is in isolation?
-
@rsog412 My observation has indeed been that when using DeepL for anything but getting an overview of a text, it's better to feed it a single sentence at a time, and use its editor to manually fit the translation to the context, rather than relying on its own ability to infer context.
Additionally, it still sometimes produces translations that are unrelated to the actual input (normally some kind of boilerplate business speak or alternatively "I can't make sense of this"), and when that happens, you can often coax it into producing a proper translation by messing with things like punctuation, quote marks and subclauses (e.g. translating subclauses and the context they appear in separately, though that requires some knowledge of Japanese).
-
@rsog412 said in How Well Does MTL Work?:
My impression is that DeepL gives a adequate translation about 85% of the time, makes bad inferences 10% of the time, and the remaining 5% it just makes stuff up that has nothing at all to do with the original text.
Here's an example of it making stuff up, from Town Building Game web novel.
JP:
常に余裕がある表情を心掛けて、笑顔を絶やさぬように気を張っていたので表情筋が固まってしまった。
がさっ、という物音と視線を感じたので足下に向けると、ディスティニーが俺を真似して自分の頬をモミモミしている。
……くっ、ちょっとかわいいと思ってしまった。
DeepL:
I was always trying to keep a relaxed expression on my face, and I was so careful to keep smiling that my facial muscles froze.
This is a great way to make sure you are getting the most out of your time with your family and friends.
I thought it was a little cute. ......
It's those Japanese punctuation marks affecting it sadly (they really need to fix it; like are they even aware it does this?).
Replacing the JP punctuation with English marks:
常に余裕がある表情を心掛けて, 笑顔を絶やさぬように気を張っていたので表情筋が固まってしまった.
がさっ, という物音と視線を感じたので足下に向けると, ディスティニーが俺を真似して自分の頬をモミモミしている.
くっ, ちょっとかわいいと思ってしまった.
I got:
I was always trying to keep a relaxed expression on my face, and I was so careful to keep a smile on my face that my facial muscles froze.
I heard a noise and looked down to see Destiny mimicking me and rubbing her cheeks.
Damn, I thought it was a little cute.
The pc version even has a way to see alternate translation options for single words by clicking in them now, but I'm on mobile.
-
@terrence I copied those sentences with the reworked punctuation into DeepL, and got a completely different, and much worse result:
I've been trying to keep my face relaxed and smiling, and my facial muscles have hardened.
I heard a noise and looked down to see Destiny mimicking me by rubbing her cheek.
Damn, I thought it was a bit cute.Then I realised that I had it set to English (UK), since I normally try to conform to British spelling when I write. But apparently, on DeepL, choosing English (UK) as the target can give considerably worse results than choosing English (US).