AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2005.02.02 09:56 "[Tiff] Tiff differences", by Duarte Costa
2005.02.02 15:35 "Re: [Tiff] Tiff differences", by Bob Friesenhahn
2005.02.02 15:43 "Re: [Tiff] Tiff differences", by Andy Cave
2005.02.02 17:37 "RE: [Tiff] Tiff differences", by Ed Grissom

2005.02.02 17:37 "RE: [Tiff] Tiff differences", by Ed Grissom

According to the original poster, IRasC reads the second image fine, just slow. Since I helped write the code, I know that IRasC can read single strip files, and since the file is uncompressed, it should succeed even with very large files.

The basic problem is that IRasC likes to work with tiles, and will be most efficient when the image is tiled. In both cases, the internal memory manager will buffer up 256 or 512 scanlines and then cut those up into tiles so that the engine can work on them. The best solution would be to convert the file to a tiled tiff.

"Tiff 1" file has

> MinSampleValue (1 Short): 0
> MaxSampleValue (1 Short): 1

what does this mean? Single-bit RGB?

"Tiff 2" file has

> RowsPerStrip (1 Short): 31728

Since this is neither the image_length, nor 3 x image_length, nor (-1), I am not sure what the writer is trying to convey here. Hmmmm.... it is 3 x the image_WIDTH. I'm thinking that this is an error and may have something to do with the problem.

D.Costa - you might be able to use the ImageStation Raster Utilities (ISRU) that comes with IRasC to tile and add overviews to both of these files and get much better performance. Failing that, "tiffcp" might be able to at least "tile" the files. As far as the "residuals" go, I am not sure what is causing that error, but it likely has nothing to do with the TIFF format or this list. Please contact your Intergraph support person if the suggestions above do not resolve this issue.

ed grissom
ed.grissom@intergraph.com