LoginSign upSite Map
Creare sito web

Skype us

My status

Twitter Updates

qabiria: @multilingualmag: TOIN redesigns website - The first update has been done on the English language version with oth... http://t.co/gPgt5HiScU
qabiria: How To Network Like A Pro - Business Insider http://t.co/LKbewOWWFs
qabiria: How I made 6 figures as a freelance writer in 2011 [Useful advice for translators too] http://t.co/1dRDQARKlZ

Subscribe

Read the articles with your RSS reader

Add to Google Reader or Homepage

Read the articles in your e-mail

Request a quotation

If you would like us to provide a proposal for any project you may have, please use the contact form.
BC_64px
img_04.jpg
Text-to-Speech (TTS) as a translator quality control tool Print
(4 votes, average 4.50 out of 5)
Written by Marco Cevoli   
Thursday, 13 November 2008 19:00

You are alone and you want your translation read back to you? Text-to-speech systems can help you.

D

uring the years we have spent working as translators, we have noticed that a large number of colleagues are accustomed to reading aloud the final version of their translations. This way, their translation can be checked by listening as well as by reading. This is an effective method to identify possible mistakes. Some professionals, when they have the opportunity, prefer having somebody reading the translation back to them. Listening to somebody else reading the text enables translators to easily spot imperfections and correct them. Unfortunately, however, as we all know, translators are often working alone. They cannot always count on a friend or a family member helping them.
Once more, technology comes to our rescue. Vocal synthesis systems, better known as TTS - text-to-speech, can be a very useful tool.
TTS artificially reproduce the human voice. The conversion is operated through specific software that uses digital voices. Without entering into technical details, we will just indicate that there are several synthesis engines that use voices of varying quality. The sound of Windows XP default voices, for instance, does not sound particularly realistic. Lately, however, technology has leaped forward and high-quality, realistic sounding voices are available. Obviously, they are not free. Some of them, however, are sold at affordable prices (below 50 Euro).

The main problem, however, is not cost. Not every synthesis engine manufacturer provides applications for end users. Moreover, not all of them produce their own digital voices. We should also consider that generic users are not their business target. Their systems are usually intended for people with sight handicaps, for whom hearing a written text is a primary need, not a complementary benefit. Since the market is not driven by the generic users, it is relatively difficult to find and choose the best applications. The best programs, however, can read Word and PDF files, websites and they can also save audio files in WAV or MP3 format. Besides desktop dedicated programs, the web offers a large range number of on-line tools that can convert written texts into down-loadable audio files, or create podcast from websites or RSS feeds.

How can you find your way to the best solution for you and - more importantly - use it as a translation support? It is not a difficult task, provided that you are clear about your objectives and the frequency of your intended use. There are two possible options. If you are going to use TTS on a regular basis, it is worth investing some hundred euros in a dedicated software and in high quality voices. If  you use the system as a translation quality check tool, you will only need the voice in your target language. We were positively impressed by TextAloud for Windows, as well as by the Nuance RealSpeak Solo voice quality. TextAloud can read e-mails, websites and other types of text and save them as MP3, Windows Media files, or even batch files. TextAloud has even an option called Proofreading Hotkey that can open TextAloud from another application like Word and play the seleced text.

Mac users should take a look at GhostReader, that uses Acapela voices and has the typical Apple simple, user-friendly interface. You will also find some programs with fantascientific on-line demos, as in the case of Loquendo's system. Loquendo is an Italian company that has developed voice synthesis and recognition systems for Telecom Italia. Companies as this one, however, do not sell their products to end users. They offer their technology to a reseller or other program developers. This is a pity, as the quality of Loquendo's voice is really unmatched.

At this time there are very few voices available other than in English. The only quality Italian voices we were able to find, besides Loquendo's, are RealSpeak's and Acapela's. The voice quality, however, lacks the realism and intonation of the best available English voices. Besides English voices (mostly American), there are a few Spanish voices (South American), as well as German, Dutch, Polish, Romanian, Swedish, etc.

If  you only need using TTS technology on an occasional basis, you should definitely opt for an on-line solution. There are several you can choose from, but they all present the limitations mentioned before: very few support Italian, and the length of texts they can convert is limited. IMTranslator, provided by Smart Link, can read text up to 1000 characters with an acceptable Italian voice. Smart Link Corporation is associated with Prompt, the well-known automated translation provider. You can find the link on Qabiria's toolbar. Leaving aside the various Chinese course and mobile phone banners, as well as the avator mimicking the spoken text, this is an effective free web-based tool.

If you do not need Italian, we recommend Spoken Text.Once you have registered, you can create audio files from PDF, Word, e-mails, websites, or copy and paste your own text. Shortly after you have selected the source text to convert into an audio file, the program provides a link where you can listen to your text or download it in mp3 format or as iPod ready to be uploaded on the player. At his time, this system only supports English, Spanish, German and French. However, you can contact the author, make a donation and request other voices. Another similar web-based tool is  YakiToMe.

Trackback(0)

TrackBack URI for this entry

Comments (2)

Subscribe to this comment's feed
...
0
I have some experiences with text to speech software programs, I recommend Panopreter Plus to Windows 7 or vista users. The software reads text and converts the text to mp3 and wave audio files with the voice of Microsoft Anna. Best of all, it's easy to use, and useful to ESL teacher and students.
Tom , August 20, 2010
...
Marco Cevoli
A good freeware for TTS is Balabolka. All computer voices installed on your system are available to Balabolka. The on-screen text can be saved as a WAV, MP3, MP4, OGG or WMA file. The program can read the clipboard content, view the text from DOC, EPUB, FB2, HTML, ODT, PDF and RTF files, customize font and background colour, control reading from the system tray or by the global hotkeys.
http://www.cross-plus-a.com/balabolka.htm
Marco Cevoli , December 15, 2010

Write comment

smaller | bigger

busy
 
© 2007-2009 Qabiria Studio SLNE - Credits - Disclaimer - facebook_bwtwitter_bw