AT&T Natural Voices Review
Raising my Voice
You can buy it at RegNow for that price.
It can read text from word, PDF txt or html files in desirable voices and speeds. Fully integrates with other applications and operating systems.
Compares poorly against more realistic voices in the market
Oftentimes I don’t feel like reading a text or my eyes are tired from staring at the monitor for prolonged periods. At these times I like the luxury of the text being read back to me. Microsoft has a built in voice, but more often than not it sounds monotonous and unnatural and can lead to flawed interpretation and listening fatigue. The AT&T Natural Voices application gives you access to a powerful Text to Speech engine that allows you to convert any text to audio. If you are looking for something better than the Stephen Hawking monotone of Microsoft’s default text to speech software and don’t mind paying for it, you might want to try premium voices from AT&T Natural Voices™ speech engine. It comes with two US English voices Mike. It synthesizes text-to-speech from most document formats such as Microsoft Word, Microsoft Excel, PDF, text and html. It also reads web pages and email. Mp3 or wav reproduction is good. You can purchase more "Natural" voices and voices speaking in other languages as well.
Recommended system requirements are a 500Mhz Processor with 256 mb RAM and 1GB Free Disk Space.
The upside: Most words are pronounced appropriately and with clarity, with the occasional name or foreign word being humorously mispronounced. Pronunciation can be edited by the user, using custom dictionaries to modify the sound for particular words or phrases. Intonation of voice is automatically modified when posing questions or exclamations. An impressive performance was how the software gave two different (but correct) pronunciations of the same word used in different contexts. I tested the verb and adjective forms of the word ‘live’ with the sentences, “They Live in London” and “They are watching a Live show”. The program aced it. Pitch and modulation of the voices adjusts according to the tone of different sentences in a way that is pleasing to the ear. The reading speed, voice and voice can be altered to suit our hearing and understandability. It also comes with a spell check. This program would be particularly usually for the visually impaired, dyslexic, and for people who have other reading problems owing to its ease of use.
The downer: Given that it is essentially a speech synthesizer, it is bound to be judged by how perfectly it enunciates words. Here’s where I found it to be a bit lacking in comparison with its peers. The voice at 8 KHz can sound a bit flat at times. I found myself preferring the rendition of a Julie from Neospeech or a Rachel or Peter from Acapela. Price wise, Natural Voices falls in the same price bracket as Realspeak, Acapela or Cepstral, but is a lot cheaper than Linguatec's Voice Reader Studio, which of course comes equipped with powerful customization options and more realistic prosody.
It’s not the best text to speech synthesizer at these prices. Free demos are availed for most TTS engines. Listen to them and make your choice. If budget is not a concern, Linguatec is recommended.