[Scribus] Transparent tif bug and related thoughts

avox avox
Thu Mar 1 15:47:47 CET 2007

Aurelien Roux wrote:
> Le Thu, Mar 01, 2007 at 04:29:22AM -0800, avox a ?crit :
>> Given the different filesizes it's no surprise that the times also
>> differ.Pure binary data is (84 * 300 / 2.54) * (59.4 * 300 / 2.54) * 4 =
>> 278383260B,so it looks as if Gimp is correct here. My guess is that
>> Scribus refuses to upsample your data and leaves that tothePDF renderer,
>> which is probably a good choice (it just takes storage withoutproviding
>> any more information; and if the PDF renderer chooses to resamplethe data
>> again for some reason, the result will be worse)
> Ah, OK, that's very interesting. I'd thought that it should be possible
> that scribus just write the correct resolution in the pdf, but I
> couldn't see how it finally worked.
> That's quite clever, hu !
> OK. However, the bug is here, and I've to make it resample by Gimp, at
> the moment.
> Just a final question : when it's so, does scribus force the pdf renderrer
> to resample anyway ? Cause the printer agency with which I'm gonna work
> is not very -I don't know the word, but they don't say much-, and didn't
> show
> me they were technically very accomplished, so, as they want 300 dpi for
> this poster (that's even something strange, hu !), I've to be sure they
> will find the poster @ 300 dpi.

PDF and PostScript always specify the area which the image shall cover and
bitmap data. Any PDF renderer must make sure that the bitmap data is scaled
to cover that area.

Printer agencies usually specify a minimum resolution to avoid ugly raster
when the document is printed at high resolution. If you upsample your images
from 150dpi to 300dpi manually the result shouldn't be the better as if you
the images at 150dpi in the first place (unless the print agency does some
strange stuff with your PDF).
I guess they just said 300dpi because you loose quality if you have less and
you dont
gain much quality if you provide a higher resolution (just wasting memory).

So your approach will likely produce some rasterisation effects in the final
If you want to check beforehand, you can print a section (eg. one of the
in original size on an A4 color printer and look at it from a distance.

View this message in context: http://www.nabble.com/Transparent-tif-bug-and-related-thoughts-tf3326142.html#a9250873
Sent from the Scribus mailing list archive at Nabble.com.

More information about the scribus mailing list