A little Mafalda animation, suggest ways to improve

Artwork Finished works
1

I made this little animation with Mafalda characters :slight_smile: it took me 24 hours (over 4 days) in Synfig and Krita.

My workflow is something of a Frankenstein monster, because I was feeling a little insecure about Synfig’s ability of doing what I want:

  1. I use Krita to make a cut-out doll of the main character, with carefully placed rotation center markers, and “filled arc” extensions, that would fill the gaps when a limb folds from 0 degrees to 90 degrees. Each limb is a different color to make sure where one ends and other starts.
  2. I use Synfig to animate the doll alone with bones. I made 2 walking animation versions: for a mostly sideways view of the character, and for a mostly frontal view. Then 1 animation for standing poses. The doll is articulated articulated in the ankles and toes too :slight_smile:
  3. I exported those animations in separate folders. I used ImageMagick to convert from color to black-and-white. I still can’t figure out how to remove the rotation markers: using a different layer in Krita would produce a risk of mis-alignment. With a different color, to be removed with ImageMagick, leaves a faint empty black circle ← I would like suggestions here.
  4. With each walking gate and standing poses in three different folders, I make a new project in Synfig, add the background, the front layer of background (tree), and the secondary character, with just a time-loop animation, playing the trumpet.
  5. I import the main character 3 times: one with the sideways walking frames, one with the frontal walking frames, and one with the standing frames.
  6. I animated the frontal walking, adding stops where I want the character to pause and do something standing in place.
  7. At that place in the timeline, I make the sideview opacity=0.0 and the standing animation=1.0, then animate the standing character by choosing frames repeatedly until the 32 frames are done.
  8. At the end of the 32 frames, the opacity is returned to the sideview character.
  9. Then I repeat to for the other “standing in place events”, and the frontal walking animation.
  10. I do it all with the entire background fitted in the 1280x720 canvas.
  11. I use translate and scale filter at the top of the object list to zoom in and pan.

I wonder if I need to import the character 3 times, so that I avoid messing with opacity, and select the frame sequence at each period of time. This part was much more work than I imagined, and making a mistake here would be hard to correct.

I discovered that zooming and panning with scale and translate are very convenient… although I abused it a little bit by using the scale’s origin property to add a zoom in the last moment.

I noticed that the timeline waypoint markers don’t render until I pass the mouse cursor over the area. Is this normal while animating the scale layer?

Thanks in advance if I get any suggestions :slight_smile:

hqdefault
hqdefault480×360 23.1 KB

4 Likes
2

Hey, you did pretty good, congrats! There’s a mistake in a walking cycle, the girl is sliding like she’s on ice (when moving one foot should be planted statically against the ground) but other than that it’s good.

Don’t worry, there won’t be any misalignment as long as you export markers at the same resolution. Just put markers in separate layers and when exporting to png uncheck the character’s layers, so in the end you will have two sets of png images: one with the character and the other one with markers. When done composing in Synfig, just remove the no longer needed markers.

Alternatively, you can group everything into one folder (Select all layers->Right click->Group layer) and manipulate the camera through transformation widget: Group Transformation Widget - Synfig Animation Studio
Maybe you will find it easier this way.

There’s “Switch Group” for this exact purpose: Switch Group Layer — Synfig User Manual 1.5.1 documentation
Check out the link, there’s also a video tutorial there.

Additionally, if you need opacity or any other parameter to change suddenly (on/off) instead of interpolating, use constant interpolation waypoints: https://www.youtube.com/watch?v=lMmtOocDaqQ

3

THANK YOU FOR THE FEEDBACK :slight_smile:

I didn’t think of the switch layer… I was only using it for stacks of PNG files. I will try it. I really want to stop doing it the way I did it.

Girls feet not sliding on the ground… I just found that the group layer can modify the time scale… with 1.50, it doesn’t “slide” anymore, but it looks funny. Maybe my animation should be longer in time, or the whole stage should be smaller.

I will try that idea with extra PNG layer with the rotation center marker.

Regarding zoom and pan: I will try grouping it, then animating the transformation widget.

1 Like