From clock at twibright.com Mon Jan 1 20:35:50 2007 From: clock at twibright.com (Karel Kulhavy) Date: Mon, 1 Jan 2007 21:35:50 +0100 Subject: [Ronja] Making the instructions easier In-Reply-To: <456E03F6.7020507@kbx.cz> References: <20061128185003.GA21523@kestrel.twibright.com> <20061129113636.GB16674@kestrel.twibright.com> <456E03F6.7020507@kbx.cz> Message-ID: <20070101203550.GA31407@kestrel.twibright.com> On Wed, Nov 29, 2006 at 11:04:38PM +0100, Jakub Sykora wrote: > What about replacing Models with Ronja Models? Ordinary reader asks > Models of what? What models? Adding an hyphen with model names is a good > idea indeed. I replaced it. CL< > > If you would like to confuse readers at maximum, you can use Ronja > Models' Modules or maybe better Ronja Models'n'Modules and mix the > contents together. But I think this is hell highway... > > K > > Karel Kulhavy wrote: > > On Tue, Nov 28, 2006 at 10:30:42PM +0100, J.D. Bakker wrote: > >>> Do you have an idea how the Ronja instructions could be made easier? > >> Two related points: > >> > >> - The structuring of the instructions makes it very inconvenient to > >> find specific things. Not every user will find it convenient to > >> follow the path you force them to take. For example, I have often > >> looked for particular parts of the instructions, such as the BOM for > >> the twister2, and found it very frustrating to have to hunt through > >> different sections. If you must retain the current structure, I would > >> suggest you make the 'Contents' pages (like > >> http://ronja.twibright.com/tetrapolis/) more verbose [I see you have > >> done some restructuring since I last looked, but I would still > >> suggest adding more info to the 'Contents'-page] > > > > I don't have to retain the current structure. It could be changed but it > > would probably take a lot of time, depending on how systemic change it would > > be. > > > > Now I got an idea about the "Models" and "Modules" on the title page. If you > > see "models" and "modules", you have to max out your brain to grab these > > abstract concepts, figure out the subtle difference between those two almost > > identical words, and figure out if these abstract concepts encompass the real > > thing (Twister) you are looking for. > > > > What about replacing Models with "Models - Tetrapolis, Inferno, ... where > > Tetrapolis, Inferno and ... would be links > > and Modules with Modules - Twister2, RX, TX, infra RX, ... > > > > and do the same with the contents pages like > > http://ronja.twibright.com/tetrapolis/ ? > > > > Any suggestion what more could be added to those http://ronja.twibright.com/tetrapolis/ -style > > content pages? > > > >> - Similarly, you may want to add basic navigation buttons on the top > >> and bottom of all build pages. A page like > >> http://ronja.twibright.com/twister2/tools.php might include links > >> such as 'Back - to materials', 'Forward - to PCB manufacture' and 'Up > >> - to (Building the) Ronja Twister2'. > > > > Such navigation is a bit confusing for me personally. You have multiple routes > > to the same destination. Does it have any more benefits than that it spares > > one hit of a back button? > > > > CL< > >> JDB. > >> -- > >> LART. 250 MIPS under one Watt. Free hardware design files. > >> http://www.lartmaker.nl/ > >> > >> -- > >> Twibright Ronja mailing list http://ronja.twibright.com > >> Ronja at lists.pointless.net > >> http://pointless.net/mailman/listinfo/ronja > > > > -- > Jakub S?kora > email: kubajz at kbx.cz <') > ICQ: 68976632 ( =- > mobil: +420 777 594 201 '' > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From clock at twibright.com Mon Jan 8 07:23:32 2007 From: clock at twibright.com (Karel Kulhavy) Date: Mon, 8 Jan 2007 08:23:32 +0100 Subject: [Ronja] 3D videos of Ronja parts Message-ID: <20070108072332.GA31650@kestrel.twibright.com> When making the 3D model for the Tetrax console I realized that the model is quite complicated and people could have problems understanding how the parts are mutually positioned. Therefore I made a video of the part rotating. When I made the infrastructure for the video rendering it was easy to automatically render videos of other parts too. You can see all the 10 videos at http://ronja.twibright.com/3d/ - look for small turning icons. CL< From kubajz at kbx.cz Mon Jan 8 09:10:41 2007 From: kubajz at kbx.cz (=?ISO-8859-1?Q?Kuba_S=FDkora?=) Date: Mon, 08 Jan 2007 10:10:41 +0100 Subject: [Ronja] 3D videos of Ronja parts In-Reply-To: <20070108072332.GA31650@kestrel.twibright.com> References: <20070108072332.GA31650@kestrel.twibright.com> Message-ID: <45A20A91.7060204@kbx.cz> Is the software used cabaple of WRML export? I think it is better to transfer smaller file (than rendered video) and utilize users graphic accelerator. What do you think about it? K Karel Kulhavy wrote: > When making the 3D model for the Tetrax console I realized that the model is > quite complicated and people could have problems understanding how the parts > are mutually positioned. Therefore I made a video of the part rotating. > > When I made the infrastructure for the video rendering it was easy to > automatically render videos of other parts too. > > You can see all the 10 videos at http://ronja.twibright.com/3d/ - look for > small turning icons. > > CL< > From honza at hoidekr.net Tue Jan 9 09:12:20 2007 From: honza at hoidekr.net (Jan Hoidekr) Date: Tue, 09 Jan 2007 10:12:20 +0100 Subject: [Ronja] Ronja na prodej In-Reply-To: <20061231173347.GA30198@kestrel.twibright.com> References: <4585A793.9070008@hoidekr.net> <20061231173347.GA30198@kestrel.twibright.com> Message-ID: <45A35C74.3090002@hoidekr.net> Link je zadan v galerii s poznamkou "demontovan 2006". Je zvlastni, ze se zatim neozval zadny zajemce. Doufam, ze spoj neskonci ve sberu. Byla by to skoda. Udelame akcni cenu - za odvoz ;-) pro upresneni: okres Prachatice, Zernovice jsou 5km od Prachatic Honza Ahoj. > Mame na prodej cely spoj Ronja, ktery byl umisten 3 roky v > siti na vesnici. Staveli jsme ho jako nadsenci do teto > technologie, kdy WiFi bylo drahe, a meli jsme dostatek casu. > > Spoj bezel na vzdalenost asi 150m. Vypadky byly jen pri > mlze, kdy nebylo videt na druhy konec. > > Popis: > Kompletni spoj - 2ks Twister, 2ks vysilaci a prijimaci cast > (RX/TX + mechanika) > Cocky TX 9cm, RX 10cm (vietnamci prodali 2 mensi ...) > Napajeci zdroj pro Twister > TP kabely s pripravenym napajenim > > Zavady: > - vytapeni jsme nikdy nepouzili (cocky se nemlzily ... ) > - jedna krabicka twisteru se rozpadla, ale na funkci to nema > vliv ;-) > - chybi jeden napajeci zdroj > > Dale nabizime: > - 2ks AUI interface, ktere fungovaly pred nasazenim Twisteru > > - kompletni elektroniku dalsiho spoje, ktery jsme > nedodelali. (2ks Twister, 2ks RX, 2ks TX). Twister je na > PCB, RX, TX ve hnizde. V elektronice je nejaka chyba, kdy > data chvilku prochazi a chvilku ne, staci se toho dotknout, > projit okolo apod. Vice jsem to uz neresili a lezi to ve skrini. > > Fotky na www.zernovice.net/ronja From ladmanj at volny.cz Tue Jan 9 10:29:54 2007 From: ladmanj at volny.cz (Jakub Ladman) Date: Tue, 9 Jan 2007 11:29:54 +0100 Subject: [Ronja] 3D videos of Ronja parts In-Reply-To: <20070108072332.GA31650@kestrel.twibright.com> References: <20070108072332.GA31650@kestrel.twibright.com> Message-ID: <200701091129.54649.ladmanj@volny.cz> Nejak to nerespektuje normalni zakryvani jedne trubky druhou, zda se mi, ze to tam nejak divne preskakuje z popredi do pozadi, ale je to prilis rychle a kratke, abych to mohl presne popsat. Jakub Ladman Dne pond?l? 08 leden 2007 08:23 Karel Kulhavy napsal(a): > When making the 3D model for the Tetrax console I realized that the model > is quite complicated and people could have problems understanding how the > parts are mutually positioned. Therefore I made a video of the part > rotating. > > When I made the infrastructure for the video rendering it was easy to > automatically render videos of other parts too. > > You can see all the 10 videos at http://ronja.twibright.com/3d/ - look for > small turning icons. > > CL< From ladmanj at volny.cz Tue Jan 9 10:34:55 2007 From: ladmanj at volny.cz (Jakub Ladman) Date: Tue, 9 Jan 2007 11:34:55 +0100 Subject: [Ronja] 3D videos of Ronja parts In-Reply-To: <200701091129.54649.ladmanj@volny.cz> References: <20070108072332.GA31650@kestrel.twibright.com> <200701091129.54649.ladmanj@volny.cz> Message-ID: <200701091134.55820.ladmanj@volny.cz> Uz mi to nedela, tak nevim jestli se mi to zdalo, nebo ne. Jakub Dne ?ter? 09 leden 2007 11:29 Jakub Ladman napsal(a): > Nejak to nerespektuje normalni zakryvani jedne trubky druhou, zda se mi, ze > to tam nejak divne preskakuje z popredi do pozadi, ale je to prilis rychle > a kratke, abych to mohl presne popsat. > Jakub Ladman > > Dne pond?l? 08 leden 2007 08:23 Karel Kulhavy napsal(a): > > When making the 3D model for the Tetrax console I realized that the model > > is quite complicated and people could have problems understanding how the > > parts are mutually positioned. Therefore I made a video of the part > > rotating. > > > > When I made the infrastructure for the video rendering it was easy to > > automatically render videos of other parts too. > > > > You can see all the 10 videos at http://ronja.twibright.com/3d/ - look > > for small turning icons. > > > > CL< From clock at twibright.com Tue Jan 9 13:55:10 2007 From: clock at twibright.com (Karel Kulhavy) Date: Tue, 9 Jan 2007 14:55:10 +0100 Subject: [Ronja] 3D videos of Ronja parts In-Reply-To: <45A20A91.7060204@kbx.cz> References: <20070108072332.GA31650@kestrel.twibright.com> <45A20A91.7060204@kbx.cz> Message-ID: <20070109135509.GA23233@kestrel.twibright.com> On Mon, Jan 08, 2007 at 10:10:41AM +0100, Kuba S?kora wrote: > Is the software used cabaple of WRML export? I think it is better to > transfer smaller file (than rendered video) and utilize users graphic > accelerator. Seems to: clock at kestrel:~$ g-vrml Usage: g-vrml [-v] [-xX lvl] [-d tolerance_distance (mm) ] [-a abs_tol (mm)] [-r rel_tol] [-n norm_tol] [-o out_file] [-u units] brlcad_db.g object(s) > > What do you think about it? I don't think the video should be replaced by vrml. The BRL-CAD raytracer renders reflections, refractions, and shadows by default. It is a CSG so the picture is geometrically perfect, as opposed to what graphics accelerators can only do - interpolate bitmaps over a triangular approximation of the object. But exporting into VRML too is a good idea. BRL-CAD comes by default with the g-vrml tool that is the same family as g-dxf tool I am already using for AutoCAD export. If someone can verify that the VRML generated files are meaningful then I can add it. It should be able to export into X3D too: clock at kestrel:~$ g-x3d Usage: g-x3d [-v] [-xX lvl] [-d tolerance_distance (mm) ] [-a abs_tol (mm)] [-r rel_tol] [-n norm_tol] [-o out_file] [-u units] brlcad_db.g object(s) Do you think I should add X3D output too? Do you have a VRML viewer on your PC and could you verify it? I tried to download FreeWRL to test myself. There is no package for OpenBSD, neither in OpenBSD nor on FreeWRL page. So I downloade the source. I saw in the README.html that javac is necessary. However javac doesn't work on OpenBSD 3.9 I have: clock at kestrel:~$ javac Segmentation fault (core dumped) clock at kestrel:~$ javac --version Segmentation fault (core dumped) clock at kestrel:~$ gdb /usr/local/libexec/Kaffe Kaffe.core (gdb) bt full #0 0x0a308efa in callMethodA () from /usr/local/lib/libkaffevm.so.0.0 No symbol table info available. #1 0x0a2e9e92 in processClass () from /usr/local/lib/libkaffevm.so.0.0 No symbol table info available. #2 0x0a2e996d in initBaseClasses () from /usr/local/lib/libkaffevm.so.0.0 No symbol table info available. #3 0x0a2e9739 in initialiseKaffe () from /usr/local/lib/libkaffevm.so.0.0 No symbol table info available. #4 0x0a2fa34c in JNI_CreateJavaVM () from /usr/local/lib/libkaffevm.so.0.0 No symbol table info available. #5 0x1c000c25 in main () No symbol table info available. CL< > > K > > Karel Kulhavy wrote: > > When making the 3D model for the Tetrax console I realized that the model is > > quite complicated and people could have problems understanding how the parts > > are mutually positioned. Therefore I made a video of the part rotating. > > > > When I made the infrastructure for the video rendering it was easy to > > automatically render videos of other parts too. > > > > You can see all the 10 videos at http://ronja.twibright.com/3d/ - look for > > small turning icons. > > > > CL< > > > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From clock at twibright.com Tue Jan 9 13:59:52 2007 From: clock at twibright.com (Karel Kulhavy) Date: Tue, 9 Jan 2007 14:59:52 +0100 Subject: [Ronja] 3D videos of Ronja parts In-Reply-To: <200701091129.54649.ladmanj@volny.cz> References: <20070108072332.GA31650@kestrel.twibright.com> <200701091129.54649.ladmanj@volny.cz> Message-ID: <20070109135952.GC23233@kestrel.twibright.com> On Tue, Jan 09, 2007 at 11:29:54AM +0100, Jakub Ladman wrote: > Nejak to nerespektuje normalni zakryvani jedne trubky druhou, zda se mi, ze to > tam nejak divne preskakuje z popredi do pozadi, ale je to prilis rychle a > kratke, abych to mohl presne popsat. > Jakub Ladman To ja jsem se na to koukal a nic takovyho jsem nepozoroval. CL< > > Dne pond?l? 08 leden 2007 08:23 Karel Kulhavy napsal(a): > > When making the 3D model for the Tetrax console I realized that the model > > is quite complicated and people could have problems understanding how the > > parts are mutually positioned. Therefore I made a video of the part > > rotating. > > > > When I made the infrastructure for the video rendering it was easy to > > automatically render videos of other parts too. > > > > You can see all the 10 videos at http://ronja.twibright.com/3d/ - look for > > small turning icons. > > > > CL< > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From kubajz at kbx.cz Tue Jan 9 14:22:33 2007 From: kubajz at kbx.cz (Jakub Sykora) Date: Tue, 09 Jan 2007 15:22:33 +0100 Subject: [Ronja] 3D videos of Ronja parts In-Reply-To: <20070109135509.GA23233@kestrel.twibright.com> References: <20070108072332.GA31650@kestrel.twibright.com> <45A20A91.7060204@kbx.cz> <20070109135509.GA23233@kestrel.twibright.com> Message-ID: <45A3A529.2070100@kbx.cz> Karel Kulhavy napsal(a): > On Mon, Jan 08, 2007 at 10:10:41AM +0100, Kuba S?kora wrote: >> Is the software used cabaple of WRML export? I think it is better to >> transfer smaller file (than rendered video) and utilize users graphic >> accelerator. > > Seems to: > clock na kestrel:~$ g-vrml > Usage: g-vrml [-v] [-xX lvl] [-d tolerance_distance (mm) ] [-a abs_tol (mm)] > [-r rel_tol] [-n norm_tol] [-o out_file] [-u units] brlcad_db.g object(s) > >> What do you think about it? > > I don't think the video should be replaced by vrml. The BRL-CAD raytracer > renders reflections, refractions, and shadows by default. It is a CSG > so the picture is geometrically perfect, as opposed to what graphics > accelerators can only do - interpolate bitmaps over a triangular approximation > of the object. > I dont think so too - I think it could be another format which offer some benefits. I dont think the model could be corrupted in such manner it would be unusuable. Colored mesh is viewed perfectly - and you can always switch to software renderer which is almost perfect. > But exporting into VRML too is a good idea. BRL-CAD comes by default with the > g-vrml tool that is the same family as g-dxf tool I am already using for > AutoCAD export. If someone can verify that the VRML generated files are > meaningful then I can add it. Try it and I will try to view it. > > It should be able to export into X3D too: > clock na kestrel:~$ g-x3d > Usage: g-x3d [-v] [-xX lvl] [-d tolerance_distance (mm) ] [-a abs_tol (mm)] [-r rel_tol] [-n norm_tol] [-o out_file] [-u units] brlcad_db.g object(s) > > Do you think I should add X3D output too? > I never used X3D - I will have to try it. > Do you have a VRML viewer on your PC and could you verify it? > Yes and I will do. > I tried to download FreeWRL to test myself. There is no package for OpenBSD, > neither in OpenBSD nor on FreeWRL page. So I downloade the source. I saw in the > README.html that javac is necessary. However javac doesn't work on OpenBSD 3.9 > I have: > > clock na kestrel:~$ javac > Segmentation fault (core dumped) > clock na kestrel:~$ javac --version > Segmentation fault (core dumped) > clock na kestrel:~$ gdb /usr/local/libexec/Kaffe Kaffe.core > (gdb) bt full > #0 0x0a308efa in callMethodA () from /usr/local/lib/libkaffevm.so.0.0 > No symbol table info available. > #1 0x0a2e9e92 in processClass () from /usr/local/lib/libkaffevm.so.0.0 > No symbol table info available. > #2 0x0a2e996d in initBaseClasses () from /usr/local/lib/libkaffevm.so.0.0 > No symbol table info available. > #3 0x0a2e9739 in initialiseKaffe () from /usr/local/lib/libkaffevm.so.0.0 > No symbol table info available. > #4 0x0a2fa34c in JNI_CreateJavaVM () from /usr/local/lib/libkaffevm.so.0.0 > No symbol table info available. > #5 0x1c000c25 in main () > No symbol table info available. > > CL< What a pity :/ > >> K >> >> Karel Kulhavy wrote: >>> When making the 3D model for the Tetrax console I realized that the model is >>> quite complicated and people could have problems understanding how the parts >>> are mutually positioned. Therefore I made a video of the part rotating. >>> >>> When I made the infrastructure for the video rendering it was easy to >>> automatically render videos of other parts too. >>> >>> You can see all the 10 videos at http://ronja.twibright.com/3d/ - look for >>> small turning icons. >>> >>> CL< >>> >> -- >> Twibright Ronja mailing list http://ronja.twibright.com >> Ronja na lists.pointless.net >> http://pointless.net/mailman/listinfo/ronja > -- Jakub S?kora email: kubajz na kbx.cz <') ICQ: 68976632 ( =- mobil: +420 777 594 201 '' ------------- dal?? ??st --------------- Netextov? p??loha byla odstran?na... Jm?no: kubajz.vcf Typ: text/x-vcard Velikost: 265 bytes Popis: [??dn? popis nen? k dispozici] Url : http://pointless.net/pipermail/ronja/attachments/20070109/8554980f/attachment.vcf From petr.cipis at tiscali.cz Tue Jan 9 17:30:33 2007 From: petr.cipis at tiscali.cz (Cipis) Date: Tue, 9 Jan 2007 18:30:33 +0100 Subject: [Ronja] Ronja na prodej References: <4585A793.9070008@hoidekr.net><20061231173347.GA30198@kestrel.twibright.com> <45A35C74.3090002@hoidekr.net> Message-ID: <004501c73413$ddcad8f0$5046a8c0@cipis.net> Ahoj, no, tak tou akcn? cenou jsi me navnadil :-) Akor?t to m?m k v?m nejak?ch 250km, tak prem?sl?m ... Cipis ----- Original Message ----- From: "Jan Hoidekr" To: "Twibright Ronja" Sent: Tuesday, January 09, 2007 10:12 AM Subject: Re: [Ronja] Ronja na prodej > Link je zadan v galerii s poznamkou "demontovan 2006". > > Je zvlastni, ze se zatim neozval zadny zajemce. Doufam, ze > spoj neskonci ve sberu. Byla by to skoda. > Udelame akcni cenu - za odvoz ;-) > > pro upresneni: > okres Prachatice, Zernovice jsou 5km od Prachatic > > Honza > > Ahoj. > > Mame na prodej cely spoj Ronja, ktery byl umisten 3 roky v > > siti na vesnici. Staveli jsme ho jako nadsenci do teto > > technologie, kdy WiFi bylo drahe, a meli jsme dostatek casu. > > > > Spoj bezel na vzdalenost asi 150m. Vypadky byly jen pri > > mlze, kdy nebylo videt na druhy konec. > > > > Popis: > > Kompletni spoj - 2ks Twister, 2ks vysilaci a prijimaci cast > > (RX/TX + mechanika) > > Cocky TX 9cm, RX 10cm (vietnamci prodali 2 mensi ...) > > Napajeci zdroj pro Twister > > TP kabely s pripravenym napajenim > > > > Zavady: > > - vytapeni jsme nikdy nepouzili (cocky se nemlzily ... ) > > - jedna krabicka twisteru se rozpadla, ale na funkci to nema > > vliv ;-) > > - chybi jeden napajeci zdroj > > > > Dale nabizime: > > - 2ks AUI interface, ktere fungovaly pred nasazenim Twisteru > > > > - kompletni elektroniku dalsiho spoje, ktery jsme > > nedodelali. (2ks Twister, 2ks RX, 2ks TX). Twister je na > > PCB, RX, TX ve hnizde. V elektronice je nejaka chyba, kdy > > data chvilku prochazi a chvilku ne, staci se toho dotknout, > > projit okolo apod. Vice jsem to uz neresili a lezi to ve > skrini. > > > > Fotky na www.zernovice.net/ronja > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja na lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From cuma at tyszkiewicz.info Wed Jan 10 08:21:04 2007 From: cuma at tyszkiewicz.info (Cuma Tyszkiewicz) Date: Wed, 10 Jan 2007 09:21:04 +0100 Subject: [Ronja] test points on the twister PCB Message-ID: <200701100921.04452.cuma@tyszkiewicz.info> Hello Would you tell me where can I find the list of testing points on the twister scheme. Best Regards Cuma Tyszkiewicz From clock at twibright.com Wed Jan 10 09:52:44 2007 From: clock at twibright.com (Karel Kulhavy) Date: Wed, 10 Jan 2007 10:52:44 +0100 Subject: [Ronja] test points on the twister PCB In-Reply-To: <200701100921.04452.cuma@tyszkiewicz.info> References: <200701100921.04452.cuma@tyszkiewicz.info> Message-ID: <20070110095244.GB1088@kestrel.twibright.com> On Wed, Jan 10, 2007 at 09:21:04AM +0100, Cuma Tyszkiewicz wrote: > Hello > > Would you tell me where can I find the list of testing points on the twister > scheme. Can you use the Twister2 list? The Twister is obsolete. CL< > > Best Regards > > Cuma Tyszkiewicz > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From cuma at tyszkiewicz.info Wed Jan 10 10:59:21 2007 From: cuma at tyszkiewicz.info (Cuma Tyszkiewicz) Date: Wed, 10 Jan 2007 11:59:21 +0100 Subject: [Ronja] test points on the twister PCB In-Reply-To: <20070110095244.GB1088@kestrel.twibright.com> References: <200701100921.04452.cuma@tyszkiewicz.info> <20070110095244.GB1088@kestrel.twibright.com> Message-ID: <200701101159.22013.cuma@tyszkiewicz.info> Dnia ?roda 10 stycznia 2007 10:52, Karel Kulhavy napisa?: > On Wed, Jan 10, 2007 at 09:21:04AM +0100, Cuma Tyszkiewicz wrote: > > Hello > > > > Would you tell me where can I find the list of testing points on the > > twister scheme. > > Can you use the Twister2 list? The Twister is obsolete. > > CL< > > > Best Regards > > > > Cuma Tyszkiewicz > > > > -- > > Twibright Ronja mailing list http://ronja.twibright.com > > Ronja at lists.pointless.net > > http://pointless.net/mailman/listinfo/ronja I have two Twisters. Both of them passed the Sanity Check and Packetloss on wire loopback tests. Next I tried to test both Twisters connecting them with crossovered TP. This test failed and after that both twisters are unable to pass the Sanity Check and Packetloss on wire loopback tests. I want to localize the corruption so I need to know where shall I measure the voltage on the Twister and what the proper voltage levels are. Can you help me? Cuma Tyszkiewicz From clock at twibright.com Wed Jan 10 13:34:59 2007 From: clock at twibright.com (Karel Kulhavy) Date: Wed, 10 Jan 2007 14:34:59 +0100 Subject: [Ronja] 3D videos of Ronja parts In-Reply-To: <45A3A529.2070100@kbx.cz> References: <20070108072332.GA31650@kestrel.twibright.com> <45A20A91.7060204@kbx.cz> <20070109135509.GA23233@kestrel.twibright.com> <45A3A529.2070100@kbx.cz> Message-ID: <20070110133459.GA29659@kestrel.twibright.com> On Tue, Jan 09, 2007 at 03:22:33PM +0100, Jakub Sykora wrote: > > Karel Kulhavy napsal(a): [...] > >But exporting into VRML too is a good idea. BRL-CAD comes by default with > >the > >g-vrml tool that is the same family as g-dxf tool I am already using for > >AutoCAD export. If someone can verify that the VRML generated files are > >meaningful then I can add it. > > Try it and I will try to view it. Sometimes it fails. The perpendpart hasn't failed and the tetrax has failed. See both in attachment. The error message suggested that the yellow part in tetrax will be missing. The BRL-CAD contains some bugs. Recently I fixed a segfault in 'rt' during animation rendering. Then I fixed another bug in 'rtweight' where it calculated random number weights when one of the parts was declared with zero density (to exclude it from the weight calculation). CL< > > > > >It should be able to export into X3D too: > >clock at kestrel:~$ g-x3d > >Usage: g-x3d [-v] [-xX lvl] [-d tolerance_distance (mm) ] [-a abs_tol > >(mm)] [-r rel_tol] [-n norm_tol] [-o out_file] [-u units] brlcad_db.g > >object(s) > > > >Do you think I should add X3D output too? > > > > I never used X3D - I will have to try it. > > >Do you have a VRML viewer on your PC and could you verify it? > > > Yes and I will do. > > >I tried to download FreeWRL to test myself. There is no package for > >OpenBSD, > >neither in OpenBSD nor on FreeWRL page. So I downloade the source. I saw > >in the > >README.html that javac is necessary. However javac doesn't work on OpenBSD > >3.9 > >I have: > > > >clock at kestrel:~$ javac > >Segmentation fault (core dumped) > >clock at kestrel:~$ javac --version > >Segmentation fault (core dumped) > >clock at kestrel:~$ gdb /usr/local/libexec/Kaffe Kaffe.core > >(gdb) bt full > >#0 0x0a308efa in callMethodA () from /usr/local/lib/libkaffevm.so.0.0 > >No symbol table info available. > >#1 0x0a2e9e92 in processClass () from /usr/local/lib/libkaffevm.so.0.0 > >No symbol table info available. > >#2 0x0a2e996d in initBaseClasses () from /usr/local/lib/libkaffevm.so.0.0 > >No symbol table info available. > >#3 0x0a2e9739 in initialiseKaffe () from /usr/local/lib/libkaffevm.so.0.0 > >No symbol table info available. > >#4 0x0a2fa34c in JNI_CreateJavaVM () from /usr/local/lib/libkaffevm.so.0.0 > >No symbol table info available. > >#5 0x1c000c25 in main () > >No symbol table info available. > > > >CL< > > What a pity :/ > > > > >>K > >> > >>Karel Kulhavy wrote: > >>>When making the 3D model for the Tetrax console I realized that the > >>>model is > >>>quite complicated and people could have problems understanding how the > >>>parts > >>>are mutually positioned. Therefore I made a video of the part rotating. > >>> > >>>When I made the infrastructure for the video rendering it was easy to > >>>automatically render videos of other parts too. > >>> > >>>You can see all the 10 videos at http://ronja.twibright.com/3d/ - look > >>>for > >>>small turning icons. > >>> > >>>CL< > >>> > >>-- > >>Twibright Ronja mailing list http://ronja.twibright.com > >>Ronja at lists.pointless.net > >>http://pointless.net/mailman/listinfo/ronja > > > > -- > Jakub S?kora > email: kubajz at kbx.cz <') > ICQ: 68976632 ( =- > mobil: +420 777 594 201 '' > begin:vcard > fn;quoted-printable:Jakub S=C3=BDkora > n;quoted-printable:S=C3=BDkora;Jakub > adr;quoted-printable:;;=C3=9Adoln=C3=AD 1273;Praha 4;;14200;Czech Republic > email;internet:kubajz at kbx.cz > tel;cell:+420 777 594 201 > url:http://kubajz.kbx.cz > version:2.1 > end:vcard > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja -------------- next part -------------- A non-text attachment was scrubbed... Name: tetrax.wrl Type: x-world/x-vrml Size: 17384 bytes Desc: not available Url : http://pointless.net/pipermail/ronja/attachments/20070110/f52dcb2f/attachment-0002.bin -------------- next part -------------- A non-text attachment was scrubbed... Name: perpendpart.wrl Type: x-world/x-vrml Size: 50865 bytes Desc: not available Url : http://pointless.net/pipermail/ronja/attachments/20070110/f52dcb2f/attachment-0003.bin From clock at twibright.com Wed Jan 10 13:40:19 2007 From: clock at twibright.com (Karel Kulhavy) Date: Wed, 10 Jan 2007 14:40:19 +0100 Subject: [Ronja] test points on the twister PCB In-Reply-To: <200701101159.22013.cuma@tyszkiewicz.info> References: <200701100921.04452.cuma@tyszkiewicz.info> <20070110095244.GB1088@kestrel.twibright.com> <200701101159.22013.cuma@tyszkiewicz.info> Message-ID: <20070110134019.GA17187@kestrel.twibright.com> On Wed, Jan 10, 2007 at 11:59:21AM +0100, Cuma Tyszkiewicz wrote: > Dnia ?roda 10 stycznia 2007 10:52, Karel Kulhavy napisa?: > > On Wed, Jan 10, 2007 at 09:21:04AM +0100, Cuma Tyszkiewicz wrote: > > > Hello > > > > > > Would you tell me where can I find the list of testing points on the > > > twister scheme. > > > > Can you use the Twister2 list? The Twister is obsolete. > > > > CL< > > > > > Best Regards > > > > > > Cuma Tyszkiewicz > > > > > > -- > > > Twibright Ronja mailing list http://ronja.twibright.com > > > Ronja at lists.pointless.net > > > http://pointless.net/mailman/listinfo/ronja > > I have two Twisters. Both of them passed the Sanity Check and Packetloss on > wire loopback tests. Next I tried to test both Twisters connecting them with > crossovered TP. This test failed and after that both twisters are unable to Did you connect the RJ45 interfaces of the Twisters with the crossovered TP or did you connect the terminal blocks? > pass the Sanity Check and Packetloss on wire loopback tests. I want to > localize the corruption so I need to know where shall I measure the voltage > on the Twister and what the proper voltage levels are. Can you help me? Are you able to use the Twister2 measurement point list? The Twister2 circuit is the same as the Twister one. The only difference is in the power supply. I don't want to spend time writing testpoints for Twister when this is already obsolete, if it's possible to use the Twister2 points. CL< > > Cuma Tyszkiewicz > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From clock at twibright.com Thu Jan 11 08:45:20 2007 From: clock at twibright.com (Karel Kulhavy) Date: Thu, 11 Jan 2007 09:45:20 +0100 Subject: [Ronja] 3D models for the Tubular Head Message-ID: <20070111084519.GA7298@kestrel.twibright.com> I stuffed the 130mm tubular head building instructions with 3D models, including videos: http://ronja.twibright.com/tubular_head_130/building.php Do you think it helps the user understand where to drill the holes and how to assemble it together? Now the tubular head building instructions page is quite long. 17 pages in my browser. Do you think I should split it into several subpages (for example for each manufactured part one page)? CL< From petr.cipis at tiscali.cz Sat Jan 13 17:37:25 2007 From: petr.cipis at tiscali.cz (Cipis) Date: Sat, 13 Jan 2007 18:37:25 +0100 Subject: [Ronja] ronja c. 2 v tisnove Message-ID: <000501c73739$7d59f720$5046a8c0@cipis.net> Ahojte, tak dneska se podarilo rozchodit druhy spoj v Tisnove: http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/ Je to na ~370m, slouzi to k pripojeni k providerovi misto pomaleho tranga. Je to uz treti lokace tohoto spoje, predtim byl nekde u Trnavy na dvou mistech, koupil jsem to od znameho, co to sundal. Jestli by to chtel nekdo kopirovat, tak durazne nedoporucuju!!!!!!!!!!!!! Je to na *pip* prace, mel jsem 100 chuti to zkopat ze strechy dolu ... Cocky jsou 90mm, moduly nejspis klasika (akorat s jinym RSSI, pri saturaci to dalo 1V, na spoji je 250mV (nekolikrat tam byly i hodnoty pres 300mV, ale diky nestabilni konstrukci neopakovatelne)), twister ten prvni, ten je teda nastesti uz moje nebo kamosova prace, takze to je asi jedine, co je v poradku. Pripojene je to na obou stranach do upraveneho switchu, taky moje prace. Cipis From cd930 at centrum.cz Sat Jan 13 19:03:58 2007 From: cd930 at centrum.cz (-=RYS=-) Date: Sat, 13 Jan 2007 20:03:58 +0100 Subject: [Ronja] ronja c. 2 v tisnove In-Reply-To: <000501c73739$7d59f720$5046a8c0@cipis.net> References: <000501c73739$7d59f720$5046a8c0@cipis.net> Message-ID: <200701132003.6560@centrum.cz> Gratulace...... :) A je to znat na latenci a jitteru? Prece jenom...je to FD a ne wifinovej HD spoj (at uz se jmenuje treba Trango). Pripadne napis jaky je prumerny traffic. Martin? -=RYS=- ______________________________________________________________ > Od: petr.cipis na tiscali.cz > Komu: "Twibright Ronja" > Datum: 13.01.2007 18:38 > P?edm?t: [Ronja] ronja c. 2 v tisnove > >Ahojte, >tak dneska se podarilo rozchodit druhy spoj v Tisnove: >http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/ >Je to na ~370m, slouzi to k pripojeni k providerovi misto pomaleho tranga. >Je to uz treti lokace tohoto spoje, predtim byl nekde u Trnavy na dvou >mistech, koupil jsem to od znameho, co to sundal. > >Jestli by to chtel nekdo kopirovat, tak durazne nedoporucuju!!!!!!!!!!!!! >Je to na *pip* prace, mel jsem 100 chuti to zkopat ze strechy dolu ... > >Cocky jsou 90mm, moduly nejspis klasika (akorat s jinym RSSI, pri saturaci >to dalo 1V, na spoji je 250mV (nekolikrat tam byly i hodnoty pres 300mV, ale >diky nestabilni konstrukci neopakovatelne)), twister ten prvni, ten je teda >nastesti uz moje nebo kamosova prace, takze to je asi jedine, co je v >poradku. Pripojene je to na obou stranach do upraveneho switchu, taky moje >prace. > > >Cipis > > >-- >Twibright Ronja mailing list http://ronja.twibright.com >Ronja na lists.pointless.net >http://pointless.net/mailman/listinfo/ronja > ------------- dal?? ??st --------------- HTML p??loha byla odstran?na... URL: http://pointless.net/pipermail/ronja/attachments/20070113/6b001ff4/attachment.html From petr.cipis at tiscali.cz Sun Jan 14 22:53:29 2007 From: petr.cipis at tiscali.cz (Cipis) Date: Sun, 14 Jan 2007 23:53:29 +0100 Subject: [Ronja] ronja c. 2 v tisnove References: <000501c73739$7d59f720$5046a8c0@cipis.net> <200701132003.6560@centrum.cz> Message-ID: <001c01c7382e$cee42de0$5046a8c0@cipis.net> Zatim to jede par hodin, tak jeste neni cas na nejaky zaver. Teda, krome toho na samotnou konstrukci pojitka - nikdy vice. Stelovani zamereni cert to vem, je to jen 370 metru, ale uchyceni modulu v tubusech - ble smrt. To budu muset predelat. Nekdy :-) traffic: http://www.cipis.net/cgi-bin/mrtg-rrd.cgi/eth0.html (je to v Bajtech!) ping: http://www.cipis.net/cgi-bin/mrtg-rrd.cgi/ping.html (to zelene) Nemame s providerem jeste domluvenou rychlost, nicmene pri testech jsme mu ucpali uplink do bodu, odkud si to posilame ronjou :-) Podle logu jsme z toho v noci vytahli 1206 KB/sec, tj. 9,65 Mbps. Cipis ----- Original Message ----- From: "-=RYS=-" To: Sent: Saturday, January 13, 2007 8:03 PM Subject: Re: [Ronja] ronja c. 2 v tisnove > Gratulace...... :) > A je to znat na latenci a jitteru? > Prece jenom...je to FD a ne wifinovej HD spoj (at uz se jmenuje treba > Trango). > Pripadne napis jaky je prumerny traffic. > Martin -=RYS=- > > ______________________________________________________________ >> Od: petr.cipis na tiscali.cz >> Komu: "Twibright Ronja" >> Datum: 13.01.2007 18:38 >> P?edm?t: [Ronja] ronja c. 2 v tisnove >> >>Ahojte, >>tak dneska se podarilo rozchodit druhy spoj v Tisnove: >>http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/ >>Je to na ~370m, slouzi to k pripojeni k providerovi misto pomaleho tranga. >>Je to uz treti lokace tohoto spoje, predtim byl nekde u Trnavy na dvou >>mistech, koupil jsem to od znameho, co to sundal. >> >>Jestli by to chtel nekdo kopirovat, tak durazne nedoporucuju!!!!!!!!!!!!! >>Je to na *pip* prace, mel jsem 100 chuti to zkopat ze strechy dolu ... >> >>Cocky jsou 90mm, moduly nejspis klasika (akorat s jinym RSSI, pri saturaci >>to dalo 1V, na spoji je 250mV (nekolikrat tam byly i hodnoty pres 300mV, >>ale >>diky nestabilni konstrukci neopakovatelne)), twister ten prvni, ten je >>teda >>nastesti uz moje nebo kamosova prace, takze to je asi jedine, co je v >>poradku. Pripojene je to na obou stranach do upraveneho switchu, taky moje >>prace. >> >> >>Cipis >> >> >>-- >>Twibright Ronja mailing list http://ronja.twibright.com >>Ronja na lists.pointless.net >>http://pointless.net/mailman/listinfo/ronja >> > > > -------------------------------------------------------------------------------- > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja na lists.pointless.net > http://pointless.net/mailman/listinfo/ronja > From clock at twibright.com Mon Jan 15 07:59:23 2007 From: clock at twibright.com (Karel Kulhavy) Date: Mon, 15 Jan 2007 08:59:23 +0100 Subject: [Ronja] Now Ronja has better videos than anyone else Message-ID: <20070115075923.GA23173@kestrel.twibright.com> I have developed and algorithm for improved video encoding called Twibright Hyperluma and deployd it on the automatically rendered 3D videos of Ronja. Algorithm: http://ronja.twibright.com/hyperluma.php Ronja 3D videos: http://ronja.twibright.com/3d/ CL< From petr.cipis at tiscali.cz Mon Jan 15 18:51:09 2007 From: petr.cipis at tiscali.cz (Cipis) Date: Mon, 15 Jan 2007 19:51:09 +0100 Subject: [Ronja] ronja v Tisnove References: <000701c6f236$042dbd20$5046a8c0@cipis.net> Message-ID: <001b01c738d6$1ec2d3f0$5046a8c0@cipis.net> vyhrabal jsem dalsi fotky: http://www.cipis.net/ronja/foto/cipis.net/01_provizorka/ Cipis ----- Original Message ----- From: "Cipis" To: "Twibright Ronja" Sent: Tuesday, October 17, 2006 10:48 PM Subject: [Ronja] ronja v Tisnove > ahojte, > podarilo se mi udelat fotku druheho konce spoje: > http://www.cipis.net/ronja/foto/cipis.net/01_provizorka/PA170168.JPG > jeste to neni vsechno dozapojovane, prijde tam jeste par veci ... > > Cipis > > _______________________________________________ > Ronja mailing list > Ronja na lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From petr.cipis at tiscali.cz Tue Jan 16 18:42:37 2007 From: petr.cipis at tiscali.cz (Cipis) Date: Tue, 16 Jan 2007 19:42:37 +0100 Subject: [Ronja] ronja v Tisnove References: <000701c6f236$042dbd20$5046a8c0@cipis.net> <001b01c738d6$1ec2d3f0$5046a8c0@cipis.net> Message-ID: <001801c7399e$1876bfc0$5046a8c0@cipis.net> todle nepropalime: http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/mlha/ Cipis ----- Original Message ----- From: "Cipis" To: "Twibright Ronja" Sent: Monday, January 15, 2007 7:51 PM Subject: Re: [Ronja] ronja v Tisnove > vyhrabal jsem dalsi fotky: > http://www.cipis.net/ronja/foto/cipis.net/01_provizorka/ > > Cipis From kubajz at kbx.cz Wed Jan 17 05:51:28 2007 From: kubajz at kbx.cz (Jakub Sykora) Date: Wed, 17 Jan 2007 06:51:28 +0100 Subject: [Ronja] ronja v Tisnove In-Reply-To: <001801c7399e$1876bfc0$5046a8c0@cipis.net> References: <000701c6f236$042dbd20$5046a8c0@cipis.net> <001b01c738d6$1ec2d3f0$5046a8c0@cipis.net> <001801c7399e$1876bfc0$5046a8c0@cipis.net> Message-ID: <45ADB960.5070007@kbx.cz> I just love these Star Wars :] K Cipis wrote: > todle nepropalime: > > http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/mlha/ > > Cipis > > > ----- Original Message ----- > From: "Cipis" > To: "Twibright Ronja" > Sent: Monday, January 15, 2007 7:51 PM > Subject: Re: [Ronja] ronja v Tisnove > > >> vyhrabal jsem dalsi fotky: >> http://www.cipis.net/ronja/foto/cipis.net/01_provizorka/ >> >> Cipis > > -- Jakub S?kora email: kubajz na kbx.cz <') ICQ: 68976632 ( =- mobil: +420 777 594 201 '' From nikola at beogrid.net Wed Jan 17 17:30:22 2007 From: nikola at beogrid.net (nikola at beogrid.net) Date: Wed, 17 Jan 2007 18:30:22 +0100 Subject: [Ronja] 100mbps Message-ID: <000c01c73a5d$2b01b450$035a640a@heavy> Hi to all Where can i find some 100mbps ronja shematic or similar? I want to make this for college projects - 100mbps laser link sorry for my english hi -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pointless.net/pipermail/ronja/attachments/20070117/20d376ea/attachment.html From anmic at fmg.sk Tue Jan 16 21:41:43 2007 From: anmic at fmg.sk (anMic) Date: Tue, 16 Jan 2007 22:41:43 +0100 Subject: [Ronja] ronja v Tisnove In-Reply-To: <001801c7399e$1876bfc0$5046a8c0@cipis.net> References: <000701c6f236$042dbd20$5046a8c0@cipis.net> <001b01c738d6$1ec2d3f0$5046a8c0@cipis.net> <001801c7399e$1876bfc0$5046a8c0@cipis.net> Message-ID: <45AD4697.1050805@fmg.sk> Cipis napsal(a): > todle nepropalime: > > http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/mlha/ > > Cipis > Velice pekne snimky! My jsme meli ve Slavkove peknou mlhu zrovna na Silvestra: http://fotky.janmatyas.net/?dir=ronja-mlha (Fotky nic moc, foceno mob. telefonem). JM From seligr at sh.cvut.cz Wed Jan 17 20:11:09 2007 From: seligr at sh.cvut.cz (Petr Seliger) Date: Wed, 17 Jan 2007 21:11:09 +0100 Subject: [Ronja] ronja v Tisnove In-Reply-To: <45AD4697.1050805@fmg.sk> References: <000701c6f236$042dbd20$5046a8c0@cipis.net> <001b01c738d6$1ec2d3f0$5046a8c0@cipis.net> <001801c7399e$1876bfc0$5046a8c0@cipis.net> <45AD4697.1050805@fmg.sk> Message-ID: <1169064669.45ae82dd58fa2@desitka.sh.cvut.cz> Cituji z emailu od anMic : > Cipis napsal(a): > > todle nepropalime: > > > > http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/mlha/ > > > > Cipis > > > > Velice pekne snimky! > > My jsme meli ve Slavkove peknou mlhu zrovna na Silvestra: > http://fotky.janmatyas.net/?dir=ronja-mlha > (Fotky nic moc, foceno mob. telefonem). > > JM > No, zrovna jak byla chumelenice v patek pred silvestrem taxme zamerovali dva spoje v Zatci. Naprudilo me to tolik ze jsem postavil udelatko pro komfortni zamereni pomoci notebooku. Pri trose sikovnosti staci na cely spoj jeden clovek. Fotky jeste musim pretridit. BTW. Mohlo by byt nejake FTP.twibright.com na upload fotek? From schum at seznam.cz Wed Jan 17 20:17:28 2007 From: schum at seznam.cz (Schumann Miroslav) Date: Wed, 17 Jan 2007 21:17:28 +0100 Subject: [Ronja] ronja v Tisnove In-Reply-To: <1169064669.45ae82dd58fa2@desitka.sh.cvut.cz> References: <000701c6f236$042dbd20$5046a8c0@cipis.net> <001b01c738d6$1ec2d3f0$5046a8c0@cipis.net> <001801c7399e$1876bfc0$5046a8c0@cipis.net> <45AD4697.1050805@fmg.sk> <1169064669.45ae82dd58fa2@desitka.sh.cvut.cz> Message-ID: <45AE8458.7030102@seznam.cz> Petr Seliger napsal(a): > Cituji z emailu od anMic : > > >> Cipis napsal(a): >> >>> todle nepropalime: >>> >>> http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/mlha/ >>> >>> Cipis >>> >>> >> Velice pekne snimky! >> >> My jsme meli ve Slavkove peknou mlhu zrovna na Silvestra: >> http://fotky.janmatyas.net/?dir=ronja-mlha >> (Fotky nic moc, foceno mob. telefonem). >> >> JM >> >> > > No, zrovna jak byla chumelenice v patek pred silvestrem taxme zamerovali dva > spoje v Zatci. Naprudilo me to tolik ze jsem postavil udelatko pro komfortni > zamereni pomoci notebooku. Pri trose sikovnosti staci na cely spoj jeden clovek. > Fotky jeste musim pretridit. > > BTW. Mohlo by byt nejake FTP.twibright.com na upload fotek? > > > Ahoj, a jake je to udelatko???? Mirek From asteri_x at freemail.hu Thu Jan 18 11:41:23 2007 From: asteri_x at freemail.hu (Martin) Date: Thu, 18 Jan 2007 12:41:23 +0100 (CET) Subject: [Ronja] 100mbps In-Reply-To: <000c01c73a5d$2b01b450$035a640a@heavy> Message-ID: Actually search the archive, because I had some Ideas. Did not yet finish testing them, due to lack of time. So it is free to you to generate other better stuff. Basically it is a simple current amplifier, and receiver. Good luck! Be prepared, that you will not finish this in 2 months. You will make a lot of try-error loops. Please send us a copy of the project when you are finished. :) Bye, Martin nikola at beogrid.net ?rta: > Hi to all > Where can i find some 100mbps ronja shematic or similar? > I want to make this for college projects - 100mbps laser link > > sorry for my english hi ____________________________________________________________________________________ Utasbiztos?t?s - Online szerz?d?sk?t?s ?s fizet?s, azonnali k?tv?nyk?ld?s e-mailben! http://www.biztositas.hu/origo/utas From cd930 at centrum.cz Thu Jan 18 16:22:19 2007 From: cd930 at centrum.cz (-=RYS=-) Date: Thu, 18 Jan 2007 16:22:19 -0000 Subject: [Ronja] ronja v Tisnove In-Reply-To: <1169064669.45ae82dd58fa2@desitka.sh.cvut.cz> References: 000701c6f236$042dbd20$5046a8c0@cipis.net> <45AD4697.1050805@fmg.sk> <1169064669.45ae82dd58fa2@desitka.sh.cvut.cz> Message-ID: <200601030134.31433@centrum.cz> Ahoj Petre, muzu se zeptat kde v Zatci jste to zamerovali a pro koho to bylo? Diky Martin ______________________________________________________________ > Od: seligr na sh.cvut.cz > Komu: Twibright Ronja > Datum: 17.01.2007 21:11 > P?edm?t: Re: [Ronja] ronja v Tisnove > >Cituji z emailu od anMic : > >> Cipis napsal(a): >> > todle nepropalime: >> > >> > http://www.cipis.net/ronja/foto/cipis.net/02_k_providerovi/mlha/ >> > >> > Cipis >> > >> >> Velice pekne snimky! >> >> My jsme meli ve Slavkove peknou mlhu zrovna na Silvestra: >> http://fotky.janmatyas.net/?dir=ronja-mlha >> (Fotky nic moc, foceno mob. telefonem). >> >> JM >> > >No, zrovna jak byla chumelenice v patek pred silvestrem taxme zamerovali dva >spoje v Zatci. Naprudilo me to tolik ze jsem postavil udelatko pro komfortni >zamereni pomoci notebooku. Pri trose sikovnosti staci na cely spoj jeden clovek. >Fotky jeste musim pretridit. > >BTW. Mohlo by byt nejake FTP.twibright.com na upload fotek? > > >-- >Twibright Ronja mailing list http://ronja.twibright.com >Ronja na lists.pointless.net >http://pointless.net/mailman/listinfo/ronja > ------------- dal?? ??st --------------- HTML p??loha byla odstran?na... URL: http://pointless.net/pipermail/ronja/attachments/20070118/d465e0f1/attachment.html From clock at twibright.com Sun Jan 21 08:57:57 2007 From: clock at twibright.com (Karel Kulhavy) Date: Sun, 21 Jan 2007 09:57:57 +0100 Subject: [Ronja] IDA approval for Ronja Message-ID: <20070121085757.GA2591@kestrel.barix.local> I have asked IDA (an organization fighting against light pollution) and they said Ronja is OK: http://ronja.twibright.com/ecologically.php CL< From waqas_saeed14am at yahoo.com Mon Jan 22 17:33:17 2007 From: waqas_saeed14am at yahoo.com (waqas saeed) Date: Mon, 22 Jan 2007 09:33:17 -0800 (PST) Subject: [Ronja] building ronja confusions Message-ID: <678993.17361.qm@web36502.mail.mud.yahoo.com> my question is that 1) how we can interface LED with transmitter while bulding ronja 2) is there any block diagramm for the whole system --------------------------------- Everyone is raving about the all-new Yahoo! Mail beta. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pointless.net/pipermail/ronja/attachments/20070122/b4d7f75e/attachment.html From lawnmowerman at o2.pl Mon Jan 22 19:05:48 2007 From: lawnmowerman at o2.pl (Kuba) Date: Mon, 22 Jan 2007 20:05:48 +0100 Subject: [Ronja] TX RX in ABS Boxes?? Message-ID: <004b01c73e58$542edcf0$6500a8c0@baker> Hi. Can I use a hermetic ABS boxes with a pice of copper outside (thermal control) instead of metal boxes?? (TX ,RX) It will be easier and faster to build for me . Sorry for my english :( Thanks for any answer. Baker -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pointless.net/pipermail/ronja/attachments/20070122/11ac393d/attachment.html From clock at twibright.com Mon Jan 22 20:30:01 2007 From: clock at twibright.com (Karel Kulhavy) Date: Mon, 22 Jan 2007 21:30:01 +0100 Subject: [Ronja] TX RX in ABS Boxes?? In-Reply-To: <004b01c73e58$542edcf0$6500a8c0@baker> References: <004b01c73e58$542edcf0$6500a8c0@baker> Message-ID: <20070122203001.GA20817@kestrel.barix.local> On Mon, Jan 22, 2007 at 08:05:48PM +0100, Kuba wrote: > Hi. > > Can I use a hermetic ABS boxes with a pice of copper outside (thermal > control) instead of metal boxes?? (TX ,RX) No - you need metal for electromagnetic shielding. CL< > It will be easier and faster to build for me . > Sorry for my english :( > > Thanks for any answer. > > Baker > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From clock at twibright.com Mon Jan 22 20:31:50 2007 From: clock at twibright.com (Karel Kulhavy) Date: Mon, 22 Jan 2007 21:31:50 +0100 Subject: [Ronja] building ronja confusions In-Reply-To: <678993.17361.qm@web36502.mail.mud.yahoo.com> References: <678993.17361.qm@web36502.mail.mud.yahoo.com> Message-ID: <20070122203150.GA4762@kestrel.barix.local> On Mon, Jan 22, 2007 at 09:33:17AM -0800, waqas saeed wrote: > my question is that > > 1) how we can interface LED with transmitter while bulding ronja The LED is a part of the transmitter, there is no interface. > > 2) is there any block diagramm for the whole system No but I have it in the TODO list people already wanted it. CL< > > > --------------------------------- > Everyone is raving about the all-new Yahoo! Mail beta. > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From lawnmowerman at o2.pl Mon Jan 22 22:08:58 2007 From: lawnmowerman at o2.pl (Kuba) Date: Mon, 22 Jan 2007 23:08:58 +0100 Subject: [Ronja] TX RX in ABS Boxes?? References: <004b01c73e58$542edcf0$6500a8c0@baker> <20070122203001.GA20817@kestrel.barix.local> Message-ID: <000601c73e71$eabbc7f0$6500a8c0@baker> Hello And what aboute aluminium foil?? I will past them inside the box. It will be a EM shielding?? :) Thanks a lot for the replay Baker. ----- Original Message ----- From: "Karel Kulhavy" To: "Twibright Ronja" Sent: Monday, January 22, 2007 9:30 PM Subject: Re: [Ronja] TX RX in ABS Boxes?? > On Mon, Jan 22, 2007 at 08:05:48PM +0100, Kuba wrote: >> Hi. >> >> Can I use a hermetic ABS boxes with a pice of copper outside (thermal >> control) instead of metal boxes?? (TX ,RX) > > No - you need metal for electromagnetic shielding. > > CL< > >> It will be easier and faster to build for me . >> Sorry for my english :( >> >> Thanks for any answer. >> >> Baker >> -- >> Twibright Ronja mailing list http://ronja.twibright.com >> Ronja at lists.pointless.net >> http://pointless.net/mailman/listinfo/ronja > > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja > From hnizdil at gmail.com Mon Jan 22 23:39:44 2007 From: hnizdil at gmail.com (Honza Hnizdil) Date: Tue, 23 Jan 2007 00:39:44 +0100 Subject: [Ronja] TX RX in ABS Boxes?? In-Reply-To: <000601c73e71$eabbc7f0$6500a8c0@baker> References: <004b01c73e58$542edcf0$6500a8c0@baker> <20070122203001.GA20817@kestrel.barix.local> <000601c73e71$eabbc7f0$6500a8c0@baker> Message-ID: On 1/22/07, Kuba wrote: > Hello > > And what aboute aluminium foil?? I will past them inside the box. It will be > a EM shielding?? :) I think someone has already asked about that. The aluminium foil is unsufficient, I mean. But maybe it will work. I'm not sure:) > > Thanks a lot for the replay > > Baker. > > ----- Original Message ----- > From: "Karel Kulhavy" > To: "Twibright Ronja" > Sent: Monday, January 22, 2007 9:30 PM > Subject: Re: [Ronja] TX RX in ABS Boxes?? > > > > On Mon, Jan 22, 2007 at 08:05:48PM +0100, Kuba wrote: > >> Hi. > >> > >> Can I use a hermetic ABS boxes with a pice of copper outside (thermal > >> control) instead of metal boxes?? (TX ,RX) > > > > No - you need metal for electromagnetic shielding. > > > > CL< > > > >> It will be easier and faster to build for me . > >> Sorry for my english :( > >> > >> Thanks for any answer. > >> > >> Baker > >> -- > >> Twibright Ronja mailing list http://ronja.twibright.com > >> Ronja at lists.pointless.net > >> http://pointless.net/mailman/listinfo/ronja > > > > > > -- > > Twibright Ronja mailing list http://ronja.twibright.com > > Ronja at lists.pointless.net > > http://pointless.net/mailman/listinfo/ronja > > > > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja > From clock at twibright.com Tue Jan 23 09:00:32 2007 From: clock at twibright.com (Karel Kulhavy) Date: Tue, 23 Jan 2007 10:00:32 +0100 Subject: [Ronja] TX RX in ABS Boxes?? In-Reply-To: References: <004b01c73e58$542edcf0$6500a8c0@baker> <20070122203001.GA20817@kestrel.barix.local> <000601c73e71$eabbc7f0$6500a8c0@baker> Message-ID: <20070123090032.GA10235@kestrel.barix.local> On Tue, Jan 23, 2007 at 12:39:44AM +0100, Honza Hnizdil wrote: > On 1/22/07, Kuba wrote: > > Hello > > > > And what aboute aluminium foil?? I will past them inside the box. It will be > > a EM shielding?? :) > > I think someone has already asked about that. The aluminium foil is > unsufficient, I mean. But maybe it will work. I'm not sure:) You could glue copper foil inside the box and solder it together along the corners. Then install the components. Then add foil over it, solder it too, so now it would be completely soldered up. Then install the lid. But if you needed to repair or fix a bug from the manufacture, you would have to desolder the foil. But I think the metal box is easier to make than this. CL< From clock at twibright.com Tue Jan 23 09:07:25 2007 From: clock at twibright.com (Karel Kulhavy) Date: Tue, 23 Jan 2007 10:07:25 +0100 Subject: [Ronja] TX RX in ABS Boxes?? In-Reply-To: <000601c73e71$eabbc7f0$6500a8c0@baker> References: <004b01c73e58$542edcf0$6500a8c0@baker> <20070122203001.GA20817@kestrel.barix.local> <000601c73e71$eabbc7f0$6500a8c0@baker> Message-ID: <20070123090725.GA19923@kestrel.barix.local> On Mon, Jan 22, 2007 at 11:08:58PM +0100, Kuba wrote: > Hello > > And what aboute aluminium foil?? I will past them inside the box. It will be > a EM shielding?? :) No - the shield has to be without gaps. And overlapping foil counts as a gap - the foil doesn't make a good contact when just touching. And soldering aluminium foil requires a vvery special fluid and probably extra experience, if possible at all - I never did it. CL< > > Thanks a lot for the replay > > Baker. > > ----- Original Message ----- > From: "Karel Kulhavy" > To: "Twibright Ronja" > Sent: Monday, January 22, 2007 9:30 PM > Subject: Re: [Ronja] TX RX in ABS Boxes?? > > > > On Mon, Jan 22, 2007 at 08:05:48PM +0100, Kuba wrote: > >> Hi. > >> > >> Can I use a hermetic ABS boxes with a pice of copper outside (thermal > >> control) instead of metal boxes?? (TX ,RX) > > > > No - you need metal for electromagnetic shielding. > > > > CL< > > > >> It will be easier and faster to build for me . > >> Sorry for my english :( > >> > >> Thanks for any answer. > >> > >> Baker > >> -- > >> Twibright Ronja mailing list http://ronja.twibright.com > >> Ronja at lists.pointless.net > >> http://pointless.net/mailman/listinfo/ronja > > > > > > -- > > Twibright Ronja mailing list http://ronja.twibright.com > > Ronja at lists.pointless.net > > http://pointless.net/mailman/listinfo/ronja > > > > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From lawnmowerman at o2.pl Tue Jan 23 11:37:32 2007 From: lawnmowerman at o2.pl (Kuba) Date: Tue, 23 Jan 2007 12:37:32 +0100 Subject: [Ronja] TX RX in ABS Boxes?? References: <004b01c73e58$542edcf0$6500a8c0@baker><20070122203001.GA20817@kestrel.barix.local><000601c73e71$eabbc7f0$6500a8c0@baker> <20070123090725.GA19923@kestrel.barix.local> Message-ID: <000801c73ee2$df0ab3a0$6500a8c0@baker> Hi again :( ... soryy I'm watching for a ready-made product I found alu boxes in the neighbour city http://tme.pl/obudowa-aluminiowa-28x72x102mm/arts/pl/a30/oa1.html Size: H:28 x D:72 x W:140 mm or H:44 x D:72 x W:102 mm (the TX PCB is smaller then box :( here is a problem??) Can I solder PCB like this http://www.fotosik.pl/pokaz_obrazek/pelny/9148021f48420c8d.html ???? It cost 13.90 PLN = 3.62 EUR = 100 CZK Can I use them?? Thanks !! Baker. ----------------------------------------------------------------------------------------------- ----- Original Message ----- From: "Karel Kulhavy" To: "Twibright Ronja" Sent: Tuesday, January 23, 2007 10:07 AM Subject: Re: [Ronja] TX RX in ABS Boxes?? > On Mon, Jan 22, 2007 at 11:08:58PM +0100, Kuba wrote: >> Hello >> >> And what aboute aluminium foil?? I will past them inside the box. It will >> be >> a EM shielding?? :) > > No - the shield has to be without gaps. And overlapping foil counts as a > gap > - the foil doesn't make a good contact when just touching. And soldering > aluminium foil requires a vvery special fluid and probably extra > experience, > if possible at all - I never did it. > > CL< >> >> Thanks a lot for the replay >> >> Baker. >> >> ----- Original Message ----- >> From: "Karel Kulhavy" >> To: "Twibright Ronja" >> Sent: Monday, January 22, 2007 9:30 PM >> Subject: Re: [Ronja] TX RX in ABS Boxes?? >> >> >> > On Mon, Jan 22, 2007 at 08:05:48PM +0100, Kuba wrote: >> >> Hi. >> >> >> >> Can I use a hermetic ABS boxes with a pice of copper outside (thermal >> >> control) instead of metal boxes?? (TX ,RX) >> > >> > No - you need metal for electromagnetic shielding. >> > >> > CL< >> > >> >> It will be easier and faster to build for me . >> >> Sorry for my english :( >> >> >> >> Thanks for any answer. >> >> >> >> Baker >> >> -- >> >> Twibright Ronja mailing list http://ronja.twibright.com >> >> Ronja at lists.pointless.net >> >> http://pointless.net/mailman/listinfo/ronja >> > >> > >> > -- >> > Twibright Ronja mailing list http://ronja.twibright.com >> > Ronja at lists.pointless.net >> > http://pointless.net/mailman/listinfo/ronja >> > >> >> >> -- >> Twibright Ronja mailing list http://ronja.twibright.com >> Ronja at lists.pointless.net >> http://pointless.net/mailman/listinfo/ronja > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja > From clock at twibright.com Wed Jan 24 09:34:29 2007 From: clock at twibright.com (Karel Kulhavy) Date: Wed, 24 Jan 2007 10:34:29 +0100 Subject: [Ronja] TX RX in ABS Boxes?? In-Reply-To: <000801c73ee2$df0ab3a0$6500a8c0@baker> References: <20070123090725.GA19923@kestrel.barix.local> <000801c73ee2$df0ab3a0$6500a8c0@baker> Message-ID: <20070124093429.GA17503@kestrel.barix.local> On Tue, Jan 23, 2007 at 12:37:32PM +0100, Kuba wrote: > Hi again :( ... soryy > > I'm watching for a ready-made product > I found alu boxes in the neighbour city > http://tme.pl/obudowa-aluminiowa-28x72x102mm/arts/pl/a30/oa1.html > Size: H:28 x D:72 x W:140 mm or H:44 x D:72 x W:102 mm These alu boxes are not suitable becuase they don't have good enough contact between the lid and the box. Do you have IKEA in Poland? Go to IKEA and buy this: EMU 0507" or "EMU0518" - tin plated steel tin box, 2 pieces one inside other, price CHF 9.90 in Zu:rich, size of a shoebox, suitable thickness. Then it's easy to make exactly the right box using this template, which is included in the building instructions: http://ronja.twibright.com/drawings/rx_t0.png CL< > > (the TX PCB is smaller then box :( here is a problem??) > Can I solder PCB like this > http://www.fotosik.pl/pokaz_obrazek/pelny/9148021f48420c8d.html ???? > > It cost 13.90 PLN = 3.62 EUR = 100 CZK > > Can I use them?? > > Thanks !! > > Baker. > > ----------------------------------------------------------------------------------------------- > > ----- Original Message ----- > From: "Karel Kulhavy" > To: "Twibright Ronja" > Sent: Tuesday, January 23, 2007 10:07 AM > Subject: Re: [Ronja] TX RX in ABS Boxes?? > > > > On Mon, Jan 22, 2007 at 11:08:58PM +0100, Kuba wrote: > >> Hello > >> > >> And what aboute aluminium foil?? I will past them inside the box. It will > >> be > >> a EM shielding?? :) > > > > No - the shield has to be without gaps. And overlapping foil counts as a > > gap > > - the foil doesn't make a good contact when just touching. And soldering > > aluminium foil requires a vvery special fluid and probably extra > > experience, > > if possible at all - I never did it. > > > > CL< > >> > >> Thanks a lot for the replay > >> > >> Baker. > >> > >> ----- Original Message ----- > >> From: "Karel Kulhavy" > >> To: "Twibright Ronja" > >> Sent: Monday, January 22, 2007 9:30 PM > >> Subject: Re: [Ronja] TX RX in ABS Boxes?? > >> > >> > >> > On Mon, Jan 22, 2007 at 08:05:48PM +0100, Kuba wrote: > >> >> Hi. > >> >> > >> >> Can I use a hermetic ABS boxes with a pice of copper outside (thermal > >> >> control) instead of metal boxes?? (TX ,RX) > >> > > >> > No - you need metal for electromagnetic shielding. > >> > > >> > CL< > >> > > >> >> It will be easier and faster to build for me . > >> >> Sorry for my english :( > >> >> > >> >> Thanks for any answer. > >> >> > >> >> Baker > >> >> -- > >> >> Twibright Ronja mailing list http://ronja.twibright.com > >> >> Ronja at lists.pointless.net > >> >> http://pointless.net/mailman/listinfo/ronja > >> > > >> > > >> > -- > >> > Twibright Ronja mailing list http://ronja.twibright.com > >> > Ronja at lists.pointless.net > >> > http://pointless.net/mailman/listinfo/ronja > >> > > >> > >> > >> -- > >> Twibright Ronja mailing list http://ronja.twibright.com > >> Ronja at lists.pointless.net > >> http://pointless.net/mailman/listinfo/ronja > > > > -- > > Twibright Ronja mailing list http://ronja.twibright.com > > Ronja at lists.pointless.net > > http://pointless.net/mailman/listinfo/ronja > > > > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja From lawnmowerman at o2.pl Wed Jan 24 14:22:57 2007 From: lawnmowerman at o2.pl (Kuba) Date: Wed, 24 Jan 2007 15:22:57 +0100 Subject: [Ronja] Description of function (EN) Message-ID: <000801c73fc3$254a56f0$6500a8c0@baker> Is there any description of function in English like in this publications http://ronja.twibright.com/education_projects.php ?? Thanks. Kuba. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pointless.net/pipermail/ronja/attachments/20070124/fcb149fb/attachment.html From clock at twibright.com Sun Jan 28 10:51:33 2007 From: clock at twibright.com (Karel Kulhavy) Date: Sun, 28 Jan 2007 11:51:33 +0100 Subject: [Ronja] Twibright Hyperluma 2 and Luminaplex Message-ID: <20070128105133.GA19708@kestrel.barix.local> After finding out that the Twibright Hyperluma algorithm is already covered by a patent I developed two better algorithms - Hyperluma 2 and Luminaplex - for video encoding and used them for encoding the 4:2:0 Y'CbCr Ronja video loops. The Twibright Luminaplex basically encodes the video in the best possible manner, minimizing the error - there is not much more to improve by design :) Now I am publishing them as a prior art to prevent software patents http://ronja.twibright.com/hyperluma.php CL< From clock at twibright.com Sun Jan 28 11:53:50 2007 From: clock at twibright.com (Karel Kulhavy) Date: Sun, 28 Jan 2007 12:53:50 +0100 Subject: [Ronja] Luminaplex and Hyperluma 2 description Message-ID: <20070128115350.GA30690@kestrel.barix.local> Software patents suck and this is an example how to fight against them :) You don't have to read this I am sending it here to have a published description on a mailing list which is on someone else's server (Ronja mailing list) to have a proof of prior art.

I am publishing the algorithm description at the moment of release on an independent server (Ronja mailing list hosted at pointless.net) to have a reasonable proof of prior art and resulting patent invalidity in case someone tries to patent my idea.

Hyperluma and Luminaplex - algorithms for improved video encoding

Twibright Labs have developed a family of algorithms to encode ordinary 4:2:0 and 4:2:2 Y'CbCr video with unprecedented colour detail quality. There are three algorithms - Hyperluma 1, Hyperluma 2 and Luminaplex, in order of increasing picture quality. The algorithms solve the so called chroma bleed problem which is a result of the mathematically incorrect idea of chroma subsampling in analog and digital video broadcast, computer video encoding and similar applications. The chroma bleed is best visible as dark contours along edges of bright colours in the picture. It can be seen on the edge between green and magenta and between red and blue on a TV testcard. However it exists for any kind of colour detail, and contributes to overall corruption of the image details.

The corruption - the black bar between green and magenta and between red and blue - is very visible on TV testcards.

5 vertical strips of the screenshot show the quality of individual encoding agorithms compared with the original. Note how the edge corruption decreases from left to right. Use a CRT to view this image properly, LCDs usually display false artifacts on the eges. Nearest neighbour chroma interpolation is assumed in the decoder.

Twibright Hyperluma 1

Hyperluma 1 internally simulates the receiver (decoder) of the signal and then adjusts the luma channel until the effect is canceled. The result - these artifacts, which could confuse the viewer in resolving details, are gone.

During chroma subsampling we are presented with 3 pixels, each containing R, G, and B values. We are supposed to produce four Y' (luma) values, one Cb (blue chroma) and one Cr (red chroma value). The original ancient idea of the TV system was that the luma values would describe perceived brightness (luminance), which is perceived by human eye with twice as sharp resolution than the colour information. However, as the luma and chroma are calculated from gamma-corrected (nonlinear) R' G' B' values, luma doesn't represent just luminance (perceived brightness), but is also influenced by the chroma values. To reconstruct luminance properly, both luma and chroma values are necessary - luminance depends on both.

When the chroma values are subsampled and therefore changed, the luminance then suffers too, because depends partially also on the chroma. The first step in ordinary encoding system is to convert R' G' B' values into Y' Cb Cr, the second one is to subsample the Cb Cr. Hyperluma instead converts R' G' B' into gL Cb Cr, where gL is gamma-transformed luminance. gL represents accurately the brightness as perceived by the viewer. In the second step, the Cb and Cr are subsampled as usually. Hyperluma then adds third step, where gL Cb Cr are used to calculate Y' - a luma that would produce with the given Cb Cr the brightness perception that was in the picture at the given pixel position.

The core of Hyperluma - the Hyperluma Table

The core to the Hyperluma calculation is to solve the following mathematical problem: given Cb, Cr, and gL, what is the Y', that with Cb and Cr produces the given gL?

Because the function transforming Y'CbCr into gL (gamma-transformed luminance) cannot be easily mathematically reversed, the software uses binary division to search for the solution. As this is 8 times slower than the forward transformation and would slow down the video encoding, it precalculates the reverse transformation into an universal lookup table 12.96 megabytes big. This table can be compressed down to just 29kB with bzip and then easily distributed with the program. When the program starts, it loads the table into the RAM, allowing fast video processing without a need to wait on program startup until the lookup table is precalculated.

After I developed the algorithm, published it and consulted with specialists, it came out that the algorithm is patented. That was an impulse to develop Hyperluma 2 and ultimately Luminaplex, which improve the quality even more and don't seem to be covered by any patent.

The problem of Hyperluma 1 is that it removes the problem of nonconstant luminance by providing a constant luminance, however the chrominance stays the same. By the process of chroma filtering and nonlinearities in chroma definition, the chrominance in the picture suffers. This is partially solved by Hyperluma 2.

Twibright Hyperluma 2

This produces 4:2:0 Y'CbCr encoded with Hyperluma 2:


  * Input: 3x 16 bits RGB
  * Calculate a subsampled RGB, in both dimensions 1:2 subsampling
  * Calculate subsampled R'G'B' from the subsampled RGB by applying the gamma function.
  * Calculate Cb and Cr from the R'G'B'
  * Calculate 16-bit luminance Y for each from 4 pixels from the original high-resolution RGB
  * Calculate Yc, a companded luminance by applying gamma function on the luminance Y (Yc has nothing to do with Y'!)
  * Use Yc, Cb and Cr as an input for the hyperluma table to calculate luma Y' for each of four pixels.
  * Send the resulting 6 numbers - Y', Y', Y', Y', Cb, Cr - to the output

4:2:2 encoding is analogical. Note that the algorithm doesn't depend on what subsampling filter you use.

Hyperluma 2 improves the chrominance by calculating the chroma not by chroma filtering, but by physically correct photon averaging. The luminance is corrected completely like in Hyperluma 1, the chrominance is corrected partially. But the chrominance correction is not perfect - the luma shift caused by the luma correction by the hyperluma table causes the colours to drift from their ideal values.

Hyperluma 1 and Hyperluma 2 are algorithms of constant luminance. But the viewer doesn't watch the luminance only. He cares also about the chrominance! I took a completely different approach and set a goal to get both constant luminance and constant chrominance. The result is the Twibright Luminaplex algorithm, an algorithm that produces the ultimately correct result.

Twibright Luminaplex

The ultimate idea of encoding RGB or 4:4:4 Y'CbCr into 4:2:0 or 4:2:2 Y'CbCr is strikingly simple. You want the viewer to see the same on the TV as what is in the input signal. And that's what lead me to the Luminaplex algorithm, the algorithm that provides the ultimate quality.

How to tell whether the viewer sees the same

Sees the same doesn't mean the TV shows the same. The human eye is alleged to have half resolution in colour than in brightness. But how to quantify this, when the error is perceived according to a power function, while the colour mixing goes linearly? I cut this Gordic knot with a simple algorithm that I still believe is accurate. This is what the perceived_error function in the source code calculates:

The perceived error measure




  * Operates on 2x2 pixel blocks
  * Decodes the Y'CbCr data like a decoder with nearest neighbourhood interpolation would do
  * Calculates 16-bit linear RGB that appear on the screen phosphors
  * Calculates luminance Y from RGB for each of the 4 pixels.
  * Calculates gamma-companded luminance Yc from each Y by just applying the gamma function (do not confuse Yc and Y'!)
  * Calculates gamma-companded luminance Yc for each pixel (that't the brightness perception
  * Averages the RGB (linear) values of all 4 pixels together. Then applies gamma so R'G'B' come out. That's the colour perception.
  * We are having 7 numbers - Yc, Yc, Yc, Yc, R', G', B'. Calculate the Yc, Yc, Yc, Yc, R', G', B' that would be displayed from the original RGB signal without 4:2:0 Y'CbCr encoding. Take a difference for each of the 7 channels, square it and sum the squares. Take a square root. This is called Root Mean Square (RMS) and that says how crappy the viewer considers the picture.
  * I defined it as RMS because it has a property that when the same absolute value of difference is spread into multiple pixels, it is considered less crappy. I think this corresponds to the reality of perception. A single screaming bug in the image is worse than some small imperfection uniformly spread in the picture.

The Luminaplex algorithm itself




  * Operates on blocks of 2x2 pixels
  * Calculate 6 values - Y', Y', Y', Y', Cb, Cr as for Hyperluma 2. This is our starting point.
  * Now we are going to continuously monitor the quality of our data by calculating the perceived error from these 6 numbers - Y', Y', Y', Y', Cb, Cr
  * Take some dimension (for example the first Y') and try to increase the value. Does the error go up? If yes, try going the other way. If no, increase until the error would go up.
  * Take another dimension, then another until you exhaust all the 6 dimensions.
  * Go another round and another until you make a round where you actually didn't do any adjustment. This is the final result.
  * You can see the algorithm proceeds like a satellite dish viewer who wants to get the best picture. He starts with some picture. Then he adjust the azimuth for the best picture. Then he adjust the elevation for the best picture while keeping the improved azimuth. Then he repeats until he realizes there is nothing more to improve. Then he sits down into the couch and watches.

Note that this algorithm, unlike Hyperluma 1 and 2, are defined with the assumption that the receiver uses a nearest neighbour filter for the chroma upsampling. There is actually no need to use any more complicated filter than this. The transmission chain with Twibright Luminaplex encoder and nearest neighbour decoder preserves the luminance and the chrominance in the passpand of human eye. Some spatial frequency mirrors are generated in the chrominance area above the capability of human eye to discern colours, because the average-4-pixels filter doesn't cut off above the Nyquist frequency well. However it doesn't matter, because human eye cannot see these mirrors anyway.

Running a decoder with more sophisticated chroma filter on a Luminaplex output is not only a waste of CPU time, but also degrades the picture somewhat. The Luminaplex output was optimized for the simplest filter. For example mplayer on png output (where I can study the output exactly) uses this simplest filter.

How to optimize Luminaplex output for decoders with sophisticated filters?

But what if we need to encode with Twibrigtht Luminaplex for a decoder that uses a more sophisticated filter, like bilinear or even bicubic? Then the task becomes more complicated, but not impossible.

Instead of optimizing blocks 2x2 now we have to optimize the whole picture as a single block, because changing one luma or chroma value influences the errors in several neighbouring pixels. It's the same algorithm, but instead of 6 dimensions we have several hundreds thousands or millions dimensions.

Each iteration we don't have to minimize the global image error, but we need to calculate the error only for those pixels that are influenced in the output by the changed value. The size of this area is given by the size of the filter core in the decoder. It will surely slow down the calculation, several times, depending on the core size. But it can still be practical for one-time encoding of videos for distribution for example on a website or on a DVD. I didn't implement this variant into pix-y4m.c because it would be more complicated and slower and I didn't intend to use it on the Ronja project.

The Luminaplex algorithm is not suitable for broadcast because theoretically one could feed a poisonous picture that would contain a lot of difficult squares that would slow down the decoder and cause a stream dropout. The Hyperluma 2 doesn't have this problem, the execution always takes the same time there.

Some 8-bit R'G'B' triplets are not representable by 8-bit Y'CbCr - example seems to be 192,64,192. In such case Luminaplex sometimes produces some kind of 2x2 dithering to minimize the error. Very slight dithering artifacts on constant edges or constant colours result. If this is not acceptable for some reason, Hyperluma 2 must be used instead. Sometimes the colour transition is so sharp that Y'CbCr would have to go out of range to produce the correct result. It was originally programmed this way but then it came out that mplayer seems to clip Y'CbCr into the 16-235,16-240,16-240 range before processing and the result was horrible - a green tint on black. So I removed this and allowed only Y'CbCr values within the range to be used. This degraded the SNR on the numeric test (below) only by about a decibel, from 45dB to 44dB.

Numeric quality comparison

I have made a test with randomly generated pixel content. Such content exercises all possible colour transitions in the linear RGB cube homogenously. The RMS error is calculated for LSB per pixel - if it comes out as N, it means you would need to increase all 8-bit R'G'B' channels in a picture by N to get the same level of perceived error. I also provide a SNR value in dB from this SNR. 0 dB means error excursion of 127.5LSB - the same excursion from the average as the strongest possible signal - white/black bars. 400,000 randomly filled pixel squares were used for each run. Iterations per pixel are indicated, that counts how many times the perceived_error calculation is performed per pixel (from which 0.25 is for evaluation purpose only). The time/pixel is measured on 1.5GHz Pentium M with OpenBSD 4.0, gcc 3.3.5, and optimization flags -g -O2 -fomit-frame-pointer -funroll-loops -fstrength-reduce.


Algorithm - RMS error [LSB] - SNR [dB] - Time/pixel [usec] - Iterations/pixel
Ordinary - 9.075 LSB - 22.95 dB - 0.20 usec - 0.25 iter.
Hyperluma 1 - 5.635 LSB - 27.09 dB - 0.18 usec - 0.25 iter.
Hyperluma 2 - 2.800 LSB - 33.17 dB - 0.18 usec - 0.25 iter.
Luminaplex - 0.731 LSB - 44.83 dB - 2.01 usec - 6.96 iter.

Source codes

pix-y4m.c is the source code. The Hyperluma 1 is disabled in the code. As it is now the code implements Twibright Luminaplex and the inaccessible Hyperluma 1 code is there only for illustration.

Usage in real world

Ronja 3D videos which are used to illustrate how to put Ronja together, are now all encoded with the Luminaplex algorithm. It is indicated by a small Luminaplex logo in the upper left corner. I think it was a good idea to develop the Luminaplex algorithm for this application, because the models are brightly coloured for comprehensibility and then the chroma bleed problem is the strongest. The chroma bleed disease attacks details of the picture and people could have problem identifying all those small nuts, bolts and holes in the Ronja models.

Patents

As I already mentioned, the Hyperluma 1 algorithm seems to be covered by some patents. Hyperluma 2 and Luminaplex are not, I hope. On January 28, 2007 I published these algorithms and their GPL free software implementation to constitute prior art and to prevent someone from stealing them by patenting them and then being able to prevent people from using this algorithm. I have sent a note to various mailing lists and respectable individuals so a plausible proof of prior art can be constructed to challenge a patent if someone tried to abuse the patent system this way.

This of course applies only in the case the Hyperluma 2 or Luminaplex algorithms doesn't accidentally happen to be already covered by a patent as happened in the case of Hyperluma 1. I didn't conduct a patent search because these are expensive and checked just the patents connected with Hyperluma 1. If you are curious you can conduct a patent search yourself. Should you find a patent that infringes either algorithm, let me know.

Software patents are harmful for software development, especially free software. If you look into some public patent database you can see the patent system is full of simple algorithms performing simple tasks, effectively banning free usage of some independently invented innovative ideas.

Bibliography




  * The magnitude of nonconstant luminance errors (Acrobat PDF format) by Charles Poynton
  * Poynton's Color FAQ
  * Chroma Subsampling on Wikipedia
  * http://poynton.com/PDFs/YUV_and_luminance_harmful.pdf by Charles Poynton
  * Mach Bands
  * Cornsweet illusion
  * GB2293514 "Luminance signal coding; correcting for failure of constant luminance"
  * WO9609724 "Luminance signal coding; correcting for failure of constant luminance"
  * US4999702 "Method and apparatus for processing component signals to preserve high frequency intensity" From clock at twibright.com Mon Jan 29 18:26:49 2007 From: clock at twibright.com (Karel Kulhavy) Date: Mon, 29 Jan 2007 19:26:49 +0100 Subject: [Ronja] Hehe experiments with lens Message-ID: <20070129182649.GA23340@kestrel.barix.local> http://www.explosm.net/db/files/Comics/rsun.jpg CL< From clock at twibright.com Tue Jan 30 18:15:14 2007 From: clock at twibright.com (Karel Kulhavy) Date: Tue, 30 Jan 2007 19:15:14 +0100 Subject: [Ronja] Description of function (EN) In-Reply-To: <000801c73fc3$254a56f0$6500a8c0@baker> References: <000801c73fc3$254a56f0$6500a8c0@baker> Message-ID: <20070130181514.GA18287@kestrel.barix.local> On Wed, Jan 24, 2007 at 03:22:57PM +0100, Kuba wrote: > Is there any description of function in English like in this publications http://ronja.twibright.com/education_projects.php ?? On the Ronja page there is only "How Ronja works". That contains some technical notes. But it's only fragments of a complete description. The user actually doesn't have to know how Ronja works to be able to build one that works on the first try http://ronja.twibright.com/technotes/how.php CL< From gartjohn at gmail.com Tue Jan 30 19:48:21 2007 From: gartjohn at gmail.com (john gartner) Date: Tue, 30 Jan 2007 17:48:21 -0200 Subject: [Ronja] led option Message-ID: <95bc04d80701301148p6a7d5479s4e854ddcee17fbc9@mail.gmail.com> Hi: I found the led HPWT-BD00-00000 for $0.40 each (need to buy 60 units) and the led HPWT-BD00-F4000 for $3.00 each (!) (need to buy 60 units). The diference in performance between them is worth the diference in price ? How far the cheaper one can go using the lens ? Thanks, Helio. PS: I can sell a few if anybody in the list wants to buy. It's very hard to get those leds. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pointless.net/pipermail/ronja/attachments/20070130/f9e769a6/attachment.html From arunk at speedpost.net Tue Jan 30 21:22:17 2007 From: arunk at speedpost.net (Arun Krishnan) Date: Wed, 31 Jan 2007 02:52:17 +0530 Subject: [Ronja] led option In-Reply-To: <95bc04d80701301148p6a7d5479s4e854ddcee17fbc9@mail.gmail.com> References: <95bc04d80701301148p6a7d5479s4e854ddcee17fbc9@mail.gmail.com> Message-ID: <1170192137.19837.1172049913@webmail.messagingengine.com> Hello, Download the datasheet for that family from lumileds.com and check their light output. The 00000 is very VERY low brightness. You won't be able to get much of any usability from it. Also, I think the light output angle is too wide for Ronja use. Datasheets spill all! Regards, Arun ----- Original message ----- From: "john gartner" To: ronja at lists.pointless.net Date: Tue, 30 Jan 2007 17:48:21 -0200 Subject: [Ronja] led option Hi: I found the led HPWT-BD00-00000 for $0.40 each (need to buy 60 units) and the led HPWT-BD00-F4000 for $3.00 each (!) (need to buy 60 units). The diference in performance between them is worth the diference in price ? How far the cheaper one can go using the lens ? Thanks, Helio. PS: I can sell a few if anybody in the list wants to buy. It's very hard to get those leds. From gartjohn at gmail.com Tue Jan 30 23:03:26 2007 From: gartjohn at gmail.com (john gartner) Date: Tue, 30 Jan 2007 21:03:26 -0200 Subject: [Ronja] led option In-Reply-To: <1170192137.19837.1172049913@webmail.messagingengine.com> References: <95bc04d80701301148p6a7d5479s4e854ddcee17fbc9@mail.gmail.com> <1170192137.19837.1172049913@webmail.messagingengine.com> Message-ID: <95bc04d80701301503x4db75ff9g229e92a704e4f03c@mail.gmail.com> Hi: Thanks for the info. I saw that there's a huge difference between them. Do you think that 3 bucks for each led (HPWT-BD00-F4000) is a reasonable price ? Thanks, Helio. On 1/30/07, Arun Krishnan wrote: > > Hello, > > Download the datasheet for that family from lumileds.com and check their > light output. The 00000 is very VERY low brightness. You won't be able > to get much of any usability from it. Also, I think the light output > angle is too wide for Ronja use. > > Datasheets spill all! > > Regards, > Arun > > > ----- Original message ----- > From: "john gartner" > To: ronja at lists.pointless.net > Date: Tue, 30 Jan 2007 17:48:21 -0200 > Subject: [Ronja] led option > > Hi: > > I found the led HPWT-BD00-00000 for $0.40 each (need to buy 60 units) > and the led HPWT-BD00-F4000 for $3.00 each (!) (need to buy 60 units). > > The diference in performance between them is worth the diference in > price ? > > How far the cheaper one can go using the lens ? > > Thanks, > Helio. > > > PS: I can sell a few if anybody in the list wants to buy. It's very hard > to get those leds. > > -- > Twibright Ronja mailing list http://ronja.twibright.com > Ronja at lists.pointless.net > http://pointless.net/mailman/listinfo/ronja > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pointless.net/pipermail/ronja/attachments/20070130/0400a34e/attachment.html From kubajz at kbx.cz Wed Jan 31 08:36:06 2007 From: kubajz at kbx.cz (Jakub Sykora) Date: Wed, 31 Jan 2007 09:36:06 +0100 Subject: [Ronja] led option In-Reply-To: <95bc04d80701301503x4db75ff9g229e92a704e4f03c@mail.gmail.com> References: <95bc04d80701301148p6a7d5479s4e854ddcee17fbc9@mail.gmail.com> <1170192137.19837.1172049913@webmail.messagingengine.com> <95bc04d80701301503x4db75ff9g229e92a704e4f03c@mail.gmail.com> Message-ID: <45C054F6.7020501@kbx.cz> No it is not. Ondrej Tesar is able to send you few pieces abroad. He sells one piece for nearly a buck (25 CZK), but raise it with the shipping costs... http://wiki.twibright.com/index.php/GettingRonjaElectronics#Czechia K john gartner wrote: > Hi: > > Thanks for the info. I saw that there's a huge difference between them. > Do you think that 3 bucks for each led (HPWT-BD00-F4000) is a reasonable > price ? > > Thanks, > Helio. > > On 1/30/07, *Arun Krishnan* > wrote: > > Hello, > > Download the datasheet for that family from lumileds.com > and check their > light output. The 00000 is very VERY low brightness. You won't be able > to get much of any usability from it. Also, I think the light output > angle is too wide for Ronja use. > > Datasheets spill all! > > Regards, > Arun > > > ----- Original message ----- > From: "john gartner" > > To: ronja na lists.pointless.net > Date: Tue, 30 Jan 2007 17:48:21 -0200 > Subject: [Ronja] led option > > Hi: > > I found the led HPWT-BD00-00000 for $0.40 each (need to buy 60 > units) > and the led HPWT-BD00-F4000 for $3.00 each (!) (need to buy 60 units). > > The diference in performance between them is worth the diference in > price ? > > How far the cheaper one can go using the lens ? > > Thanks, > Helio. > > > PS: I can sell a few if anybody in the list wants to buy. It's very hard > to get those leds. > > -- > Twibright Ronja mailing list http://ronja.twibright.com > > Ronja na lists.pointless.net > http://pointless.net/mailman/listinfo/ronja > > -- Jakub S?kora email: kubajz na kbx.cz <') ICQ: 68976632 ( =- mobil: +420 777 594 201 '' From gartjohn at gmail.com Wed Jan 31 16:17:32 2007 From: gartjohn at gmail.com (john gartner) Date: Wed, 31 Jan 2007 14:17:32 -0200 Subject: [Ronja] HSDL-4230 or HSDL-4220 Message-ID: <95bc04d80701310817n48d6a427wee13aa2cb6f942f2@mail.gmail.com> Hi: Do you know were I can get those babys ? HSDL-4230 or HSDL-4220 Someone could sell me a few units ? Thanks, . -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pointless.net/pipermail/ronja/attachments/20070131/b966a279/attachment.html