The Zebra Talker is a hardware and software combination that will allow you to synthesize human speech. The system is based around the Votrax SC-01 Speech Synthesizer IC which produces 45 basic speech sounds and links them together to make speech. Approximately 1400 words are able to be created using this device. A phoneme editor and text-to-speech software come with the package.
The hardware portion of the Zebra Talker is basically an I/O port (191) connected to the Votrax speech synthesizer. The interface has it’s own power supply so as to not use any of the computer’s power. A sound output cable leads away to be connected up to your monitor audio input or some other amplifier (the output is low level and requires some type of amplification). A resistor-capacitor combination acts as an clock oscillator for the synthsizer IC. There is no on-off switch, power being supplied by plugging in the cable.
The Z Talker comes with the Votrax speech dictionary, so even if you had no software, you could create speech by just OUTing the proper values to port 191 and then use PRINT IN 191. To make this easier, a phoneme editor program is supplied to let you easily “build” your own speech.
The editor lets you build up your words one sound at a time using the Votrax dictionary or your own guessing. After you have built up a word(s), you can print out the sequence of sounds along with the Votrax symbol, the hex and decimal equivalent, and the 2068 character. For example, if you want it to say “computer”, you would enter the following sequence of sounds as given in the Votrax dictionary:
k uh1 m p y1 iu u1 t er
Notice the different “vowel” sounds. There are 36 in all. A table containing special vowel-phoneme combinations is also given in the dictionary. This is not new; if you open any good sized dictionary. they will usually have phonetic representations at the bottom of the page. This idea is used here to create speech from a set of common sounds!
The second program contains a text to speech routine that allows you to just type in words and the computer will speak them! This is the best part of the package! Using stream #4, all text is sent to the synthesizer. The machine code can be easily added to your own program at location 59000 and is 6000 bytes long. The use is very easy and almost com- pletely transparent to the user. All you have to do is enter your text into your programs with PRINT #4 statements. For it to say “computer talks”, all you would have to do is enter into your program:
10 PRINT #4;"computer talks"
The text to speech program will handle about 908 of all words entered. The program checks for combinations of letters and creates the proper sequence of sounds to produce the spoken word. Due to certain unusual combinations and so many rules in the English language, some words come out pronounced wrong. In the last example, the word “computer” comes out sounding like “compooter”. To correct the pronunciation, you need to “fudge” the spelling. To make “computer” sound right you need to enter “compu ter”.
The sound from the Z Talker is very intelligible and can be changed in pitch by a simple POKE. There are some sounds it has trouble with though. It has trouble with the “p” sound. It sounds more like a quiet “t” in most instances. But some sounds are perfect like when you have it say “six”.
If you really want to change the pitch drastically, you can change capacitor C2 to a higher or lower value and make it range from sounding like Jabba the Hut to Kermit the frog!
I would like to see a nice case for the interface instead of the two pieces of cardboard stuck on each side. Also, I would like to see the power drawn from the computer instead of a separate supply and I wish that they provided a longer audio cable so it would reach the back of my monitor. One last thing, it should have a 2068 wide feedthrough connector so as to add other 2068 interfaces.
Over all, I’m impressed. The construction of the interface, the quality of the programs, even the sound is pretty good. Using the text-to-speech, it’s fun to see how well it pronouces all the words you tell it to and to see what happens when you make words up. And remember, it will sound the same no matter what computer it’s on. Available from: Zebra Systems, Inc., 78-06 Jamaica Avenue, Woodhaven, NY 11421 (718) 296-2385