Home See here – how to hear and see

“All shall be well, and all shall be well and all manner of thing shall be well.” – Julian of Norwich

Sea here

Text and audio linked {oo}

I often listen to a speaker say something significant, and it is quite difficult to find it again in the audio file. 

Here are links to files I have collected, where the text and the audio match:

OneDrive

You may need to open a Microsoft account in order to view or download files on OneDrive, but it is free.

The four characters {oo}  – two eyes and two ears,

have been used to mark folders and files where text and audio are both available.

YouTube videos often have transcripts that you can view by opening the video and clicking on the 3 dots.

These transcripts are often inaccurate, but better than nothing. 

If you open the .txt or .docx file in my OneDrive and search (Ctrl-F) you can find the nuggets of wisdom.

You should be able to search for a word or phrase within a .docx file.

Converting text to audio files.

Artificial intelligence is the Sorcerer’s Apprentice that may or may not destroy us, but in the meantime we might as well make use of it.

I have found that Amazon Polly neural voices provide the best quality text-to-speech.  There is a 100,000 character limit for each conversion, and it has to be paid for. Audio files created this way are marked with {pn} for neural voices, {ps} for standard. SSML tags can add to the ease of listening.

It is easy to divide a text into segments of 100,000 characters or less, creating a series of audio files.

Using my Android phone, @Voice Aloud Reader works well, and is my favourite TTS app. 

Google Wavenet voices are also good quality. Audio files created this way are marked with {gw}

Balabolka provides a free alternative that is easier to use, but the sound quality is not so good.

Open Library has a built-in text-to-speech facility, with a choice of voices. You can capture the whole output of a book using Audacity.  Choose Windows WASAPI audio host – guidance. The software is clever enough to leave out page numbers, but does read out bibliographic reference numbers. Here is an example.

Files and folders on my OneDrive containing synthesised speech I have marked with {oo}s 

More on using text-to-speech: