LibTiff Mailing List

Richard,

Joris wrote:
> > TIFFReadScanline and that sort of access might be more optimally
> > streaming. But support for it depends on the particular subcodec,
> > and I seem to remember not all subcodecs support it. It does not
> > work on tiled images, either, and since there's no actual
> > definition in the man pages of what should be returned on
> > subsampled images, there's a bit of a mess in that case. So I
> > essentially regard this scanline based access a broken concept in
> > LibTiff.

Richard Nolde wrote:
>   I know that you know much more about TIFF and libtiff than I do, but
> I found the basic functions of reading by tile or scanline to work
> extremely well in my application tiffcrop. I know that it is being
> used by folks running fax services and by at least one scanner
> /digital camera programmer in addition to my own adaptation in a high
> volume automated bar code reader.  I think it might be more correct
> to say that the scanline interface is limited in its handling of
> color spaces and possibly certain compression schemes, but I've yet
> to receive any complaints on performance or failure to handle a TIFF
> file. I do have the advantage of a large multiprocessor machine with
> 64 GB of memory, but there are hundreds of concurrent users and most
> of our files only run in the  250 page range, which with G4
> compression on Bilevel images means they are only 20 to 30 MB each.

That seems about right. You're not using the scanline access on tiled 
images. If some subcodecs use larger buffers then is necessary for scanline 
access, they still work for you. And you've probably not seen subsampled 
images through this interface yet, and/or you deal with them as if they were 
broken.

I didn't mean to say TIFFReadScanline is useless. I meant to say that, as a 
concept, it's broken, as it has not been maintained into the tiles update of 
TIFF, the level of support on the part of the subcodec is not guaranteed, 
and we haven't bothered (yet?) to clear up the mess of scanline based access 
to subsampled images. It would be interesting to see if and how an 
unsuspecting interface survives such images, so if you're interested to run 
a test I'll post you some if you like.

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


AWARE [SYSTEMS]		Imaging expertise for the Delphi developer
		TIFF and LibTiff Mailing List Archive
	LibTiff Mailing List TIFF and LibTiff Mailing List Archive January 2008 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	Thread 2008.01.11 18:01 "Scanline access in libtiff", by Richard Nolde 2008.01.12 09:32 "Re: Scanline access in libtiff", by Joris Van Damme 2008.01.12 09:32 "Re: Scanline access in libtiff", by Joris Van Damme Richard, Joris wrote: > > TIFFReadScanline and that sort of access might be more optimally > > streaming. But support for it depends on the particular subcodec, > > and I seem to remember not all subcodecs support it. It does not > > work on tiled images, either, and since there's no actual > > definition in the man pages of what should be returned on > > subsampled images, there's a bit of a mess in that case. So I > > essentially regard this scanline based access a broken concept in > > LibTiff. Richard Nolde wrote: > I know that you know much more about TIFF and libtiff than I do, but > I found the basic functions of reading by tile or scanline to work > extremely well in my application tiffcrop. I know that it is being > used by folks running fax services and by at least one scanner > /digital camera programmer in addition to my own adaptation in a high > volume automated bar code reader. I think it might be more correct > to say that the scanline interface is limited in its handling of > color spaces and possibly certain compression schemes, but I've yet > to receive any complaints on performance or failure to handle a TIFF > file. I do have the advantage of a large multiprocessor machine with > 64 GB of memory, but there are hundreds of concurrent users and most > of our files only run in the 250 page range, which with G4 > compression on Bilevel images means they are only 20 to 30 MB each. That seems about right. You're not using the scanline access on tiled images. If some subcodecs use larger buffers then is necessary for scanline access, they still work for you. And you've probably not seen subsampled images through this interface yet, and/or you deal with them as if they were broken. I didn't mean to say TIFFReadScanline is useless. I meant to say that, as a concept, it's broken, as it has not been maintained into the tiles update of TIFF, the level of support on the part of the subcodec is not guaranteed, and we haven't bothered (yet?) to clear up the mess of scanline based access to subsampled images. It would be interesting to see if and how an unsuspecting interface survives such images, so if you're interested to run a test I'll post you some if you like. 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

LibTiff Mailing List

Thread

2008.01.12 09:32 "Re: Scanline access in libtiff", by Joris Van Damme