Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!apple!well!jef
From: jef@well.sf.ca.us (Jef Poskanzer)
Newsgroups: comp.graphics
Subject: Re: Scan line lengths in image file formats
Message-ID: <21667@well.sf.ca.us>
Date: 14 Nov 90 22:45:31 GMT
References: <1579@ncrwat.Waterloo.NCR.COM>
Reply-To: Jef Poskanzer <jef@well.sf.ca.us>
Organization: Paratheo-Anametamystikhood Of Eris Esoteric, Ada Lovelace Cabal
Lines: 23

In the referenced message, dgs@swdev.Waterloo.NCR.COM (David G. Schwartz) wrote:
}                                          Does there exist any standard
}or convention that an image input device should pad the length of a scan
}line to a byte, word or double-word boundary? I'm thinking specifically
}about monochrome (i.e. 1 bit/pel) images.

Some formats specify padding.  Some formats specify no padding.  Some
formats can go either way, because they store bytes per line separately
from pixels per line.  There is no standard.

My preference is for no padding.  I can see that there might be a speed
advantage to padding in very specialized circumstances (homogenous
hardware / software environment, and you want to do image I/O between
the disk and the screen with no extra data copies).  But that's a very
rare situation, and most of the time padding has no perceptible speed
advantage, and encourages non-portable implementations.  And the
either-way formats are useless, since you can't take advantage of the
padding if you're not sure it will be there.
---
Jef

  Jef Poskanzer  jef@well.sf.ca.us  {ucbvax, apple, hplabs}!well!jef
     "Say yur prayers, yuh flea-pickin' varmint!" -- Yosemite Sam