HTML question about italics and bold

[jackie_w]

Originally Posted by ElMiko

<snip>If however, the majority—or a significant plurality—of these TTS programs (particularly those used in ebook platforms) already incorporate semantic differentiation... <snip>

I suspect one of the reasons the TTS feature of a typical e-reading app (e.g. on Android) makes no distinction between <i> and <em> is because of the way the text content to be read aloud is extracted from the epub html. It may be unsophisticated, but by far the simplest and fastest way is to take each html file and use regex to delete all the tags, i.e. everything between a '<' and '>'. Then read aloud everything that remains.

You could test the theory by styling some of the text with CSS 'hidden' or 'display:none' and see how the app handles it.

The only time I hear TTS 'expression' in a typical e-reading app is due to punctuation like '?' or '!'. How good/bad/indifferent this is probably depends, to some extent, on which 'voice' you use to do the reading aloud.

The specialist screen reading apps aimed at visually impaired users are probably much better at interpreting emphasis tags. If/when that tech will filter down to everyday e-reading apps is anyone's guess.

[JSWolf]

Originally Posted by Turtle91

You can change how emphasized words are spoken using css assuming the device supports the standards... you can also change how emphasized words are displayed with css.

The defaults of <i>/<em> and <b>/<strong> are the same (italics and bold respectively) but they can be changed. So Wolfe's assertion that there is no difference is a little misleading. There is a semantic difference. <i> is for a visual only presentation (however you want it to look), while <em> is whether the word is supposed to be emphasized (typically when read aloud, but even when you are visually reading the word - I can assign an emphatic voice in my mind while reading...) DNSB gave a good example in #9.

Whether or not anyone has used a device that sufficiently differentiates the semantic coding does not invalidate the difference in coding...

I also choose to code my books with the assumption that meeting accessibility standards, as well as proper use of semantic tags, will just make the books better as/when devices improve their support... I also won't have to go fix the books if a major portion of the world passes laws requiring book producers to follow accessibility standards (like the EU did)....

What is the CSS to change <em> so it's different from <i> when read by TTS but the same visually? Also, what TTS software reads <em> and <i> differently?

[JSWolf]

Originally Posted by DNSB

The ePub reader used by one of my neighbours who is legally blind is a Victor (??) Reader Stream device supplied by the CNIB. It does handle quite a bit of semantic content and I used them to test some of accessible ePubs I created.

For me, it was more or less that it took little effort for me to make my ePubs more accessible and I liked the idea of doing so.

Does Victor do anything different with an ePub2 and an ePub3 (with the accessibility info in the OPF)?

[icallaci]

How many ebooks actually use both <i> and <em> in the same book? Maybe I haven't been paying attention but it seems that either one or the other is used, but not both. So even if TTS software can make a distinction, it doesn't matter, at least currently. I'm curious if anyone has seen both forms used in one ebook.

[Turtle91]

Originally Posted by JSWolf

What is the CSS to change <em> so it's different from <i> when read by TTS but the same visually?

Same visually, different in voice:

i {font-style:italic}
em {
  font-style:italic;
  voice-stress: moderate;
}

But it could be different:

em {
  font-variant:small-caps;
  font-style:normal;
  font-family:serif;
  color:purple;
  text-decoration:underline;
  voice-family: female;
  voice-pitch: medium;
  voice-stress: moderate;
  voice-rate: fast;
  voice-volume: soft;
  pause-after: strong;
  voice-balance: left;
}

Of course these are just examples. You can find more in the CSS Speech Spec (https://www.w3.org/TR/css-speech-1/)

Originally Posted by JSWolf

Also, what TTS software reads <em> and <i> differently?

Read the other responses in this thread, and the responses in every other thread where you ask this question.

[Turtle91]

Originally Posted by icallaci

How many ebooks actually use both <i> and <em> in the same book? Maybe I haven't been paying attention but it seems that either one or the other is used, but not both. So even if TTS software can make a distinction, it doesn't matter, at least currently. I'm curious if anyone has seen both forms used in one ebook.

There are books in the wild that have them....but not many. My guess is that this stems partly from the fact that this is a relatively recent change and the number of books produced since the change is still small compared to books produced up to the time of the change. Mostly it stems from the fact that <i> and <b> were deprecated in favor of <strong> and <em> so publishers either did one or the other. Later, mostly due to accessibility reasons, TPTB un-deprecated <i> and <b> and instead clarified the semantic difference and usage of one over the other. So the relatively low number of books that use both properly can be explained with the saying: "You can't teach an old dog new tricks." and "people don't like to change unless they have to"...

[JSWolf]

Originally Posted by icallaci

How many ebooks actually use both <i> and <em> in the same book? Maybe I haven't been paying attention but it seems that either one or the other is used, but not both. So even if TTS software can make a distinction, it doesn't matter, at least currently. I'm curious if anyone has seen both forms used in one ebook.

I've seen books use <i> or <em> and italic in the CSS.

[JSWolf]

Originally Posted by Turtle91

Same visually, different in voice:

i {font-style:italic}
em {
  font-style:italic;
  voice-stress: moderate;
}

But it could be different:

em {
  font-variant:small-caps;
  font-style:normal;
  font-family:serif;
  color:purple;
  text-decoration:underline;
  voice-family: female;
  voice-pitch: medium;
  voice-stress: moderate;
  voice-rate: fast;
  voice-volume: soft;
  pause-after: strong;
  voice-balance: left;
}

Of course these are just examples. You can find more in the CSS Speech Spec (https://www.w3.org/TR/css-speech-1/)


Read the other responses in this thread, and the responses in every other thread where you ask this question.

But how many programs support ePub3 speech given that it's new? Also, how many do you think would eventually support CSS speech?

[DNSB]

Originally Posted by JSWolf

What is the CSS to change <em> so it's different from <i> when read by TTS but the same visually? Also, what TTS software reads <em> and <i> differently?

No CSS is needed. If the ePub to speech is standards compliant, the <i> and <em> tags are all that is needed. Visually, both default to using italic unless you deliberately modify that behaviour.

As I mentioned the Victor device does not care if it is reading ePub2 or ePub3. If you are reading ePub3, you can use SSML and a pronunciation lexicon to improve the quality. You may want to take a look at the CSS Speech Module chunk of the ePub3 specification mentioned by @Turtle91 or the CSS Speech article in Daisy's knowledge base articles.

[DNSB]

Originally Posted by JSWolf

But how many programs support ePub3 speech given that it's new? Also, how many do you think would eventually support CSS speech?

SSML has been around for ~15 years. Hardly what I would consider new. And I totally confused by what you are referring to as "CSS speech". Are you referring to the CSS from the CSS Speech specification? Or???

[JSWolf]

Originally Posted by DNSB

SSML has been around for ~15 years. Hardly what I would consider new. And I totally confused by what you are referring to as "CSS speech". Are you referring to the CSS from the CSS Speech specification? Or???

The CSS Speech in ePub3.

https://www.w3.org/TR/epub-tts-10

[DiapDealer]

It's not like everybody didn't foresee this thread going down this particular rabbit-hole, right?

Jon, you have about two more posts to get your already extremely well-known opinions on this subject repeated for the umpteenth time. Then I'm closing it down. This is why we can't have nice things.

[ElMiko]

@Turtle91 — in your examples you have speech direction specified in the css, which I didn't even know existed! I guess there is a difference between those of us that do this in our pajamas and those that are actually paid to do it...

Couple of questions:

• Does that css work for other tags (*cough* <i> *cough*), or will it only "work" if it's attached to <em>/<strong>?
• Is that css necessary to get any stress effect at all, or do the TTS programs that apply semantic rendering automatically apply stress to <em> tags, even in the absence of specific css direction?

EDIT:
@DiapDealer — sorry! didn't realize i was opening a can of worms...

[DNSB]

Originally Posted by ElMiko

@Turtle91 — in your examples you have speech direction specified in the css, which I didn't even know existed! I guess there is a difference between those of us that do this in our pajamas and those that are actually paid to do it...

Couple of questions:

Is that css necessary to get any stress effect at all, or do the TTS programs that apply semantic rendering automatically apply stress to <em> tags, even in the absence of specific css direction?

The <em> and <strong> tags should automatically trigger the stress. The modifications in the CSS can be used to modify the default behaviour.

Originally Posted by ElMiko

EDIT:
@DiapDealer — sorry! didn't realize i was opening a can of worms...

Jon seems to believe that if an ePub does not perform optimally with ADE 2, then it should not have been created and he has no fear in repeating his opinions ad infinitum. As in this thread where despite the existence of people whose vision is too poor to allow reading an ePub and so will use a TTS reader, adding semantics and accessibility features does nothing for him personally and so are malum in se.

[DiapDealer]

Originally Posted by ElMiko

@DiapDealer — sorry! didn't realize i was opening a can of worms...

You didn't do a thing wrong.