AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2006.11.27 21:46 "[Tiff] Bugzilla is operational again", by Frank Warmerdam
2007.01.18 20:01 "[Tiff] tif_getimage - using colormap even if photometric interp wrong", by Frank Warmerdam
2007.01.18 19:47 "RE: [Tiff] tif_getimage - using colormap even if photometric interpwrong", by Chris Cox
2007.01.18 20:47 "Re: [Tiff] tif_getimage - using colormap even if photometricinterpwrong", by Joris
2007.01.18 21:25 "Re: [Tiff] tif_getimage - using colormap even if photometricinterpwrong", by Frank Warmerdam
2007.01.22 10:13 "Re: [Tiff] tif_getimage - using colormap even ifphotometricinterpwrong", by Gerben Vos
2007.01.18 23:33 "Re: [Tiff] tif_getimage - using colormap even if photometricinterpwrong", by Bob Friesenhahn
2007.01.18 23:54 "Re: [Tiff] tif_getimage - using colormap even if photometricinterpwrong", by Joris
2007.01.18 20:29 "Re: [Tiff] tif_getimage - using colormap even if photometric interp wrong", by Bob Friesenhahn

2007.01.18 23:54 "Re: [Tiff] tif_getimage - using colormap even if photometricinterpwrong", by Joris

Bob,

The main thing is, you can rightly make the codec behave robust on wrong images. But it is not right to make it behave wrong on whatever image. That creates more problems further down the road, then it solves. Taking a Colormap into account when Photometric is MINISBLACK is just that, it is just wrong.

There are really interesting things you can do with an extra colormap. For example, even if the original image is grayscale, it can be viewed "in color" by applying an extra colormap. The fact that useful applications exist for attaching a colormap to the image which is not (by default) used to view the image means that it should not be used by default if it exists.

Agreed.

I would stress it is quite natural to disregard tags that don't apply. We disregard private tags we don't know about. We disregard T4Options tag in JPEG compressed TIFFs, Subsampling in RGB TIFFs, and we've no other choice then to disregard Colormaps in RGB TIFFs. I see no reason why that general rule wouldn't apply to a Colormap tag in grayscale TIFFs.

The fact that there are actual good, private, non-default uses and reasons to accompany grayscale or even RGB images with a Colormap, doesn't change this, of course, nor does it lead you to other conclusions then does my reasoning. I still wouldn't accompany a grayscale image by a Colormap myself, but still, as you say, there might be reasons to consider doing so, and doing so certainly shouln't interfere with the default grayscale interpretation.

As to "really interesting things you can do with an extra colormap", I seem to remember a full-color BMP file is allowed to contain a palette, and one private, non-default interpretation is to take this as default optimal palette for the image. My memory on this dates back quite a while, likely to the time where most were using 256 color displays at best, and the fast availability of an optimal palette for an image was more important.

Best regards,

Joris Van Damme
info@awaresystems.be
http://www.awaresystems.be/
Download your free TIFF tag viewer for windows here:
http://www.awaresystems.be/imaging/tiff/astifftagviewer.html