Dedicated read aloud / text-to-speech support

[amirsol81]

As inspired by the discussions in #166 .
I would like to suggest adding a fully implemented Read Aloud (Text-to-Speech) feature to Paperback.
Text-to-speech functionality would provide an audio option for text-based content, greatly benefiting users with visual impairments, dyslexia, or other reading challenges. For this to be effective, the app should support:

• Seamless playback.
• Proper language detection and handling (including respecting language tags).
• Visual highlighting / synchronized text focus.
• Playback controls.

Of course, currently Paperback does not handle language switching, even when screen readers are used for reading.
I believe a dedicated Speech menu might be appropriate. It could offer items such as:

• Play.
• Pause.
• Play from top.
• Rewind.
• Fast-forward.
• Speech profiles (if profile-based settings are implemented).

Additionally, Paperback Options should include a new section for Speech-related configuration. This could include:

• Speech engine / Change engine.
• Select voice.
• Speech rate.
• Pitch.
• Volume.
• Additional pause at sentence end (ms).
• Additional pause at paragraph end (ms).
• End-of-page pause (ms).
• End-of-section pause (ms).

For blind users, this feature has strong value: it would allow them to use high-quality SAPI 5 TTS engines directly in Paperback, without having to change the primary speech engine used by their screen reader. This provides flexibility, variety, and a more pleasant reading experience overall.

[trypsynth]

Hi Amir,
I agree, this is definitely something I want. I also want to make the TTS interface very extendable from a development standpoint, so any TTS engine can be integrated in to Paperback with nothing more than a DLL. That said, upon reading your first draft of this feature, I've got a couple thoughts:

• first, I'd make play/pause be one item that dynamically updates, I think, unless you can suggest a strong use-case for not doing so.
• Secondly, how would rewind/fast-forward work? I.e. what would they move you by? According to the text control, one line?

Thanks a ton for opening this!

[amirsol81]

Hi Amir, I agree, this is definitely something I want. I also want to make the TTS interface very extendable from a development standpoint, so any TTS engine can be integrated in to Paperback with nothing more than a DLL.

Hi Quin. That sounds really interesting and ambitious!

• first, I'd make play/pause be one item that dynamically updates, I think, unless you can suggest a strong use-case for not doing .

I like the idea. A single Play/ Pause works fine.

• Secondly, how would rewind/fast-forward work? I.e. what would they move you by? According to the text control, one line?

Hmm. One line sounds good to me. Of course, it means we're moving by paragraph, not sentence. But as you said elsewhere, defining sentence boundaries is difficult. Even Bookworm's rewind/ fast-forward is inconsistent. Sometimes it's sentence-based, and sometimes it's paragraph-based. So I think that line unit would be a safe one.

Thanks a ton for opening this!

Always a pleasure! And huge thanks for making Paperback such a versatile and feature-rich book reader.

[irrah68]

Many thanks for the most excellent epub reader I have found for a long time. This is already a very good and intuitive software and the journey has just began.

I have some thoughts about Sapi support and how it could be implemented or something. I'm not a programmer, just an ordinary user.

I use a Finnish Sapi 5 speech synthesizer called Mikropuhe. It works well with JAWS and NVDA. I also use it with a locally developed program, which supports sapi synthesizers. This program can be used without a screen reader but can also be used whit it.

The Finnish language is spoken as it is written. Many words are pronounced according to their spelling. Mikropuhe is very good and exact in reading Finnish text. It reads what is on the screen and doesn't care about how words of foreign languages should be pronounced. It says them as they are written. And people like this way, because they can see (hear) how these words should be written. For example I read English in Finnish. :) This is a bit difficult to explain, hopefully you can understand the basic idea. I have also a Braille display, so I can always read the text in Braille.

I hope that if you add a language switching feature, it doesn't mean that some words are spoken through Mikropuhe's voice and some other words are spoken by another voice. I don't know whether Mikropuhe supports these tags, I doubt it.

When reading books by using this read aloud text-to-speech system it would be a good idea that when the text goes ahead, a user can follow this from her/his Braille display. If the cursor could follow the speech, this would be great.

When reading a book, how about something like this?

• Start reading from the current position: Control + Enter
• Interrupt speech: Control

When a user reads text by using a Braille display and then decides to continue whith speech, the reading process should start (or continue) from the location to which the cursor was brought by pressing the touch cursor key of the braille display, for example.

[trypsynth]

Hi @irrah68 thanks for this, I'll definitely take this all into consideration when working on the feature, and will ask users such as yourself to test it.