AWARE SYSTEMS
TIFF and LibTiff Mail List Archive

Thread

2006.05.10 18:00 "[Tiff] Tiff throws error on tiffinfo but not with tiffdump", by Jeremy Nix
2006.05.10 18:16 "Re: [Tiff] Tiff throws error on tiffinfo but not with tiffdump", by Bob Friesenhahn
2006.05.10 19:13 "Re: [Tiff] Tiff throws error on tiffinfo but not with tiffdump", by Gerben Vos
2006.05.23 11:03 "[Tiff] Question to TIFF mailing list", by Alexey Aushev
2006.05.23 12:13 "[Tiff] Question to TIFF mailing list", by Gerben Vos

2006.05.10 19:13 "Re: [Tiff] Tiff throws error on tiffinfo but not with tiffdump", by Gerben Vos

I have a tiff image that has been giving me problems with JAI (Java Advanced Imaging).

Group3Options (292) LONG (4) 1<0>

  Compression Scheme: None

Any ideas on why these commands worked differently?

tiffdump is a lower-level utility than tiffinfo, so it tends to give less error messages and warnings.

Is there something inherently wrong with the tiff image?

There is nothing really wrong, although it is a bit strange that the image has a Group 3 options tag while it isn't compressed, which is what tiffinfo complains about. Also, while I calculated the correct strip byte count to check if it's really uncompressed (ceil(2544 / 8) * 3295), I noticed that the strips contain 14 bytes too many, which is a bit strange too. Probably has been rounded up to a multiple of 16. Neither of these strangenesses should be anything to worry about.

Fill order 2 (lsb-to-msb) isn't very standard. JAI should still be able to read the image, but if it doesn't support this fill order, you'll get a vertically striped effect (the bits in every byte have been reversed). Or this field may be set to 2 while it really should be 1 and JAI is reversing the bits for you, which you don't want. That would be an error in the TIFF. Maybe the fill order and group 3 options have been wrongly copied over while uncompressing the image.

I'm not sure if the values of the page number fields are correct; I've never used them, but the way I read the standard they should be <0 2> and <1 2>. Most software ignores this field.

That's all I can make of this data. Hope it helps!

Cheers,

Gerben Vos.