On 4 Jul 2008, at 5:35 PM, Robin wrote:

> On Fri, Jul 4, 2008 at 3:47 PM, Robin <[EMAIL PROTECTED]> wrote:
>> Hi,
>>
>> I have some figures prouced from Matlab using imagesc command. An
>> example is here:
>> http://acrids.robince.net/robince/simA_theta1.pdf
>>
>> This should be rectangular grayscale blocks. It shows up correctly in
>> acrobat. However in Preview and Skim the image is very blurred - the
>> boundaries between the rectangular regions are no longer clear.
>> Preview is much worse than skim (completely smoothed image) - with
>> Skim you can make out some of the structure but it is blurred.
>>

Really weird Skim and Preview give different results, as there is no  
setting in PDFKit other than the ones exposed in the preferences (in  
particular, whether it should anti-alias). So Preview must make some  
internal overrides to PDFKit to interpolate even more (though the only  
way I could think of makes absolutely no difference). It seems Skim is  
much closer to Adobe Reader (at least, Adobe Reader 9)

>> A more knowledgable friend informs me:
>>
>>> Seems matlab tries exporting the centres of imagesc graphs as a  
>>> teeny
>>> tiny bitmapped graphic which then scaled up. This is pretty much the
>>> definition of a pathological case of bitmap scaling for any  
>>> algorithm
>>> - with more pathology the smarter the algorithm - and betrays
>>> imagesc's origins as part of the bitmap manipulation toolchain.
>>
>> Since there is a considerable difference between Preview and Skim I
>> was wondering if there might be a hidden option for pdfkit somewhere
>> that could cause either Preview or Skim to display the image properly
>> - or whether anything else might be done to get either application to
>> show imagesc plots correctly?
>
>
> Some updated information from my friend:
>
>> hmmn, looks like os x ignores the "Interpolate" dictionary entry for
>> postscript images. I've tried explicitly setting it to "false", the
>> default, by hacking up a file exported from matlab and it merrily
>> ignores it. My postscript fu is weak these days, so I may be wrong.
>>
>> As pretty much nothing sets this flag, ignoring it is almost  
>> certainly
>> the right thing to do to make all pdf and ps files look beautiful...
>> except when it isn't. From reading the red book the reason for the
>> interpolation flag was to be able to explicitly turn /on/
>> interpolation in newer devices that supported it as it is pretty
>> brutal computationally (by 1990s standards) and would hammer old
>> hardware like printers. The specs sort of imply that in an ideal  
>> world
>> interpolation would always be used, so MATLAB's dependence on its
>> absence is pretty bogus.
>
> So I guess the question is - is it possible to make Skim respect the
> 'Interpolate' setting?
>
> Cheers
>
> Robin

No. Rendering the PDF is entirely up to PDFKit.

Christiaan


-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
_______________________________________________
Skim-app-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/skim-app-users

Reply via email to