Hi Mark, Thanks for your help! You have got me on the right track. There are just a few points which I am having problems with when moving the example to a real application.
1 - I'm trying to implement this using Graphics2D, as I have some quite complex geometry stored as Java2D GeneralPath objects. 2 - There will also be multiple transparent gradients on top of each other. I have worked up another short test using Java2D & multiple transparent gradients here: http://al.chemy.org/misc/TransparentGradient2.java This is the correct Java2D output as a PNG: http://al.chemy.org/misc/TransparentGradient2DesiredOutput.png And here is what my code using itext renders: http://al.chemy.org/misc/TransparentGradient2CurrentOutput.pdf There are several issues with this example: 1 - The clipping paths/groups do not seem to be ordered correctly, they seem to be nestled. This is most likely a simple error on my part somewhere? But it is causing the following error when I try to open the file in Adobe Illustrator "An unknown shading type was encountered." It also gets rendered incorrectly (the same in Adobe Reader): http://al.chemy.org/misc/TransparentGradient2Illustrator.png 2 - The ordering of the blue triangle gradient seems to be inverted - something to do with the y axis being flipped in Java2D? But when I invert the gradient colours even the white area of the triangle goes red - this could be related to the clipping path/group problem. Thanks again for your help. Best Regards, Karl On 2009/04/10, at 3:09, Mark Storer wrote: > I'm not sure how to /fix/ it, but I can see what the problem is. > > You want a consistent color with varying alpha, composited with the > background. What you have is consistent alpha with varying color, > written over the background. > > Your shading is just used for color. Not sure why. And you want a > range [0] to [1] alpha, not [1 0 0] (red) to [1 1 1] (white) RGB. > > Only that's not right. If it were written entirely over the > background, you wouldn't see the blue rect behind at all... but the > white area IS stomping it entirely. So I'm confused to. > > Ah! You're using Luminosity. White is opaque, red less so. That > makes sense. > > So how does one specify alpha? First, you'll need to use an Alpha > group, not a Luminosity one... > > Aha! > > Your alpha comes from the graphics state's SMask entry, for which > you're using your shading XObject. Okay. But use "Alpha" instead > of luminosity, and I think you can get away with DeviceRGB [0] to [1]. > > Your red box is also coming from that shading, which is part of the > issue... Lets see if I can hack on your current output to get what > you want... > > Bingo. I stuck with DeviceRGB to make my hackage a bit easier. > Also, the Adobe version's color ranges are from 0.12-ish to 1, > rather than 0 to 1, as in the attached file. > > Change your code thusly: > > Drop: > // Create a gradient to use as the mask > PdfShading shading = PdfShading.simpleAxial(cb.getPdfWriter(), 0, 0, > 0, 100, Color.RED, Color.WHITE); > template.paintShading(shading); > //Place template > cb.addTemplate(template, 0, 0); > > Replace: > // Create a gradient to use as the mask > PdfShading shading = PdfShading.simpleAxial(cb.getPdfWriter(), 0, 0, > 0, 100, Color.BLACK, Color.WHITE); > template.paintShading(shading); > > cb.setColorFill( Color.RED ); > cb.rectangle( 0, 0, 100, 100 ); > cb.fill(); > > > That'll get you in the right ballpark. You might need to mess with > the shading's axis, or swap BLACK & WHITE to get the direction right. > > --Mark Storer > Senior Software Engineer > Cardiff.com > > #include <disclaimer> > typedef std::Disclaimer<Cardiff> DisCard; > > > >> -----Original Message----- >> From: Alchemy [mailto:alch...@al.chemy.org] >> Sent: Thursday, April 09, 2009 6:44 AM >> To: Post all your questions about iText here >> Subject: Re: [iText-questions] GradientPaint & Transparency >> >> >> Hi Again, >> >> I have been playing with this today and found a possible way to >> implement a transparent gradient working from the example here: >> http://1t3xt.info/examples/browse/?page=example&id=414 >> >> This is the output I am getting with following code: >> http://al.chemy.org/misc/TransparentGradientCurrentOutput.pdf >> http://al.chemy.org/misc/TransparentGradient.java >> >> And this is what I am aiming to get: >> http://al.chemy.org/misc/TransparentGradientDesiredOutput.pdf >> >> As you can see there is quite a difference. >> I am not sure if this is the right way to go about it, or even if a >> gradient can be used as an SOFT MASK? >> >> Thanks in advance, >> >> Karl >> >> >> >> On 2009/04/09, at 15:07, Alchemy wrote: >> >>>>>>> When drawing to Java Graphics2D It seems like iText does not >>> support >>>>>>> GradientPaint gradients with transparency? >>>>>> You're right, I don't think that's supported. >>>>> >>>>> I had a look at the relevant iText code and lucky for me I >>> managed to >>>>> get it going by only changing one line :) >>>>> I'm using only RGB at this point so I changed >>>>> com.lowagie.text.pdf.PdfShading line 145 to also return >> the alpha >>> value. >>>>> The full function is below and the new and improved PDF output >>> from my >>>>> previous example is here: >>>>> http://al.chemy.org/misc/graphics2D-2.pdf >>>>> Beautiful isn't it! >>>> >>>> As far as I can see, the change breaks other examples... >>> Yes you are right. >>> It worked straight away in Apple's default PDF reader Preview, but >>> unfortunately I didn't test it fully on Windows etc... >>> Different PDF readers seem to react differently. >>> The worst case scenario is that the files will not even open, >>> otherwise only the transparency may be lost and the colour >> maintained. >>> >>> I am still interested in implementing this, and have mocked up a >>> transparent gradient in Adobe Illustrator and checked the >> pdf output. >>> In general the output is very bloated compared to iText, but it >>> looks like a shape is drawn with a transparent mask on top >> of it (as >>> opposed to using the straight ARGB colours I plugged in last time). >>> The relevant tags in the pdf go something like this: >>> <</I false/K false/CS/DeviceRGB/S/Transparency>> >>> .. >>> <</OPM 1/BM/Normal/CA 1.0/OP false/SMask 18 0 R/ca 1.0/AIS >> false/op >>> false/Type/ExtGState/SA true>> >>> .. >>> <</G 17 0 R/S/Luminosity/Type/Mask/BC 19 0 R>> >>> .. >>> <</Subtype/Form/Length 67/Matrix[1.0 0.0 0.0 1.0 0.0 >> 0.0]/Group 8 0 >>> R/Resources<</XObject<</Fm0 16 0 R>>/ExtGState<</GS0 21 0 R>>>>/ >>> BBox[455.781 456.219 579.219 332.781]>>stream >>> >>> The file itself is here (a 123x123 box at x456 y456): >>> http://al.chemy.org/misc/GradientTransparencyAiTest.pdf >>> >>> So my question is... what do you think is the best way to >> implement >>> this functionality using Graphics2D? >>> Either with a workaround in the application or by extending iText >>> itself. >>> >>> Thank you again for your time! >>> I really appreciate it. >>> >>> Best, >>> Karl >>> >>> - >>> Alchemy >>> An open drawing project >>> http://al.chemy.org >>> - >>> >>> >>> >> >> >> -------------------------------------------------------------- >> ---------------- >> This SF.net email is sponsored by: >> High Quality Requirements in a Collaborative Environment. >> Download a free trial of Rational Requirements Composer Now! >> http://p.sf.net/sfu/www-ibm-com >> _______________________________________________ >> iText-questions mailing list >> iText-questions@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/itext-questions >> >> Buy the iText book: http://www.1t3xt.com/docs/book.php >> >> > <TransparentGradientCurrentOutputTWEAK.pdf> ------------------------------------------------------------------------------ This SF.net email is sponsored by: High Quality Requirements in a Collaborative Environment. Download a free trial of Rational Requirements Composer Now! http://p.sf.net/sfu/www-ibm-com _______________________________________________ iText-questions mailing list iText-questions@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/itext-questions Buy the iText book: http://www.1t3xt.com/docs/book.php
