Translation efficiency for subtitling Thread poster: José Henrique Lamensdorf
|
|---|
Last week a colleague - who has been translating for subtitling longer than me - called me, she was in a fix. She had an appointment, and a good client of hers needed a video from YouTube translated and spotted for subtitling in two hours. It was a short thing, less than 3 minutes long. I also had an appointment, however I thought it was quite possible within the available time.
We had a chance to briefly chat on m.o., or how I would do it. It came up that some subtitling courses te... See more Last week a colleague - who has been translating for subtitling longer than me - called me, she was in a fix. She had an appointment, and a good client of hers needed a video from YouTube translated and spotted for subtitling in two hours. It was a short thing, less than 3 minutes long. I also had an appointment, however I thought it was quite possible within the available time.
We had a chance to briefly chat on m.o., or how I would do it. It came up that some subtitling courses teach to do the entire job, translating and subtitling at once, in Subtitle Workshop, the market standard, in spite of being freeware.
For the sake of speed, I tried to do it that way, and halted after the first subtitle, feeling that my usual way was faster. I'll never know, however I was finished before 15 minutes had elapsed, the deadline for my appointment. No, there was no script available.
So what's my usual way?
I translate first, using Express Scribe, freeware too to play the video, and the Windows Notepad, both side-by side on the screen to write the subtitles. My standard subtitle is up to 2 lines x 32 chars, so I write:
12345678901234567890123456789012|12345678901234567890123456789012
on Notepad, set it to wrap around text, and adjust its window width to fit just that line above, not one extra char before breaking. This will make word wrap warn me every time a subtitle is too long.
I write my subs using the | ("pipeline") character for the line break within a subtitle, and save it as a TXT file. I've never questioned my reasons, as it works, but I open this TXT file with Media Subtitler, freeware, just to add the zeroed timecodes, and save it to SRT, SSA, or any other Subtitle Workshop will open. Then I do the spotting (and spellchecking too!) in Subtitle Workshop.
Though it may appear messier than doing it all the way in one same program, for me this has been very efficient. I'd like to hear comments from colleagues who might have found even more efficient ways to do it. ▲ Collapse
| | | | | Messy and time-consuming | Jun 23, 2011 |
Very time-consuming and imprecise (how do you calculate reading speed? Number of words/character per minute? Frames before/after shot changes? Min/max duration of subtitles?).
3 minutes of video do not require 2 hours of time to be subtitled. It might take 1 hour, if the script is not availabe and the speech is particularly difficult to understand.
And SubtitleWorkshop is not a professional subtitling tool. I wonder what the subtitling companies I work for would say, if... See more Very time-consuming and imprecise (how do you calculate reading speed? Number of words/character per minute? Frames before/after shot changes? Min/max duration of subtitles?).
3 minutes of video do not require 2 hours of time to be subtitled. It might take 1 hour, if the script is not availabe and the speech is particularly difficult to understand.
And SubtitleWorkshop is not a professional subtitling tool. I wonder what the subtitling companies I work for would say, if I begin submitting files executed with a freeware software.
I know how freeware software work, I've been playing with them quite a lot when I was a student.
There is no way one can produce a high quality subtitled file using one of these things. Industry standards impose a much higher quality.
Just my 2 cents. ▲ Collapse
| | | | Cristina Popescu (X) 
Romania
Local time: 12:55
English to Romanian
+ ...
I think the method described is very complicated and I don't see how the result could observe industry standards regarding reading speed, rhythm and synchronicity. I find it most efficient to do spotting and translation at once, using professional subtitling software. Sometimes the client requires the spotted file first and then I get to use my own spotting as template, but I always find myself splitting subtitles or merging them while translating, so I end up changing a lot of what I did in the... See more I think the method described is very complicated and I don't see how the result could observe industry standards regarding reading speed, rhythm and synchronicity. I find it most efficient to do spotting and translation at once, using professional subtitling software. Sometimes the client requires the spotted file first and then I get to use my own spotting as template, but I always find myself splitting subtitles or merging them while translating, so I end up changing a lot of what I did in the first place and it's not very efficient. ▲ Collapse
| | | | | Some interesting opinions | Jun 24, 2011 |
Ciao, Diana,
Diana Battaglio wrote:
Very time-consuming and imprecise (how do you calculate reading speed? Number of words/character per minute? Frames before/after shot changes? Min/max duration of subtitles?).
Yes, such calculations and concerns struck me as interesting, when I made a sort of "cameo appearance" at a friend's subtitling course. I had never thought about it, however I know exactly why I never bothered. The fact is that I spent almost 18 years translating video for dubbing before dipping my toes into subtitling. To translate video for dubbing, one must catch and follow the rhythm of the video, which became second-nature to me.
You may watch a sample of how my translation for dubbing comes out here. These are clips from a training video, originally in English, here dubbed in Brazilian Portuguese. It's from an old VHS tape, somewhat downgraded to spare bandwidth, yet I like to show it because of the many close-ups, and this specific dubbing studio being always so brilliant in their work.
Diana Battaglio wrote:
3 minutes of video do not require 2 hours of time to be subtitled. It might take 1 hour, if the script is not availabe and the speech is particularly difficult to understand.
But of course! It took me less than 15 minutes to do it all and upload it. I had an appointment a bit later than the friend who outsourced it to me.
Diana Battaglio wrote:
And SubtitleWorkshop is not a professional subtitling tool. I wonder what the subtitling companies I work for would say, if I begin submitting files executed with a freeware software.
I'd have several examples to quote, this is the third one I chose, not to overextend myself. My elder son is an IT pro. I had written all the content for my first web site ever, and given him an idea on what I wanted. It took him eighteen months to find time to build that site. When he finally got to it, he sat on my computer, and wrote all the code directly, using... the Windows Notepad! Everybody was amazed at the results, and it stayed online for several years until I decided that I'd do a whole new one myself thereon, using specific, "professional" software (because Í don't know squat about web programming). It's not as good as the first one, however I'm able to update it on my own whenever I want to.
So there is no "professional" nor "amateur" software; there may be a professional or an amateur using any kind of software.
The author of the excellent AVI.net (freeware) once wrote on his web site a valuable lesson. He challenged any commercial software to render comparable results upon converting VOB to AVI. (Btw, he was dead right.) Yet he commented that freeware usually entails an unfriendly interface, because the developer was more concerned with its funcionality. It also usually does not include any kind of user support other than from peers in a newsgroup. Bottom line is that anyone using freeware must know exactly what they are doing.
Diana Battaglio wrote:
I know how freeware software work, I've been playing with them quite a lot when I was a student.
There is no way one can produce a high quality subtitled file using one of these things. Industry standards impose a much higher quality.
Quite a biased assertion! All the non-proprietary subtitle file formats are pure TEXT, interspersed with some parameters. Try opening SRT, SSA, SUB, TXT, any of them with Windows Notepad, and you'll see. Therefore it would require patience, knowledge, and skill, but it's perfectly possible to use the Notepad to create a subtitles file for, say, Adobe Encore or Final Cut, which you'd probably rate as "professional software". The modest, free, Subtitle Workshop also creates subtitle files for them. It all depends on who's using it.
Industry standards require quality upon burning the subtitles. I tested several programs you'd rate as "professional", and yet the sharpest video with the sharpest burnt-on subtitles came from... VirtualDub! (freeware)
To make you happy, there is no decent freeware for editing video, so I use Sony Vegas. Likewise, there is no decent DVD authoring freeware, so I use Sony DVD Architect. For audio, though Audacity (free) is quite acceptable, I prefer Acoustica, though sometimes I use Sony SoundForge.
As I know what I'm doing, I don't depend on user-friendly interfaces nor customer support. I can focus on delivering higher quality than my clients want. Yet I wonder if any plain user, by means of all those config wizards and presets often found on what you call "professional" software reach the same level.
| | |
|
|
|
kmtext
United Kingdom
Local time: 10:55
English
+ ...
| That method sounds unnecessarily complicated. | Jun 24, 2011 |
Life's too short and the rates are currently too low to spot, transcribe, translate and review separately, so I tend to do everything in one pass and then review. I've been doing it for nearly 20 years though, and I've done live monolingual subtitling and simultaneous interpretation, so, depending on the subject matter, I can just about manage simultaneous translation while subtitling if I have to.
| | | | | SubtitleWorkshop lacks of linearity and technical settings | Jun 25, 2011 |
José Henrique Lamensdorf wrote:
Yes, such calculations and concerns struck me as interesting, when I made a sort of "cameo appearance" at a friend's subtitling course. I had never thought about it, however I know exactly why I never bothered. The fact is that I spent almost 18 years translating video for dubbing before dipping my toes into subtitling. To translate video for dubbing, one must catch and follow the rhythm of the video, which became second-nature to me.
Well, these "calculations and concerns" are not just a matter of rhythm and synchronicity.
They are precise technical parameters which must be respected when producing subtitles for certain companies. Otherwise the reviewer will send all your files back with a few notes of complaint. Reading speed is not related to synchronicity: it is based on the audience's reading ability. Subtitles for children will have a lower reading speed (normally 80/90 words pm) then subtitles for an adult audience (180/190 words per minute) and professional software calculate that during the subtitling process. Sorry to insist, but SubtitleWorkshop doesn't come with all these technical settings.
José Henrique Lamensdorf wrote:
All the non-proprietary subtitle file formats are pure TEXT, interspersed with some parameters. Try opening SRT, SSA, SUB, TXT, any of them with Windows Notepad, and you'll see. Therefore it would require patience, knowledge, and skill, but it's perfectly possible to use the Notepad to create a subtitles file for, say, Adobe Encore or Final Cut, which you'd probably rate as "professional software". The modest, free, Subtitle Workshop also creates subtitle files for them. It all depends on who's using it.
I never said you can't do it.
José Henrique Lamensdorf wrote:
As I know what I'm doing, I don't depend on user-friendly interfaces nor customer support. I can focus on delivering higher quality than my clients want. Yet I wonder if any plain user, by means of all those config wizards and presets often found on what you call "professional" software reach the same level.
I'm certain you deliver higher quality than your clients want.
I'm just saying that SubtitleWorkshop is not suitable for professional DVD or broadcast subtitling, at least not when dealing with European subtitling companies and their strict guidelines involving many technical aspects. Plus, from your first post, it is evident that it doesn't allow you to work as linearily as you would like.
Now, I don't know what you mean by "plain user", as I don't know any "plain user" who works with a professional subtitling software (WinCAPS, Swift, Spot, EZTitles...), but... YES, as far as I've seen, "by means of all those config wizards and presets" often found on what I call "professional" software, subtitlers DO reach the same level.
I'm convinced that there is an enormous difference between software like SubtitleWorkshop and the previously mentioned professional software and I can see it while working.
Then, I wonder why subtitling companies bother about training their in-house subtitlers to work with what "I" call professional software or about hiring freelancers who use these software only.
José Henrique Lamensdorf wrote:
I'd like to hear comments from colleagues who might have found even more efficient ways to do it.
In my opinion this process of transcribing, translating and spotting separately is complicated and not as efficient as doing everything with one software.
I don't know any other more efficient way of doing everything in one go.
| | | | | At last, some learning! | Jun 25, 2011 |
Ciao, Diana,
This is getting good! IMHO we can possibly learn some useful lessons to improve our methods without necessarily changing them.
Diana Battaglio wrote:
Well, these "calculations and concerns" are not just a matter of rhythm and synchronicity.
As I said, I came from dubbing, so I admit that I kept the same tools for translation, just discovered that the major difference lies in using a completely different frame of mind.
Dubbing calls for rebuilding the audio track entirely, however mouth movements visible on the screen are a pervasive constraint that must be respected at all times. Incidentally, I am Brazilian, my country is ranked as having one of the best - if not the best - dubbing industry worldwide. I guess you are Italian, and I don't know if it has improved since, however years ago I heard that Italy was known for having the worst dubbing standards. In fact, once in Rome, I guess in the mid 1980s, I briefly (many better things to do there than watch TV) saw a dubbed episode of Kojak. I recall it made me wonder if European TVs had separate channel selectors for audio and video, since the audio in Italian there made sense, yet it was so out of sync that it seemed to be from some other channel.
I envision the subtitling goal as enabling someone to watch a movie in a strange language with a bilingual friend who is a demon typist giving them the gist, the essence, of what is said there as concisely as possible. The less they have to read to get the whole idea, the better. The constraint here - in lieu of the dubbing lip-sync - is the time it takes an "average" spectator to read that before that sub goes off.
Diana Battaglio wrote:
They are precise technical parameters which must be respected when producing subtitles for certain companies.
I'd challenge this statement, on the grounds that such "precise" parameters would have to be determined by too many program-specific variables.
One is targeted public age, as you aptly mentioned below.
Film genre should have its influence, though somewhat limited, too. Though it sounds biased, it would be arguable that slapstick comedy appeals to more action-oriented people, while those more intellectual movies would be interesting to more book-wormish types. Yet any conclusion on reading speed from these inputs would be loaded with prejudice.
Another one would be language. I remember seeing US films on Israeli TV. Audio was in English, of course, yet since both Hebrew and Arab are extremely compact languages in writing (most vowels are implied, not written), they managed to put TWO one-liner subtitles on the same screen, and these seldom used more than half the screen width.
So would the truly professional subtitling firms have a formula which, based on a specified set of parameters, to determine your "precise technical parameter" for each specific film/language pair? I honestly doubt it.
Diana Battaglio wrote:
Otherwise the reviewer will send all your files back with a few notes of complaint.
Not few, many. After so many years translating corporate video, only a few feature films (all for dubbing), I recently translated a TV series for subtitling, using a prominent studio's own software. It had a quite handy feature to warn me whenever a subtitle was too long for the time alotted. To my surprise, my "unaided" average was about one "too long" subtitle every 7 minutes, pretty good, in spite of the usual "swell" from EN to PT being up to 20%.
Then suddenly their software stopped working on my computer. They couldn't solve the issue, and said it happens more often than they'd like, but it does. So I had to go back to my old way. On the first episode I tackled mixing their output format with my old way, I was so engrossed into making it work that I got careless on metrics. So their system (not their reviewer) sent me back a report on all subs I overextended myself, which I quickly fixed. On the next and later ones, I had no problem with this. It's a matter of paying attention to it.
I requested a post-review subtitles file, and they obliged. From what I saw, my seasoned TV subtitles reveiewer had no reason to complain.
Diana Battaglio wrote:
Reading speed is not related to synchronicity: it is based on the audience's reading ability. Subtitles for children will have a lower reading speed (normally 80/90 words pm) then subtitles for an adult audience (180/190 words per minute) and professional software calculate that during the subtitling process. Sorry to insist, but SubtitleWorkshop doesn't come with all these technical settings.
No doubt about it! Yet films for children will have less 'dense' content as well. I recall having translated (for dubbing) a couple of series of educational documentaries for what I assume was Junior High school. Narration was so slow paced that I gave them a substantial discount on my standard rates, which also 'felt good' as my contribution ro education.
Subtitle Workshop has several automation features. FYI I use none of them; I just use SW to format my subs.
I'll tell you a story. Many, many years ago, in the late 1960s, Ford lanched one of the first front-wheel-drive cars in Brazil. It soon became known as quickly wearing the front tires on the inside. One rather modest mechanic found the solution, and all such cars after having been to his shop would wear the tires evenly. Ford had adopted wrong wheel alignment standards, he calculated the right ones (and advised them).
That mechanic's wheel alignment stand was pretty spartan, while much more advanced units already existed. So I once asked him about it. He told me, As the machine is simple, I can keep it absolutely accurate myself. This and any other such machine will only measure what is there. *I* must decide exactly what is to be done, where, how, and how much, to get the suspension and steering exactly the way it should be.
So automation is only about having machines doing accurately what a human thinks that should be done. If machines do the decision-making, they'll only be replicating - faster, of course - some human reasoning process.
Diana Battaglio wrote:
I'm certain you deliver higher quality than your clients want.
I'm just saying that SubtitleWorkshop is not suitable for professional DVD or broadcast subtitling, at least not when dealing with European subtitling companies and their strict guidelines involving many technical aspects. Plus, from your first post, it is evident that it doesn't allow you to work as linearily as you would like.
This is at the core of what I considered discussing on this thread. Those who wrote here with some (justifiable!) contempt for my methods prefer the all-at-once approach. I take the Henry Ford approach, the production line. First I wear my translator hat, then I wear two hats at once, timespotter+proofreader, and so on.
I know many, really many, subtitle translators who don't do spotting. The reason I know them is that I only translate EN/PT, however I speak (and understand, of course) IT/FR/ES to spot subtitles in any pair among these five languages, and take them all the way to a finalized DVD, if requested. (However I don't translate these last three.)
It's unlikely I'd work for European subtitling companies, as I translate into PT-BR, and EN-US. Nevertheless, Hollywood might have enough work for me.
Diana Battaglio wrote:
Now, I don't know what you mean by "plain user", as I don't know any "plain user" who works with a professional subtitling software (WinCAPS, Swift, Spot, EZTitles...), but... YES, as far as I've seen, "by means of all those config wizards and presets" often found on what I call "professional" software, subtitlers DO reach the same level.
Maybe involuntarily you gave me a BIG answer on why some apparently reputable subtitling studios pay such despicably low rates to translators. Apparently their "professional" software replaces most of the skill/talent required for the job, so cheap labor will cut it. Such software is probably coupled with Trados, so mass-production becomes feasible. All this high-tech entourage fails to prevent frequent translation bloopers from appearing onscreen (mostly on cable TV), yet it's costwise effective.
Diana Battaglio wrote:
I'm convinced that there is an enormous difference between software like SubtitleWorkshop and the previously mentioned professional software and I can see it while working.
Definitely! It serves to make the process failproof, outcomes impervious to human failure. It enables subtitling studios to hire labor instead of craftmanship.
Diana Battaglio wrote:
Then, I wonder why subtitling companies bother about training their in-house subtitlers to work with what "I" call professional software or about hiring freelancers who use these software only.
Now you've spilled the beans! In-house subtitlers are clearly "labor", Chaplin's Modern Times style.
It should have dawned upon me long before I ever entered subtitling. Now and then a prestigious dubbing studio would hire me to translate ONE feature film, or ONE episode of a series. One of them told me the truth... they wanted it to show as a sample of their quality to land a juicy contract with a distributor or network. As I learned, one such contract involved 300 films. After they got it, possibly thanks to my quality in translation (and theirs in direction and dubbing too, of course) they could have the whole package thereon translated by "labor".
Yet corporate clients don't want their training, indoctrination, product launch, or institutional videos translated by "labor", so the subtitling studios get me to translate them. Most often these corporate clients use small video production studios to do the subtitling. To my surprise, all of these just want the subtitles translated in DOC (MS Word) files, which they will reportedly implement one by one, via copy&paste, using FinalCut, very professional software.
So the conclusion is clear, there are professional methods and there is professional software. To use both at the same time would be an overkill.
Diana Battaglio wrote:
In my opinion this process of transcribing, translating and spotting separately is complicated and not as efficient as doing everything with one software.
I don't know any other more efficient way of doing everything in one go.
Thanks. Apart from all this discussion, which brought some interesting insights, you gave me the answer. If I ever need - by market demand - to increase my efficiency, I'm convinced that my present dubbing-inherited method cannot be improved any further to this regard. Its reason to be stems mostly from a script, transcript, or template, seldom - if ever - being provided, as it's the case in corporate video. The way to go to increase efficiency would be to use some "professional" all-at-once subtitling software. Working at significantly lower pay rates, effectiveness will be assigned to the reviewer, as I won't go anywhere twice to improve what I've already done. From what I see on TV, the reviewer may often be a figure of speech, as some translation bloopers appear correctly later in the same show.
Yet it's the trend. As so many subtitle translators complain about low rates, labor efficiency seems to be what the market is looking for.
| | | | | Labour?!? Really? | Jun 25, 2011 |
José Henrique Lamensdorf wrote:
Apparently their "professional" software replaces most of the skill/talent required for the job, so cheap labor will cut it. Such software is probably coupled with Trados, so mass-production becomes feasible. All this high-tech entourage fails to prevent frequent translation bloopers from appearing onscreen (mostly on cable TV), yet it's costwise effective.
José Henrique Lamensdorf wrote:
Definitely! It serves to make the process failproof, outcomes impervious to human failure. It enables subtitling studios to hire labor instead of craftmanship.
José Henrique Lamensdorf wrote:
Now you've spilled the beans! In-house subtitlers are clearly "labor", Chaplin's Modern Times style.
This sounds quite offensive to me. Many colleagues of mine work as in-house subtitlers and they're the most brilliant, talented and meticulous translators I've ever met. And they certainly aren't paid nuts.
Same for many freelancers:
José Henrique Lamensdorf wrote:
Yet it's the trend. As so many subtitle translators complain about low rates, labor efficiency seems to be what the market is looking for.
Not all of them complain about low rates: I establish my rates myself and subtitling companies are free to decline them, if they think I'm too expensive.
They wouldn't get back to these expensive craftsmen, if they preferred labour efficiency.
José Henrique Lamensdorf wrote:
Thanks. Apart from all this discussion, which brought some interesting insights, you gave me the answer.
Happy this has been of help, somehow.
[Edited at 2011-06-25 14:31 GMT]
| | |
|
|
|
| Thank goodness! | Jun 25, 2011 |
José Henrique Lamensdorf wrote:
Now you've spilled the beans! In-house subtitlers are clearly "labor", Chaplin's Modern Times style.
Diana Battaglio wrote:
This sounds quite offensive to me. Many colleagues of mine work as in-house subtitlers and they're the most brilliant, talented and meticulous translators I've ever met. And they certainly aren't paid nuts.
Same for many freelancers:
Of course, everything has its exceptions. Yet 9 out of 10 subtitling studios pay less than half my rates. It took me 8 years in the trade to find the one "exception" outside my home country that I was possibly entitled to.
José Henrique Lamensdorf wrote:
Yet it's the trend. As so many subtitle translators complain about low rates, labor efficiency seems to be what the market is looking for.
Diana Battaglio wrote:
Not all of them complain about low rates: I establish my rates myself and subtitling companies are free to decline them, if they think I'm too expensive.
An in-house subtitler gets a fixed salary & benefits, plus a production goal. If they don't achieve the goal or don't think the compensation is adequate, their employer will find someone else who does.
Diana Battaglio wrote:
They wouldn't get back to these expensive craftsmen, if they preferred labour efficiency.
Corporate video goes for fair-priced craftsmen. As I said, one out of ten subtitling studios does it too (of course, at somewhat lower rates than corporate, but offering additional facilities, such as scripts, their software etc.)
Recent evidence shows that they (corporate) still want the craftsmen. Last week I was hired to redo an entire corporate video series translation, because it had been done by someone possibly using professional software, but nothing else.
If you found more than one out of ten, you are lucky!
| | | | To report site rules violations or get help, contact a site moderator: You can also contact site staff by submitting a support request » Translation efficiency for subtitling | CafeTran Espresso | You've never met a CAT tool this clever!
Translate faster & easier, using a sophisticated CAT tool built by a translator / developer.
Accept jobs from clients who use Trados, MemoQ, Wordfast & major CAT tools.
Download and start using CafeTran Espresso -- for free
Buy now! » |
| | Trados Studio 2022 Freelance | The leading translation software used by over 270,000 translators.
Designed with your feedback in mind, Trados Studio 2022 delivers an unrivalled, powerful desktop
and cloud solution, empowering you to work in the most efficient and cost-effective way.
More info » |
|
| | | | X Sign in to your ProZ.com account... | | | | | |
Login to reply/comment 
Translation news
