This is fantastic! Never in my life could I input my signature with a mouse, and have it come out this close to how it looks like the pen-and-paper version. Mindblowing, as it worked like so out of the box.
Using a trackball I couldn't get this to work noticeably better than just plain paintbrush input in MS Paint. Conversely the one from TFA took me a little more time to get the hang of but I got a lot better control with it.
I feel like if you're modifying the signature after they sign it and approve it - that could be a problem. As long as the modifications are applied in real-time (or with explicit user confirmation after modification), I think it is morally okay (and probably legally, but I'm not a lawyer).
You may find users who get mad if your settings are too aggressive though, and if they're unable to get a signature that they approve of.
Given it a bit of thought (as I was also puzzled), I think the comment was about the modification compared to a normal way of doing it. So if I would normally write a scribble with angular lines for a signature, and it might be consistent across different places, this brush could smooth it out and modify what it might be.
This technique was also used in mid-late non-motion-plus Wii games to smooth out the pointer movement! Early games hadan incredibly twitchy pointer because they were simply mapping the IR data 1:1 to cursor movement, whereas later ones have an invisible circle around the cursor and only move the cursor itself once the circle edges start "dragging" it.
This is so satisfying. These types of experiments are something I really love about the open-web, and part of what bums me out about how most social networks tend to throttle links.
The dragging behavior is so intuitive – it's funny because usually if you create this kind of resistance in a UI it can be confusing, but in this context it works so well.
> this kind of resistance in a UI it can be confusing
it's actually intuitive because it mimics a real life physical dragging of an object by a rope, which most people have a feel for. Skeuomorphism can be quite intuitive imho.
It’s not a tool for signatures. It’s also hard to sign documents with a paint roller.
But this UI is much better at, e.g. drawing a smooth, symmetrical heart symbol, with a crisp turn.
Different UI for different tasks, and it’s very cool to see something that intuitively lets you control something that is normally hidden under your finger.
Skeuomorphs retain purely ornamental components of the thing it’s mimicking. If it was a graphic of a rope rather than a dashed line, or if it looked like the line was tied up around the thing in a bow, those would be skeuomorphic elements. But graphical interface elements always retain ‘some’ functional connection to the physical world and trigger ‘some’ abstracted existing metal model about how the world works. GUIs themselves conceptually mimic the idea of a physical space because it’s easier for people to reason about than a bunch of text. This is no more skeuomorphic than the reply button below the comment box.
I don't think I've ever seen this feature ever before (keeping in mind that the purpose of the tool is to draw smooth lines, and there would probably be another tool for drawing signatures). It's quite brilliant!
Does it have the leading drag handle? That's the brilliant part. My ancient copy of Photoshop has a smoothed brush; but it doesn't have the leading drag handle.
Kind of. You can click a gear and set various options on it, including a "pulled string mode", which works similarly to this, where it has a "smoothing radius" that movement within will not change cursor direction, and then a "stroke catch up" mode where the stroke is at a somewhat constant speed, regardless of how you move your mouse.
I wonder what Duo Lingo are using behind the scenes. I've been busy with the Chinese and Japanese courses, and one thing I quickly noticed is how there are two different 'grades' of practising drawing hanzi/kana: when you first learn how to write it (with your finger, in the app) it constrains your movements so they perfectly follow the guidelines. But then later, it relaxes the constraints totally, or almost totally, so your hiragana can really be a shorthand squiggle. They then assess it quite liberally too, so as long as the general gist is correct, your kana will pass -- but it's also easy to fail it if you're just guessing.
This is very nice, not just for finger/mouse painting! I tried it on my Cintiq and it was actually a lot better for me than brush stabilization usually is - I think the logic is the same as seen in e.g. Krita, but the visualization of the cursor and where the paint will appear is very helpful. Usually painting software doesn't have such an indicator of where the actual stroke will be placed and when it will move.
I believe a logic similar to this was used to enact the "Gestures" system in Black and White 1. Breaking down the mouse-movements into vectors following a guide-point. (https://blackandwhite.fandom.com/wiki/Gesture).
Damn, what a flashback. I forgot about that game, it was quite something for its time. I remember the gesture spell casting system not working very consistently, but it was still a ton of fun.
Wow, this is amazing! I see you've been building this on GitHub for 7 years - that's truly impressive dedication. What keeps you motivated to stick with this product for so long?
I think it would be much better to use some kind of Kalman filtering so you don't have the huge lag. You don't need to commit to permanent output immediately so it is possible to draw with zero lag and then smooth the output afterwards.
The cursor lags behind so your finger doesn’t cover it. That helps for drawing carefully, but it means you need to use an exaggerated motion when changing direction. It’s a little unnatural at first, but might be better than alternatives with practice?
Very nice! I love the way the leash droops from gravity when you give it some slack, to unobtrusively and intuitively show users what's really going on, how the control system actually works, so it's tangible and physical, not a mysterious unpredictable black box. True honest wysiwyg "direct manipulation", no invisible magic.
>The program Dynadraw implements a dynamic drawing technique that applies a simple filter to mouse positions. Here the brush is modeled as a physical object with mass, velocity and friction. The mouse pulls on the brush with a synthetic rubber band. By changing the amount of friction and mass, various kinds of strokes can be made. This kind of dynamic filtering makes it easy to create smooth, consistent calligraphic strokes.
>The C source for the dynadraw demo program is available. You can save this onto your IRIS, compile it, and give it a try. If you're a Mac or PC hacker, you might be able to port this program easily to your own platform.
// This is a rudimentary p5.js 'port' of Paul Haeberli's
// legendary and monumentally influential program "Dynadraw",
// which is described at: http://www.sgi.com/grafica/dyna/index.html
// Originally created in June 1989 by Paul Haeberli (@GraficaObscura)
// Ported to Processing January 2004 by Golan Levin (@golan)
// Ported to p5.js September 2021 by Golan Levin.
Golan does a mind blowing amount of amazingly creative stuff:
I also love "Rouen Revisited" (1996)̀, a "monumental" architectural visualization that Golan Levin and another artist/engineer Paul Debevec collaborated on at Interval Research Labs, which led to Paul's later work at USC:
Was looking for the DynaDraw shoutout. As a calligrapher, it’s the way to go for something more expressive than fixed lag.
Haeberli used a simple simulation of Hookes law,
Where F=-kx
F is the force applied to the spring.
k is spring constant or stiffness.
x is extension distance.
https://perfect-freehand-example.vercel.app/
(I honestly am baffled that things like this have to be said.)
If you're concerned about people modifying a document after you share it out, consider using a digital signature or a hash.
You may find users who get mad if your settings are too aggressive though, and if they're unable to get a signature that they approve of.
https://drawmote.app/
And apparently without asking any permissions on the phone. What sorcery is this?
The dragging behavior is so intuitive – it's funny because usually if you create this kind of resistance in a UI it can be confusing, but in this context it works so well.
it's actually intuitive because it mimics a real life physical dragging of an object by a rope, which most people have a feel for. Skeuomorphism can be quite intuitive imho.
But this UI is much better at, e.g. drawing a smooth, symmetrical heart symbol, with a crisp turn.
Different UI for different tasks, and it’s very cool to see something that intuitively lets you control something that is normally hidden under your finger.
Friction: 0.04
Brush radius: 13
-> clear
-> draw your signature
The former looks like this https://i.imgur.com/TGE7N1z.png
Perfect freehand is the right way to solve this.
I wonder what Duo Lingo are using behind the scenes. I've been busy with the Chinese and Japanese courses, and one thing I quickly noticed is how there are two different 'grades' of practising drawing hanzi/kana: when you first learn how to write it (with your finger, in the app) it constrains your movements so they perfectly follow the guidelines. But then later, it relaxes the constraints totally, or almost totally, so your hiragana can really be a shorthand squiggle. They then assess it quite liberally too, so as long as the general gist is correct, your kana will pass -- but it's also easy to fail it if you're just guessing.
There's even a program called Lazy Nezumi that adds global mouse smoothing, rulers etc. to Windows.
https://en.wikipedia.org/wiki/Hysteresis
It’s ~20 commits done in two batches. 2018 for the initial release then some more work on it in 2023.
Zastai: you could have that kind of progress on any lib you wish to release, and it’s a nice feeling to have something out in the wild.
https://en.wikipedia.org/wiki/Direct_manipulation_interface
It reminds me of artist/engineer's Paul Haeberli's legendary and monumentally influential "DynaDraw" which he wrote at SGI.
DynaDraw: A Dynamic Drawing Technique (June 1989):
https://web.archive.org/web/19970605062552/http://www.sgi.co...
>Here's a really fun and useful hack.
>The program Dynadraw implements a dynamic drawing technique that applies a simple filter to mouse positions. Here the brush is modeled as a physical object with mass, velocity and friction. The mouse pulls on the brush with a synthetic rubber band. By changing the amount of friction and mass, various kinds of strokes can be made. This kind of dynamic filtering makes it easy to create smooth, consistent calligraphic strokes.
>The C source for the dynadraw demo program is available. You can save this onto your IRIS, compile it, and give it a try. If you're a Mac or PC hacker, you might be able to port this program easily to your own platform.
Source Code (for SGI GL):
https://web.archive.org/web/19970727185417/http://www.sgi.co...
Lots more great stuff from Paul Haeberli on his "Graphica Obscura" page:
https://web.archive.org/web/19970706205455/http://www.sgi.co...
https://en.wikipedia.org/wiki/Paul_Haeberli
He inspired another artist/engineer Golan Levin to write an implementation of DynaDraw with Processing and p5.js:
https://editor.p5js.org/golan/sketches/cZPRgx6q9
Golan does a mind blowing amount of amazingly creative stuff:http://www.flong.com/archive/index.html
https://en.wikipedia.org/wiki/Golan_Levin
I love his Double-Taker (Snout) (2008):
http://www.flong.com/archive/projects/snout/index.html
He explained that and more in this Ted talk "Golan Levin makes art that looks back at you":
https://www.youtube.com/watch?v=1G0MzlfMPuM
>As Joy Mountford once said, "The mouse is probably the narrowest straw you could try to suck all of human expressions through."
https://en.wikipedia.org/wiki/Joy_Mountford
I also love "Rouen Revisited" (1996)̀, a "monumental" architectural visualization that Golan Levin and another artist/engineer Paul Debevec collaborated on at Interval Research Labs, which led to Paul's later work at USC:
http://www.flong.com/archive/projects/rouen/index.html
Paul Debevec teaches and researches at USC and Eyeline Studios (powered by Netflix):
https://www.pauldebevec.com/
https://en.wikipedia.org/wiki/Paul_Debevec
Paul Debevec's "Light Stage" was featured in The Congress (2013), in the dramatically riveting and technically accurate scan scene:
https://www.youtube.com/watch?v=pPAl5GwvdY8
More links:
https://news.ycombinator.com/item?id=34953477
Haeberli used a simple simulation of Hookes law, Where F=-kx F is the force applied to the spring. k is spring constant or stiffness. x is extension distance.
DynaDraw also added damping IIRC
Thx for the links