Podcasting

I have finally decided I need to podcast the new version of Time Crystal, the third edition. Podcasts are the medium readers like best. Nobody has time or patience to read these days but anyone can listen to a podcast while they drive to work or do the shopping.

I remind myself of the comments by listeners to the podcasts of the earlier version and realise that I made a number of errors. They hated the way I tried to characterise the voices. They hated the sound effects I added for aliens’ voices. They hated the music. Although they did seem to like my voice.

But I would prefer to use speech synthesis software to generate the new podcasts if possible, since it would save me a huge amount of time recording and even more editing the recording. In addition, by using speech synthesis I can quickly edit and listen again in an interactive editorial mode, something far more difficult if you are recording your own voice. So podcasting using speech synthesis can, I believe, greatly enhances the quality of the text.

Yet I am not sure whether listeners would accept a computer generated podcast. I believe the quality of the TextAloud voices I use is high, but is it high enough to allow listeners to listen for half an hour at a time, week after week?

So, as an experiment, I have produced a sample podcast of Episode 1. The fact that I knew people were going to listen to it seemed to focus my mind on the text. While preparing the podcast by listening to the spoken words, I not only corrected mispronunciations (within TextAloud) but I also heard faults with the text itself and with its rhythm, faults which I had simply not noticed before.

I rephrased many sections, sometimes adding whole paragraphs. For example the idea of directly addressing the Cosmic Monopole instead of (as hitherto) referring to it in the third person, was completely new and came naturally as I listened.

I even realised something fundamental about the scenario described in the Sacred Book which I had never thought of before, something which will almost certainly change the whole course of the story later on.

Once more I am forcibly struck by the necessity of listening to text spoken, and for me listening to a computerised voice is better than listening to myself, for if a computer can make the text sound interesting then it must have intrinsic quality.

Now I need YOUR feedback, gentle reader. Please listen to the podcast and tell me what you think. It lasts 27 minutes, but you don’t have to listen to it all if you don’t have the time. Five minutes will suffice to tell you if you like or loath the voice.

You can find it here and either download or listen on line.

If that link does not work try http://timecrystal.co.uk/audio/TC_01.mp3

So please, dear gentle reader, the next time you need to do something boring such as building a garden shed or waiting for the kettle to boil, please just spend a few minutes listening to this podcast and let me know whether you find the voices are acceptable.

Please write to me know at wyken@timecrystal.co.uk

And of course, as always, I would also love to know what you think about the story.

Finally you can download the TextAloud speech synthesis system yourself if you would like to try it out. It is useful for proof-reading letters or reading web pages, and it’s free. Just click

http://www.nextup.com/download.html

Note that the voices I use cost a bit extra but they are worth it. They are all AT&T voices, and you get two of them included when you buy the optional upgrade. My advice is to try the free version first and see if you find it useful.

Technical Stuff

If you happen to be interested in writing or podcasting then you might want to know a bit about the technical aspects of producing this podcast. One aim of this newsletter is not just to describe the art of writing but to describe the mechanics.

If the response from your feedback is positive then I will need to set about creating a new Podiobooks publication, repeating this process for all 52 episodes of Part 1.

I believe that podcasting is a part of the author’s toolkit, but it takes a huge amount of time. I podcast via Podiobooks which has several benefits:

• Access to larger audiences
• Feedback from listeners
• Automatic feed into iTunes

However there are disadvantages. Evo Terra who runs Podiobooks has stringent standard for the format he requires. Therefore one of my worries when producing this podcast via TextAloud was whether it was able to produce the format which Evo wants, namely an MP3 file with the following properties:

• Bit Rate: 128 kbps
• Rate: 44.1 kHz
• Stereo Mode/Channel: Joint Stereo

Luckily TextAloud 3 is able to produce files with exactly these features, making the production process a dream compared to the nightmare of recording and editing which used to take me days.

Production Method

The method of production is as follows.

I write the text in a single Word document. Modern versions of Word are easily able to handle a 170,000 word file (in sharp contrast to earlier versions).

For the purposes of this trial I embedded TextAloud commands directly into the document and formatted them as Heading 3. I have a Word macro which will hide these when I need to publish the text and I could write a new one to delete all text of this kind if required.

I then listened to TextAloud reading the text. This required adding pronunciations to the dictionary and editing the text itself as outline above.

Having got the text in a good state I used a Word macro to select a complete Episode which I then copied into TextAloud and added header and footer commands which will speak the intro and the outro.

The voices I used were the AT&T Crystal and Mike. I have bought several other voices in the past but I find these are easier to understand than some of the others.

I also ensured that TextAloud would add the MP3 tags which Evo wants as far as possible. I exported the file and used a program called Mp3tag to tweak the tags. Once the procedure is set up it’s relatively quick.