introduced "relative-change" for pitch alterations, using Hz and semitones.

Author: danielweck

css3-speech/Overview.html

@@ -1749,8 +1749,8 @@ <h2 id=voice-char-props><span class=secno>10. </span>Voice characteristic
     <tr>
      <td width="20%"><em>Value:</em>
 
-     <td>&lt;number&gt; | &lt;percentage&gt; | x-low | low | medium | high |
-      x-high | inherit
+     <td>&lt;number&gt; | &lt;percentage&gt; | &lt;relative-change&gt; |
+      x-low | low | medium | high | x-high | inherit
 
     <tr>
      <td width="20%"><em>Initial:</em>
@@ -1794,7 +1794,21 @@ <h2 id=voice-char-props><span class=secno>10. </span>Voice characteristic
 
    <dt><strong>&lt;percentage&gt;</strong>
 
-   <dd>Specifies a relative change to the inherited value.
+   <dd>Specifies a multiplicative change to the inherited value. Thus 150%
+    means one and a half times the inherited average pitch for this voice.
+
+   <dt><strong>&lt;relative-change&gt;</strong>
+
+   <dd>Specifies a relative change (decrement or increment) to the inherited
+    value. The syntax of allowed values is a &lt;number&gt; preceded by "-"
+    or "+", and followed by either of "Hz" (for Hertz) or "st" (for
+    semitones). A semitone is half of a tone (a half step) on the standard
+    diatonic scale. Note that a semitone doesn't correspond to a fixed value
+    in Hertz: instead, the ratio between two consecutive frequencies
+    separated by exactly one semitone is approximately 1.05946 (the actual
+    arithmetics involved are beyond the scope of this specification, please
+    refer to existing literature on that subject). Examples of valid values:
+    "+0.5st", "+5st", "-2st", "+10Hz", "-5.5Hz".
 
    <dt><strong>x-low</strong>, <strong>low</strong>, <strong>medium</strong>,
     <strong>high</strong>, <strong>x-high</strong>
@@ -1803,10 +1817,6 @@ <h2 id=voice-char-props><span class=secno>10. </span>Voice characteristic
     implementation and voice specific.
   </dl>
 
-  <p class=issue>ISSUE: should we also allow for relative changes in terms of
-   semitones as permitted by SSML? (this would require adapting both "number"
-   and "percentage" to disambiguate uses of "st" and "Hz" values)
-
   <table class=propdef summary="name: syntax">
    <tbody>
     <tr>
@@ -1817,8 +1827,8 @@ <h2 id=voice-char-props><span class=secno>10. </span>Voice characteristic
     <tr>
      <td width="20%"><em>Value:</em>
 
-     <td>&lt;number&gt; | &lt;percentage&gt; | x-low | low | medium | high |
-      x-high | inherit
+     <td>&lt;number&gt; | &lt;percentage&gt; | &lt;relative-change&gt; |
+      x-low | low | medium | high | x-high | inherit
 
     <tr>
      <td width="20%"><em>Initial:</em>
@@ -1866,7 +1876,21 @@ <h2 id=voice-char-props><span class=secno>10. </span>Voice characteristic
 
    <dt><strong>&lt;percentage&gt;</strong>
 
-   <dd>Specifies a relative change to the inherited value.
+   <dd>Specifies a multiplicative change to the inherited value. Thus 25%
+    means a quarter of the inherited pitch range for this voice.
+
+   <dt><strong>&lt;relative-change&gt;</strong>
+
+   <dd>Specifies a relative change (decrement or increment) to the inherited
+    value. The syntax of allowed values is a &lt;number&gt; preceded by "-"
+    or "+", and followed by either of "Hz" (for Hertz) or "st" (for
+    semitones). A semitone is half of a tone (a half step) on the standard
+    diatonic scale. Note that a semitone doesn't correspond to a fixed value
+    in Hertz: instead, the ratio between two consecutive frequencies
+    separated by exactly one semitone is approximately 1.05946 (the actual
+    arithmetics involved are beyond the scope of this specification, please
+    refer to existing literature on that subject). Examples of valid values:
+    "+0.5st", "+5st", "-2st", "+10Hz", "-5.5Hz".
 
    <dt><strong>x-low</strong>, <strong>low</strong>, <strong>medium</strong>,
     <strong>high</strong> and <strong>x-high</strong>
@@ -1875,10 +1899,6 @@ <h2 id=voice-char-props><span class=secno>10. </span>Voice characteristic
     implementation and language dependent.
   </dl>
 
-  <p class=issue>ISSUE: should we also allow for relative changes in terms of
-   semitones as permitted by SSML? (this would require adapting both "number"
-   and "percentage" to disambiguate uses of "st" and "Hz" values)
-
   <table class=propdef summary="name: syntax">
    <tbody>
     <tr>
@@ -2581,8 +2601,8 @@ <h2 class=no-num id=property-index>Appendix B &mdash; Property index</h2>
     <tr valign=baseline>
      <td><a class=property href="#voice-pitch">voice-pitch</a>
 
-     <td>&lt;number&gt; | &lt;percentage&gt; | x-low | low | medium | high |
-      x-high | inherit
+     <td>&lt;number&gt; | &lt;percentage&gt; | &lt;relative-change&gt; |
+      x-low | low | medium | high | x-high | inherit
 
      <td>medium
 
@@ -2597,8 +2617,8 @@ <h2 class=no-num id=property-index>Appendix B &mdash; Property index</h2>
     <tr valign=baseline>
      <td><a class=property href="#voice-pitch-range">voice-pitch-range</a>
 
-     <td>&lt;number&gt; | &lt;percentage&gt; | x-low | low | medium | high |
-      x-high | inherit
+     <td>&lt;number&gt; | &lt;percentage&gt; | &lt;relative-change&gt; |
+      x-low | low | medium | high | x-high | inherit
 
      <td>implementation dependent
 
@@ -3043,6 +3063,8 @@ <h2 class=no-num id=changes>Appendix F &mdash; Changes from previous draft</h2>
     SSML element given that the CSS "content" replacement functionality
     addresses the same requirement.
 
+   <li>Added support for semitones in pitch alterations.
+
    <li>Added reference to "time" values syntax (s, ms) for &lsquo;<a
     href="#voice-duration"><code
     class=property>voice-duration</code></a>&rsquo;.

css3-speech/Overview.src.html

@@ -1338,7 +1338,7 @@ <h2 id="voice-char-props"><span class="secno">10. </span>Voice
 </tr>
 <tr>
 <td width="20%"><em>Value:</em></td>
-<td>&lt;number&gt; | &lt;percentage&gt; | x-low | low | medium
+<td>&lt;number&gt; | &lt;percentage&gt; | &lt;relative-change&gt; | x-low | low | medium
  | high | x-high | inherit</td>
 </tr>
 <tr>
@@ -1379,7 +1379,19 @@ <h2 id="voice-char-props"><span class="secno">10. </span>Voice
 
 <dt><strong>&lt;percentage&gt;</strong></dt>
 
-<dd>Specifies a relative change to the inherited value.</dd>
+<dd>Specifies a multiplicative change to the inherited value. Thus 150% means one and a half times the inherited average pitch for this voice.</dd>
+
+<dt><strong>&lt;relative-change&gt;</strong></dt>
+
+<dd>Specifies a relative change (decrement or increment) to the inherited value.
+The syntax of allowed values is a &lt;number&gt; preceded by "-" or "+", and followed by either of "Hz" (for Hertz) or "st" (for semitones).
+A semitone is half of a tone (a half step) on the standard diatonic scale.
+Note that a semitone doesn't correspond to a fixed value in Hertz: instead,
+the ratio between two consecutive frequencies separated by exactly one semitone
+is approximately 1.05946 (the actual arithmetics involved are beyond the scope of this specification,
+please refer to existing literature on that subject).
+Examples of valid values: "+0.5st", "+5st", "-2st", "+10Hz", "-5.5Hz".
+</dd>
 
 <dt><strong>x-low</strong>, <strong>low</strong>,
 <strong>medium</strong>, <strong>high</strong>,
@@ -1389,9 +1401,6 @@ <h2 id="voice-char-props"><span class="secno">10. </span>Voice
 that are implementation and voice specific.</dd>
 </dl>
 
-<p class="issue">ISSUE: should we also allow for relative
-changes in terms of semitones as permitted by SSML? (this would require adapting both "number" and "percentage" to disambiguate uses of "st" and "Hz" values)</p>
-
 <table class="propdef" summary="name: syntax">
 <tbody>
 <tr>
@@ -1400,7 +1409,7 @@ <h2 id="voice-char-props"><span class="secno">10. </span>Voice
 </tr>
 <tr>
 <td width="20%"><em>Value:</em></td>
-<td>&lt;number&gt; | &lt;percentage&gt; | x-low | low | medium
+<td>&lt;number&gt; | &lt;percentage&gt; | &lt;relative-change&gt; | x-low | low | medium
  | high | x-high | inherit</td>
 </tr>
 <tr>
@@ -1447,7 +1456,19 @@ <h2 id="voice-char-props"><span class="secno">10. </span>Voice
 
 <dt><strong>&lt;percentage&gt;</strong></dt>
 
-<dd>Specifies a relative change to the inherited value.</dd>
+<dd>Specifies a multiplicative change to the inherited value. Thus 25% means a quarter of the inherited pitch range for this voice.</dd>
+
+<dt><strong>&lt;relative-change&gt;</strong></dt>
+
+<dd>Specifies a relative change (decrement or increment) to the inherited value.
+The syntax of allowed values is a &lt;number&gt; preceded by "-" or "+", and followed by either of "Hz" (for Hertz) or "st" (for semitones).
+A semitone is half of a tone (a half step) on the standard diatonic scale.
+Note that a semitone doesn't correspond to a fixed value in Hertz: instead,
+the ratio between two consecutive frequencies separated by exactly one semitone
+is approximately 1.05946 (the actual arithmetics involved are beyond the scope of this specification,
+please refer to existing literature on that subject).
+Examples of valid values: "+0.5st", "+5st", "-2st", "+10Hz", "-5.5Hz".
+</dd>
 
 <dt><strong>x-low</strong>,
 <strong>low</strong>,
@@ -1459,9 +1480,6 @@ <h2 id="voice-char-props"><span class="secno">10. </span>Voice
 that are implementation and language dependent.</dd>
 </dl>
 
-<p class="issue">ISSUE: should we also allow for relative
-changes in terms of semitones as permitted by SSML? (this would require adapting both "number" and "percentage" to disambiguate uses of "st" and "Hz" values)</p>
-
 <table class="propdef" summary="name: syntax">
 <tbody>
 <tr>
@@ -2083,6 +2101,7 @@ <h2 class="no-num" id="changes">Appendix F &mdash; Changes from previous draft</
 XML to speech" in the abstract.</li>
 <li>Removed unused normative links to CSS3 Modules (actually moved to informative references), now the only dependency is CSS3 Values and Units.</li>
 <li>Removed issue about the 'sub' SSML element given that the CSS "content" replacement functionality addresses the same requirement.</li>
+<li>Added support for semitones in pitch alterations.</li>
 <li>Added reference to "time" values syntax (s, ms) for 'voice-duration'.</li>
 <li>Moved "content" outside of "phonetics", as the ::before and ::after use-cases do not relate to pronunciation rules (this is actually more similar to audio cues, only applied with text rather than audio files)</li>
 <li>Added prose to explicitly support alphabet other than IPA, via the "x-" vendor-specific prefix.</li>