Return-Path: Received: from atl4mhob16.myregisteredsite.com ([209.17.115.54] verified) by media-motion.tv (CommuniGate Pro SMTP 4.2.10) with ESMTP id 5267324 for AE-List@media-motion.tv; Mon, 28 Oct 2013 21:33:06 +0100 Received: from mailpod.hostingplatform.com ([10.30.71.210]) by atl4mhob16.myregisteredsite.com (8.14.4/8.14.4) with ESMTP id r9SKYM7P011957 for ; Mon, 28 Oct 2013 16:34:22 -0400 Received: (qmail 15640 invoked by uid 0); 28 Oct 2013 20:34:22 -0000 X-TCPREMOTEIP: 60.225.197.206 X-Authenticated-UID: chris@chriszwar.com Received: from unknown (HELO ?192.168.0.3?) (chris@chriszwar.com@60.225.197.206) by 0 with ESMTPA; 28 Oct 2013 20:34:22 -0000 Content-Type: text/plain; charset=iso-8859-1 Mime-Version: 1.0 (Mac OS X Mail 6.6 \(1510\)) Subject: Re: [AE] Quick AE Render question. From: Chris Zwar In-Reply-To: Date: Tue, 29 Oct 2013 07:34:19 +1100 Content-Transfer-Encoding: quoted-printable Message-Id: <2DAE7279-EE02-4FC6-84E8-71AB8F367F57@chriszwar.com> References: To: "After Effects Mail List" X-Mailer: Apple Mail (2.1510) On 28/10/2013, at 11:41 PM, Robert Kjettrup wrote: >=20 > Thanks Chris for the tip about NOT using LZW compression on 16bit tiff = files. Never thought about that could be a problem, i always leave this = setting on. But dosent this more sound like a bug? It's not a bug, it's just the way the algorithm was designed. LZW is a = very old and primitive form of compression and was specifically designed = for 8 bit files. It looks for patterns or values that repeat themselves = a lot, and builds up a table of these repeated values. In an 8 bit file = where each value is restricted to 0 - 255 there's a high probability = that sequences of numbers will repeat themselves, especially for things = that have an underlying structure like language. LZW can still do very = well on text files, for such an old algorithm. But in a 16 bit file the = range of values is so much bigger (0-65535) that the probability of = having specific sets of numbers that repeat falls dramatically, and the = table becomes an overhead. I imagine that for simple images, such as mattes, then LZW would still = offer impressive file size advantages in 16 bit mode. But for natural = images such as photographs then it just doesn't work very well, and it's = due to the 8 bit nature of the algorithm's origins. I suppose it's a case of YMMV, but in general LZW isn't recommended for = 16 bit tiffs. -Chris=