<?xml version="1.0" encoding="UTF-8"?><!-- generator="wordpress/2.0.4" -->
<rss version="2.0" 
	xmlns:content="http://purl.org/rss/1.0/modules/content/">
<channel>
	<title>Comments on: Colour space, and the many pitfalls</title>
	<link>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/</link>
	<description>An insider's look at the world of low-budget visual effects.</description>
	<pubDate>Thu, 20 Nov 2008 21:50:09 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.0.4</generator>

	<item>
		<title>by: Corrección de color &#124; CG News</title>
		<link>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/#comment-6394</link>
		<pubDate>Sun, 16 Mar 2008 15:49:09 +0000</pubDate>
		<guid>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/#comment-6394</guid>
					<description>[...] Links: Espacios de color Corrección de color secundaria [...]</description>
		<content:encoded><![CDATA[<p>[&#8230;] Links: Espacios de color Corrección de color secundaria [&#8230;]
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Hugh Macdonald</title>
		<link>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/#comment-4635</link>
		<pubDate>Sun, 12 Aug 2007 21:21:51 +0000</pubDate>
		<guid>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/#comment-4635</guid>
					<description>When doing the grain in comp, you're right, it's not actually in the film, but what we're trying to replicate is the grain that was being done in-celluloid, as opposed to the light hitting the film, which was scene-based, and therefore linear.

The FilmGrain node in Shake was designed to work in Log, although, if you can get the same result in Linear, then great, but I'm pretty sure that it would be a much harder task.

The grain in film is the slight variations in the density of the film (which is what Log approximates) as opposed to variations in the light hitting the film (which is what Linear trys to emulate)</description>
		<content:encoded><![CDATA[<p>When doing the grain in comp, you&#8217;re right, it&#8217;s not actually in the film, but what we&#8217;re trying to replicate is the grain that was being done in-celluloid, as opposed to the light hitting the film, which was scene-based, and therefore linear.</p>
<p>The FilmGrain node in Shake was designed to work in Log, although, if you can get the same result in Linear, then great, but I&#8217;m pretty sure that it would be a much harder task.</p>
<p>The grain in film is the slight variations in the density of the film (which is what Log approximates) as opposed to variations in the light hitting the film (which is what Linear trys to emulate)
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: sjefferson</title>
		<link>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/#comment-4634</link>
		<pubDate>Sun, 12 Aug 2007 20:58:34 +0000</pubDate>
		<guid>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/#comment-4634</guid>
					<description>Good article, but I'm a little confused about:

"The one exception is adding film grain. Because this is actually happening inside the film, this is always done in Log."

If you're adding grain in shake, presumably just to comped in elements in order to match the grain to a plate, then doing a filmout, the grain you add isn't 'inside the film' any more than any other digitally added element, right? So does it really matter whether it's done in Log or Lin space, as long as it visually matches the plate's grain?</description>
		<content:encoded><![CDATA[<p>Good article, but I&#8217;m a little confused about:</p>
<p>&#8220;The one exception is adding film grain. Because this is actually happening inside the film, this is always done in Log.&#8221;</p>
<p>If you&#8217;re adding grain in shake, presumably just to comped in elements in order to match the grain to a plate, then doing a filmout, the grain you add isn&#8217;t &#8216;inside the film&#8217; any more than any other digitally added element, right? So does it really matter whether it&#8217;s done in Log or Lin space, as long as it visually matches the plate&#8217;s grain?
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: lewis</title>
		<link>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/#comment-3788</link>
		<pubDate>Sat, 19 May 2007 23:40:59 +0000</pubDate>
		<guid>http://www.brokenpipefilms.com/unframedvfx/2007/05/colour-space-and-the-many-pitfalls/#comment-3788</guid>
					<description>Really good post.  Wish info like this had been out there when I was figuring all this stuff out on my own :)

About the grading.  If you're simulating sticking a filter in front of the camera lens what you have is quite right.  But these days people tend to do rather more horrible and complex things in their grades, like shadows and midtones adjustments :)  So doing things the "wrong" way can do quite normal and accepted things - such as tinting the highlights without affecting the shadows and midtones like a filter would, which you can do with a Mult in log or video space.  Grading is always (as far as I know) done in a perceptually uniform-ish rather than physically accurate space because it's easier to grab hold of bits of the tonal range that we see as being distinct - so usually three bands, shadows mids and highlights.  Traditionally done with the classic three balls and three rings of a TK desk, which are Add, Gamma and Mult, in video or log space.  Gamma is especially horrible to use in Linear... if you stick our friend Marcie through Shake's ColorCorrect and look at what all the knobs do to the histogram, in linear it's hard to get seperate control over the blacks and the midtones while in log that works pretty well.

wow i'm rambling and it's bedtime...

L</description>
		<content:encoded><![CDATA[<p>Really good post.  Wish info like this had been out there when I was figuring all this stuff out on my own <img src='http://www.brokenpipefilms.com/unframedvfx/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' /> </p>
<p>About the grading.  If you&#8217;re simulating sticking a filter in front of the camera lens what you have is quite right.  But these days people tend to do rather more horrible and complex things in their grades, like shadows and midtones adjustments <img src='http://www.brokenpipefilms.com/unframedvfx/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' />   So doing things the &#8220;wrong&#8221; way can do quite normal and accepted things - such as tinting the highlights without affecting the shadows and midtones like a filter would, which you can do with a Mult in log or video space.  Grading is always (as far as I know) done in a perceptually uniform-ish rather than physically accurate space because it&#8217;s easier to grab hold of bits of the tonal range that we see as being distinct - so usually three bands, shadows mids and highlights.  Traditionally done with the classic three balls and three rings of a TK desk, which are Add, Gamma and Mult, in video or log space.  Gamma is especially horrible to use in Linear&#8230; if you stick our friend Marcie through Shake&#8217;s ColorCorrect and look at what all the knobs do to the histogram, in linear it&#8217;s hard to get seperate control over the blacks and the midtones while in log that works pretty well.</p>
<p>wow i&#8217;m rambling and it&#8217;s bedtime&#8230;</p>
<p>L
</p>
]]></content:encoded>
				</item>
</channel>
</rss>
