I like being in the "zone", where my mind engages with programming projects and I lose track of time.
However, more than once, I've looked up, thinking it was lunchtime, and found it was nearing or past 5pm. I decided I needed occasional, scheduled interruptions that would connect me with the world at large.
This lead me to look into text-to-speech, and the options that were available at first were ESpeak and Festival. At the hour and half-hour, I had a program saying "The time is [TIME IN 12-HOUR AM/PM FORMAT] and the current temperature is [CURRENT LOCAL TEMP IN DEGREES FAHRENHEIT]".
Eventually, I found a voice I liked in ESpeak, and I lived with that for quite some time. This changed recently, when I found that Amazon has an AWS service called Polly, which allows you generate an MP3 with your desired text.
Since I tend to use time and temperature, I made a bash loop through temperatures between -20 and 120°F (figuring that, in Indiana, temperatures beyond that range would be unlikely) and times of [1-12]:[0-59] [am/pm] and used mpg123 to play them. RMS argues that I should not trust services on other people's computers, so it's better to generate these things on my own computer, so I'm considering hard what's the better way to go forward, but, when it comes to quality, you can't argue that Polly's way nicer.
Plus, you don't have to munge the pronunciation with a pseudo-English pidgin in order to get the pronunciation as much with Polly.
After originally posting this, Curtis Smith of ECN brought up his TTS solution, named Cepstral. He has provided me with an example file, which I have added below.
Audio samples (using words from HAL9000 in 2001: A Space Odyssey) below. If you use another tool for Text-to-Speech and want to add a sample, send me an audio file (MP3 preferred, but I can convert) using the following text.
I enjoy working with people. I have a stimulating relationship with Dr. Poole and Dr. Bowman. My mission responsibilities range over the entire operation of the ship, so I am constantly occupied.