AWARE [SYSTEMS] Imaging expertise for the Delphi developer
AWare Systems, Imaging expertise for the Delphi developer, Home TIFF and LibTiff Mailing List Archive

LibTiff Mailing List

TIFF and LibTiff Mailing List Archive
March 2006

Previous Thread
Next Thread

Previous by Thread
Next by Thread

Previous by Date
Next by Date

Contact

The TIFF Mailing List Homepage
This list is run by Frank Warmerdam
Archive maintained by AWare Systems



Valid HTML 4.01!



Thread

2006.03.20 18:04 "3.8.1 and JPEG problem", by Mikhail Kruk
2006.03.20 18:13 "Re: 3.8.1 and JPEG problem", by Bob Friesenhahn
2006.03.20 19:00 "Re: 3.8.1 and JPEG problem", by Andrey Kiselev
2006.03.20 20:32 "Re: 3.8.1 and JPEG problem", by Mikhail Kruk
2006.03.25 08:07 "Re: 3.8.1 and JPEG problem", by Joris Van Damme
2006.03.25 18:08 "Re: 3.8.1 and JPEG problem", by Joris Van Damme
2006.03.26 02:52 "Re: 3.8.1 and JPEG problem", by Mikhail Kruk
2006.03.26 03:25 "Re: 3.8.1 and JPEG problem", by Joris Van Damme
2006.03.26 23:48 "Re: 3.8.1 and JPEG problem", by Mikhail Kruk
2006.03.27 00:56 "Re: 3.8.1 and JPEG problem", by Joris Van Damme
2006.03.27 03:56 "Re: 3.8.1 and JPEG problem", by Mikhail Kruk
2006.03.27 04:54 "Re: 3.8.1 and JPEG problem", by Joris Van Damme
2006.03.27 14:39 "Re: 3.8.1 and JPEG problem", by Ed Grissom
2006.03.27 14:54 "Re: 3.8.1 and JPEG problem", by Mikhail Kruk

2006.03.26 23:48 "Re: 3.8.1 and JPEG problem", by Mikhail Kruk

> > The snapshot I have doesn't read or write JPEG TIFFs (output is all
> > black)... I'm testing using tiffcp, is it possible that the problem is
> > with it, not the lib?  Reading JPG works fine with my own program I
> > build against 3.8.0.  Do you want me to try linking my app with the
> > CVS snapshot?
> > I also tried tiffcp from 3.8.0 and it doesn't work.
> > I guess the next logical step is to assume that tiffcp is broken and
> > test with my app, but I want to make sure that I'm looking at the
> > right CVS since you sound like you've made some changes and I don't
> > see them.
>
> I've not tested tiffcp, but the results you were getting make perfect
> sense with the bugs I've corrected in several other files, so I would
> assume there's nothing wrong with tiffcp as such. It may *seem* that way
> though... There are several ways to access JPEG compression in TIFF, and
> only the recommendable way, without use of the pseudotag JPEGCOLORMODE,
> was broken (as far as I know). So, one app may have been suffering,
> while another may have been behaving perfectly OK.
>
> When I update CVS, I see what I have comitted. Frank was waiting for the
> bugfixes, so I assume he too, will soon update CVS and check out my
> fixes. Since I know very little about CVS systems beyond there basic
> use, all I can suggest at this stage is
> - wait a couple of hours maybe, and check again?
> - wait for Frank to report back?
> - check the changelog file in the root level. The top few entries should
> read the following:

All right, I got the CVS problem partly figured out.  I don't
understand why diff wasn't showing any differences, but a new checkout
gave me all your changes.
I tested again using tiffcp.  The only TIFF file which is
JPEG-compressed I have is in YCBCR color space.  tiffcp didn't process
it correctly, I believe there is a bug in it, but when I brute forced
it to do
TIFFSetField(in, TIFFTAG_JPEGCOLORMODE, JPEGCOLORMODE_RGB);
it worked (tiffcp -none jpg.tiff out.tiff produced an unencoded image
which looks correctly).
Don't have time to compe up with correct tiffcp patch right now).

I then tried encoding a regular (packbits) TIFF into JPEG (tiffcp -c
jpeg test.tiff jpg.tiff) and the result shows up as black in the
viewer (the same viewer shows my original YCBCR-encoded file normally,
so it knows how to do JPEG; the viewer is Microsoft OFfice Document
Imaging, it's pretty good, I've managed to confuse it only once with a
100x20 file which had DPI of 100 specified in it, but I digress).
I then used tiffcp to convert that black image back to unencoded and
got a good result.