Voice Message Text-to-Speech Guidelines

This topic covers general and specific guidelines for using the text-to-speech technology when sending a voice message. This technology is employed by your system when you enter text in the Voice Message Body field when composing a message.

Carefully consider the information and level of detail to include when composing a message to be delivered as a phone call. The system employs text-to-speech technology that does not include voice inflection and can run your thoughts together, making it harder for recipients to understand the message. Using these guidelines, you can add "natural" pauses and take other steps to improve how your messages are read.

General Guidelines

You can control the speed at which the message is voiced by inserting the code \!rdxx in the message, where “xx” is the percentage of the standard speed. For example, inserting \!rd75 at the beginning of a message will cause it to be read at 75% of the standard speed.

You can also add pauses to your message. You can add a pause by inserting the command \!Pxxx, where “xxx” is the length of the pause in milliseconds. For example, inserting \!P500 before a phrase provides a 500 millisecond (.5 second) pause prior to that phrase. This type of command can be inserted anywhere within the message where a pause is needed.

In the following example, the first sentences will be voiced at 85% of the standard speed and the telephone number at 75%. The example also includes four 500 millisecond pauses:

\!rd85 The responder management system \!P500 has been activated due to a multiple casualty incident. \!P500 Your services may be needed for this incident. \!P500 Please respond with your availability. \!P500 You can also call the Emergency Operations Center at \!rd75 4125552122.

Detailed Guidelines

In most cases, the “General Guidelines” provide the information you need to create understandable voice messages. However, if you need to know specific information or have questions, this section provides details on how the system interprets specific types of text.*

*The detailed guidelines in this section are a subset of those listed in the document SVOX Speech Output SDK: Text Preprocessing Documentation – en-US 6.0.0, © 2001-2011 SVOX AG.

Acronyms

If an acronym is not part of an exception list, sequences of 2 to 3 uppercase letters are spelled out. Sequences that contain 4 or more uppercase letters are pronounced. Examples:

Email Addresses and URLs

Domain names are pronounced properly, being either fully pronounced or spelled out. All punctuation marks included in the string are fully pronounced. Examples:

Phone Numbers

Generally, a phone number is read digit by digit including the country and area codes. Brackets around the area code and the symbol “+” (which introduces a country code) trigger a short pause. Characters used to separate groups of digits, such as a slash, hyphen, or whitespace, are not pronounced; a short pause is generated instead. Eleven digits is the longest sequence that is recognized as a telephone number. Examples:

Dates

Short dates should be separated with dots or slashes, as in 1/1/16. Avoid hyphens as the numbers are likely to be interpreted as a range rather than a date. For months abbreviated to 3 characters, as in Jan or Feb, the full month will be pronounced. For days abbreviated to 3 or 4 characters, as in Mon or Thur, the full day will be pronounced. Examples:

Time

Time expressions are recognized whenever the number in the hour position is not greater than 24, and the numbers within the minutes and the seconds positions are not greater than 59. Seconds are optional. Time units (hours-minutes-seconds) must be separated by a colon, or an “h” when seconds are omitted.

A time expression can also include an indication about the part of the day or the time zone:

Time of the day:

●           AM or am is spelled out

●           PM or pm is spelled out

Time zones:

●           EST is read Eastern Standard Time

●           CDT is read Central Daylight Time

Examples:

Measurements

Common measurement units are supported if they follow a number. Some exceptions apply, such as:

●           In the case of very short units such as those that are typically one letter long.

●           Where the reading might be ambiguous such as A, either the letter A or ampere; V, either letter V or volt.

●           In the case of units that conflict with common acronyms and abbreviations.

Numbers supported are cardinal numbers, floats, signed numbers, and fractions. Examples:

Periods and Line Breaks

To ensure periods are not read as dot, copy and paste the unformatted voice message and then remove any line breaks.

Special Characters

The system recognizes some special characters and emoticons. However, it is best to avoid using these in alert messages.

Numbers: Cardinal

Numbers that are 0 through 9 digits in length are read as full numbers. Those with 9 digits or more are spelled out, or each digit is pronounced in isolation. Examples:

Numbers that include separators, such as a space, apostrophe, comma, or period, are recognized and pronounced correctly. Separators are allowed after each third digit, counting from right to left. Examples:

Numbers: Ranges

Range expressions are recognized when two numbers are connected by a hyphen. The hyphen is read as “to.” Only numbers up to 999 are supported. Examples:

Numbers: Float

The part before the decimal point is read following the rules for cardinal numbers. Numbers to the right of the decimal point are always pronounced digit by digit. The decimal point may be written as a comma or as a period. Examples:

Numbers: Ordinal

Numbers up to 999 followed by an ordinal suffix are pronounced as ordinals. The suffix, such as st, nd, rd, th, must be attached to the number. Examples:

Numbers: Roman Numerals

Roman numerals not followed by a dot are read as cardinal numbers, while those followed by a dot are pronounced as ordinals. Several exceptions apply due to ambiguities and frequent conflicts with acronyms such as I, V, C, D, L, M, X, CD, XL. Examples: