Semantic markup question!

[JSWolf]

And before <em> was a thing, nobody complained that <i> was not read correctly.

[Quoth]

Originally Posted by JSWolf

And before <em> was a thing, nobody complained that <i> was not read correctly.

They complained the TTS was robotic and lacked emotion. I was surprised by the TTS on the Kindle DXG. It was worst than my DOS TTS system, which could also have markup. Curiously TTS on my Android 6 Huawei (bought cheap NOS last year) using Pocket book on Android is better than the TTS on my Android 10 tablet or Android 11 phone (all using Google Speech engine and Pocketbook). The Huawei phone comes with an alternate speech engine which is maybe slightly better than Google's install.
Maybe Google are downgrading the Android one because they have a new TTS for making audio books for sale from epubs? I've not tried that. I'm sceptical it can compete with a half decent <em>experienced</em> human reader.

TTS has improved since the stupid HTML release that depreciated <i> and <b> and the next release after that admitted you need both <i> and <em> as well as <b> and <strong> because most textual bold and italic in English is just visual. But other languages do need <em> and <strong> and English sometimes need it.
HTML does now have both a lot of historic tags and newer tags that are more to do with web pages than ebooks. Some historic tags were because there was no CSS. Later someone decided then they can be used semantically.

[Quoth]

I gave the Kindle DXG to a partially sighted person and he gave it back.
"No way can I listen to book on that!"

[SomeSteve]

A question from a person who knows very little about such things (i.e., me):

Some books contain big chunks of text in italics that are not meant to be stressed. For instance, a novel might commence with a kind of prelude, all italicized, that could go on for several pages. Or a character's thoughts might be italicized, sometimes for a paragraph or more at a time. (I'm not advocating that anybody go wild with italics this way. I find such things a chore to read, and a temptation to skip.) Occasionally, in such expanses of italicized text, a word will emerge that is not in italics, to indicate that it should be stressed. How should this be coded so that a machine reader will emphasize it without changing the appearance of the text? If you enclose the word to be stressed in <em> tags, it will be italicized and appear indistinguishable from the text that surrounds it. If you enclose it in <strong> tags, it will be displayed in bold, instead of in a normal typeface (and I don't know if a machine reader would emphasize it, anyway). Can it be done?

Here a brief example, where the italics indicate a character's thoughts, and the word "not" should be stressed:

Aha, so the cow did not jump over the moon!

[Hitch]

Originally Posted by JSWolf

And before <em> was a thing, nobody complained that <i> was not read correctly.

That's because by and large, Jon, dear, before "em" existed, semantic markup wasn't really a thing and those listening to books on tape (or whatever media) took what they could get.

As "SomeSteve" mentions, there are crap-tons of books that have large swaths of italicized text, meant for (for example) inner dialogue or telepathic communications or what-have-you. That would have no way to emphasize sokme portion of that inner dialogue or whatever, without em, wouldja?

Hitch

[JSWolf]

But i there any TTS program that can take a block of italics and read it without emphasis but read any non-italics words emphasized?

[Quoth]

Originally Posted by SomeSteve

Some books contain big chunks of text in italics that are not meant to be stressed. For instance, a novel might commence with a kind of prelude, all italicized, that could go on for several pages. Or a character's thoughts might be italicized, sometimes for a paragraph or more at a time. (I'm not advocating that anybody go wild with italics this way. I find such things a chore to read, and a temptation to skip.)

Actually I mentioned that un-emphasised italics are more common. Also poetry, song and telepathy or mind-speech.

Originally Posted by SomeSteve

Occasionally, in such expanses of italicized text, a word will emerge that is not in italics, to indicate that it should be stressed. How should this be coded so that a machine reader will emphasize it without changing the appearance of the text? If you enclose the word to be stressed in <em> tags, it will be italicized and appear indistinguishable from the text that surrounds it. If you enclose it in <strong> tags, it will be displayed in bold, instead of in a normal typeface (and I don't know if a machine reader would emphasize it, anyway). Can it be done?

Yes, you have a CSS class for <em> used in a block of italics. The CSS can visually use normal (my preferred) or bold (some prefer).
<em class="in-italic-block">not</em>
The CSS will over-ride default behaviour of the visual rendering engine, but is usually ignored by a TTS system. I've not tested it, but that's what I'd do. I think that can even be done by a Character Style in LO Writer, so that no editing of the epub made by Calibre from an extra Save As in docx is needed.

[Quoth]

Originally Posted by JSWolf

But i there any TTS program that can take a block of italics and read it without emphasis but read any non-italics words emphasized?

Many seem to ignore italics altogether.

That's a separate issue. What is really bad is Random House of Penguins using <em> and <strong> for all italics and bold.

[JSWolf]

Originally Posted by Quoth

Many seem to ignore italics altogether.

That's a separate issue. What is really bad is Random House of Penguins using <em> and <strong> for all italics and bold.

I've not see any eBooks that use <i> and <em>. It's one or the other. Same for <b> and <strong>.

[Turtle91]

SomeSteve:

One way to do that would be to use an <i> to indicate the visual italics and use <em> to indicate an emphasized word or phrase.

CSS:
  i    {font-style:italic}
  i em {font-style: normal}

HTML:
<p><i>Aha, so the cow did <em>not</em> jump over the moon!</i></p>

If you have “large swaths” (my cutoff is anything more than 1 paragraph) then I will use a <div> to the default styling using the css. This eliminates the need for all the extra html tags - reducing code-bloat.

CSS:
div.telepathy {font-style: italic}
div.telepathy em {
          background-color: red; 
          color: green; 
          text-decoration: underline; 
          font-variant: small-caps;
          font-style: normal;
          font-weight: bold; 
          TTS-voice: stress-the-heck-out-of-it;
  **or whatever you want to style your emphasized section**
}


HTML:
<div class="telepathy">
  <p>Aha, so the cow did <em>not</em> jump over the moon!</p>
  <p>It <em>must</em> have been the spoon!</p>
  <p>Because <em>most</em> people realize that emphasis is
     <em>NOT</em> just a visual slanting of the letters!!</p>
</div>

[Hitch]

Originally Posted by Turtle91

SomeSteve:

One way to do that would be to use an <i> to indicate the visual italics and use <em> to indicate an emphasized word or phrase.

(snipping code for brevity)

I'd also add that it's commonplace--and we do it all the time at my shop--to de-emphasize/de-italicize the "emphasized" portion. it's like quoting a book title in a footnote, for example, in which the footnote text is italicized. You remove the italics from the title--it's nothing more than a visual cue, like every other "rule" in typography and layout.

The entire freaking point is to say to the reader, "HEY READER! Here is something that you should pay attention to." The order of things--italics for the entire paragraph, regular face for the book title, or the other way 'round, doesn't matter.

What matters is bloody CLARITY. That's what is important. Doing something consistently, conforming your uses thereof to all be the same, is what matters.

NOT arguing over this nonsense of whether or not there's some goddamned eBook out there that uses BOTH em and i. Of course they exist--I know that personally.

Hitch

[SomeSteve]

Thanks, especially to Quoth and Turtle91, for the explanations and examples.

[Turtle91]

Originally Posted by Turtle91

TTS-voice: stress-the-heck-out-of-it;

Just for clarification, when I used the voice stress example above I just made it up...it is not the correct format.

It should be something along the lines of:

voice-stress: strong

You can find the current (as of 14 Feb 2023) Draft standard for CSS speech on the W3 website.

Cheers!

[JSWolf]

Originally Posted by Turtle91

Just for clarification, when I used the voice stress example above I just made it up...it is not the correct format.

It should be something along the lines of:

voice-stress: strong

You can find the current (as of 14 Feb 2023) Draft standard for CSS speech on the W3 website.

Cheers!

And what are they going to do to make a right mess of it?

[Turtle91]

How about you read it. Then you can come back with a valid question and/or comment about what is actually happening instead of just being a curmudgeon???