The best reviews I have found of speech synthesizers (the voices) is by a Finnish guy, Veli-Pekka Tätilä, who has limited sight:
http://www.student.o...f_speech_synths.htmlHe also has a review of TextAloud:
http://www.student.o...ew_of_textaloud.htmlThis may be a proxy server that I have access to because I am also a student. If you have trouble accessing the links, let me know.
When I tried the free version of Natural Reader, I did not like that they had some graphics at the bottom of the page that would obstruct the last few sentences of text when the voice was reading. I did like that they highlighted the sentence being read and also the individual word that was being read.
I also object to their pricing. They have three or four levels of software. They are running a special price now, which makes their software more in line with other programs. The free version, like all free versions, does not include any voice that has to be licensed. All licensed voices have a cost. The Microsoft voices are free. The voices that are licensed are generally better sounding.
Their next version called the Professional ($49.50) can record .wav files only, has 2 voices (voices often cost $25 to $30 each so that is included), and I believe there is a file size limit. The Professional version also includes a Microsoft Office add-in that appears in the menu bar of Office programs. I did not get to see how it works and wonder if the sentence and words are highlighted as they are in their reader. I hope one of you who is using the program can shed some light on this question.
The enterprise version ($99.50) includes 2 more voices, an audio recorder and editor, a way to combine audio files into one larger file, XML support, and a batch file converter. I did not try this version. I only tried the free version. You don't need the audio recorder and editor or converter. Audacity does a wonderful job of all these tasks and more. But then what you are really paying for is the four voices. They push the AT&T Natural Voices but also have the NeoSpeech voices.
When I compared this data with TextAloud ($29.95 with only the free Microsoft voices, you pay for any addtional voices you want), I found that you get just about everything that Natural Reader has in their enterprise package. After checking all the sound files that the Finnish guy, Vellu, made of different synthesizers (voices), I knew I wanted Acapela's Heather voice. The AT&T Audrey voice is also nice. By the way, he had each voice read the same section of literature for his comparisons. Most of the voices that TextAloud offers on their web site will work with any software once you install the voice on your computer. The Acapela voices, however, are only licensed to be used with TextAloud. If you want to use them with other software, you have to make another arrangement with Acapela. I downloaded the free version, then decided to hear it with Heather. When I purchased the Heather voice, I was offered the opportunity to buy TextAloud with the voice with a $5.00 discount. Nice touch!
When I downloaded the Free version of TextAloud, I also downloaded the manual which comes with audio-recorded files that read the whole text. By the way, the free version is the same as the paid version with a time limit. TextAloud saves files as .wav, .mp3, and .wma right out of the box. You select the speak-to-file button to record. Again, if I need a recorder/editor, I would use Audacity, which is an excellent open source program. TextAloud installs itself on the browser menu bar with a speak and stop button, making it very convenient. You highlight what you want read and hit speak. Other controls such as speed, volume, pitch, and selection of voice are set within the TextAloud program where you go to make changes.
When you listen to a reading inside TextAloud, the only thing tracked is the word being spoken (the sentence is not highlighted as in Natural Reader). When Text is spoken from the browser, there is no tracking. I sometimes lose the place, so I miss this. If a lot of distracting action is going on around me, I copy the text, a pop up box asks me if I want to put the text in TextAloud as a continuing or new document. Then I listen inside of TextAloud to take advantage of the tracking. Also, TextAloud has a feature where the reading will begin at the cursor. This is a very handy feature when you are proofreading, as you can stop the reading, make corrections, and put the cursor where you want to pick up and continue listening.
If one of you has been using Natrual Reader or any other reader, I would like to know if they track the words outside of the main software. Also, do other programs highlight the sentence as the reading is going on with the individual word also being highlighted (as Natural Reader does). The Finnish guy Vellu, also suggests that a good program would have easy navigation for moving through a paragraph sentence by sentence and moving through a document paragraph by paragraph. I liked his suggestions and recommended them to NextUp.
The business of picking voices takes some study. Not only do different voices come from different companies, but also they come as 8 or 16 kHz. TextAloud says you should use an 8 kHZ voice if you will be using the voice over the telephone, such as for delivering your answering message. Also, the voices come as SAPI-4 or SAPI-5 compliant. TextAloud handles both SAPI standards. You can set the SAPI to be only 4 or only 5 or both. SAPI-5 voices are supposed to be pitch adjustable, but Heather is not. I am not sure why, but maybe that is part of the licensing arrangement with TextAloud. Since the synthesizer voice is what you will be listening to for long periods of time, you should take some time to study the voices. The TextAloud voice page is
http://www.nextup.co...chEngine/voices.htmlAt the beginning of this entry, I gave the site of Vellu (Finnish guy), but I noticed today that I had to give my proxy server info to access his page. Since I usually already have the proxy info in the browser, I had not realized it was necessary before today.
One thing I definitely do not like about TextAloud is that they offer multiple readers for various content. Crazy!! If it reads, it should read everything: news, weather, stocks, RSS. I should not have to purchase separate programs by content. A person who needs to have the software to speak for them should not have to buy a separate software. Natural Reader also writes that their software can be used by people who need to be able to speak and can't. In this case, to speak over the telephone, it appears that an 8kHz voice needs to be purchased.
Because of what was said in this forum about Experssivo [http://www.expressivo.com/experience/ideas_and_description.html], I looked at it today. It is very impressive. I have not heard the voice Jennifer before, but it was very pleasant. I will have to look at it later so I can spend more time with it. All three of the programs mentioned here have a good rating of 5 cows at TUCOWS. However, TUCOWS does not have the latest version of TextAloud.
I look forward to hearing what some of you have found in using other programs.
Marjorie