Tweening 2d Hand sketches?

Hi everyone, I was wondering if it was possible for hand sketches to be filled in? I sort of want to create a cartoon that’s mostly handrawn 2d stuff and limited 3d modeling stuff. Will the skeleton work? The thing is, I see that in many tutorials many body parts are isolated and stuff. I would like that too since I’ll keep certain parts, however, I don’t really plan on moving them as the tutorial has them. I will keep certain static parts so I can hand sketch the moving parts.

Can Synfig interpolate the movement between each hand drawn frame/part ? I will keep it like 4-5 frames per seconds, but can I count on Synfig to generate the rest in between? I will try to keep the movement small too, but I’d like it flow a lot.

It’s almost like those anime cells. and then filling in the gaps as much as possible.

I honestly hope that animators do not have to draw those supposed 15 frames per second and that parts are tweened too. I would feel really really bad for them, but also respect them too.

Or is there another program out there that can guess the movement and all? Or maybe it can guess, and I can add in more details in each generated frame.


Hi junwen28,
I have to leave for a trip right now. I would like to reply properly when I return.
Please be so kind of wait a few days.
Cheers and beautiful drawings!

The quick answer is yes, if you use Synfig to draw and edit the lines. If you’re talking about scanning in hand-drawn pictures from paper, then no, Synfig can’t interpolate bitmap drawings (jpgs, bmps,pngs, etc)

I can see how some of those shots could be done very easily in Synfig, other shots would be much harder (e.g. the head turn which would require many unique drawings)

Do you think it’s possible one day a computer will be able to interpolate hand drawn art? I’ll of course edit it inside photoshop to make the lines and colors much much more differentiable ? Any programs?

Thank you.

It depends not on the complexity of the drawing but on the complexity of movement within your animation.

Here’s something you could try to test out your art with Synfig:

First - take two key drawings that you would like to TWEEN.

  1. Scan your line drawings, clean up and enhance the contrast then save as bitmap or PNG images.
  2. Import the image into INKSCAPE ( )
  3. Use INKSCAPE to “Trace Bitmap” at two levels so you get black line art and white areas.
  4. Delete the white areas and save the file.
  5. Import the saved INKSCAPE file into SYNFIG.

Now you have a vector graphics version of your first key drawing at FRAME 0. So create a new KEY position at Frame five, for example, and see if you animate your drawing.

This will work with simple movements. More complex actions will require more drawings but it will tween some for you and should reduce your workload.

If it works, you can use INKSCAPE to colour your artwork as well.


d try to test out your art with Synfig:

First - take two key drawings that you would like to T

Could you explain how to or point to a tutorial about how to draw two separate shots in synfig and then let synfig tween between the two? Thank you.

Here is a practical case:

  1. Insert your keyframes at the correct time imported in Synfig Studio as image list.
  2. Study the keyframes, shapes and lines to construct the vector layers in Synfig studio.
  3. Draw a draft (outlines) and refine timing.
  4. Repeat 2 and 3 as much as needed
  5. Once happy create the full artwork over the imported images and matching the keyframe timing.
  6. Rework the step 6 as needed.

This is all a huge help, but I have a question. When the head turns around, at first their are no eyes and at the end we see the eyes. How do those lines appear? I know how an animation like the walk cycle tutorial would work, but how do you make lines such as the ones for the eyes appear? I don’t understand how a shape can be there at one point but not at another. Thanks for the help!

The basic idea is that the shapes are there all the time, you just need to change its z-depth at the correct time with the correct shape to tween correctly. Notice that this is a rare case of a smooth head turn with full details. Now that I see (since long time ago I traced the keyframes) I think that it is too slow. It looks not real to me. To emulate a head turn you can use other techniques that would involve less number of keyframes but a good preparation, like the example.
See this quick head turn:
Before do the quick flip, prepare the keyframes carefully
On quick turns, eyes are closed (unnoticeable sometimes but it is there)
Use counteractions or secondary motions (tongue, hair, ears, it depends on the character)

I hope you find it useful.

Okay, so when something new is shown in a Synfig animation, it is actually not new. Meaning, all the objects are present throughout an entire scene, they just aren’t shown for the entire scene. Is this correct?

Right, they can be hidden by other objects or have its amount value (transparency of the object) set to zero. Also they can be out of the scene frame.
There is one option for BLines where you can hide and show blinepoints on the timeline. It is a bit tricky but works reasonabely well for intermediate points. Right click on one blinepoint and select “Set active point ON” or “Set active point OFF”.

I really like the look of the final colored frames from Pokemon that are in the link on the original post of this thread. How can we use synfig to get this kind of look. It looks not so crisp as vector art. It looks like it has some graininess or texture. How can we make our synfig renders look like this? Thanks!

You can emulate some sort of grainy texture using a noise layer and blending it correctly over the final work. Take a look to the 0.62.02 splash screen sifz file and see how it is achieved: … f9972faef2


I’m wondering when rendering in Synfig, which method of “hiding” a layer renders fastest:

  • moving the layer off of the stage
  • moving the layer behind other layers
  • setting the layer amount Amount to 0

Or are they all the same? Is there a best practice for hiding layers?

Setting its amount to 0 will increase it. The other options actually renders the layer although it is not visible. Placing it outside of the canvas dimension can be fast too, but you can’t be sure that it has some layer which could affect other visible parts so the safest way is set its amount to zero.

I know that it is a bummer to animate the amount value and also to set its waypoint type to constant. You can have a workaround that is convert the Amount parameter to a Switch type. Then set the Link On to 0.0 and the Link Off to 1.0 and just animate the switch parameter normally. It would show/hide the layer easily.

Some day we have to solve this annoying thing…

Won’t the noise layer jitter when rendered? So the draft pencil is designed for quick sketching. Does it work for frame by frame timed sketching. That makes sense. I do wish it was saved internally. Is the draft pencil raster or vector? That would be nice if my understanding of the tool is correct since I wouldn’t have to use another app for roughing.

If the ‘Animation Speed’ parameter is 0.0 it will stay stationery. Try to give something bigger than 0.0 to see it in action.

No it doesn’t

It is only saved externally.


Synfig saves me so much time in other ways that this is only a small problem. However it is good to know that setting the Amount to 0 is best in terms of rendering performance. Perhaps this would be a good to add to the FAQ.

Setting the Amount to a Switch is a good tip as well - I’ll do that in future.

Many thanks for your help again Genete!

I guess what you want it similar to this: Potentially it could have a great effect, but the images still aren’t vectorised, which is what you want for animation.