RDI's Speech Processing Technologies

Hafss^©

Spoken Language/Tajweed Interactive Self Learning Assistant

The goal of this innovative technology is to afford an effective assistance in the self learning of spoken language. Hafss^© which has proved effective in helping people learn Tajweed of the holy Qur’an is also helpful in learning spoken Arabic and other languages as well. In fact, RDI is cooperating with European partners about adapting it to help in learning the less famous spoken European languages like Sweden.

After its first apparition in GITEX 2004, Hafss^© proved itself in a large project of building Qur’an web encyclopedia for the Kuwaiti ministry of endowments. The challenging feature about Hafss^© is that it can detect and fully diagnose the faults – if any – in the learner’s utterance of given examples. This closes the feedback loop without which the learning process can never be completed.

After GITEX^© 2006 and Cairo Int'l Book Fair 2007, an end-user version of Hafss^© is released for the Arabic market esp. in the Gulf and Egypt with other ones following soon. For that version with all its rich material and integrated interactive auxiliary tools, RDI has been awarded the prize of e-Content Award first prize in the e-Learning track from the Egyptian IT Industry Development Agency "ITIDA" for the year 2006.

Moreover, on Sept. 15th, 2007, RDI's Hafss^© technology for the interactive learning of spoken language applied on the Tajweed of the hloy Qur'an has been declared first in the 2007 World Summit Award contest for best e-content/e-Learning track where about 850 other candidates from 160 countries had participated.
 

Where to buy Hafss^©? (JPG, 468 KB)

Buy Hafss^© via the Internet.

Watch AhKam Al-Tajweed episodes taught with the assistance of Hafss(c) by Mr. Ahmed Ragheb

For more detailed info...

Essay on Hafss in PC magazine, Mar. 2007.

(Arabic PDF, 522 KB)

Essay on Hafss in Lughat Al-Assr.

(Arabic WinRAR'ed JPG's, 216 KB)

ArabTalk^©

Arabic Text-to-Speech (TTS)

This Arabic TTS from RDI has a hybrid concatenative/parametric speech synthesizer which affords a compact system within few tens of mega bytes of space. Moreover, RDI's automatic large-scale Arabic phonetic transcriptor (diacritizer) is the other distinctive corner stone of ArabTalk^©.

Along with several diacritization options, many speech effects in both male and female voices are available for flexibility and high performance in real-time.

SDK's for desktop applications, web services, as well as client-server applications are available for developers. Moreover, male/female Arabic TTS speaker data bases can be built for certain real-speakers chosen by the customer.

For more detailed info..
 

Arabic Flash movie on Arabic Text-to-Speech.

(Arabic flash, 3,057KB)

Demo of RDI's Arabic Talking Book.

(WinZipped Installable EXE, 110 MB)

Paper on ArabTalk^©.
(PDF, 96 KB)

Essay on RDI's Arabic Phonetic Transcriptor for Arabic TTS.
(Arabic PDF, 456 KB)

Speech Compression^©

RDI has been known for long of its pioneering CELP vocoding speech compression system that has been awarded the Best-of-GITEX 1996 prize.

In fact, this system that could achieve compression ratios within 15-to-25 at a good quality, has remained for years one of RDI's corner stones for producing its encyclopedic Arabic multimedia that dominated the market in the late 1990's.

Today, RDI is working on a novel technology for very low bit rate speech coding via a speech re-synthesis process for mobile devices and telecom. applications.