Hi all, Release is getting closer, and we need to decide what parts of Powerstroke we are going to release. After release, it will be very annoying for users if we change the output of Powerstroke for certain input parameters. So we have to avoid having to do that.
Ideas: - remove reflection-based extrapolated miters (my original extrapolation), add the arc extrapolation miter - remove CubicBezierFit and CubicBezierJohan interpolators, add Catmull-Rom interpolator
thanks for input. I hope to work on this soon.
-Johan
Hi Johan and all comunity, one -probaly not wanted- opinion, could be posible a "simple" interpolator like the green one. Think this is the type of interpolator normaly artist want on a design. It make more saccesful the use of powerstroke -less precise but more fast to use and with kindly results-
I use some type of stroke with changer in my previous program- Creature House Expesssion-, it have less options but the result "always feet" as the artist user expect.
I think a goal in power stroke be: easy to use with one knot or more. now with more than one knot on a path the design process slow down.
Maybe is to late or i'm opening pandora box, sorry if is as it.
Regards, Jabier.
El mié, 03-09-2014 a las 23:36 +0200, Johan Engelen escribió:
Hi all, Release is getting closer, and we need to decide what parts of Powerstroke we are going to release. After release, it will be very annoying for users if we change the output of Powerstroke for certain input parameters. So we have to avoid having to do that.
Ideas:
- remove reflection-based extrapolated miters (my original
extrapolation), add the arc extrapolation miter
- remove CubicBezierFit and CubicBezierJohan interpolators, add
Catmull-Rom interpolator
thanks for input. I hope to work on this soon.
-Johan
Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Another opinion: For this kind of interpolator -or for all- a knot -not necessary all- can have two modes, interpolated or not -cusp-, this add more power to PS.
Regards, Jabier.
El jue, 04-09-2014 a las 00:27 +0200, Jabiertxo Arraiza Cenoz escribió:
Hi Johan and all comunity, one -probaly not wanted- opinion, could be posible a "simple" interpolator like the green one. Think this is the type of interpolator normaly artist want on a design. It make more saccesful the use of powerstroke -less precise but more fast to use and with kindly results-
I use some type of stroke with changer in my previous program- Creature House Expesssion-, it have less options but the result "always feet" as the artist user expect.
I think a goal in power stroke be: easy to use with one knot or more. now with more than one knot on a path the design process slow down.
Maybe is to late or i'm opening pandora box, sorry if is as it.
Regards, Jabier.
El mié, 03-09-2014 a las 23:36 +0200, Johan Engelen escribió:
Hi all, Release is getting closer, and we need to decide what parts of Powerstroke we are going to release. After release, it will be very annoying for users if we change the output of Powerstroke for certain input parameters. So we have to avoid having to do that.
Ideas:
- remove reflection-based extrapolated miters (my original
extrapolation), add the arc extrapolation miter
- remove CubicBezierFit and CubicBezierJohan interpolators, add
Catmull-Rom interpolator
thanks for input. I hope to work on this soon.
-Johan
Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On 4-9-2014 21:44, Jabiertxo Arraiza Cenoz wrote:
I give a deep try to power-stroke. Sorry for the noise.
No worries Jabier. Powerstroke certainly could be much improved (also in usability). But for now, I'd like to work towards a solid release. After release, we are free to work on improving it and testing new ideas.
regards, Johan
Hi to all again.
Searching throught powerstroke code i find the problems i find is not about interpolator, is in the next step fitting the "interpolated plane curve" to the final curved ones.
This bugs/gaps happens also in the pattetn along path.
I find this document, maybe too old and also not enought good or appropiate to fix the problem... http://people.csail.mit.edu/sarasu/pub/cgim02/cgim02.pdf Maybe searching in Gimp code for brush...
I can help on it when 0.91 is on and finaly end my PointWise and Fillet-Chamfer LPE.
Regards, Jabier.
El vie, 05-09-2014 a las 00:14 +0200, Johan Engelen escribió:
On 4-9-2014 21:44, Jabiertxo Arraiza Cenoz wrote:
I give a deep try to power-stroke. Sorry for the noise.
No worries Jabier. Powerstroke certainly could be much improved (also in usability). But for now, I'd like to work towards a solid release. After release, we are free to work on improving it and testing new ideas.
regards, Johan
Also a document could be interesting. Regards, Jabier.
El vie, 05-09-2014 a las 19:34 +0200, Jabiertxo Arraiza Cenoz escribió:
Hi to all again.
Searching throught powerstroke code i find the problems i find is not about interpolator, is in the next step fitting the "interpolated plane curve" to the final curved ones.
This bugs/gaps happens also in the pattetn along path.
I find this document, maybe too old and also not enought good or appropiate to fix the problem... http://people.csail.mit.edu/sarasu/pub/cgim02/cgim02.pdf Maybe searching in Gimp code for brush...
I can help on it when 0.91 is on and finaly end my PointWise and Fillet-Chamfer LPE.
Regards, Jabier.
El vie, 05-09-2014 a las 00:14 +0200, Johan Engelen escribió:
On 4-9-2014 21:44, Jabiertxo Arraiza Cenoz wrote:
I give a deep try to power-stroke. Sorry for the noise.
No worries Jabier. Powerstroke certainly could be much improved (also in usability). But for now, I'd like to work towards a solid release. After release, we are free to work on improving it and testing new ideas.
regards, Johan
Hi Johan.
Working whith the attached svg, i nottice 2 things can ?easy changed:
Top one: when the cap is not 0 width, the result is strange if not knot at start or end position.
Second one: catmull-rom give strange results whith consecutive knots at same Y -distance to the center.
First one not sure how his best implement but second one could be something posibol at "calc_bezier" function, mimmic the part: 261::// safety check for repeated points but for move a bit the Y position.
Regards, Jabier.
On Wed, Sep 3, 2014 at 3:27 PM, Jabiertxo Arraiza Cenoz <jabier.arraiza@...2893...> wrote:
Hi Johan and all comunity, one -probaly not wanted- opinion, could be posible a "simple" interpolator like the green one.
Is your thought that the green would have been achieved by a single powerstroke knot?
Cheers, Josh
No by two like the power-stroked black one -at the same position- Green is only a bad aproximation of probabily the result the user want with this two knots. In my opinion of course.
Regards, Jabier.
El mié, 03-09-2014 a las 15:49 -0700, Josh Andler escribió:
On Wed, Sep 3, 2014 at 3:27 PM, Jabiertxo Arraiza Cenoz <jabier.arraiza@...2893...> wrote:
Hi Johan and all comunity, one -probaly not wanted- opinion, could be posible a "simple" interpolator like the green one.
Is your thought that the green would have been achieved by a single powerstroke knot?
Cheers, Josh
On Wed, Sep 3, 2014 at 2:36 PM, Johan Engelen <jbc.engelen@...2592...> wrote:
- remove CubicBezierFit and CubicBezierJohan interpolators, add
Catmull-Rom interpolator
It appears to me (and I could be wrong) that the smoothness parameter only affects the Johan interpolator. Is that correct? I'd be interested to see the Catmull-Rom one of course.
Cheers, Josh
On 4-9-2014 0:52, Josh Andler wrote:
On Wed, Sep 3, 2014 at 2:36 PM, Johan Engelen <jbc.engelen@...2592...> wrote:
- remove CubicBezierFit and CubicBezierJohan interpolators, add
Catmull-Rom interpolator
It appears to me (and I could be wrong) that the smoothness parameter only affects the Johan interpolator. Is that correct? I'd be interested to see the Catmull-Rom one of course.
Correct. I'll remove it too. My main goal is to remove bells and whistles.
-Johan
On 4-9-2014 0:52, Josh Andler wrote:
On Wed, Sep 3, 2014 at 2:36 PM, Johan Engelen <jbc.engelen@...2592...> wrote:
- remove CubicBezierFit and CubicBezierJohan interpolators, add
Catmull-Rom interpolator
It appears to me (and I could be wrong) that the smoothness parameter only affects the Johan interpolator. Is that correct? I'd be interested to see the Catmull-Rom one of course.
The Centripetal Catmull-Rom interpolator is now in trunk. I think it is justified to add it this late in dev-cycle, because afaik it is the main contender for future SVG's variable stroke width paths. It has certain features that I think are sometimes desired, sometimes not desired: it can result in non-function curves (multiple values at certain time coords), which means the stroke will "back up on itself". Don't know how to explain quickly. Build latest trunk and play with it to see what I mean.
cheers, Johan
So I messed around a bit just to see what the various results were. Unfortunately Inkscape locked up while dragging a power stroke knot so a particularly odd one with catmull isn't included. Anyway, here are 4 examples with each type of interpolator. I would consider keeping Fit as an option as it has an interesting handling of 3 and the only one that doesn't double up on 4.
After you open this, you need to select all of the number 4 examples and press an arrow key, as it doesn't render the end cap correctly on load. https://www.dropbox.com/s/525yp1fagn0fb96/powerstroke.svg?dl=0
Cheers, Josh
On Thu, Sep 4, 2014 at 3:05 PM, Johan Engelen <jbc.engelen@...2592...> wrote:
On 4-9-2014 0:52, Josh Andler wrote:
On Wed, Sep 3, 2014 at 2:36 PM, Johan Engelen <jbc.engelen@...2592...> wrote:
- remove CubicBezierFit and CubicBezierJohan interpolators, add
Catmull-Rom interpolator
It appears to me (and I could be wrong) that the smoothness parameter only affects the Johan interpolator. Is that correct? I'd be interested to see the Catmull-Rom one of course.
The Centripetal Catmull-Rom interpolator is now in trunk. I think it is justified to add it this late in dev-cycle, because afaik it is the main contender for future SVG's variable stroke width paths. It has certain features that I think are sometimes desired, sometimes not desired: it can result in non-function curves (multiple values at certain time coords), which means the stroke will "back up on itself". Don't know how to explain quickly. Build latest trunk and play with it to see what I mean.
cheers, Johan
participants (3)
-
Jabiertxo Arraiza Cenoz -
Johan Engelen -
Josh Andler