Better alignment tools?
Thread poster: Jing Nie
Jing Nie
Jing Nie
China
Local time: 18:28
Member (2011)
English to Chinese
+ ...
Apr 7, 2011

I am making a big TMX using Alignment method. I have tried Winalign, but it is not useful. For example , there are two sentences in the original language, but the two sentences are translated into one sentence, and vice versa. It took me a long time to revise the result from Winalign.

So I searched online for better alignment tools, and find some scholars in Universities have build some prototype tools that their alignment rules are based on the machine translation results rather th
... See more
I am making a big TMX using Alignment method. I have tried Winalign, but it is not useful. For example , there are two sentences in the original language, but the two sentences are translated into one sentence, and vice versa. It took me a long time to revise the result from Winalign.

So I searched online for better alignment tools, and find some scholars in Universities have build some prototype tools that their alignment rules are based on the machine translation results rather than just format of two documents. But I do not find the real software, I only find their papers about it.


I think that google, Microsoft and several other companies have mastered the alignment technology based on machine translation. But they do not provide alignment softwares to public. Maybe the market for the alignment software is too small to notice?
Collapse


 
Soonthon LUPKITARO(Ph.D.)
Soonthon LUPKITARO(Ph.D.)  Identity Verified
Thailand
Local time: 17:28
English to Thai
+ ...
SDL Align Apr 7, 2011

In Trados Suite 2007 I use SDLX>SDL Align. It is much more efficient than Trados 2007 WinAlign e.g. on segment relocating, segment splitting and joining, segment moving to end etc. It also gives quicker processing time even with many segments. It gives you TM out of the fly [no need to convert in complicate processes].

Soonthon Lupkitaro


 
Michael Beijer
Michael Beijer  Identity Verified
United Kingdom
Local time: 10:28
Member (2009)
Dutch to English
+ ...
3 suggestions Apr 7, 2011

I would suggest trying:

1. FarkasAndras' open source aligner (http://sourceforge.net/projects/aligner/)(free)
2. AlignFact
... See more
I would suggest trying:

1. FarkasAndras' open source aligner (http://sourceforge.net/projects/aligner/)(free)
2. AlignFactory light (http://terminotix.com/index.asp?name=AlignFactoryLight&content=item&brand=1&item=11&lang=en)
3. memoQ's built in aligner

Regarding what you said about, "alignment rules based on machine translation results rather than just the format of the two documents", the memoQ aligner can use terms in your glossaries as so-called "anchors" to assist in the alignment process. I'm not sure how this works, but this would seem to be in the direction of using MT results...



Michael
Collapse


 
FarkasAndras
FarkasAndras  Identity Verified
Local time: 11:28
English to Hungarian
+ ...
Autoaligner Apr 7, 2011

Well, I completely agree with your analysis that aligners without a built-in autoaligner are basically useless, and that there are far too few convenient autoaligners available. WinAlign doesn't have an autoaligner, nor does SDL Align.

There are a couple out there, though. The MemoQ package includes one, and Alignfactory and the Abbyy Aligner also have a built-in autoaligner.

These are the paid ones. Free & open source options start with Hunalign, which is the heart and
... See more
Well, I completely agree with your analysis that aligners without a built-in autoaligner are basically useless, and that there are far too few convenient autoaligners available. WinAlign doesn't have an autoaligner, nor does SDL Align.

There are a couple out there, though. The MemoQ package includes one, and Alignfactory and the Abbyy Aligner also have a built-in autoaligner.

These are the paid ones. Free & open source options start with Hunalign, which is the heart and soul of my open source aligner.
As far as I know, LF Alinger is the only fee aligner that autoaligns .doc, .docx and similar formats and generates TMX files out of them. It has a command line interface.


By the way, Microsoft did publish an open source autoaligner, although I'm pretty sure it's not the one they use now. Numerous universities and research groups also published their own autoaligners. These are all "raw alignment engines" i.e. they take usually only take txt files as input and generate txt files as their output. They take quite a bit of skill to use - which is why I picked hunalign, which I thought was the best, and added a more user friendly interface and feature set to it.

Edit: Michael beat me to it... to add some info about "use terms as anchor" et al:
Autoaligners work on the basis of one (or a combination) of a couple of principles. the Gale-Church algorithm is perhaps the oldest and most versatile; it works on the basis that corresponding sentences should roughly correspond in length. The upside is that no input data is needed apart from the texts themselves. Of course the quality of the alignment can be further improved if you supply some extra information, such as a dictionary/glossary, which can help identify corresponding sentences. This is how MemoQ uses the termbases, and LF Aligner does this, too, if you supply a dictionary in your language pair.

[Edited at 2011-04-07 10:04 GMT]
Collapse


 
Jing Nie
Jing Nie
China
Local time: 18:28
Member (2011)
English to Chinese
+ ...
TOPIC STARTER
Amazing Apr 7, 2011

I tried your freeware with two small word file.
I do not use terms at all, but the result is 100% correct.

I will try its more functions. BTW, could you please make it more beautiful? I am not used to MS-DOS interface.


FarkasAndras wrote:



[Edited at 2011-04-07 10:04 GMT]


 
Stanislav Pokorny
Stanislav Pokorny  Identity Verified
Czech Republic
Local time: 11:28
English to Czech
+ ...
Transit NXT aligner Apr 7, 2011

Though a proprietary tool, I have very good experience with Transit NXT's aligner. If you have a termbase (which is mostly my case), it will use it to make the aligning process more accurate. Very efficient indeed.

 
Diana Coada (X)
Diana Coada (X)  Identity Verified
United Kingdom
Local time: 10:28
Portuguese to English
+ ...
@Andras Apr 7, 2011

Could you please give some step-by-step instructions on how to use it? I read the how-to instructions, I chose the texts, but I cannot see any output file and no report shows up.

Am I doing something wrong?

This could be a great and simple tool to use, thanks for letting us know about it!


 
FarkasAndras
FarkasAndras  Identity Verified
Local time: 11:28
English to Hungarian
+ ...
Troubleshooting Apr 7, 2011

Diana Coada wrote:

Could you please give some step-by-step instructions on how to use it? I read the how-to instructions, I chose the texts, but I cannot see any output file and no report shows up.

Am I doing something wrong?


Probably. It's hard to tell without a detailed description of what you did and what messages showed up in the console window, and without seeing the log file. Perhaps you left the console window stuck waiting for your input?
I'd suggest running the aligner on the files in the "sample" folder and following the instructions in the howto file all the way through. If you get to the "The aligner has terminated" message without errors, then the output files will be in the same folder where your input files were.


On making the aligner prettier: I might build a graphical interface at some point to make it more user-friendly, but that's not a trivial task.


 
Diana Coada (X)
Diana Coada (X)  Identity Verified
United Kingdom
Local time: 10:28
Portuguese to English
+ ...
@Andras Apr 7, 2011

As soon as I choose the target file the window simply disappears. That's it. But I will try your suggestion out.

Thanks!


 
FarkasAndras
FarkasAndras  Identity Verified
Local time: 11:28
English to Hungarian
+ ...
window Apr 8, 2011

It's probably just a case of the console window being hidden under other windows. Just Alt-Tab your way back to it. If it's not that, check the log file and try running it without a gui.

 
Jing Nie
Jing Nie
China
Local time: 18:28
Member (2011)
English to Chinese
+ ...
TOPIC STARTER
I just find an online abbyy aligner Apr 9, 2011

http://aligner.abbyyonline.com/en

It supports files up to 1MB, it is free.

You may try. Unfortunately, it do not support my language pair.

[修改时间: 2011-04-09 15:06 GMT]


 
FarkasAndras
FarkasAndras  Identity Verified
Local time: 11:28
English to Hungarian
+ ...
Sample Apr 9, 2011

That's not a working aligner you can use, it just provides an evaluation sample.
Quoting ABBYY: "you can upload a file up to 1 MB in size and get up to 50 aligned text strings"

1MB is tiny already, but the limitation to 50 segment pairs makes the thing useless.


 
Lena Grainger
Lena Grainger  Identity Verified
United Kingdom
Local time: 10:28
English to Russian
+ ...
Exporting? Aug 13, 2013

FarkasAndras wrote:

Well, I completely agree with your analysis that aligners without a built-in autoaligner are basically useless, and that there are far too few convenient autoaligners available. WinAlign doesn't have an autoaligner, nor does SDL Align.

There are a couple out there, though. The MemoQ package includes one, and Alignfactory and the Abbyy Aligner also have a built-in autoaligner.

These are the paid ones. Free & open source options start with Hunalign, which is the heart and soul of my open source aligner.
As far as I know, LF Alinger is the only fee aligner that autoaligns .doc, .docx and similar formats and generates TMX files out of them. It has a command line interface.
[Edited at 2011-04-07 10:04 GMT]


FarkasAndras,
I downloaded your aligner and tried to use it. I did not yet have a chance to explore it very deeply, but the actual segmentation in it is far superior to both PlusTools and Trados's WinAlign.
A very stupid question here: how do I export, or at least print, the result (which looks like a 2-column table, but cannot be copied)?
Perhars it is really, really simple, so I'll delete this message when you reply)


 
Post removed: This post was hidden by a moderator or staff member for the following reason: empty post


To report site rules violations or get help, contact a site moderator:


You can also contact site staff by submitting a support request »

Better alignment tools?







Anycount & Translation Office 3000
Translation Office 3000

Translation Office 3000 is an advanced accounting tool for freelance translators and small agencies. TO3000 easily and seamlessly integrates with the business life of professional freelance translators.

More info »
Wordfast Pro
Translation Memory Software for Any Platform

Exclusive discount for ProZ.com users! Save over 13% when purchasing Wordfast Pro through ProZ.com. Wordfast is the world's #1 provider of platform-independent Translation Memory software. Consistently ranked the most user-friendly and highest value

Buy now! »