Re: [Fractint] whatever
"""I assume you know that you can render to disk at even larger sizes, then use a graphic editor like IrfanView to resize/resample down to half the original, so as to get a form of anti-aliasing applied to your image. It makes the edges even smoother, and usually ends up adding additional colors from the basic 256 (8 BitsPerPixel) mode to the 16.7 million (24 BitsPerPixel) mode.""" Why do I always find out about these things so late in life ........ ? .
Subject-Was: Re: [Fractint] whatever
Paul N. Lee """ JackOfTradeZ@comcast.net said:
"""I assume you know that you can render to disk at even larger sizes, then use a graphic editor like IrfanView to resize/resample down to half the original, so as to get a form of anti-aliasing applied to your image. It makes the edges even smoother, and usually ends up adding additional colors from the basic 256 (8 BitsPerPixel) mode to the 16.7 million (24 BitsPerPixel) mode."""
Why do I always find out about these things so late in life ........ ?
Rendering at a higher resolution than your device and downsampling is not a necessary trick. It is also something that you might guess and observe. I would do it with text when my destination printer is an ink jet (some of them hav more than sixteen levels of ink per pixel), and for reasons of compatibility with operator of that ink jet, I am doing the whole page as a 34 megapixel PNG or TIFF. Usually, though, I produce a PDF, so that operators hav nothing to do but run Acrobat; no formatting, colour corrected with -dUseCIEColor (in my GhostScript ps2pdf.bat). At 600 dpi, you just about need to be printing on transparencies, inspecting work with a microscope, or just zooming into actual size (printer pixel==screen pixel) to appreciate such things as anti-aliasing from a downsample. In the usual scheme of things, to save myself time and trouble, I render at seventeen megapixels and resample for the screen: 672by900 works out to a full page, according to Internut Explorer concerning background images, just over half a megapixel. Seventeen megapixels is a nice size of starting resolution for laser printing fractals that are ten inches square at Office Depot or Staples; 300 dpi. is nine megapixels at that size. 1400by1400 (times 3by3 in batch files)==4200 pixels square (17 megapixels) is a size in fractint.cfg that I orijinally wrote for the Fractal Universe Calendar contest, which is down. Their payout, US$400, was far short of royalities, though. So, getting into that calendar was about becoming known. I am probably better off making a connection with a marketing house in Edmonton -- i.e. formatting a calendar, myself, and getting someone to promote it under the banner of someone's business, OR doing the whole job of getting a sponsor, myself, and using the marketing house pretty much as a print shop for calendars. Offset presses typically still run at 300dpi and four colours per pixel, because it is hard to tune four colour-separation presses more precisely than that. Sales is not my strength. In fractint.cfg: AF2 ,Disk/RAM 'Video' , 3, 0, 0, 0, 11, 1400, 1400,256,3x3 10...14" square on laser _______ http://ecn.ab.ca/~brewhaha/ BrewJay's Babble Bin
Jay Litwyn wrote:
Subject-Was: Re: [Fractint] whatever
Paul N. Lee """ JackOfTradeZ@comcast.net said:
"""I assume you know that you can render to disk at even larger sizes, then use a graphic editor like IrfanView to resize/resample down to half the original, so as to get a form of anti-aliasing applied to your image. It makes the edges even smoother, and usually ends up adding additional colors from the basic 256 (8 BitsPerPixel) mode to the 16.7 million (24 BitsPerPixel) mode."""
Why do I always find out about these things so late in life ........ ?
Rendering at a higher resolution than your device and downsampling is not a necessary trick. It is also something that you might guess and observe. I would do it with text when my destination printer is an ink jet (some of them hav more than sixteen levels of ink per pixel), and for reasons of compatibility with operator of that ink jet, I am doing the whole page as a 34 megapixel PNG or TIFF. Usually, though, I produce a PDF, so that operators hav nothing to do but run Acrobat; no formatting, colour corrected with -dUseCIEColor (in my GhostScript ps2pdf.bat). At 600 dpi, you just about need to be printing on transparencies, inspecting work with a microscope, or just zooming into actual size (printer pixel==screen pixel) to appreciate such things as anti-aliasing from a downsample.
For perspective, high-quality printed color output (like you see in glossy magazines like National Geographic) is done at 300 lines per inch. -- David gnome@hawaii.rr.com authenticity, honesty, community
...get a form of anti-aliasing applied to your image -P.N.L.
I've sometimes calculated fractal images 4 times larger in X and Y and downsized them. This may be a bit of overkill, but I feel like this provides the ultimate in image quality of a Fractint fractal that you can reasonably do. I only use 4 times 'over-sizing' and 'quarter down-sizing' for creating desktop wallpaper from .GIF fractals. It's important to have the output from the resizing be a JPG image -- and make sure the JPG encoding 'quality' setting is high enough to not discard some of the smoothing that you have just gained by the re-sizing... Also, make sure that the re-sizing algorithm used actually combines pixels -- there are a few apps that just discard unused pixels and output a smaller percentage of your original pixels when downsizing images! Some apps let you set the re-sizing algorithm: - Don't use 'Nearest Neighbor'! - Bicubic is fine (& a typical default) and better than Bilinear - Even better (but only by a small amount) than Bicubic is Lanczos The reason you need JPG output is as follows: Consider a re-sizing algorithm reducing an image in size by 1/2. When the algorithm needs to combine an white input pixel with an black input pixel (for example) to write out a single pixel of a particular shade of gray, the needed gray is highly likely to *not* be in the 256 colors in the GIF color map. (Unless the entire image is all shades of gray.) This problem is true for all pairs of colors that the re-sizing algorithm needs to combine. The JPG pixel color encoding algorithm can create and write out the exact needed color to best smooth your image. --------------------- A closely related issue that can inadvertently degrade images is poor monitor gamma adjustment. Unfortunately, your fractal can be correctly anti-aliased, but degraded during viewing on your monitor. --------------------- A quick look at the Wikipedia article: http://en.wikipedia.org/wiki/Anti-aliased on anti-aliasing can give a feeling for what is going on at the pixel level. Especially see Figure 3 in the article. Wikipedia says: "In digital signal processing, anti-aliasing is the technique of minimizing the distortion artifacts known as aliasing when representing a high-resolution signal at a lower resolution. [This is what is being done when we calculate a finite resolution fractal image using an 'infinite resolution' fractal formula. -hhl] Anti-aliasing is used in digital photography, computer graphics, digital audio, and many other applications. Anti-aliasing means removing signal components that have a higher frequency than is able to be properly resolved by the recording (or sampling) device. This removal is done before (re)sampling at a lower resolution. When sampling is performed without removing this part of the signal, it [can] cause undesirable artifacts such as the black-and-white noise near the top of figure 1-a below." Note that these artifacts are only generated by particular patterns in an image. In practice, many of the fractals created by Fractint do not have these patterns and thus, do not suffer from the artifacts when re-sizing is used to approximate true anti-aliasing. ------------------------ A quick look at the Wikipedia article: http://en.wikipedia.org/wiki/Gamma_correction on monitor gamma is useful if you're not familiar with this. The article's pictures show what a strong influence gamma has on the displayed image. Wikipedia says: "...image files are almost always stored on computers and communicated across the Internet with gamma encoding. ...still image files ([such] as JPEG) are explicitly encoded (that is, they carry gamma-encoded values, not linear intensities), as are motion picture files (such as MPEG)." - Hal Lane ######################### (\__/) # hallane@earthlink.net # (='.'=) ######################### (")_(") -----Original Message----- From: fractint-bounces+hallane=earthlink.net@mailman.xmission.com On Behalf Of JackOfTradeZ@comcast.net Sent: Wednesday, March 31, 2010 12:43 PM To: fractint@mailman.xmission.com Subject: Re: [Fractint] whatever """I assume you know that you can render to disk at even larger sizes, then use a graphic editor like IrfanView to resize/resample down to half the original, so as to get a form of anti-aliasing applied to your image. It makes the edges even smoother, and usually ends up adding additional colors from the basic 256 (8 BitsPerPixel) mode to the 16.7 million (24 BitsPerPixel) mode.""" Why do I always find out about these things so late in life ........ ? .
Hal Lane wrote:
...get a form of anti-aliasing applied to your image -P.N.L.
I've sometimes calculated fractal images 4 times larger in X and Y and downsized them. This may be a bit of overkill, but I feel like this provides the ultimate in image quality of a Fractint fractal that you can reasonably do.
I only use 4 times 'over-sizing' and 'quarter down-sizing' for creating desktop wallpaper from .GIF fractals.
Even multiples are best, since you're talking about individual bits arranged in a basically-square grid. So 4x oversizing gives you 4 pixels to use in determining the color of the 1 pixel in the downsized version. Next step up would be 16x. Fractint can generate, what, 32768x32768 images now, or bigger? Generate it that size, downsize it to 1/16 that size, and you've still got a high rez fractal ...
It's important to have the output from the resizing be a JPG image -- and make sure the JPG encoding 'quality' setting is high enough to not discard some of the smoothing that you have just gained by the re-sizing...
You don't need the lossiness of JPG to encode 24-bit color; PNG supports that quite well. For archival storage of source images for future rework, write a fractal generator that produces 48-bit color and stores them as TIFF files. I think what GIF supports that the others don't is the ability to essentially include the parameters in the GIF file itself? I seem to recall reading that it used text blocks that the GIF format allowed, but that most graphics programs would clear when they saved a GIF file ... I don't think JPG supports that- does PNG? -- David gnome@hawaii.rr.com authenticity, honesty, community
Subject-Was: Re: [Fractint] get a form of anti-aliasing applied to your image ----- Original Message ----- From: "david" <gnome@hawaii.rr.com> (...)
I think what GIF supports that the others don't is the ability to essentially include the parameters in the GIF file itself? I seem to recall reading that it used text blocks that the GIF format allowed, but that most graphics programs would clear when they saved a GIF file ... I don't think JPG supports that- does PNG? (...)
wrJPGcom (I used a DOS package for converting tga to jpeg when PSP did not support progressive scan) writes comments into JPEGs. I think EXIF data in JPEGs from cameras is comments. I might still hav JPEGs into which I piped fractint formulas and parameter files; not that a lot of people knew they were there. I think if you want to publish (or otherwise archive) formulas or parameter files, then it is better to do that separately, so that people who are not interested in math can easily skip it, and people who are so can find it. Then comes that handy web construct called an anchor (link). On my local system, though, all of my published fractals hav their own directories. That way, I can call one file chromo and one file lumo in each directory, plus batch files for jenerating them (ignoring the one that fractint writes, because it will be the same) if I want to change colours. And that way, I can also write a batch file for rejenerating everything, say if I come across more frames for 25" by 22". If you are any good with key-operated software, you run XP, and you know what "program /?[enter]" does, what a readme.htm file is for, or what the F1 key usually does (help menus), then you need fractint in your path to really appreciate some of the throughput on this list. Control Panel/System/Advanced/Environment Variables/Path/Edit [end];c:\fractint (or wherever you extracted a zip file to--Square brackets indicate a single key). Winfract works, though I can't say it has all the same features, or that it understands what to do when I double-click on a .par file. Here and in sci.fractal, people are much more likely to be interested in software. Everywhere else, I skip the math or sketch it out in English. To answer your last question, David, http://wiki.tcl.tk/10564 That is how you write comments into a PNG. I do not see a utility for doing it. That is source code. I think I am getting a lot of noise in my search from web pages that invite comments.
Jay Litwyn wrote: [snip]
To answer your last question, David, http://wiki.tcl.tk/10564 That is how you write comments into a PNG. I do not see a utility for doing it. That is source code. I think I am getting a lot of noise in my search from web pages that invite comments.
Jason Summer's TweakPNG, http://entropymine.com/jason/tweakpng, is able to manipulate (some) data chunks of PNG images. I've never tested it, though. Regards, Gerald
Jay Litwyn wrote:
Subject-Was: Re: [Fractint] get a form of anti-aliasing applied to your image
----- Original Message ----- From: "david" <gnome@hawaii.rr.com> (...)
I think what GIF supports that the others don't is the ability to essentially include the parameters in the GIF file itself? I seem to recall reading that it used text blocks that the GIF format allowed, but that most graphics programs would clear when they saved a GIF file ... I don't think JPG supports that- does PNG? (...)
wrJPGcom (I used a DOS package for converting tga to jpeg when PSP did not support progressive scan) writes comments into JPEGs. I think EXIF data in JPEGs from cameras is comments.
True. But JPG is a lossy format, I think not suitable for fractals.
I might still hav JPEGs into which I piped fractint formulas and parameter files; not that a lot of people knew they were there. I think if you want to publish (or otherwise archive) formulas or parameter files, then it is better to do that separately, so that people who are not interested in math can easily skip it, and people who are so can find it. Then comes that handy web construct called an anchor (link). On my local system, though, all of my published fractals hav their own directories. That way, I can call one file chromo and one file lumo in each directory, plus batch files for jenerating them (ignoring the one that fractint writes, because it will be the same) if I want to change colours. And that way, I can also write a batch file for rejenerating everything, say if I come across more frames for 25" by 22".
If you are any good with key-operated software, you run XP,
Nope, don't run Windows here.
To answer your last question, David, http://wiki.tcl.tk/10564 That is how you write comments into a PNG. I do not see a utility for doing it. That is source code. I think I am getting a lot of noise in my search from web pages that invite comments.
Then Fractint could presumably write its formulas into PNG files. -- David gnome@hawaii.rr.com authenticity, honesty, community
I don't think anti-aliasing is really worth it. As in I see no point in going beyond 2:1 (4:1 on area). And, I hav some fractals with fine textures; single outer-set points that disappear if I zoom in on them. If I downsample them, then they will fade against a black background. That can't be right (in any case, I do not want that), so for Fiktoid, Rainfly, and TigerDilly, the best output should come from exactly printer resolution. Only on an inkjet will pixel colour be very close to what it is in a file, and so many colours are in those textures that I think luminance is more important. I already hav 600 dpi in some of my files because of ink jets. At some print shops that means more time, even for 300 dpi. My theory is that fractals hav arbitrary resolution, and in some cases, if you downsample them, then you add noise. In a best case, that noise will be invisible, like what a JPEG does at 95% quality. In a worst case, adding that noise will consume a lot more time, and ink bleed will hav much the same effect. I believe in anti-aliasing for text and cartoons. _______ http://ecn.ab.ca/~brewhaha/Fractal_Gallery.HTM
participants (5)
-
david -
Gerald K. Dobiasovsky -
Hal Lane -
JackOfTradeZ＠comcast.net -
Jay Litwyn