I am improving deep zooming in Fracton. Until now, there was a maximum magnification limit of about 1e300. Now that has been increased to 1e5000. While testing the new capability, I tried a deep zoom into a fractal on the FractInt Deep Zoom web page. The fractal was called 233 because it had a magnification of 1e233. I zoomed into a "two way symmetry" until I found the two armed oval this added to the fractal. That wasn't quite deep enough to test the new capability so I kept zooming until I got to this 4 armed oval at a magnification of 4e355. Here is a link to an image: http://dl.dropbox.com/u/33642054/image/for_deep_sake_800_5.jpg The parameter file used to make the image: For_Deep_Sake { ; Exported from Fracton. ; Very deep zoom into 233 by Henry Birdseye ; time = 25 min 52 sec at 640 x 480 with no aa ; on a Mac Pro 2 x 2.8 Ghz Quad-Core Xeon ; by Mike Frazier, public domain reset=2004 type=mandel passes=1 float=y center-mag=-0.732029929828136198455293015271057305\ 56951672230492293928276495641157718102978454448408\ 12404043441296278932140339697145134072822677086003\ 31825208146977091604215917875596703200044518808133\ 12801254326908977150030223083472061881359083259002\ 83828081904425741653511148955679003264312972988485\ 99604251927584934955556100991250337284373373831428\ 61012978245857780794715230007269210452612877056774\ 14739098258016795/0.362254943051205664189634244517\ 66242823791702343502563424963735040314093089208022\ 27682889214854224732520160831281692440946744540034\ 20702120965466655270667981376175896195517699326525\ 26234883861539394603456149411457895604893495093606\ 28103863820000164681452390320212490757742403863463\ 36423030849760452187370470065159297457100639142719\ 24057038297466507954806920894911320212951920345452\ 97489894023152920291521/0.43185401e356/0.1e1/-58/0 params=0/0/0/0/0/0/0/0/0/0 maxiter=10000 inside=1 colors=pH6rH5vH3zH0vH2rG3mE2hD2cB2ZA2U81P71K51F41A\ 30510000330760BA1ED1IG1LJ1PN2SQ2WT2_W2b_3fb3ie3mh3\ ql4tq3wu1zz0ww1tt2pp3mm5jj6gg7cc8``9YYAXVBURBROBOK\ BLHBJDBGABD6CA3C80CC4CHH8MJ9RLBWNC`PEeSFkWInZHqaHt\ cGwfGziFwfFtdFqaEnZEkWEhUEeREaODZMDWJDTGDQDCNBCK8C\ 413000443775BB8EFAIIDLMFPQISUKWXNZ`PbdSegUikXloZpr\ asvcsvcsvcsvcqsaoq`mnZlkXjiWhfUfdTdaRbZP`XO_UMYRKW\ PJUMHVKGWJGXHFYGFZEF_DEaBEbAEd9Dg9Cj8Bl8Ao89r78u77\ r67p67n68l68i59g59e59c5A`5BZ6CW7EU8FR8GP9IMAJKBLMC\ IOFKQHNSKPUNSWQUYSX_VZ`Y`b`cdcfffihiljlolornrulprk\ mnikkgihffdddabbZa`W_YSYWPWUMVRITPFRNCQK8OI5RK7TN9\ WPBYRD`TFbWHeYJg_LiaNkcPneRpgTsiVulYxn_zpawn`tk_pi\ YmfXjdWgaVcZT`WSYURVRQROPOLOKJNHGMDDLAAKADMAGNAIPA\ LQAOSARUBUWBXYB__CbaCecChfCkhDnjDqlDtnCqlBmjBjiAfg\ 9cf8_d8Xc6SY5NS4IM3EH29B145000311632943D54G75J87MA\ 8QB9TCAWEBZFCbHEfHBjH9nH7 } -- Mike Frazier www.fracton.org
Mike wrote:
I zoomed into a "two way symmetry" until I found the two armed oval this added to the fractal. That wasn't quite deep enough to test the new capability so I kept zooming until I got to this 4 armed oval at a magnification of 4e355.
Just curious, how do you know this image is not similar to another at a much shallower zoom? From extensive, albeit not recent, explorations I did years ago, I recall that it is devilishly difficult to find deep zoomed images that are "new". I don't mean that they are rare, but casual zooming without letting images fully develop, which is almost a requirement because of slowness of aribrary precision, can easily fall into vortices of self similarity. Since a lot of time has passed since I tried this, it's possible that images are formed much faster than before, and it's easiere to find new (e.g. not self similar to shallower) images. Tim
On Sat, Oct 22, 2011 at 3:44 PM, Tim Wegner <twegner@swbell.net> wrote:
Mike wrote:
I zoomed into a "two way symmetry" until I found the two armed oval this added to the fractal. That wasn't quite deep enough to test the new capability so I kept zooming until I got to this 4 armed oval at a magnification of 4e355.
Just curious, how do you know this image is not similar to another at a much shallower zoom? From extensive, albeit not recent, explorations I did years ago, I recall that it is devilishly difficult to find deep zoomed images that are "new". I don't mean that they are rare, but casual zooming without letting images fully develop, which is almost a requirement because of slowness of aribrary precision, can easily fall into vortices of self similarity.
Since a lot of time has passed since I tried this, it's possible that images are formed much faster than before, and it's easiere to find new (e.g. not self similar to shallower) images.
Tim
Hi Tim. I did read the comments you posted many years ago about the self similarity of the original "233" fractal and you are undoubtedly correct. However, I don't think a zoom into a two way symmetry point in a self similar image yields exactly the same fractal that can be found at a shallower depth. The reason I say that is that the distance you zoom into a spiral affects the spacing between the rings much deeper near the minibrot. At least that is what I have noticed other times when I have repeatedly zoomed into a spiral before picking some symmetry point to stop the spiral. Anyway, if I am wrong it was still a good test fractal. The reason I picked it, was that it started at a magnification of 1e233 and that saved me a lot of time zooming. I was also curious to see what was in there. Thanks so much for writing FractInt. It inspired a lot of people, me included. I am currently adding some new things to Fracton. I wanted to have Fracton be able to draw the deepest FractInt images so I fixed the magnification limitation. I am also currently working on adding arbitrary precision arithmetic to all formulas. I wrote arbitrary precision versions of all the trig functions and I have a lot of it working. The deep images are really interesting and I think there are going to be some good images there. It is very slow though. The speeds of the formulas I have tested range from 30 to 1000 times slower than the regular double precision math. -- Mike Frazier www.fracton.org
Nice image. I love the way Fracton renders the colors. It would be nice to have a Linux version of it! Mike Frazier wrote:
I am improving deep zooming in Fracton. Until now, there was a maximum magnification limit of about 1e300. Now that has been increased to 1e5000. While testing the new capability, I tried a deep zoom into a fractal on the FractInt Deep Zoom web page. The fractal was called 233 because it had a magnification of 1e233. I zoomed into a "two way symmetry" until I found the two armed oval this added to the fractal. That wasn't quite deep enough to test the new capability so I kept zooming until I got to this 4 armed oval at a magnification of 4e355.
Here is a link to an image:
http://dl.dropbox.com/u/33642054/image/for_deep_sake_800_5.jpg
The parameter file used to make the image:
For_Deep_Sake { ; Exported from Fracton. ; Very deep zoom into 233 by Henry Birdseye ; time = 25 min 52 sec at 640 x 480 with no aa ; on a Mac Pro 2 x 2.8 Ghz Quad-Core Xeon ; by Mike Frazier, public domain reset=2004 type=mandel passes=1 float=y center-mag=-0.732029929828136198455293015271057305\ 56951672230492293928276495641157718102978454448408\ 12404043441296278932140339697145134072822677086003\ 31825208146977091604215917875596703200044518808133\ 12801254326908977150030223083472061881359083259002\ 83828081904425741653511148955679003264312972988485\ 99604251927584934955556100991250337284373373831428\ 61012978245857780794715230007269210452612877056774\ 14739098258016795/0.362254943051205664189634244517\ 66242823791702343502563424963735040314093089208022\ 27682889214854224732520160831281692440946744540034\ 20702120965466655270667981376175896195517699326525\ 26234883861539394603456149411457895604893495093606\ 28103863820000164681452390320212490757742403863463\ 36423030849760452187370470065159297457100639142719\ 24057038297466507954806920894911320212951920345452\ 97489894023152920291521/0.43185401e356/0.1e1/-58/0 params=0/0/0/0/0/0/0/0/0/0 maxiter=10000 inside=1 colors=pH6rH5vH3zH0vH2rG3mE2hD2cB2ZA2U81P71K51F41A\ 30510000330760BA1ED1IG1LJ1PN2SQ2WT2_W2b_3fb3ie3mh3\ ql4tq3wu1zz0ww1tt2pp3mm5jj6gg7cc8``9YYAXVBURBROBOK\ BLHBJDBGABD6CA3C80CC4CHH8MJ9RLBWNC`PEeSFkWInZHqaHt\ cGwfGziFwfFtdFqaEnZEkWEhUEeREaODZMDWJDTGDQDCNBCK8C\ 413000443775BB8EFAIIDLMFPQISUKWXNZ`PbdSegUikXloZpr\ asvcsvcsvcsvcqsaoq`mnZlkXjiWhfUfdTdaRbZP`XO_UMYRKW\ PJUMHVKGWJGXHFYGFZEF_DEaBEbAEd9Dg9Cj8Bl8Ao89r78u77\ r67p67n68l68i59g59e59c5A`5BZ6CW7EU8FR8GP9IMAJKBLMC\ IOFKQHNSKPUNSWQUYSX_VZ`Y`b`cdcfffihiljlolornrulprk\ mnikkgihffdddabbZa`W_YSYWPWUMVRITPFRNCQK8OI5RK7TN9\ WPBYRD`TFbWHeYJg_LiaNkcPneRpgTsiVulYxn_zpawn`tk_pi\ YmfXjdWgaVcZT`WSYURVRQROPOLOKJNHGMDDLAAKADMAGNAIPA\ LQAOSARUBUWBXYB__CbaCecChfCkhDnjDqlDtnCqlBmjBjiAfg\ 9cf8_d8Xc6SY5NS4IM3EH29B145000311632943D54G75J87MA\ 8QB9TCAWEBZFCbHEfHBjH9nH7 }
-- Mike Frazier www.fracton.org <http://www.fracton.org>
------------------------------------------------------------------------
_______________________________________________ Fractint mailing list Fractint@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/fractint
-- David gnome@hawaii.rr.com authenticity, honesty, community
On Sat, Oct 22, 2011 at 5:58 PM, david <gnome@hawaii.rr.com> wrote:
Nice image. I love the way Fracton renders the colors. It would be nice to have a Linux version of it!
Thanks. You can anti-alias FractInt images too by drawing them larger and shrinking them in an external image manipulation program. Hal Lane does this on his web site. Converting Fracton to Linux would be very difficult. It is 30,000 lines of Objective C code now. A lot of the code interacts with the user interface and that is totally Mac centric. The image was 5x5 anti-aliased which means it took 25 times as long to draw as a non-antialiased one. I ran it on my old 2 core PowerPC Mac that I use for long rendering jobs. It took about a week on that machine. -- Mike Frazier www.fracton.org
Mike Frazier wrote:
On Sat, Oct 22, 2011 at 5:58 PM, david wrote:
Nice image. I love the way Fracton renders the colors. It would be nice to have a Linux version of it!
Thanks. You can anti-alias FractInt images too by drawing them larger and shrinking them in an external image manipulation program. Hal Lane does this on his web site.
And here I thought Fracton was generating the images that way ... does Fracton work in 256-colors like Fractint does?
Converting Fracton to Linux would be very difficult. It is 30,000 lines of Objective C code now. A lot of the code interacts with the user interface and that is totally Mac centric.
If the code is cleanly separated between UI classes and fractal calculation classes, I think all you'd need is a documented API for using the fractal calculation classes. Then a Linux UI could be coded separately using any of the common Linux graphics frameworks (QT, GTK, whatever).
The image was 5x5 anti-aliased which means it took 25 times as long to draw as a non-antialiased one. I ran it on my old 2 core PowerPC Mac that I use for long rendering jobs. It took about a week on that machine.
Sounds like JackOTradz setup for generating fractal animations. -- David gnome@hawaii.rr.com authenticity, honesty, community
On Sun, Oct 23, 2011 at 3:53 PM, david <gnome@hawaii.rr.com> wrote:
Mike Frazier wrote:
On Sat, Oct 22, 2011 at 5:58 PM, david wrote:
Nice image. I love the way Fracton renders the colors. It would be nice to have a Linux version of it!
Thanks. You can anti-alias FractInt images too by drawing them larger and shrinking them in an external image manipulation program. Hal Lane does this on his web site.
And here I thought Fracton was generating the images that way ... does Fracton work in 256-colors like Fractint does?
Fracton has 24 bit color and does the anti-aliasing inside the program. I was just suggesting it does exactly the same thing that can be done with FractInt and an image manipulation program. Converting Fracton to Linux would be very difficult. It is 30,000 lines of
Objective C code now. A lot of the code interacts with the user interface and that is totally Mac centric.
If the code is cleanly separated between UI classes and fractal calculation classes, I think all you'd need is a documented API for using the fractal calculation classes. Then a Linux UI could be coded separately using any of the common Linux graphics frameworks (QT, GTK, whatever).
It is separated that way and uses a model, view, controller architecture but there is still so much that would be difficult to port. For example, all the multiprocessor stuff is totally Mac centric since you have to have the OS handle that. The equation compiler uses tons of Cocoa string object constructs. There are dozens of Cocoa objects that are used everywhere. It might be easier to just start over for a lot of it. Anyway, isn't there already a Linux version of FractInt? -- Mike Frazier www.fracton.org
Mike Frazier wrote:
On Sun, Oct 23, 2011 at 3:53 PM, david <gnome@hawaii.rr.com <mailto:gnome@hawaii.rr.com>> wrote:
Mike Frazier wrote:
On Sat, Oct 22, 2011 at 5:58 PM, david wrote:
Nice image. I love the way Fracton renders the colors. It would be nice to have a Linux version of it!
Thanks. You can anti-alias FractInt images too by drawing them larger and shrinking them in an external image manipulation program. Hal Lane does this on his web site.
And here I thought Fracton was generating the images that way ... does Fracton work in 256-colors like Fractint does?
Fracton has 24 bit color and does the anti-aliasing inside the program. I was just suggesting it does exactly the same thing that can be done with FractInt and an image manipulation program.
Converting Fracton to Linux would be very difficult. It is 30,000 lines of Objective C code now. A lot of the code interacts with the user interface and that is totally Mac centric.
If the code is cleanly separated between UI classes and fractal calculation classes, I think all you'd need is a documented API for using the fractal calculation classes. Then a Linux UI could be coded separately using any of the common Linux graphics frameworks (QT, GTK, whatever).
It is separated that way and uses a model, view, controller architecture but there is still so much that would be difficult to port. For example, all the multiprocessor stuff is totally Mac centric since you have to have the OS handle that.
Could be worked around.
The equation compiler uses tons of Cocoa string object constructs. There are dozens of Cocoa objects that are used everywhere. It might be easier to just start over for a lot of it.
Maybe, maybe not. I'm not a programmer, although I've done my share of programming (wrote a mandelbrot grapher in 6502 Assembler for the C64 a few ages ago).
Anyway, isn't there already a Linux version of FractInt?
There is (that's what I use), but it has the same 256-color limitation as DOS Fractint. -- David gnome@hawaii.rr.com authenticity, honesty, community
Mike Frazier wrote:
I am improving deep zooming in Fracton. Until now, there was a maximum magnification limit of about 1e300. Now that has been increased to 1e5000. While testing the new capability, I tried a deep zoom into a fractal on the FractInt Deep Zoom web page. The fractal was called 233 because it had a magnification of 1e233. I zoomed into a "two way symmetry" until I found the two armed oval this added to the fractal. That wasn't quite deep enough to test the new capability so I kept zooming until I got to this 4 armed oval at a magnification of 4e355.
[snip] Not so long ago I rummaged through old images of Spanky at the Waybackmachine, just to see if they had any original GIFs with embedded Fractint parameters (since Noel Giffin didn't extract the latter when switching to PNG). At the bottom of the Fractint deepzoom page are three (very) small images by Minoru Morikawa with high magnifications I was able to get the GIFs from (provided you need further test cases): M1280CM {;Period = 1280, original by Minoru Morikawa. ;Parameter extracted from .gif of Spanky Fractal Database ;via WaybackMachine. reset=1920 type=mandel center-mag=-1.9999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999658333731\ 427438790145077247121847974138182233607055090006418913464666623825200902\ 155469267606911089906942691124679201168459760973710944897246128552776529\ 257685328240100080237998003833424261830949072944282449278696055187392357\ 683804713719163678356272241943553833135337893951888658749096082084285053\ 555457760598498046920986968370288634889628983724485904342751959658269359\ 498331427854724345390358490172382231264180563748664331807986467844778158\ 158377270189634175632861186201380920422544975453692340192775113583837308\ 812893227918466404755168396106046516402553176888619654978752532269376363\ 118604895912513512333424234635506995308169804308417754367835394894867048\ 398976938524724576747952067109371244038108742369661630611685459110402718\ 46686949227460530920915399561672380621965455/0.0/1.585247e+1539 params=0/0 float=y maxiter=40000 inside=0 viewwindows=4.2/0.75/yes/0/0 colors=@default.map } M1280CM {;Period = 1280, original by Minoru Morikawa. ;Parameter extracted from .gif of Spanky Fractal Database ;via WaybackMachine. reset=1920 type=mandel center-mag=-1.9999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999658333731\ 427438790145077247121847974138182233607055090006418913464666623825200902\ 155469267606911089906942691124679201168459760973710944897246128552776529\ 257685328240100080237998003833424261830949072944282449278696055187392357\ 683804713719163678356272241943553833135337893951888658749096082084285053\ 555457760598498046920986968370288634889628983724485904342751959658269359\ 498331427854724345390358490172382231264180563748664331807986467844778158\ 158377270189634175632861186201380920422544975453692340192775113583837308\ 812893227918466404755168396106046516402553176888619654978752532269376363\ 118604895912513512333424234635506995308169804308417754367835394894867048\ 398976938524724576747952067109371244038108742369661630611685459110402718\ 46686949227460530920915399561672380621965279874/0.0/4.058231e+1542 params=0/0 float=y maxiter=40000 inside=0 viewwindows=4.2/0.75/yes/0/0 colors=000ZTcZQb_MaTia<3>CqG8sA3v4XmTXsFXz1Zhb<3>ioNkqJnsFquAgefrbgVbf<3\
MFfJ9eVhd<3>MoXKqVHsTai`<2>poHVgf<3>KegHdhWcf<3>RLcPGbaeg<3>xUoYcb<2>aR\ R_N2UgMUaPRW6_ga<3>ljFZad<3>hEWUfd<3>H`UDZSAYP6WM_X`<2>k0JV_iTRmRIq`gf<3\ rjcwkbVUYTGPR1FUmfRsfNyfVg`<3>Jd9uIaJXuMogBwhcWhjKkr8nbcch_`oWXUhfRjfOl\ g`f`deUhdNXZaXQXXae<3>XCYY5WUgc<3>IiONeQCc9_QOc84Ydf<3>cTieQjgMkZ`d<3>g6\ T`kYeoPjtF`c`d_VhVO_hf<3>mofpqftsgXYZ<2>V2BRVeZwnYcb<3>dLNfGJhBFj6BXid<3\ sgg<3>eRWaNTZIQVENR9KNUDJn6<3>kuMrwQzyV<3>kwLhwJdvG`vEXuB<3>EvC9wDHuJ<2\ eocj_`<3>fflehodjrclubny<3>UphRpcPq_MrV<3>mrVtrUpqR<3>`kFXjCTh9<5>7NeHU\ c<2>DU` }
M320PAF {;Period = 320, original by Minoru Morikawa. ;Parameter extracted from .gif of Spanky Fractal Database ;via WaybackMachine. reset=2004 type=mandel center-mag=-1.9999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999999999999999\ 999999999999999999999999999999999999999999999999999999999999967551563204\ 016794975048689625419617134198762521084576514482332215352782310319946823\ 796412815029605503843219603114657605645820515330288131871104025361457604\ 2963548464247337044885580413808058348126413/0.0/1.124847e+386 params=0/0 float=y maxiter=20000 inside=0 colors=000Q_QAm8XVeOcdElcYHcQC`I7YZNWROKJQ8kOkrQpySvbHKpkDXZjOkoFytnSUxY\ HjNZpPQFBwaNWYOLUQ9dHj<2>`0xbIY_DOX8EaXUYgGUr2dXgchhati_OQUQ9hVTldFpn1od\ lyxrXW_<2>3yDTFjF8n11rWKeMIdBGc_Md<2>FOXlk`kXfrhfyteeUY<2>fs5_a_TrSiBjn0\ oYH`<2>A2HnaRwqBCE2mL`vJVlQYsVOz_EeJ_<2>b9DiMa<2>wPLUF_H7TSZYDkPkVcqc_xm\ WhSbkYYlZbskZzxUODkd`Pco6vFjjRgoXidBmc0uYPa<2>8`KZZVfVegccilaSZiDkldSbbY\ Z`cVZiRZKb<2>EERbESZ6CWVVLdJAn68QJcTka`pZhueNj<2>cTyuIFiMj<2>wPwjMepMdvN\ caNc<2>ORSkKgrIhaIYYDOT8EoSVzZJjMf<2>zPiVFgJ8h70jZVb<2>CxQY_RPmBkSWrZLye\ AM_N1n2_JiTGmMDq_WgTehMpjiRTnWFsa0oMKaLeYKdVLbUJcQHaHoNYkCog0Fmx<2>pm0RR\ p<3>Al1<3>DpaOnVZlN<2>Ws9FHE5F43dBSUWEbKVRWJWKmY_viSoObyQYeTf<2>dnhkY_qi\ TCGWUX`IgU6rN`RY<2>Kh7aRTXWFS`1aYgXihZzdmFWv8LaKWYIKUG8VbQJt9 } Regards, Gerald
Not so long ago I rummaged through old images of Spanky at the Waybackmachine, just to see if they had any original GIFs with embedded Fractint parameters (since Noel Giffin didn't extract the latter when switching to PNG).
At the bottom of the Fractint deepzoom page are three (very) small images by Minoru Morikawa with high magnifications I was able to get the GIFs from (provided you need further test cases):
Those are great because they are over 1e300 and have comparison images. I tried the first two and they are so slow it will take a while to see it they work. The last one makes an image in about 30 sec with a very tiny window. There was a difference in a portion of the image so I need to see why. Thanks for posting them. -- Mike Frazier www.fracton.org
participants (4)
-
david -
Gerald K. Dobiasovsky -
Mike Frazier -
Tim Wegner