

                                   
                                  
                                                  
                 H V S S  
                                                    
                       HUMAN VOICE SPEECH SYSTEM  
                      

                            Version 1.00 (1)

                    Copyright (C) 2001 Sebastian Mate

                         D O C U M E N T A T I O N
  
 Table of contents
 -----------------
  1. Introduction - Welcome to HVSS!
  2. TERMS AND CONDITIONS
  3. Configuration of HVSS
  4. How to use HVSS
  5. FAQ (Frequently Asked Questions)
  6. Credits/Links
  7. Contact the programmer

 1. Introduction - Welcome to HVSS!
 ----------------------------------
 HVSS is - as it's full name says (Human Voice Speech System) - a program
 that can read a text-file with a (more or less) true, human voice. There
 are several TTS (text-to-speech) programs out there, and some of them are
 really good, some not. This is the first TTS written in QuickBasic that
 produces sound-output you can understand while listening to it ;-)

 TTS is a pretty "untached" piece of software - there are only a few pro-
 grammers who ever have tried this not very easy task: To make a speaking
 computer. Basically, there are two types of TTS-systems:

 Rule-based TTS: The program reads a word and with its algorithms, it tries
 to find out how to pronounce the word. For some languages, you need more
 than 100 rules.
 For example: the current word is "Hello". The program knows that it has
 to pronounce the "e" before the double-"l" fast, that it has to ignore the
 second "l". Instead of saying "H-E-L-L-OH" it says "H-AE-L-OW", which is
 the phonetic style of writing this word.
 Rule-based TTS systems are very hard to program and the source is very
 complex.

 Database-based TTS: The program comes with a phonetic database which should
 contain all words of a language with their phonetic writing. HVSS is a
 database-based TTS-system and is based upon the free, open CMU phonetic
 dictionary (see credits!).

 How HVSS works:

 1. Get a word from the text-file,
 2. convert - if required - numbers to text-style,
 3. get the phonetic symbols for the word (in CMU dictionary),
 4. generate the word as wave-file (add the phonetic sounds together),
 5. and play the word on your soundcard (or output it in a WAV-file).

 Features:
  -> Rule based TTS-system for North American English
  -> Knows about 125,000 English and non-English words
  -> Real-Time DMA sound playback
  -> Fast Number-To-Text alorithm

 This program is a very early release with many bugs. This software has
 not been tested by many people. Please report me any bugs you can find
 or just give me feedback! My (e-mail-) address is below! Thank you!


 2. TERMS AND CONDITIONS
 -----------------------
 Please read the following text. Before you use this software, you have
 to agree to it:

 WARRANTY: THE PROGRAMMER ASSUMES NO RESPONSIBILITY FOR ANY HARM OR COSTS
 THAT COMES FROM USING THE MATERIAL CONTAINED IN THESE FILES AND TO YOU,
 YOUR COMPUTER, OR ANYTHING RELATING TO YOUR EXISTENCE. NO WARRANTY IS
 PROVIDED OR IMPLIED ON THESE FILES! THIS FILE IS PROVIDED AS IS. TRADEMARKS:
 ALL NAMES AND TRADEMARKS BELONG TO THEIR OWNERS. COPYRIGHT: THE COPYRIGHT OF
 THIS SOFTWARE HAS THE PROGRAMMER, HE IS OWNER. YOU MAY COPY AND DISTURB
 THESE FILES FOR FREE. YOU MAY ADD A "FLOPPY"-FEE, IF REQUIRED. IF YOU WANT
 TO USE YOUR PROGRAM WRITTEN WITH THIS SOURCE (IF SO), YOU MUST MAKE SURE,
 THAT YOU DISTURB THE PROGRAMMER'S FILE FOR FREE. CREDITS: IF YOU USE PARTS
 OF THIS SOFTWARE (IF ALLOWED), GIVE THE PROGRAMMER CREDITS. FILES: BINARY
 FILES: YOU ARE NOT ALLOWED TO DISASSEMBLE ANY OF THEM. IF NO SOURCE IS
 INCLUDED, ASK THE PROGRAMMER IF HE GIVES YOU THE SOURCE. SOURCECODE: YOU
 MAY USE THE SOURCE-CODE AS MENTIONED IN THE FILE (THE SOURCE-CODE FILE).
 OTHER COPYRIGHT: IF THE PROGRAMMER INFRIGES UPON ANY OTHER COPYRIGHTS FROM
 ANY OTHER SOFTWARE OR PRODUCT, PLEASE INFORM HIM. USAGE: BY USING THIS
 SOFTWARE OR DOCUMENTATION, YOU AGREE TO THE ABOVE TERMS AND CONDITIONS.

 3. Configuration of HVSS
 ------------------------
 Requirements:
 -> IBM PC, Pentium highly recommend
 -> Harddisk with about 15 MB free disk space
 -> A Sound Blaster compatible sound card

 Please unpack the ZIP-file to a good directory, like C:\HVSS.
 
 TIP: Run SMARTDRV to make this program faster. The program has a very
      high disk-access rate which slows the program down if there is no
      caching program (SmartDrive). Go to DOS and type "SMARTDRV /X"!

 Before you can run this program, run CONFIG.EXE which creates the sound
 files and splits the CMU-dictionary up in several small files (required
 for faster access). This can take several minutes!

 There is no installation required.

 4. How to use HVSS
 ------------------
 Go to the directory where HVSS is installed. Please run Smartdrive (type
 SMARDRV /X). Now you have the choice between two different programs:

 1. Real-time reader: This program directly reads the text in real-time.
    Type "READ FILENAME.TXT". To use this program, you need to have a sound-
    card that supports DOS.

 2. Text-to-wave file converter: This programs does not output the sound on
    a soundcard, it writes the data into the wave-file OUT.WAV. Type
    "TXT2WAV FILENAME.TXT". If the program has finished, you should be able
    to play the wave file with several sound programs.

 VERY IMPORTANT:
 -> You must call the programs from the HVSS base directory!
 -> You should load "SMARTDRV /X", otherwise the program is pretty slow!
 -> You need to run CONFIG.EXE before you can run HVSS.

 5. FAQ (Frequently Asked Questions)
 -----------------------------------
 QUESTION: I do not hear any sound or the sound is very noisy, etc.
 ANSWER:   Your soundcard seems not to support the DMA sound player of
           HVSS. Please try the TXT2WAV program instead and play OUT.WAV
           with a differnt wave player (see below)

 QUESTION: I can't play the OUT.WAV file. My wave player says something
           like "File corrupt", "File damaged" etc.
 ANSWER:   This happens for example it you try to open the wave with the
           wave-recorder of Windows. There seems to be a small bug in the
           source which generates the wave's header.
           You can play the Wave-file successfully with RealPlayer, WinAMP,
           the Media Player of Windows (that old one which comes with WIN95)
           or GIMI (the powerful GUI system for DOS: www.sm-gimi.de).

 QUESTION: The program does not say some of the words in my text!
 ANSWER:   HVSS only knows how to speak 125,000 English words, your word
           seems to be unknown.
 TIP:      Use for example "home page" or "desktop computer" instead of
           "homepage" and "desktopcomputer".

 QUESTION: The program crashes or it has completed the file without any
           result.
 ANSWER:   1. Run CONFIG.EXE.
           2. Run the programs always from HVSS' base directory.

 QUESTION: I have a different question. Can you help me?
 ANSWER:   Yes. Please contact me. See below!


 6. Credits/Links
 ----------------
 This program would not exist without the Carnegie Mellon University (CMU)
 and their powerful spelling dictionary.

 Here's what CMU says about it:
 
 "About the CMU dictionary

 The Carnegie Mellon University Pronouncing Dictionary is a machine-readable
 pronunciation dictionary for North American English that contains over
 125,000 words and their transcriptions.
 
 This format is particularly useful for speech recognition and synthesis,
 as it has mappings from words to their pronunciations in the given phoneme
 set. The current phoneme set contains 39 phonemes, for which the vowels
 may carry lexical stress."
 
 I love to see software developers who make their software free and available
 for everyone - as the CMU does! The CMU dictionary is completely free and
 open for everyone! Thank you!

 The dictionary file contains this text:

 "The Carnegie Mellon Pronouncing Dictionary [cmudict.0.6] is Copyright 1998
 by Carnegie Mellon University. Use of this dictionary, for any research or
 commercial purpose, is completely unrestricted.  If you make use of or
 redistribute this material, we would appreciate acknowlegement of its
 origin."

 Of course, I like to tell everyone its origine: You can find it's homepage
 here:   http://www.speech.cs.cmu.edu/cgi-bin/cmudict/

 Thanks to all the people of http://www.qb45.com, the very best programming
 page. They helped me very much and gave me a huge amount of information.
 I will not write down any names - the list would be too long :-)

 7. Contact the programmer
 -------------------------
 Please send feedback and bug-reports to

                      JMate@t-online.de

 Also, please visit my homepage

                     http://www.sm-gimi.de

 and the official HVSS homepage at:

                 http://www.gimi.f2s.com/hvss

 There, you will find very special BASIC programs which have never
 been written before (direct MIDI, language translation, speech etc.).

 Your Sebastian Mate





































