[REL] Mariken (2000)
- pillowbaker
- Posts: 3019
- Likes: 6194
- Joined: Mon Mar 07, 2022 4:05 am
Re: [REL] Mariken (2000)
Suggestion taken, David. I have often seen Notepad++ recommended, and it seems I have nothing to lose.
So, I have been working, when I can, on the Mariken children's novel by Peter van Gestel. I thought it would be more of a breeze, since it translated so clearly using any translation tool. But comparing every line and considering word choices has been a lot more time consuming than I thought! I am keeping at it, though, and it is coming out better than I hoped.
I have completed the first 3 chapters will full editing. Right now, I am just making text files, but I will take suggestions for other file types, from those interested, since I want to carry over the small illustrations from the original pdf.
Hope it is enjoyable. Chapter 4 is underway.
edit: I've just noticed a few errors in chapter 3. Will be corrected next chapter.
edit: Now on chapter 9!
- now 11!
- now almost done with chapter 12. I'd say about 65% of the way done.
- on chapter 14 now
- On chapter 15
So, I have been working, when I can, on the Mariken children's novel by Peter van Gestel. I thought it would be more of a breeze, since it translated so clearly using any translation tool. But comparing every line and considering word choices has been a lot more time consuming than I thought! I am keeping at it, though, and it is coming out better than I hoped.
I have completed the first 3 chapters will full editing. Right now, I am just making text files, but I will take suggestions for other file types, from those interested, since I want to carry over the small illustrations from the original pdf.
Hope it is enjoyable. Chapter 4 is underway.
edit: I've just noticed a few errors in chapter 3. Will be corrected next chapter.
edit: Now on chapter 9!
- now 11!
- now almost done with chapter 12. I'd say about 65% of the way done.
- on chapter 14 now
- On chapter 15
You do not have the required permissions to view the files attached to this post.
Last edited by pillowbaker on Thu Jul 11, 2024 7:48 am, edited 5 times in total.
Re: [REL] Mariken (2000)
I will wait until the end for the full text. How many chapters is it, again? I must have the Dutch version saved somewhere. I don't know if there is an easy way to edit a PDF that is already set to be NOT edited so that the Dutch text can be replaced with English and the pictures are with it. I suppose the pictures can be copied from the PDF, with a Snip tool as necessary, and then deposited in a new PDF with the new English text. I have never made a PDF, or any other document with text AND pictures. Any experts here? I think you have enough work with the text translation, and you should keep that separate, and folks can refer to the original for the pictures. Unless you like a lot of work and you want to go into editing and publishing!
- pillowbaker
- Posts: 3019
- Likes: 6194
- Joined: Mon Mar 07, 2022 4:05 am
Re: [REL] Mariken (2000)
Work has been coming along well with the book. It has been more difficult in areas than I thought it would, or at least, in order to get it right, it has taken more time and research. I suppose I could run it through AI and get something nearly as good. I am taking notes on any interesting obstacle I come across and how I figured out a decision, and I will post those for any language translation/editing nerds.
Point of this post: some of the passages from the Mariken novel are just so cute!!
I would like to share a couple with those of you who enjoyed Mariken. No spoilers, just cuteness.
Point of this post: some of the passages from the Mariken novel are just so cute!!
Spoiler: |
Spoiler: |
Re: [REL] Mariken (2000)
I thought for certain you would use an AI translation for purposes of comparison.
I see you are at least up to Ch 10. How many chapters do you have left?
I see you are at least up to Ch 10. How many chapters do you have left?
- pillowbaker
- Posts: 3019
- Likes: 6194
- Joined: Mon Mar 07, 2022 4:05 am
Re: [REL] Mariken (2000)
You know, I have not looked at how many total chapters there are! It's going to be a surprise.
The first 7 chapters were pretty short, and get longer from there. I'd imagine, since this is a children's novel of decent length, there will be possibly between 25 and 35.
As far as AI goes, I am using the newer GPT models as the base for the translation (two different "brands"), and using google and deepL for cross referencing and making sure the GPTs don't skip any lines. I guess my statement above could suggest that I am not using AI at all, but I meant that I could very well just churn it through AI and paste out the results in a couple days without editing, and end up with something almost as good.
You would just end up with moments like I had with "Princ od papira", where the older neighbor boasts that he "never bothered with girlfriends, but had lots of girlfriends." Or moments where a dialogue response just doesn't work very well with the dialogue leading up to it. My most recent obstacle resulted from cultural and word differences over kissing. I had to let it brew in the back of my head for a week to eventually come up with a simple (yet slightly inadequate) fix. Can't wait to show you the writeup.
The first 7 chapters were pretty short, and get longer from there. I'd imagine, since this is a children's novel of decent length, there will be possibly between 25 and 35.
As far as AI goes, I am using the newer GPT models as the base for the translation (two different "brands"), and using google and deepL for cross referencing and making sure the GPTs don't skip any lines. I guess my statement above could suggest that I am not using AI at all, but I meant that I could very well just churn it through AI and paste out the results in a couple days without editing, and end up with something almost as good.
You would just end up with moments like I had with "Princ od papira", where the older neighbor boasts that he "never bothered with girlfriends, but had lots of girlfriends." Or moments where a dialogue response just doesn't work very well with the dialogue leading up to it. My most recent obstacle resulted from cultural and word differences over kissing. I had to let it brew in the back of my head for a week to eventually come up with a simple (yet slightly inadequate) fix. Can't wait to show you the writeup.
Re: [REL] Mariken (2000)
That's funny, I always turn to the last page first.
But yeah, they are very useful and can produce quite readable results.
I doubt it. I don't trust unsupervised computers any further than I can throw them.pillowbaker wrote: Wed Jun 12, 2024 7:56 pmI meant that I could very well just churn it through AI and paste out the results in a couple days without editing, and end up with something almost as good.
But yeah, they are very useful and can produce quite readable results.
Re: [REL] Mariken (2000)
Night457 wrote: Wed Jun 12, 2024 9:08 pm That's funny, I always turn to the last page first.![]()
I doubt it. I don't trust unsupervised computers any further than I can throw them.pillowbaker wrote: Wed Jun 12, 2024 7:56 pmI meant that I could very well just churn it through AI and paste out the results in a couple days without editing, and end up with something almost as good.
But yeah, they are very useful and can produce quite readable results.
At this point AI can generate results that at least make all the dialogue comprehensible. Colloquialisms still trip them up though. And frankly those can be hard to correct manually. You might have to research local expressions and slang, either from the time period the movie is set or the one where it came out.
Without knowing context, I'd say the line "never bothered with girlfriends, but had lots of girlfriends" meant to say "never bothered with a wife, but had lots of girlfriends".
Re: [REL] Mariken (2000)
This is true. It is also true that bald-faced lies are "comprehensible". I am interested in accuracy when that is possible.At this point AI can generate results that at least make all the dialogue comprehensible.
AI understanding colloquialisms would be expecting too much. The human research to do that can be interesting but it is tremendously time-consuming! That is for someone really dedicated to a text.
- pillowbaker
- Posts: 3019
- Likes: 6194
- Joined: Mon Mar 07, 2022 4:05 am
Re: [REL] Mariken (2000)
deadman wrote: Sat Jul 13, 2024 8:32 pm Colloquialisms still trip them up though. And frankly those can be hard to correct manually. You might have to research local expressions and slang, either from the time period the movie is set or the one where it came out.
Without knowing context, I'd say the line "never bothered with girlfriends, but had lots of girlfriends" meant to say "never bothered with a wife, but had lots of girlfriends".
Surprisingly, it seems that the GPT AIs are getting pretty skilled at recognizing idioms and colloquialisms. This could possibly only apply to certain languages that it has more interaction with.
As for that quotation about the girlfriends, the translation issue derived from the Russian and Serbian language's feminine form for the work "friend". (as in ami or amie in French, amigo or amiga in Spanish).
The line came from the older neighbor of the mysterious artist, who the main girl and her friend Peggy were searching for. The neighbor had asked the boy if she was his "girlfriend", but he replied no, she's just his friend. The neighbor boasted he never bother with "friends" (feminine form, as in female friends), and only accumulated many girlfriends (in Russian, this was a completely separate word).
As for catching colloquialisms, GPT 4-o managed to catch a good amount of them in my work so far on the Mariken novel. Recent memory:
"Dear brother," said Master Aesculaap, "let her pray in her own way first. Then we will know what kind of meat we have in the barrel [or tub]."
4-o simply changed it to "that way we'll know what we're dealing with [or what we're up against]." I still went and researched the expression in the original Dutch. It's a common expression that's used when sizing up other people (their character, etc). I have been comparing with another AI, to ensure I am creating a well-rounded translation, it however didn't bother to modify the expression, and yet, perhaps it recognized that it wasn't really necessary.
And suggestions for English expressions that mean the same thing but are more colorful, like this one?
On chapter 15 now.
Re: [REL] Mariken (2000)
I love that "meat in the barrel" expression! I think the unadorned but metaphorical GPT 4-o translation is a good choice, although I also would be curious about any colorful English equivalents.