Sintetizator govora - sinteza - Text To Speech (TTS) - Tekst u govor - softver, program za slepe

Sinteza - sintetizator govora veštački proizvodi ljudski govor. Sintetizator govora pretvara pisani tekst u govor (Text To Speech - TTS) i može biti realizovan softverski i hardverski.

Sintezovan govor se stvara spajanjem delova snimljenog govora. Sintetizatori se razlikuju po veličini snimljenih delova. Najčešće se koriste sistemi koji proizvode govor iz glasova i dvoglasova. Ovakvi sintetizatori su brzi, zahtevaju manje računarske resurse, ali im je jasnoća i prirodnost slabija. Za specifične upotrebe koriste se cele snimljene reči, pa čak i rečenice. U sintetizator se mogu dodati i neke specifičnosti ljudskog glasa i govora kao što su: boja i visina glasa, ženski-muški-dečiji glas, brzina govora, raspoloženje i drugo.

Sintetizator govora omogućava osobama sa oštećenim vidom (slepima i slabovidima) i osobama sa teškoćama u čitanju da slušaju napisane tekstove.Takođe, sintetizator govora omogućava slepima i slabovidima da koriste računar tako što izgovara ono što čitača ekrana (screen readera) "pročita" na ekranu i pretvori u tekst.

izvor wikipedia.org

 

Saznajte više:

- Istorija sintetizatora govora
- Text-to-Speech Software: What It Is and How It Works

Istorija sintetizatora govora (Text To Speech (TTS) - Tekst u govor)

Mehanički sintetizatori

Mnogo pre nastanka računara bilo je pokušaja da se naprave mašine koje će proizvoditi ljudski govor. Rane primerke "glava koje govore" napravili su:
- Gerbert od Aurilaka
- Gerbert of Aurillac (Pope Sylvester II (946 – 1003)),
- Albert Veliki
- Albertus Magnus (1198–1280)
- Rodžer Bejkon
- Roger Bacon (1214–1294).

1779. danski naučnik Kristijan Kratenstajn, radeći za Rusku akademiju nauka, napravio je modele ljudskog vokalnog trakta koji je mogao da proizvede pet dugih vokala ([aː], [eː], [iː], [oː] i [uː]).

Nakon toga napravljena je "akustičko-mehanička govorna mašina" koja je radila na principu mehova, koju je napravio Volfgang von Kempelen iz Beča. Opisao ju je u svom radu iz 1791. godine. Ovoj mašini dodati su modeli jezika i usana, omogućivši joj da proizvodi kako konsonante tako i vokale.

1837. Čarls Vitstoun napravio je "govornu mašinu" zasnovanu na von Kempelenovom dizajnu, a 1857. M. Faber je napravio "Eufoniju". Vitstounov dizajn je ponovo uveden 1923. od strane Padžeta.

30-tih godina 20. veka, u Bel Laboratorijama napravljen je VOKODER, elektronski analizator i sintetizator govora kojim se upravljalo preko tastature i koji je bio dosta razumljiv. Homer Dadli je usavršio ovaj uređaj u VODER, i izložio ga na Njujorškom Svetskom Sajmu 1939.

Reprodukciju ovog modela su napravili dr Frenklin S. Kuper i njegovi saradnici u Haskins Laboratorijama kasnih 1940-ih, a završili su ga 1950. godine. Bilo je nekoliko raznih verzija ovog hardverskog uređaja ali samo jedan je opstao i do danas. Ova mašina konvertuje slike akustičkih uzoraka govora u formi spektograma u zvuk. Koristeći ovaj uređaj Alvin Liberman i njegove kolege uspeli su da otkriju akustičke signale za percepciju fonetskih segmenata (konsonanata i vokala).

Rani elektronski sintetizatori govora su zvučali poput robota i često bili jedva razumljivi. Vremenom se kvalitet sintetizovanog govora toliko poboljšao da je savremene sisteme za sintezu govora ponekad teško razlikovati od pravog ljudskog govora.

Kako bude opadao odnos cena/performanse, odnosno kako sintetizatori govora budu postajali kvalitetniji, a jeftiniji, tako će veći broj osoba, posebno slepih imati koristi od njih.

Elektronski uređaji

Prva sinteza govora zasnovanu na radu računara napravljena je kasnih 1950-ih godina, a prvi kompletan sistem tekst-u-govor je završen 1968. godine. 1961. godine, fizičar Džon Lari Keli Junior i njegov kolega Luis Grestman koristili su računar IBM 704 za sintezu govora, što je bio jedan od značajnijih događaja u istoriji Bel Laboratorija. Kelijev sintisajzer snimanja glasa (vokoder) reprodukovao je pesmu "Daisy Bell" uz muzičku pratnju Maksa Metjuza. Slučajno, Artur Klark bio je u poseti kod svog prijatelja i kolege Džona Pirsa u Murej Hil ustanovi pri Bell Laboratorijama. Klark je bio toliko impresioniran demonstracijom da ju je iskoristio u uzbudljivoj sceni svog scenarija za svoju novelu 2001:Odiseja u svemiru, gde računar HAL 9000 peva istu pesmu dok ga astronaut Dejv Bovman uspavljuje.

Uprkos uspehu potpune elektronske sinteze govora i dalje se vrše istraživanja o mehaničkim sintisajzerima govora za upotrebu za čovekolike robote.

Prvi sintetizatori za ručne elektronske uređaje javljaju se 1970-tih. Jedan od prvih bio je Telesensory Systems Inc (TSI) Govor+ prenosni kalkulator za slepe proizveden 1976. Drugi uređaji su proizvedeni prvenstveno za obrazovne svrhe, kao Speak & Spell,, proizveden u Texas Instruments-u 1978. Prva multi-player igra pomoću sinteza glasa bila je Milton iz Milton Bradley Company, koja je proizvela uređaj 1980.

Text-to-Speech Software: What It Is and How It Works

Mother and daughter sharing headphones and looking at a phone screen

At a Glance

  • Text-to-speech software is an assistive technology tool.
  • Text-to-speech can be used on many different devices and in many ways.
  • It can help kids with reading issues learn more effectively.

Text-to-speech (TTS) software is an assistive technology tool that can be used to help kids who have trouble reading. The Individuals with Disabilities Education Act (IDEA) says that eligible kids with learning and attention issues should be able to access the same learning materials as other students. Using software that turns written material into audio is a way to make that happen for kids with reading issues like dyslexia.

About Text-to-Speech Software

Text-to-speech technology allows kids to see text and hear it read aloud at the same time. As the text appears on the screen, it’s spoken. Some software uses a computer-generated voice. Others use a recorded human voice. There are different types of text-to-speech software.

  • Screen readers: iPads and iPhones have text-to-speech features, as do many other smartphones. The software reads text files and even the names of programs or folders that your child points to on the screen. This is also helpful for kids who have issues with vision. Learn about how you can use built-in features to turn cell phones and laptops into assistive technology tools.
  • Text readers: This software turns written text into speech. It can read certain types of files and web pages aloud. Text readers also highlight words as they’re read, so your child can follow along. This combination of seeing and hearing to learn is called multimodal reading.
  • POET: This tool, created by the Benetech DIAGRAM Center, creates image descriptions for DAISY (Digital Accessible Information System) books. Once a user uploads and creates a description, it’s available for the next reader who downloads the DAISY file.

Using Text-to-Speech Technology

Text-to-speech tools can be used with different kinds of technology, from computers to mobile devices. It’s often used with optical character recognition (OCR). OCR is a technology that scans printed material into a computer or handheld unit and converts it to digital text. There are even portable OCR devices called reading pens that can scan and read back text.

Some devices, such as the iPad, feature apps that read digital books. Your child can choose whether to use the read-aloud function—and when. Other controls:

  • Font size and color
  • Brightness and contrast
  • Type of voice (male, female or computerized)
  • Rate of speech

Leap Frog books may be a good alternative if you don’t have access to an iPad or iPhone.

How Text-to-Speech Can Help Kids With Reading Issues

Print materials in the classroom—books, handouts and tests—can make learning harder for kids who have trouble reading. Using digital text with text-to-speech functionality can make the process easier. (Some websites have this feature, too.) Researchers have found that the combination of seeing and hearing text when reading:

  • Improves word recognition
  • Increases the ability to pay attention and remember information
  • Allows kids to focus on comprehension instead of sounding out words
  • Increases kids’ staying power for reading assignments
  • Helps kids recognize and fix errors in their own writing

Talking to Your Child’s School About Text-to-Speech

It’s a good idea to start the conversation with your child’s teacher if you think your child would benefit from text-to-speech software. If your child has an Individualized Education Program (IEP) or 504 plan, ask the team to discuss the benefits of assistive technology. You can also find low- or no-cost audiobooks to try at home.

Key Takeaways

  • Text-to-speech software provides a reading experience that combines seeing with hearing.
  • Many e-readers and digital devices offer text-to-speech apps and options.
  • Talk to the IEP team if you think your child could benefit from TTS as an assistive technology tool.

About the Author

Amanda Morin

Amanda Morin

A parent advocate and former teacher, Amanda Morin is the proud mom of kids with learning and attention issues and the author of The Everything Parent’s Guide to Special Education.

More by this author

Reviewed by Whitney Hollins May 05, 2014

Izvor: Understood

Hvala ti Rebecca Muller iz EducatorLabs za poslati izvor.

Vidi još: čitač ekrana (screen reader), Brajev red - displej