--- Log opened Wed Apr 16 00:00:08 2008
--- Day changed Wed Apr 16 2008
00:00 < genete> I guess that last thing you need to do with a new layer type is register it at makefile level, right?
00:00 < genete> If I find out the layer stuff I should make a wiki
00:00 < genete> heh
00:00 < factor> oh nerw layer type cool
00:00 < factor> yup
00:01 < genete> I'll make a wiki page right now!
00:01 < factor> well make file if you add a mod or dock about it
00:02 < genete> I asked dooglus for a "Bend" layer. Define a straight line and a curved line and the underlying context would be curved (similar to curved gradient)
00:02 < factor> i will be making some scripts if you weish to add a feature to the dock section
00:02 < genete> it would be cool
00:02 < factor> like lens
00:02 < factor> wish even
00:03 < genete> I wish a better onionskin options
00:03 < genete> onion sking should not be defined by keyframes only.
00:04 < factor> Yup
00:04 < genete> I would like something like ktoon does (or tries to do)
00:04 < factor> heh
00:05 < genete> and be able to select past, past& future or future only onionskin
00:05 < genete> even select the transparency amount
00:05 < genete> ktoon is cool but crashes too much
00:10 < factor> Ouch
00:10 < factor> i tried it did not like it too much at all
00:13 < factor> i like the ktoon documents for developers though on its web page
00:15 < factor> not too much in it though
00:16 < factor> pretty bare but has projrect management info 
00:21 < factor> i weill try pencil later today
00:21 < genete> nice, please feed back :)
00:25 < factor> ok back later
00:25 < factor> out
00:26 -!- factor [n=Factor-@32.149.147.71] has quit ["mIRGGI meni puis"]
00:28 -!- _Elk [n=Elk@ti0110a340-0355.bb.online.no] has quit []
00:30 -!- MangoFusion_ [n=jamesu@host81-132-204-68.range81-132.btcentralplus.com] has quit ["Not here"]
00:35 -!- genete [n=Genete@84.122.51.93.dyn.user.ono.com] has quit ["Abandonando"]
00:47 -!- malexmedia [n=alex@oh-71-48-76-229.dhcp.embarqhsd.net] has left #synfig []
01:51 -!- xerakko [n=Miguel@debian/developer/xerakko] has quit ["Ex-Chat"]
01:58 -!- pabspabspabs [n=pabs@d122-105-78-204.per9.wa.optusnet.com.au] has joined #synfig
02:08 -!- pabs3 [n=pabs@d122-105-78-204.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
02:08 -!- pabspabspabs is now known as pabs3
02:41 < pabs3> hmmmm, what do people think about having a planetplanet aggregator at synfig.org/planet ?
02:42 < pabs3> we could add youtube/etc rss feeds and people's blogs
02:55 -!- AkhIL [n=AkhIL@90.188.214.121] has quit [Remote closed the connection]
03:03 -!- AkhIL [n=AkhIL@90.188.214.121] has joined #synfig
03:44 -!- Yaco [n=Franco@190.188.249.48] has joined #synfig
03:58 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has joined #synfig
04:22 -!- pixelgeek [n=chatzill@c-71-59-140-184.hsd1.or.comcast.net] has joined #synfig
04:22 < AkhIL> greetings pixelgeek 
04:22 < pixelgeek> Hi AkhIL 
04:22 < AkhIL> how synfig domo reel goes?
04:23 < pixelgeek> I think I have all the ingredients - just need to assemble them in an aesthetically pleasing manner.  :)
04:23 < pixelgeek> I even got your jedi light pen converted to 30 fps
04:24 < AkhIL> nice!
04:24 < AkhIL> I want to show this reel in my university
04:26 < pixelgeek> btw - your email ended up in my spam folder.  
04:26 < AkhIL> ha ha...
04:27 < pixelgeek> I guess being from russia from someone that had never sent me email before, with large file attachments and no text triggered the warning signs.  
04:27  * AkhIL is spamer
04:33 < pixelgeek> genete, dooglus - I've seen the ppm 480x0 image size issue before as well
04:33 < pixelgeek> re: http://synfig.org/forums/viewtopic.php?f=2&t=68&sid=256db1bc2e0a095d4dd636f81b696d2e
04:46  * AkhIL still have but in synfig wiki http://img73.imageshack.us/img73/8770/synfigwikinm2.jpg
04:46 < AkhIL> s/but/bug
04:48 < pabs3> Yaco: have you been able to commit to the synfigskin git repository?
04:48 < Yaco> hi pabs3 i didnt try to do it 
04:49 < Yaco> these days i have been busy with other proejcts
04:49 < Yaco> but soon i will go back to work on that
04:49 < pabs3> cool, let me know if you have any problems with it
05:04 -!- Yaco [n=Franco@190.188.249.48] has quit [Read error: 104 (Connection reset by peer)]
05:11 -!- Yaco [n=Franco@190.188.249.48] has joined #synfig
05:26 -!- timonator [n=timonato@cl-2290.ham-01.de.sixxs.net] has quit [Remote closed the connection]
05:26 -!- timonator [n=timonato@cl-2290.ham-01.de.sixxs.net] has joined #synfig
05:53 -!- akagogo [n=carlos@201.230.50.144] has joined #synfig
05:59  * AkhIL did user page http://synfig.org/User:AkhIL
06:33 < pixelgeek> nice wanted poster!  
06:33 < pixelgeek> Want to add yourself to the People page of the wiki too?  
06:34 < pixelgeek> Want to add yourself to the People page of the wiki too?  
06:34 < pixelgeek> You'd be the first one, AkhIL 
06:35 < AkhIL> pixelgeek: I have to get nice small photo first
06:49 -!- Yaco [n=Franco@190.188.249.48] has quit [Read error: 113 (No route to host)]
07:37 -!- akagogo [n=carlos@201.230.50.144] has quit [Read error: 104 (Connection reset by peer)]
07:51 < pixelgeek> 'night all
08:17 -!- pixelgeek [n=chatzill@c-71-59-140-184.hsd1.or.comcast.net] has quit [Read error: 113 (No route to host)]
08:23  * factor trying out pencil
08:23 < factor> looks like pencil uses GL too
08:29 < AkhIL> svn version of pencil is bugy for me
08:30 < factor> I dunno not very feature full yet
08:30 < AkhIL> release version do not support linux
08:30 < factor> synfig is light years ahead in the base drawing stuff
08:30 < factor> I am on linux
08:30 < AkhIL> but for fast sketch animation is very nice
08:30 < factor> yes
08:30 < factor> I agree that seems to be its only quality is sound layer GL and sketch into vector
08:31 < factor> but dont know how to use it 
08:31 < factor> it may have them 
08:31 < AkhIL> in svn version it have camera layer
08:31 < factor> but seems way lacking compared to synfig
08:31 < factor> yes this one has camera layer
08:32 < factor> version .4.3
08:32 < factor> b
08:32 < factor> looks like it has not been followed up since july of last year
08:32 < factor> this is a july 07 rel
08:32 < AkhIL> mine is 0.4.4b
08:32 < factor> is that the non linux one
08:33 < AkhIL> when i draw something on vector layer and then go to bitmap layer image on vector layer goes right-down
08:34 < AkhIL> i had compiled it from svn
08:36 < factor> cool missing ton o useful stuff this is again uless I dont know how
08:36 < factor> stuff like zoom
08:40 < factor> I do like the dock into one view
08:42 < factor> but I still like synfig better
09:09 < factor> off to sleep cant remember whatI was going to do , but ran out of time any way
09:10 < factor> tried the pencil prog though
09:10 < factor> Z time
09:41 -!- zenoscope [i=cbab2893@gateway/web/ajax/mibbit.com/x-4f47909274e5dd87] has quit [Remote closed the connection]
10:05 -!- rore [n=rore__@d77-218-4-48.cust.tele2.fr] has quit [Read error: 101 (Network is unreachable)]
10:17 -!- rore [n=rore__@d77-218-4-48.cust.tele2.fr] has joined #synfig
11:52 -!- MangoFusion [n=jamesu@host81-132-204-68.range81-132.btcentralplus.com] has joined #synfig
13:22 < dooglus> in C++, /* ... */ comments don't 'nest' - ie. /* /* foo; */ */ doesn't work - the 2nd */ is 'seen'
13:22 < dooglus> so if you want to 'comment out' a big block of code which contains comments, you use #if 0 ... #endif around it
13:22 < dooglus> and change the 0 to a 1 to 'temporarily' uncomment it again
13:23 < dooglus> it's a way of quickly testing a block of code, or maybe switching between 2 alternative algorithms
13:23 < dooglus> problem in synfig seems to be that the code was suddenly abandoned with these #if blocks still present
13:24 < dooglus> (I have around 36 hours of logs to catch up on, so I'm going to appear to be talking about random stuff for a while...)
13:26 < Bombe> :)
13:46 < dooglus> pabs3: I like the 'planet' aggregator idea.  I don't know how many of us have blogs (but I think it might just be you?) but even that's a good enough start once you add in the google video feed.
13:47 < dooglus> genete: re. adding a layer, I'll add a simple new layer today and write a wiki page on exactly what I had to do to get it to show up.
13:57 -!- Genete [i=d90c1034@gateway/web/ajax/mibbit.com/x-672d7bf47e41d1b7] has joined #synfig
13:57 < Genete> hi dooglus
13:59 < Genete> dooglus: hey! make a wiki for adding layers rules and understand what they need would be awesome!
14:19 -!- Zelgadis [n=zelgadis@87.103.170.181] has joined #synfig
14:33 < factor> that would rock dooglus
14:35 < factor> ok back later
14:35 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has left #synfig ["Ex-Chat"]
14:37 < dooglus> ... as soon as i get out of the bath i will
14:38 < dooglus> typing with my one dry hand on a 66% dead monitor isn't conducive to coding :)
14:38 < dooglus> did you try akhil's 'phun' link?  it's fun!
14:40 < AkhIL> 8-0 what, what, what.... who called me???
14:41 < dooglus> AkhIL: you remember the 'phun' physics sim video?
14:42 < AkhIL> nope
14:42 < dooglus> doh
14:45 < dooglus> ok, so it was genete, not you...  http://es.youtube.com/watch?v=0H5g9VS0ENM
14:48 -!- AkhIL [n=AkhIL@90.188.214.121] has quit [Read error: 104 (Connection reset by peer)]
14:49 -!- AkhIL [n=AkhIL@90.188.214.121] has joined #synfig
14:51 < dooglus> I clearly remember you saying "that should be easy to add to synfig, if not more" - but the logs tell me it never happened :)
14:51 < Genete> :)
14:53 < dooglus> Genete: do you remember the conversation?
14:53 < Genete> the link was brought to me by my cousin. He's a great coder guy (and in fact he was who introduced me to linux and C++ programming)
14:53 < Genete> about phun?
14:53 < dooglus> Genete: your link wasted my last 2 days :)
14:53 < dooglus> Genete: yes
14:53 < Genete> oh dear! it wasn't my intention! :/
14:54 < AkhIL> dooglus: I sed something like this 3 hours later about another video example
14:54 < dooglus> Genete: specifically, you linked to it, and someone (I though AkhIL) said "that would be easy to add to synfig" or something?
14:54 < AkhIL> (13:55:03) AkhIL: http://www.vimeo.com/889190/  very nice VFX done in blendr. But I think synfig should have enough power to do same (13:55:37) AkhIL: to do same and even more
14:54 < AkhIL> log from 2008-04-14
14:54 < Genete> no, only show the link to factor and then wnt to bed
14:54 < dooglus> Genete: I enjoyed it.  Did you see the competitions on the forum?
14:54 < dooglus> AkhIL: ok, thanks for that.
14:55 < Genete> dooglus: no I didn't. I saw it so addictive so I bookmarked it for later enjoy ;)
14:56 < Genete> I always loved the "The incredible machine" game
14:56 < dooglus> AkhIL: I did 'open link in new tab' for all the links while catching up with the logs, then looked at them after - I must have got confused about which link you were talking about :)
14:57 < Genete> (I'm not so optimist as AkhIL is...)
14:57 < dooglus> Genete: http://www.youtube.com/watch?v=UP1wPFICdUY - someone made a fully working digital clock in phun :)
14:58 -!- AkhIL [n=AkhIL@90.188.214.121] has quit [Read error: 104 (Connection reset by peer)]
14:59 < dooglus> Genete: http://phun.cs.umu.se/forum/viewtopic.php?id=346 - one of the challenges, with some very clever solutions
15:01 -!- AkhIL [n=AkhIL@90.188.214.121] has joined #synfig
15:01 < Genete> cool
15:01 < Genete> dooglus: out for lunch. Don't dissapear because I have to ask you some questions.
15:01 < AkhIL> my pidgin goes mad
15:02 < dooglus> so yesterday I wandered away from phun, through "crayon physics deluxe" ( http://www.youtube.com/watch?v=QsTqspnvAaI ) and onto a game where you get to torture kittens in hell before cooking them ( http://www.kloonigames.com/blog/games/cacodemon )
15:02 < dooglus> that's the kind of distraction that comes my way when I boot into Windows :)
15:48 < Zelgadis> Genete: I just learned how to do realistic sky, examining your Feb entry (boat splash)!
15:49 < Zelgadis> amazing...
16:17 -!- MangoFusion_ [n=jamesu@host217-44-66-133.range217-44.btcentralplus.com] has joined #synfig
16:28 < Genete> Zelgadis: I'm glad it helped you :)
16:28 < Zelgadis> Genete: can I use your background for my animations?
16:29 < Genete> Zelgadis: the file is released free for coummunity. Just mention me as original creator. 
16:29 < Zelgadis> Yeah, according to GPL ;) Thanks!
16:30 -!- Zelgadis [n=zelgadis@87.103.170.181] has quit ["Schastlivo!"]
16:33 -!- MangoFusion [n=jamesu@host81-132-204-68.range81-132.btcentralplus.com] has quit [Read error: 110 (Connection timed out)]
16:33 -!- factor [n=Factor-@32.148.221.248] has joined #synfig
16:33 < AkhIL> Zelgadis: your avatar in people page is broken
16:39 -!- Zelgadis [n=zelgadis@87.103.170.181] has joined #synfig
16:48 < Genete> dooglus: you there?
16:51 < Zelgadis> wil back soon
16:51 -!- Zelgadis [n=zelgadis@87.103.170.181] has quit ["Schastlivo!"]
16:55 < Genete> dooglus: trying to understand the layers creation I've started from an "easy" one: Stretch
16:56 < Genete> looking stretch.cpp it says this: 
16:56 < Genete> etl::handle<Transform>
16:56 < Genete> Layer_Stretch::get_transform()const
16:56 < Genete> {
16:57 < Genete> 	return new Stretch_Trans(this);
16:57 < AkhIL> imho color correction simpler. but stretch more interesting
16:57 < Genete> }
16:59 < Genete> how can this function return a etl:handle<Transform> type if the sentence "new Stretch_Trans(this);" returns a class pointer? Is ther some such of implicit casting?
16:59 -!- Zelgadis [n=zelgadis@87.103.170.181] has joined #synfig
16:59 -!- Netsplit kubrick.freenode.net <-> irc.freenode.net quits: MangoFusion_
17:00 < Genete> AkhIL: any idea how is ti possible or do I have forgotten my C++ knowledge?
17:01 -!- Netsplit over, joins: MangoFusion_
17:23 < dooglus> Genete: hi
17:24 < Genete> hi dooglus
17:25 < dooglus> Genete: I just was working on a simple "how to add a layer" page
17:26 < dooglus> Genete: I've saved what I got done so far - it might help: http://synfig.org/Source:Adding_a_Layer
17:27 < dooglus> Genete: re. your question:  the method is declared to return a handle<Transform>, and so the Stretch_Trans* is cast to a handle<Transform> automatically
17:28 < dooglus> Genete: specifically, etl/_handle.h has:
17:28 < dooglus> template <class T>
17:28 < dooglus> class handle
17:28 < dooglus> 	typedef T* pointer;
17:28 < dooglus> 	//! Constructor that constructs from a pointer to new object
17:28 < dooglus> 	handle(pointer x):obj(x)
17:29 < dooglus> ie. give it a pointer to type T, and it constructs you a handle<T>
17:29 < Genete> I have to review my C++ knowledge :/
17:29 < Genete> Grrrr I always hated templates :(
17:30 < dooglus> Genete: just replace every T with Transform
17:30 < dooglus> Genete: then the typedef reads "typedef Transform* pointer"
17:30 < dooglus> so replace every "pointer" with "Transform*"
17:31 < dooglus> now we have a constructor "handle(Transform *x) ..."
17:31 < Genete> I know the theory. But it is difficult to read the code if you don't go so back on it. In this case you needed to go to etl/_handle.h what is "so far"...
17:31 < dooglus> ie. a constructor which can make handle<Transform> objects from Transform* pointers
17:31 < Genete> for example what the hell is GUID?
17:32 < dooglus> Genete: you could just assume that it works, and not look at the details
17:32 < Genete> it is like some sort of counter for transform layers?
17:32 < Genete> yeah! it is what I'm doing!
17:32 < dooglus> a GUID is a long string of hex digits like you see in .sifz files
17:32 < dooglus> it identifies any Node
17:33 < dooglus> <animated type="angle" id="angle" guid="5E22C99EBA260EC514042BD01EC7C249">
17:33 < Genete> cool
17:33 < Genete> let's take a look to your page.... Yuuupieeee!!   \o/
17:34 < dooglus> I think GUIDs are how linking is done - if two circles have a radius with the same GUID, then the radii aren't just the same, but they're linked
17:36 -!- Zelgadis [n=zelgadis@87.103.170.181] has quit ["Schastlivo!"]
17:38 < Genete> dooglus: wow! interesting concept for a noob like me!
17:39 < Genete> dooglus: thankyou very much for the wiki page. Too much people would appreciate it. 
17:40 < Genete> dooglus: there is a ny rule to know if a new layer is a filter, distortion, standard or what? In other words how can we know in which module should we add our layer?
17:42 < dooglus> Genete: I think it's pretty ad-hoc
17:43 < dooglus> Genete: filters modify the layers under them, geometry layers make new shapes, etc.
17:43 < dooglus> I updated the page a bit
17:47 -!- wildhostile [n=wildhost@ALamentin-104-1-81-120.w80-8.abo.wanadoo.fr] has joined #synfig
17:49 < Genete> dooglus: so get_color is the defined function for final quality render and accellerated render is the one for synfigstudio?
17:58 < dooglus> Genete: no
17:58 < Genete> just read it :/
17:58 < dooglus> Genete: get_color isn't used if the layers above don't transform us (except by the 'Info' panel)
17:58 < dooglus> Genete: try changing the set_s(0) to set_s(1) in get_color and see the effect it has - it messes up the Info panel display, but that's all
17:59 < dooglus> (updated the page again with more words)
18:00 < Genete> set_s() is set saturation for color types?
18:02 < dooglus> Genete: yes
18:02 < dooglus> I'll write about it
18:03 < Genete> dooglus: have to go now but I'll keep my two eyes on this!!! THANKS!
18:03 < dooglus> Genete: http://synfig.org/api/synfig/classsynfig_1_1Color.html#b62c2069fc43baa97fba17b33ff4d32d
18:03 -!- Genete [i=d90c1034@gateway/web/ajax/mibbit.com/x-672d7bf47e41d1b7] has quit ["http://www.mibbit.com ajax IRC Client"]
18:04 < dooglus> I don't think it's a good idea to link to such a random-looking URL though
18:39 -!- rubikcube [n=kvirc@dslc-082-082-085-145.pools.arcor-ip.net] has joined #synfig
18:57 -!- wildhostile [n=wildhost@ALamentin-104-1-81-120.w80-8.abo.wanadoo.fr] has quit [Remote closed the connection]
18:57 -!- factor [n=Factor-@32.148.221.248] has quit ["mIRGGI meni puis"]
18:57 -!- genete [n=Genete@84.122.51.93.dyn.user.ono.com] has joined #synfig
19:02 -!- MangoFusion_ [n=jamesu@host217-44-66-133.range217-44.btcentralplus.com] has quit []
19:04 < dooglus> genete: http://synfig.org/Source:Adding_a_Layer - I started writing how to add a parameter
19:04 < genete> dooglus: in your example you call pen.dec_x(x) with a not initialized value of x, is that correct?
19:05 -!- _Elk [n=Elk@ti0110a340-0355.bb.online.no] has joined #synfig
19:08 < genete> I'm happy that I'm not lost with your tutorial. :)
19:16 -!- omry [n=omry@bzq-84-108-20-56.cablep.bezeqint.net] has quit [Remote closed the connection]
19:43 < dooglus> genete: x is initialised in the inner loop
19:44 < genete> ah right, so when the call is done it has a value :)
19:45 < genete> why do we need to decrement x in the outer loop?
19:45 < dooglus> genete: each time y goes up (ie. we start a new row) we need to step x back to the beginning of the row
19:45 < dooglus> dec_x(x) takes x off the pen's x position (not 1)
19:46 < dooglus> http://synfig.org/api/ETL/classgeneric__pen.html#ee072394e747bc06d4901dcd2339ce1d
19:47 < dooglus> or, better, http://synfig.org/api/ETL/__pen_8h-source.html#l00200 - 'cos the documentation is missing!
19:50 < genete> I got it. One is with argument and other without. So one is incremented by one and the other is decremented an mount of x
19:52 -!- MangoFusion [n=jamesu@host217-44-66-133.range217-44.btcentralplus.com] has joined #synfig
19:53 < dooglus> right
19:54 < dooglus> updated http://synfig.org/Source:Adding_a_Layer again to show adding a parameter to scale the saturation
19:54 < dooglus> any more questions about it?
19:54 < dooglus> I guess I should discuss inheriting from Layer_Composite to get amount, blend_method, z_depth params
19:55 < genete> yes. And also an example that uses hit_check() fucntion and the calculation of the bounding box
19:56 < genete> and later some about inherit from Transform layer for new distortion layers :)
19:59 < dooglus> genete: to see when get_color() is used...
19:59 < dooglus> replace the body of get_color() in desaturate.cpp with:
19:59 < dooglus> return Color(1, 0, 0, 1);
19:59 < dooglus> ie. it always returns solid red
19:59 < dooglus> then load http://dooglus.rincevent.net/random/rgb3.sifz
19:59 < dooglus> it has a twirl over a desaturate
20:00 < dooglus> you'll see the red where the twirl means that the rectangle rendered by accelerated_render isn't enough
20:01 < dooglus> screenshot of studio: http://dooglus.rincevent.net/random/rgb3.png
20:01 < dooglus> result of commandline render: http://dooglus.rincevent.net/random/rgb3commandline.png
20:02 < dooglus> the red from get_color() shows up the same in both
20:03 < genete> so what does it means? that desaturate is badly written or twirl?
20:03 < genete> btw, I'v loaded the file with my current version and it loads without complain!
20:03 < dooglus> and in the gui, the Info panel shows red even when pointing in the middle: http://dooglus.rincevent.net/random/rgb3studio.png
20:04 < dooglus> genete: neither really - it just means that layers should implement get_color() as well as accelerated_render() if they want to render properly when distorted
20:05 < genete> ah
20:05 < dooglus> genete: you have a proper get_color() in your version I guess.  I broke it, getting it to always return red, so you can see where exactly it's used
20:06 < genete> look: http://i85.photobucket.com/albums/k74/Genete/synfig/Pantallazo-Capas.png
20:06 < genete> the file you have posted
20:07 < dooglus> it means you don't have the 'desaturate' layer available
20:07 < dooglus> any layer type not known will load OK, but not do anything, and show up as 'MIME'
20:08 < dooglus> I think it will save fine too, so you can edit .sif files which contain layers you don't know (in your old version of studio), and they'll save back just fine
20:08 < genete> console output:  Couldn't find layer to be activated in layer list. Rebuilding index...
20:08 < dooglus> I mean, even though you don't know what this 'desaturate' layer is, you can edit the .sif file, change the other layers and send it back to me, and the 'desaturate' layer will still be in there, even though your synfig doesn't know what exactly it is
20:09 < dooglus> genete: I see that a lot, but I don't think it's related
20:10 < genete> that would allow make our own fork layers without not problem on share files?
20:11 < dooglus> genete: yes - it lets you make your own secret layer types, and still allow me to load and save your compositions
20:11 < dooglus> although, of course, I won't see the work done by the layer types I don't have here
20:11 < genete> that's cool anyway
20:12 < genete> dooglus: the perform and unperform functions of a transform type layer, must be always defined?
20:13 < genete> what if I know the perform function but cannot calculate its inverse?
20:14 < dooglus> I'm not familiar with that code
20:15 -!- malexmedia [n=alex@oh-76-2-223-75.dhcp.embarqhsd.net] has joined #synfig
20:15 < dooglus> but I can see that both directions are needed
20:15 < dooglus> one to decide where to draw ducks
20:15 < dooglus> and the other to calculate the effect of dragging a duck on the parameters
20:16 < genete> glup
20:17 < genete> I deduced it waching stretch layer that have an easy transformation
20:19 < genete> the problem is when unperform fucntion gives you two possible resutls with the same argument
20:20 < genete> in that case I think that we should decide on one of the sides depending on other extra argument (that I cannot imagine its meaning by now)
20:29 -!- factor [n=Factor-@32.148.3.187] has joined #synfig
20:34 < malexmedia> heyo guys
20:34 < malexmedia> i and my artist are sitting down together with synfig for the first time right now
20:34 < malexmedia> we're having some difficulty with the draw tool simplifying the strokes he makes
20:35 < malexmedia> any tips?
20:35 < dooglus> malexmedia: do you see the 'tool options' panel?
20:35 < malexmedia> yeah
20:35 < factor> cool
20:35 < dooglus> select the draw tool, then in the 'tool options' panel there's a parmeter towards the bottom called 'smooth' which determines how much the draw tool simplifies your strokes
20:36 < malexmedia> the lowest it'll go is 0.010
20:36 < malexmedia> we tried setting to 0.0, and it defaulted back to that
20:36 < malexmedia> the results are all wonky 
20:36 < malexmedia> ie, my artist draws a smooth curve, and it results in two or three ducks
20:37 < malexmedia> with mostly straight lines between
20:37 < dooglus> this is with a tablet?  or mouse?
20:37 < malexmedia> tablet
20:37 < dooglus> what if you use the mouse?  is that better?
20:37 < malexmedia> he's pretty good with it in inkscape
20:38 < malexmedia> okay, trying with the mouse
20:38 < malexmedia> ooh
20:38 < malexmedia> a mouse-drawn curve is represented by an appropriate number of ducks
20:39 < malexmedia> whereas tablet-drawn results are extremely over-simplified
20:39 < genete> malexmedia: try it with and without "Local Error" anabled
20:39 < malexmedia> trying
20:39 < malexmedia> we were trying without
20:40 < genete> it changes the Smooth meaning and its lower value
20:40 < malexmedia> it seems as though local error minimizes or removes the problem
20:41 < dooglus> I wonder what it really does - maybe it could do with being renamed to something more meaningful
20:41 < malexmedia> hrm, individual strokes are less corrupted
20:42 < malexmedia> but when it comes to drawing actual objects, the results are very... blocky
20:42 < malexmedia> or splotchy
20:42 < dooglus> malexmedia: with 'local error' on too?
20:43 < malexmedia> yes
20:43 < malexmedia> local error helps, but not enough
20:43 < genete> with local error enabled try higher smooth values
20:43 < malexmedia> the resulting shapes just don't faithfully represent the strokes
20:44 < malexmedia> default was 20, trying 40
20:44 < malexmedia> 40 is way too much simplification
20:44 < factor> If you do a sine wave with high freq. Does it catch that and also with a low freq. Sine wave
20:44 < genete> 20 or 40 is too much
20:45 < factor> or do you get errors with both
20:45 < genete> what are your default units malexmedia ?
20:45 < malexmedia> hmm, turning local error on defaulted smooth to 20
20:45 < genete> File->Settings->Misc.
20:45 < malexmedia> default units are points i believe
20:45 < malexmedia> we didn't change them
20:46 < genete> for me 20 is unusable
20:46 < genete> for smooth
20:46 < malexmedia> i assumed it was jumping up to 20 because that was the minimum
20:46 < malexmedia> we're experimenting with local error on and lower smooth values
20:46 < genete> ok
20:46 < dooglus> malexmedia: there are 2 values stored - the local and the global smoothness
20:47 < dooglus> malexmedia: toggling the 'localerror' checkbox will switch between the two
20:47 < malexmedia> with local error enabled, smooth won't go any lower than 1.0
20:47 < malexmedia> hmm
20:47 < malexmedia> the strokes still look like they're being drawn by a child :-P
20:48 < genete> dooglus: I didn't know that both smooth values have effect
20:48 < dooglus> do the other tools behave properly with the tablet?
20:48 < dooglus> like if you try to draw a circle with the tablet, does it go where you request, with the requested radius?
20:48 < malexmedia> isn't 'draw' the only one with variable width support?
20:48 < factor> be cool to see a screen shot
20:48 < dooglus> it is, yes
20:48 < malexmedia> okay, testing those things
20:49 < malexmedia> yes, the circle tool follows the stylus properly
20:49 < dooglus> I don't have a tablet available here, so I can't test it unfortunately
20:50 < factor>  i lost my pen :(
20:50 < dooglus> can you say how the resulting drawn shapes are wrong?  are the ducks it does place on the draw you drew?
20:50 < malexmedia> lemme get some screenshots for reference
20:51 < genete> dooglus: I think there are two effects: the amoount of points and its curvature
20:51 < genete> the drawing speed also is important
20:51 < dooglus> malexmedia: use the 'bline tool' to draw a reference curve that you then follow with the draw tool, so I can see what you tried to follow, and what resulted
20:52 < dooglus> genete: yes - it sounds like the tangents are coming out too short.  I'm wondering if the vertices themselves are accurate at least
20:53 < genete> more vertices more short are the tangents
20:53 < genete> and its inverse
20:54 < dooglus> here's an example: http://dooglus.rincevent.net/random/drawtest.png - the black line is the bline I followed.  the red line is the result of using the draw tool with a 'smooth' of 1.0 and no 'localerror' selected
20:54 < malexmedia> okay
20:54 < dooglus> (I'm not particularly good with the mouse, so it's not a very good match - but turning the 'smooth' up would make it better
20:55 < malexmedia> um, hold on...
20:55 < malexmedia> we'll do the path follow test next
20:55 < genete> malexmedia: for long strokes few details turn local error off and set it around 1/3 of the length of the stroke (in units)
20:56 < malexmedia> madness over here, one moment please :)
20:56 < dooglus> with smooth=3.0 : http://dooglus.rincevent.net/random/drawtest2.png
20:57 < genete> for small strokes and high details turn local error on and ser smooth to  the 1/10 of the stroke length in pixels
20:57 < genete> dooglus: I'm guessing on values and results
20:58 < malexmedia> okay, we'll do the path test in a minute
20:58 < malexmedia> for the moment, here's what we've got
20:58 < malexmedia> http://misc.malexmedia.net/screenshot-synfig.png
20:58 < malexmedia> now, for reference, that's supposed to be the same general shape as a character appearing in this image: http://files.malexmedia.net/posters/Malex/audio/malex_minute/illustrations/e100_illustration_large.jpg
20:59 < genete> It seems that local error on turns smooth into pixel and when off it is turned to units
20:59 < malexmedia> bottom left hand corner, little furry monster
20:59 < malexmedia> drawn with a wacom pad in inkscape
20:59 < malexmedia> same hardware setup as the synfig screenshot
20:59 < malexmedia> it does appear to be moving the vertices
21:01 < malexmedia> haha, how do you end the bline path? :-P
21:01 < malexmedia> we can't let go :-P
21:01 < dooglus> click the 'gear' icon in the bottom of the tool options
21:01 < dooglus> or switch to a different tool
21:02 < malexmedia> okay, rockin'
21:02 < dooglus> (alt-a switches quickly to the 'normal' tool)
21:02 < malexmedia> anyway, see the significant difference between the character in the synfig snapshot and the inkscape document
21:03 < malexmedia> (while my artist is creating the path trace screenshot)
21:03 < dooglus> I can't really see the synfig version for all the 'ducks'
21:03 < malexmedia> mm
21:03 < dooglus> if you click on the background they'll all go awa
21:03 < dooglus> y
21:05 < malexmedia> working on the path trace thing now
21:05 < dooglus> one thing that you probably don't want for that guy are the 'auto' options in the draw tool - 'auto link' and 'auto extend'
21:05 < dooglus> they will join your new strokes onto old strokes if the new stroke starts or ends near an existing duck
21:06 < malexmedia> yeah, we turned all the auto options off :-P
21:06 < dooglus> ok
21:08 < malexmedia> sorry this is taking so long
21:08 < malexmedia> it would help if we were generally more familiar with the tool...
21:09 < malexmedia> hmm
21:09 < genete> malexmedia: want to see my trial?
21:09 < malexmedia> sure
21:09 < malexmedia> one of the problems we're running into here is that my artist can't trace the line
21:09 < malexmedia> i mean, he can do it slowly
21:10 < malexmedia> but when we go slower it drops more ducks, and is thus a better representation of the line
21:10 < genete> I'm not an artist but loadedd your reference file and did this: http://i85.photobucket.com/albums/k74/Genete/synfig/Pantallazo-AnimacinSynfig1Raz-1.png
21:10 < malexmedia> was that with a tablet?
21:11 < genete> yup
21:11 < genete> I forgot to do the outer line in one stroke so now is difficult to fill
21:11 < genete> I used a wacom graphire
21:12 < genete> and the settings I suggested you before
21:12 < genete> for the outer stoke I used local error off
21:12 < genete> and for the mouth and the eyes local error on
21:12 < malexmedia> how fast did you draw it?
21:12 < genete> the time I was in silence :)
21:12 < genete> oh
21:13 < genete> you mean the stroke speed?
21:13 < malexmedia> well, i mean, individual strokes
21:13 < malexmedia> yeah
21:13 < genete> it doesn't matter
21:13 < genete> well, 
21:13 < malexmedia> i think it might matter
21:14 < genete> fast but also can be slow
21:14 < genete> (not so fast)
21:14 < malexmedia> the faster my artist draws the strokes, the fewer vertices we get
21:14 < genete> it depends on the smooth values andif the local error is on or off
21:14 < malexmedia> yeah, it's not dropping vertices based on the distance moved, it's sampling at a constant rate of speed
21:15 < malexmedia> by slowing down, we're suddenly getting decent shapes
21:15 < genete> yes, that's true
21:15 < genete> if you look to the screenshot
21:15  * malexmedia looks
21:15 < genete> you can see a rounded hair on the top
21:15 < genete> between the eyebrows
21:16 < genete> it was really fast
21:16 < malexmedia> well, it's definitely able to create smooth shapes with few vertices
21:16 < genete> and the tool didn't catch the direction change
21:16 < genete> properly
21:16 < malexmedia> but smooth != faithful to the original stroke
21:16  * malexmedia nods
21:16 < dooglus> malexmedia: this sounds like a bug I've looked into before
21:17 < malexmedia> so how do we improve the situation for artists whose style is to draw very fast?
21:17 < dooglus> in order to avoid 'jags' where the artist hand shakes, synfig will avoid drawing a corner if it's less than 6 'samples' from the previous corner
21:17 < genete> malexmedia: you should not tell the artists to draw very fast for an ANIMATION design
21:17 < dooglus> this seems to be far too conservative
21:18 < malexmedia> genete: i'm not going to tell my artist how to draw ;)
21:18 < malexmedia> the precision in my artist's strokes is very high, as his inkscape images clearly demonstrates
21:18 < genete> malexmedia: if you're going to animate the monster you should not tell to an artist to draw it in studio you should tell it to the animator person
21:18 < malexmedia> i would like to translate that precision to synfig, if possible
21:19 < genete> because usually you should not "draw" to animate you should move ducks redrawing (repositioing the ducks)
21:20 < malexmedia> hmm
21:20 < genete> so if the model is quite high complex it would be so difficult to animate later
21:20 < malexmedia> i see
21:20 < malexmedia> so if the number of ducks is too high, it becomes unmanageable
21:20 < factor> right
21:20 < malexmedia> so you can't just draw multiple key frames from scratch
21:20 < malexmedia> ie, synfig can't smoothly interpolate between two unrelated frames?
21:21 < factor> working on a way to do that
21:21 < genete> you can alwais can use the smooth move tool but at the end the tangent ducks becomes impossible to control and needs individual managing
21:21 < genete> what do you understand by two unrelated frames?
21:21 < malexmedia> well
21:21 < malexmedia> i think we were hoping to onion skin a previous keyframe
21:22 < malexmedia> trace a new position
21:22 < dooglus> genete: he means to freehand-draw 2 shapes, and have them interpolated
21:22 < malexmedia> and have it interpolate between them
21:22 < factor> Yup
21:22 < malexmedia> using, ie, a vector mesh tranformation
21:22 < dooglus> malexmedia: that's not something that synfig supports at the moment
21:22 < malexmedia> although... my understanding of mesh transformations is more for 3d meshes
21:23 < malexmedia> i assume 2d vector tranformations would be similar
21:23 < factor> that is what my panel i wasweorking on will do , no eta ion that though
21:23  * malexmedia nods
21:23 < dooglus> malexmedia: see http://synfig.org/Redraw_tool - that's what you're wanting I think
21:23  * malexmedia looks
21:24 < genete> (dooglus: that "at the moment" sounds like a wish?)
21:24 < malexmedia> yes, sounds vaguely similar
21:24 < malexmedia> but at the same time
21:24  * factor hopes to make that one come true
21:25 < dooglus> genete: If wishes were horses, then beggars would ride
21:25 < malexmedia> it's not probably good to force the objects to have the same complexity
21:25 < factor> for rotoscoping out put as well
21:25 < malexmedia> for example, if you're animating a rotating object, the complexity of the 2d vectors will significantly change over time
21:25 < genete> malexmedia: you can make vertices active (visible) or not active (invisible) that is an intermediate solution...
21:26 < malexmedia> to compensate (at least in 3d meshes) you can subdivide less complex vectors to artificially match the complexity of the more complex vectors
21:26 < genete> malexmedia: that's the question. let the computer do the tweening or you draw every frame?
21:27 < malexmedia> well, in theory, the animator should be able to draw as many (or few) frames as he or she wants in order to represent the desired motion
21:27 < factor> complrexify and simplyfy :)
21:27 < factor> complexifyeven
21:27 < malexmedia> so, for example, in a case where the motion is deemed too unrealistic, a new keyframe could be drawn in the offending area to clarify the motion
21:27 < malexmedia> resulting in more accurate tweening sections
21:28 < genete> if you want the computer do the tweeining of a two different meshes you should give him some conversion rules!
21:28 < factor> yup
21:28 < genete> if not how can it do the tewwning?
21:28 < genete> tweening even
21:29 < genete> a pure hand drawn artist doesn't care on conversion rules. He simply would draw more in-betweens if needed
21:29 < malexmedia> anyway, it's the sort of thing we've constantly theorized about implementing
21:29 < malexmedia> yes, true
21:30 < malexmedia> the idea is to have an interpolation algorithm that can transform smoothly between any two vector shapes
21:30 < malexmedia> which, iirc, isn't all that hard with trangle meshes
21:31 < dooglus> the existing workflow in synfig is to draw the starting position, and then to transform it by dragging the ducks around to change its shape in future (or past) keyframes
21:31 < malexmedia> and i presume a lot of the same concepts would apply to 2d vector animation
21:31 < malexmedia> dooglus: that sounds good, just not what we were (unrealistically) hoping for :)
21:32 < dooglus> malexmedia: I don't know how unrealistic it is - it's just not what synfig offers at the moment
21:32 < dooglus> malexmedia: do you have a reference for the transformation of triangle meshes?
21:32 < malexmedia> no, sadly... just something i've been thinking about for years
21:33 < malexmedia> my dad used to work in computer graphics, so we would talk about it
21:33 < malexmedia> i think i may have attended a siggraph lecture on the subject
21:33 < malexmedia> but that would have been a few years ago
21:33 < malexmedia> but the rough concept is
21:33 < genete> malexmedia: also the problem is that given to points ther is not always a straight path to interpolate from one to other
21:34 < malexmedia> genete: well, that depends
21:34 < malexmedia> if we're talking about linear interpolation (the simplest kind) then it clearly must be a straight line from point to point
21:34 < malexmedia> the problem comes in when you have a different number of points in each target
21:35 < malexmedia> for example, in 3d meshes
21:35 < genete> that's the problem for 2D meshes too
21:35 < malexmedia> a cube can be represented with only 12 triangles
21:35 < factor> Just wondering if you can remove a duck when more than one duck gets to close to the final duck pos
21:35 < malexmedia> but a sphere needs (lots) more
21:35 < malexmedia> so, if you're transforming from a cube to a sphere
21:36 < genete> (from a aquare to a circle)
21:36 < genete> square
21:36 < malexmedia> the cube needs to be transparently (in the back-end) subdivided to have the same number of triangles as the sphere
21:36 < factor> or added complexify :)
21:36 < malexmedia> before the transformation can occur
21:36 < malexmedia> then it's pretty much a straight line for each vertice
21:37 < genete> (that means that you have to de
21:37 < genete> define a transformation rule)
21:37 < factor> yers
21:37 < malexmedia> the complexity comes in when you have to be smart about where on the surface of the cube the subdivisions happen, so the distribution of triangles is closest to the other shape
21:37 < factor> or prams as well
21:37 < genete> because you can link wrong triangles from start to end and the middle would be a mess
21:38 < genete> for example convert the north to the south and viceversa
21:38 < malexmedia> no
21:38 < malexmedia> you can't just link any triangle to any other triangle
21:38 < malexmedia> you have to pair up triangles by distance
21:39 < malexmedia> so, if you have the first triangle in triangle list A
21:39 < malexmedia> you match it up not with the first triangle in triangle list B
21:39 < malexmedia> but the _closest_ triangle in triangle list B
21:39 < genete> assuming that both meshes are a the same average position
21:40 < genete> and orientation
21:40 < malexmedia> well, this algorithm is used most often for transforming one object type to another
21:40 < malexmedia> so position and orientation are irrelevant
21:41 < malexmedia> but
21:41 < malexmedia> if applied to synfig, i see your point
21:41 < malexmedia> for example, if you have a character facing left, then draw the same character facing right
21:41 < malexmedia> the interpolation would result in a new face growing out the back of his head
21:41 < malexmedia> (clearly not what the animator had in mind)
21:42 < malexmedia> in that situation, the animator would need to draw more key frames
21:43 < genete> the main key is: more aid from the computer you want, more preparation of the model you need
21:43 < malexmedia> i'm not sure i agree
21:44 < malexmedia> the algorithm i described wouldn't actually require any special preparation
21:44 < malexmedia> it might be difficult to implement, and i admit i might be overlooking something
21:45 < genete> malexmedia: I made some sort of head turn in synfig. _Just preparing the model properly. Watch it: http://www.youtube.com/watch?v=W3ctTJQ9koc
21:45 < dooglus> malexmedia: even if the 2 shapes have the same complexity, you still need to work out which points in each correspond
21:46 < factor> yup hooks
21:46 < factor> dont need all hooked just a good amount
21:47 < factor> depends on the complexity
21:47 < malexmedia> genete: very, very nice
21:47 < dooglus> but you're right - drawing in each keyframes pretty much solves that
21:47 < dooglus> *extra* keyframes I mean
21:47 < genete> not mention if the amount of points change between shapes
21:47 < malexmedia> dooglus: yeah, the whole point is to keyframe any substantial change in the motion
21:47 < dooglus> transforming a 'C' into a 'U' by either rotating or not: http://dooglus.rincevent.net/random/c-u.gif
21:48  * malexmedia looks
21:48 < genete> malexmedia: want to have the sifz file of the head turn?
21:48 < malexmedia> genete: sure :)
21:48 < dooglus> http://dooglus.rincevent.net/random/c-u.sifz if you want to see the synfig file
21:49 < malexmedia> dooglus: yeah, the interpolation method i'm proposing wouldn't see the possibility of rotation
21:49 < malexmedia> it would just mash up the object
21:49 < malexmedia> but, wouldn't the keyframe interpolation be just one tool in the arsenal?
21:49 < dooglus> malexmedia: if you put in enough keyframes then you'd get rotation I suppose
21:49 < malexmedia> clearly you can keyframe rotation already
21:50 < malexmedia> so there would be no point in drawing the every-other-frame that would be necessary to approximate rotation
21:51 < factor> just derpends how much computer aided help you want
21:51 < malexmedia> genete: how long did the head turn test take you?
21:51 < malexmedia> approximately?
21:52 < genete> one moment WC...
21:53 < factor> afk
21:54 < dooglus> malexmedia: to get back to the draw tool for a moment - do you think it's broken for the tablet specifically?  or is it just that when you use the mouse you go slower, and that's why it works better?
21:55 < dooglus> ie. can you reproduce the same brokeness with the mouse just by moving it faster?
21:55 < genete> malexmedia: from woman drawn in gimp to the final render
21:55 -!- Netsplit kubrick.freenode.net <-> irc.freenode.net quits: MangoFusion
21:56 < genete> it took me from saturday 11:20 am to sunday 20:47
21:56 -!- Netsplit over, joins: MangoFusion
21:56 < genete> consider that I did it as a hobby
21:56 < genete> so about five full hours
21:57 < malexmedia> genete: not bad at all
21:57 < genete> and here's the file http://www.darthfurby.com/genete/synfig/woman2.sifz
21:57 < malexmedia> dooglus: we're testing now
21:57 < dooglus> genete: how much of that was working out how to do it?  how long do you think it would take to do it now, with the experience you gained?
21:58 < genete> the mouse of the other day took me about 2 hours
21:58 < malexmedia> dooglus: it's hard to tell
21:58 < malexmedia> i think we expect a lot less from the mouse
21:58 < malexmedia> it's hard to get a reference point
21:58 < genete> http://www.youtube.com/watch?v=Y7w_CPFKMSg
21:59 < genete> dooglus:  I think you didn't watched the mouse for the cahllenge
21:59 < malexmedia> ie, when we move the mouse fast, does it look jerky because the mouse is a terrible pointing device? or does it look jerky because synfig is doing something wrong?
21:59 < genete> malexmedia: animate it was pretty easy. The main work was its design
22:00 < dooglus> malexmedia: try drawing a 'z' quite quickly with the mouse and with the tablet.  I see: http://dooglus.rincevent.net/random/z-test.png that synfig won't draw 2 corners - it rounds one of them off if I move too fast
22:01 < dooglus> genete: I saw it - it reminded me of your bald-headed lady :)
22:01 < malexmedia> genete: that mouse animation is good too
22:01 < malexmedia> :)
22:01 < genete> ouch! the balded lady would cry if you compare her with a mouse ;)
22:02 < dooglus> genete: and vice versa no doubt.  but the pseudo-3d look is similar in both, right?
22:02 < malexmedia> hey, how do you change the width of the stroke at a duck?
22:02 < genete> just kidding :) ... right
22:03 < dooglus> does the 5 hours include the time taken to shave her head?  wouldn't it have been faster to leave her hair on?
22:03 < malexmedia> and/or change the width of the whole stroke
22:03 < genete> dooglus: :P
22:03 < dooglus> malexmedia: alt-5 or caret>view>ducks>width will turn on width ducks
22:03 < dooglus> malexmedia: the toolbox has a default width widget.  the circle you see there?
22:04 < dooglus> to the right of the default color selectors
22:04 < genete> and to change the width of the shole stroke modify it at the paramters of the outline layer
22:04 < malexmedia> ah, we see the width ducks now
22:04 < genete> whole even
22:04 < malexmedia> and yes, we could change the width of the initial stroke
22:04 < genete> malexmedia: have you seen the mouse ears?
22:04 < dooglus> there are 6 types of duck - alt-1 through alt-6 toggles them, or use the caret menu
22:05 < malexmedia> dooglus: can you clarify "caret menu" ?
22:05 < malexmedia> genete: i noticed the transparency
22:05 < genete> impossible to obtain properly dhape if use draw tool. 
22:05 < dooglus> malexmedia: in the drawing window, the little triangle in the top left corner, between the horizontal and vertical rulers
22:05 < genete> using the bline tool you control the curvature more efficiently
22:06 < genete> and can trace over your model more accuracyly
22:06 < genete> (that's my opinion)
22:06 < genete> malexmedia: have you got the woman file?
22:06 < dooglus> malexmedia: that 'caret' pops up a menu - the same menu you get from right-clicking in the background of the drawing area.  it's just that there's not always any background visible to right-click on
22:07 < genete> malexmedia: look at its composition. 
22:07 < genete> it is "thought to be animated"
22:07 < genete> ah! and there is another thing you nave not considered!
22:08 < genete> layer z depth!
22:08 < genete> very useful if you prepare your model previously on that feature
22:08 < malexmedia> genete: we're opening the woman2 file now
22:09 < genete> malexmedia: look her head back folder
22:09 < malexmedia> looking in head back
22:10 < genete> it is mostly hidden behind the face
22:10 < genete> but it is deformed when she turns her head
22:10 < genete> simulationg her head back
22:11 < malexmedia> i see
22:11 < genete> frame 1s
22:12 < malexmedia> it's clearly very well thought-out
22:13 < malexmedia> plenty of food for thought, gentlemen
22:13 < malexmedia> i should hop out, but thank you all for your time
22:14 < genete> inj this case the eye is hidden by the nose but in other cases the eye is placed behind the head layer (for example it would be needed for the mouse example) if want to have a 90 deg turn)
22:15 < dooglus> genete: why did you get rid of the 'rotate' layer between woman and woman2?
22:15 < genete> malexmedia: you're welcome and come back 
22:16 < genete> dooglus: rotate didn't allow to keep points linked between rotated and not rotated layers
22:17 < genete> and I needed it for the head bacl layer that is part of the neck that doesnt rotate
22:18 < genete> If I added the rotate tool to make the last movements I needed to move the points of the base of the neck to keep them static
22:18 < genete> (that's why we need bones!!!
22:21 < factor> Ahh bones would be cool
22:21 < genete> dooglus: how can I use a new feature from an old file?
22:22 < genete> in particular how can I make use of plant width for a file created before its implementation?
22:24 < dooglus> genete: load the old file into a new synfig?
22:25 < genete> I've loaded the woman2.sifz and cannot modify the plant width (eyelashes)
22:28 < genete> can you?
22:28 < dooglus> what is 'plant width'?
22:31 < dooglus> oh, I see
22:31 < dooglus> copy & paste the layer
22:31 < dooglus> the new layer will honor the width
22:31 < dooglus> the idea is that the old layer should render the same way it always did
22:32 < genete> I know the idea is keep old behavior but can we manually override it?
22:32 < dooglus> yes - copy & paste the layer, then delete the old one
22:34 < genete> cool
22:34 < factor> tada
22:35 < dooglus> and if you want to make a copy of the old layer without having the width used?  you're stuck I guess
22:36 < genete> using the (uncoded yet) clone layer option?
22:37 < dooglus> I think 'use width' should probably be a user-visible parameter
22:37 < dooglus> maybe you don't want the width of the bline to affect the plant
22:38 < genete> it could be an option (but one of those strange animatable options)
22:38 < genete> dooglus: I cannot stop thinking on a new feature :/ very useful I think
22:39 -!- bytefield [i=5367a478@gateway/web/ajax/mibbit.com/x-2ad01b83173da2a9] has joined #synfig
22:40 < genete> if I want to make a region partially outlined, current solution comes from manual link some outlines to the region to partially outline it
22:40 < genete> what if we add a parameter to the vertex that tells if the outline is active on it or not?
22:41 < dooglus> genete: I guess setting the outline width to 0 makes it 'not outlined' in parts?
22:41 < genete> if you find to neighbout or more vertices with the outline off then you don't draw the curve between them
22:41 < genete> but you cannot end with a rounded or sharp end
22:41 < dooglus> right
22:42 < dooglus> if we're adding a new feature to allow partial outlining, I'd like to allow the outline to stop part-way along a curve, too
22:42 < genete> and it is a pain to animate like just an on or off feature
22:42 < dooglus> (without having to add a new vertex)
22:42 < genete> explain better to genete the english person of the IRC
22:43 < genete> problematic*
22:43 -!- bytefield [i=5367a478@gateway/web/ajax/mibbit.com/x-2ad01b83173da2a9] has left #synfig []
22:44 < genete> dooglus: now I understand
22:44 < genete> (probably late)
22:45 < dooglus> genete: see the furry animal thing you traced earlier - the outlines don't start and stop at corners
22:45 < dooglus> they start and stop part-way along smooth curves
22:46 -!- xerakko [n=Miguel@debian/developer/xerakko] has joined #synfig
22:46 < dooglus> genete: back to the plant layer - what "strange animatable options"?
22:46 < dooglus> you mean make it a regular layer parameter?
22:46 < genete> yes I guesses you wanted that
22:49 < genete> dooglus: for example I don't understand why antialiasing is an animatable feature. 
22:49 < dooglus> just because every parameter is animatable I guess
22:50 < genete> or as we commented sometime ago Children lock
22:51 < genete> so why loop a bline is not animatable?
22:51 < genete> and activate (on/off) bline points is not a parameter?
22:53 < dooglus> hmm
22:53 < dooglus> I don't know
22:53 < genete> (I'm not complaining just commenting)
22:53 < dooglus> 'loop' isn't a parameter
22:53 < genete> even rotate order...
22:54 < dooglus> right
22:54 < dooglus> both loop (bool) and order (int) could be sub-parameters of the bline valuenode type
22:54 < genete> yup
22:54 < genete> but don't loose focus on the partial outlines
22:54 < genete> can you develop what you meant?
22:55 < genete> <dooglus> if we're adding a new feature to allow partial outlining, I'd like to allow the outline to stop part-way along a curve, too
22:55 < genete> I have not idea how to do that
22:56 < genete> but stop only at vertices I can figure how to do it roughly
22:57 < dooglus> what was Zelgadis' feature suggestion a few days ago related to this?
22:57 < dooglus> he wanted some action at a distance from a bline vertex
22:58 < genete> don't remember :/
22:58 < dooglus> genete: http://zelgadis.profusehost.net/files/width-proposal.png
22:59 < genete> yes I see and remember
22:59 < dooglus> it's similar - he wants to have width changes in places other than vertices
23:00 < genete> yes it can work if you also allow "constant" interpolation for the width in that vertex. 
23:01 < genete> or simply that virtual vertex has a width duck too
23:01 < genete> that's cool
23:01 < genete> two kind of vertices
23:02 < genete> one with tangents and other without tangents and a amount parameter to define its position in the curve
23:02 < genete> I mean the outlines would be defined by two lists
23:03 < genete> one for the bline itself and other for the thickness
23:03 < genete> one have vertex and tangent (not width) 
23:03 < genete> and the other has only amount and width
23:09 < dooglus> or a single type of vertex, but with checkboxes for "vertex & tangent" and "width" which can selectively disable either
23:10 < genete> but when you change from one to other you have to place it over the hermite curve of its neighbours and would produce some strange step 
23:11 < dooglus> yes
23:11 < genete> also we would want to drag them like the linked to blines ones
23:12 -!- rubikcube [n=kvirc@dslc-082-082-085-145.pools.arcor-ip.net] has quit ["KVIrc 3.2.6 Anomalies"]
23:12 -!- malexmedia [n=alex@oh-76-2-223-75.dhcp.embarqhsd.net] has left #synfig []
23:13 < genete> I think that you can have "width vertices" that "belong" to a pair of tangent vertices (they cannot move outside). So you should not be able to move between two blines segments
23:13 < dooglus> oh, I was thinking "male x", but he's "m alex"
23:14 < dooglus> how is that restriction useful?
23:15 < genete> to avoid a mess if you drag vertices outside, isn't it?
23:16 < dooglus> I'm not clear on it
23:16 < genete> if you accidentally cross two "width vertices" you can change drastically the width of all the bline
23:16 < dooglus> can I have 2 width vertices on a single segment?
23:16 < genete> yes you should (must)
23:17 < dooglus> so they can still cross, even with your restriction of not leaving their segment
23:17 < genete> or as many as you want
23:18 < genete> I imagined them like gems in a necklace
23:18 < genete> you can move but cannot cross it
23:18 < genete> but hey!
23:19 < genete> allow cross and give to the user freedom!
23:19 < genete> maybe I want to produce that effect!
23:20 < genete> do you like the two lists idea?
23:21 < dooglus> I don't think I understand it :)
23:21 < dooglus> the gems in a necklace are just one list.  how does the necklace work with 2 lists?
23:22 < genete> forget the gems
23:22 < genete> two lists:
23:23 < genete> A) bline definition: vertex (x,y) tangents (r, alpha). split tangent and amount (for visibility of vertices)
23:24 < genete> b) width definition: width, amount (the position over the bline)
23:25 < genete> A) is a list of bline points (without width)
23:25 < genete> B) is a list of width points that lies on the bline list of A)
23:26 < genete> for width points 'width' has a visual representation of a radious duck
23:26 < genete> and amount has the representation of a "sliding duck"
23:27 < dooglus> it would need 2 ducks wouldn't it?  one for 'width', one for 'amount'?
23:27 < genete> like liked to blines ones
23:27 < dooglus> ok
23:29 < genete> dooglus: thinking for the future those "width points" can hold also the rounded or not rounded checkbox parameter
23:29 < genete> and remove it from the whole outline layer
23:31 < genete> it means that one width point with a not rounded value would be interpolated as an straight line from and to it
23:31 < genete> and if it has a rounded value it is interpolated as a tangent to the circle it defines
23:31 < genete> it would allow lots of variations
23:32 -!- omry [n=omry@bzq-84-108-20-56.cablep.bezeqint.net] has joined #synfig
23:33 -!- xerakko [n=Miguel@debian/developer/xerakko] has quit ["Ex-Chat"]
23:42 < dooglus> genete: sounds ok.  I'm not too sure about having 2 separate lists, but maybe it's ok
23:43 < genete> I'll add a link to our conversation on Zelgadis feature request
23:59 < genete> http://synfig.org/Wish_list#Width_weigths
23:59 < genete> it seems that my local time is crazy
23:59 < genete> it says I posted it at 17:51 lol!
--- Log closed Thu Apr 17 00:00:42 2008