<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Scaling request for ippSplitScaled adn ippsJoinScaled in Intel® Integrated Performance Primitives</title>
    <link>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992763#M22523</link>
    <description>&lt;P&gt;Not sure this is the right place, but I would like to some additional functions for Split and Join that have a different scaling.&lt;/P&gt;
&lt;P&gt;ippsSplitScaled_16s32f_D2L uses a scaling that mapps -1 to 1 to Int.Min-Int.Max.&amp;nbsp; Now that might be mathematically nice but it is not nice for people doing audio.&amp;nbsp;&lt;/P&gt;
&lt;P&gt;Reason 1: &amp;nbsp; 0 maps to &amp;nbsp;0.0000152590219 not zero.&amp;nbsp; Thus if a waveform editor changes the scale, what should be zero doesn't remain zero.&amp;nbsp; Luckly 0.0f maps to 0&lt;/P&gt;
&lt;P&gt;Reason 2: &amp;nbsp; Not consistent with ippsConvert_32fxxx functions.&amp;nbsp; These map 0 to 0.&amp;nbsp; Thus to make the two work together I have to apply an ugly multiply and add on the vector; e.g. something like d = p*[1/2 - Min/(2*Max)] + 1/2 + Min/(2*Max)&lt;/P&gt;
&lt;P&gt;Reason 3:&amp;nbsp; No need for -1 to be in the range.&amp;nbsp; There are two reasons for this.&amp;nbsp; Audio is never recorded at max values, -3 dB is considered too high.&amp;nbsp; Second, A/D converter hardware can become non-linear when near the top values and often employ soft cliping.&lt;/P&gt;
&lt;P&gt;One simple thing would be to have the following scaling&lt;/P&gt;
&lt;P&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; pdest&lt;N&gt; = Min(-1, pSource&lt;N&gt;/(Max of Int16 = 32767)&lt;/N&gt;&lt;/N&gt;&lt;/P&gt;
&lt;P&gt;Thus both -32768 and -32767 map to the same value.&lt;/P&gt;
&lt;P&gt;That, and the same for 24 bit, is all I really want.&amp;nbsp;&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;However, a soft clip might me nice.&amp;nbsp; That is the map from Int16 to normalized float is (-1, 1] but if one has values, say [-10, 10], the top values are warped into the upper 3dB.&amp;nbsp; This soft clip makes audio output that is mixed too loud sound much better although one would not want to create the inverse function for it.&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;Thank Youl&lt;/P&gt;
&lt;P&gt;-john&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;</description>
    <pubDate>Tue, 18 Sep 2012 20:34:21 GMT</pubDate>
    <dc:creator>Ockham_s_Razor</dc:creator>
    <dc:date>2012-09-18T20:34:21Z</dc:date>
    <item>
      <title>Scaling request for ippSplitScaled adn ippsJoinScaled</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992763#M22523</link>
      <description>&lt;P&gt;Not sure this is the right place, but I would like to some additional functions for Split and Join that have a different scaling.&lt;/P&gt;
&lt;P&gt;ippsSplitScaled_16s32f_D2L uses a scaling that mapps -1 to 1 to Int.Min-Int.Max.&amp;nbsp; Now that might be mathematically nice but it is not nice for people doing audio.&amp;nbsp;&lt;/P&gt;
&lt;P&gt;Reason 1: &amp;nbsp; 0 maps to &amp;nbsp;0.0000152590219 not zero.&amp;nbsp; Thus if a waveform editor changes the scale, what should be zero doesn't remain zero.&amp;nbsp; Luckly 0.0f maps to 0&lt;/P&gt;
&lt;P&gt;Reason 2: &amp;nbsp; Not consistent with ippsConvert_32fxxx functions.&amp;nbsp; These map 0 to 0.&amp;nbsp; Thus to make the two work together I have to apply an ugly multiply and add on the vector; e.g. something like d = p*[1/2 - Min/(2*Max)] + 1/2 + Min/(2*Max)&lt;/P&gt;
&lt;P&gt;Reason 3:&amp;nbsp; No need for -1 to be in the range.&amp;nbsp; There are two reasons for this.&amp;nbsp; Audio is never recorded at max values, -3 dB is considered too high.&amp;nbsp; Second, A/D converter hardware can become non-linear when near the top values and often employ soft cliping.&lt;/P&gt;
&lt;P&gt;One simple thing would be to have the following scaling&lt;/P&gt;
&lt;P&gt;&amp;nbsp;&amp;nbsp;&amp;nbsp; pdest&lt;N&gt; = Min(-1, pSource&lt;N&gt;/(Max of Int16 = 32767)&lt;/N&gt;&lt;/N&gt;&lt;/P&gt;
&lt;P&gt;Thus both -32768 and -32767 map to the same value.&lt;/P&gt;
&lt;P&gt;That, and the same for 24 bit, is all I really want.&amp;nbsp;&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;However, a soft clip might me nice.&amp;nbsp; That is the map from Int16 to normalized float is (-1, 1] but if one has values, say [-10, 10], the top values are warped into the upper 3dB.&amp;nbsp; This soft clip makes audio output that is mixed too loud sound much better although one would not want to create the inverse function for it.&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;Thank Youl&lt;/P&gt;
&lt;P&gt;-john&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;
&lt;P&gt;&lt;/P&gt;</description>
      <pubDate>Tue, 18 Sep 2012 20:34:21 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992763#M22523</guid>
      <dc:creator>Ockham_s_Razor</dc:creator>
      <dc:date>2012-09-18T20:34:21Z</dc:date>
    </item>
    <item>
      <title>Hi John,</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992764#M22524</link>
      <description>Hi John, 

It is right place for submiting feature request.   I have recorded it to IPP track database. Once IPP developer team review it and i will let you know the result.   the reference ID is  DPD200299821. 

Thanks
Ying</description>
      <pubDate>Wed, 19 Sep 2012 07:04:11 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992764#M22524</guid>
      <dc:creator>Ying_H_Intel</dc:creator>
      <dc:date>2012-09-19T07:04:11Z</dc:date>
    </item>
    <item>
      <title>Thank you</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992765#M22525</link>
      <description>Thank you</description>
      <pubDate>Wed, 19 Sep 2012 18:49:28 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992765#M22525</guid>
      <dc:creator>Ockham_s_Razor</dc:creator>
      <dc:date>2012-09-19T18:49:28Z</dc:date>
    </item>
    <item>
      <title>Hi John,</title>
      <link>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992766#M22526</link>
      <description>&lt;P&gt;Hi John,&lt;/P&gt;
&lt;P&gt;Just sorry to tell, there is&amp;nbsp;no futher plan to implement the feature so far. As you see in &lt;A href="http://software.intel.com/sites/products/ipp-deprecated-features-feedback/"&gt;http://software.intel.com/sites/products/ipp-deprecated-features-feedback/&lt;/A&gt;,&amp;nbsp;IPP team&amp;nbsp;hope to streamline functions instead to expand functionality.&lt;/P&gt;
&lt;P&gt;Best Regards,&lt;/P&gt;
&lt;P&gt;Ying&lt;/P&gt;
&lt;P&gt;Ying &amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Wed, 29 May 2013 06:35:01 GMT</pubDate>
      <guid>https://community.intel.com/t5/Intel-Integrated-Performance/Scaling-request-for-ippSplitScaled-adn-ippsJoinScaled/m-p/992766#M22526</guid>
      <dc:creator>Ying_H_Intel</dc:creator>
      <dc:date>2013-05-29T06:35:01Z</dc:date>
    </item>
  </channel>
</rss>

