<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	>
<channel>
	<title>Comments on: Genius</title>
	<atom:link href="http://michaelmistretta.com/2008/genius/feed/" rel="self" type="application/rss+xml" />
	<link>http://michaelmistretta.com/2008/genius/</link>
	<description></description>
	<pubDate>Wed, 07 Jan 2009 10:16:36 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.7</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Michael Mistretta</title>
		<link>http://michaelmistretta.com/2008/genius/comment-page-1/#comment-1039</link>
		<dc:creator>Michael Mistretta</dc:creator>
		<pubDate>Sun, 14 Dec 2008 04:13:00 +0000</pubDate>
		<guid isPermaLink="false">http://michaelmistretta.com/?p=343#comment-1039</guid>
		<description>I'm pretty sure Genius is calculating things by recognizing how similar they are (by looking at artist, genres, other people's playlists, etc). There are pros and cons to each system obviously—sometimes I prefer similar artists to similar songs with the same, monotonous beat. It all depends.</description>
		<content:encoded><![CDATA[<p>I&#8217;m pretty sure Genius is calculating things by recognizing how similar they are (by looking at artist, genres, other people&#8217;s playlists, etc). There are pros and cons to each system obviously—sometimes I prefer similar artists to similar songs with the same, monotonous beat. It all depends.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: James</title>
		<link>http://michaelmistretta.com/2008/genius/comment-page-1/#comment-1033</link>
		<dc:creator>James</dc:creator>
		<pubDate>Sat, 13 Dec 2008 19:37:59 +0000</pubDate>
		<guid isPermaLink="false">http://michaelmistretta.com/?p=343#comment-1033</guid>
		<description>That's an interesting idea. I'd thought about that, but then I heard of a software used by most music labels that filters music according to beat, chords and such –it analyzes mathematically measurable parameters for every song and grades them. Then, it is fairly easy to pair them up, according to these numerical gradings. Don't you think that's more likely? 
The divergences you were referring to might be explained by the choice of parameters, ie the fact that the beat is similar doesn't necessarily mean that the melody will be compatible too...</description>
		<content:encoded><![CDATA[<p>That&#8217;s an interesting idea. I&#8217;d thought about that, but then I heard of a software used by most music labels that filters music according to beat, chords and such –it analyzes mathematically measurable parameters for every song and grades them. Then, it is fairly easy to pair them up, according to these numerical gradings. Don&#8217;t you think that&#8217;s more likely?<br />
The divergences you were referring to might be explained by the choice of parameters, ie the fact that the beat is similar doesn&#8217;t necessarily mean that the melody will be compatible too&#8230;</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Mustapha</title>
		<link>http://michaelmistretta.com/2008/genius/comment-page-1/#comment-841</link>
		<dc:creator>Mustapha</dc:creator>
		<pubDate>Sat, 13 Sep 2008 20:19:43 +0000</pubDate>
		<guid isPermaLink="false">http://michaelmistretta.com/?p=343#comment-841</guid>
		<description>Hey,

While I see how gathering information from your library makes Apple know more about you, I can't see how it helps Apple "pair" songs together

What I think the genius does is this: In addition to the statistics you spoke about, It gathers information from the users' &lt;i&gt;playlists&lt;/i&gt;, the ones they had already manually set up, and see which songs people pair together. It does this by creating a sort of compatibility coefficient between any pair of songs by seeing how often any couple of songs come together in playlists.

 So for example, if a relatively large percentage  of Itunes users have put Jack johnson's "banana pancake" with Joe Purdy's "Wash away", this pair of songs will have a relatively high compatibility coefficient and will likely come together in Genius generated playlists..

This explains the relative eccentricity of some of the results in Genius playlists (normal people often pair unrelated songs together for variable emotional reasons), and explains how the Genius becomes smarter as more and more people reveal their playlists.</description>
		<content:encoded><![CDATA[<p>Hey,</p>
<p>While I see how gathering information from your library makes Apple know more about you, I can&#8217;t see how it helps Apple &#8220;pair&#8221; songs together</p>
<p>What I think the genius does is this: In addition to the statistics you spoke about, It gathers information from the users&#8217; <i>playlists</i>, the ones they had already manually set up, and see which songs people pair together. It does this by creating a sort of compatibility coefficient between any pair of songs by seeing how often any couple of songs come together in playlists.</p>
<p> So for example, if a relatively large percentage  of Itunes users have put Jack johnson&#8217;s &#8220;banana pancake&#8221; with Joe Purdy&#8217;s &#8220;Wash away&#8221;, this pair of songs will have a relatively high compatibility coefficient and will likely come together in Genius generated playlists..</p>
<p>This explains the relative eccentricity of some of the results in Genius playlists (normal people often pair unrelated songs together for variable emotional reasons), and explains how the Genius becomes smarter as more and more people reveal their playlists.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Kevin Stauss</title>
		<link>http://michaelmistretta.com/2008/genius/comment-page-1/#comment-832</link>
		<dc:creator>Kevin Stauss</dc:creator>
		<pubDate>Thu, 11 Sep 2008 02:13:53 +0000</pubDate>
		<guid isPermaLink="false">http://michaelmistretta.com/?p=343#comment-832</guid>
		<description>Nice post. I think about how awesome Genius is when I'm playing a playlist or updating it after importing a few CDs, but you put it into a perspective I probably never would've thought about.</description>
		<content:encoded><![CDATA[<p>Nice post. I think about how awesome Genius is when I&#8217;m playing a playlist or updating it after importing a few CDs, but you put it into a perspective I probably never would&#8217;ve thought about.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Dan Philibin</title>
		<link>http://michaelmistretta.com/2008/genius/comment-page-1/#comment-831</link>
		<dc:creator>Dan Philibin</dc:creator>
		<pubDate>Wed, 10 Sep 2008 19:29:44 +0000</pubDate>
		<guid isPermaLink="false">http://michaelmistretta.com/?p=343#comment-831</guid>
		<description>You've hit the nail on the head - something I was thinking about late last night. Genius is only possible because of the amount of people that use iTunes, something that's taken years to improve and perfect. Congrats, Apple.</description>
		<content:encoded><![CDATA[<p>You&#8217;ve hit the nail on the head - something I was thinking about late last night. Genius is only possible because of the amount of people that use iTunes, something that&#8217;s taken years to improve and perfect. Congrats, Apple.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Michael Mistretta</title>
		<link>http://michaelmistretta.com/2008/genius/comment-page-1/#comment-830</link>
		<dc:creator>Michael Mistretta</dc:creator>
		<pubDate>Wed, 10 Sep 2008 17:12:41 +0000</pubDate>
		<guid isPermaLink="false">http://michaelmistretta.com/?p=343#comment-830</guid>
		<description>Or even more interesting is the idea of a Genius Sidebar that will recommend podcasts within iTunes, based on the current podcasts you subscribe to. The possibilities are endless.</description>
		<content:encoded><![CDATA[<p>Or even more interesting is the idea of a Genius Sidebar that will recommend podcasts within iTunes, based on the current podcasts you subscribe to. The possibilities are endless.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jonathan</title>
		<link>http://michaelmistretta.com/2008/genius/comment-page-1/#comment-829</link>
		<dc:creator>Jonathan</dc:creator>
		<pubDate>Wed, 10 Sep 2008 17:09:50 +0000</pubDate>
		<guid isPermaLink="false">http://michaelmistretta.com/?p=343#comment-829</guid>
		<description>I think that Genius is simply an evolved version of "Just For You" (which has since disappeared from the iTunes Store).

I'm not to worried about the privacy aspect. I think that this time next year Genius will be far more capable and accurate. I wouldn't be surprised if we saw something like this creep into other apps.

For instance, I would love to see an RSS feed Genius.

Imagine a sidebar in NetNewsWire suggesting a list of ten posts that are related and then bellow - some blogs you might enjoy. I understand that logistically, this would be very hard, but if blogs were submitted you could get rid of half the rubbish and only get the best of it.

Great post.

Jonathan</description>
		<content:encoded><![CDATA[<p>I think that Genius is simply an evolved version of &#8220;Just For You&#8221; (which has since disappeared from the iTunes Store).</p>
<p>I&#8217;m not to worried about the privacy aspect. I think that this time next year Genius will be far more capable and accurate. I wouldn&#8217;t be surprised if we saw something like this creep into other apps.</p>
<p>For instance, I would love to see an RSS feed Genius.</p>
<p>Imagine a sidebar in NetNewsWire suggesting a list of ten posts that are related and then bellow - some blogs you might enjoy. I understand that logistically, this would be very hard, but if blogs were submitted you could get rid of half the rubbish and only get the best of it.</p>
<p>Great post.</p>
<p>Jonathan</p>
]]></content:encoded>
	</item>
</channel>
</rss>
