Skip to content

Bywords#

Bywords makes words appear in a Synfig animation as they are spoken, based on a subtitle file.

It's useful in itself, but it also serves as a demonstration of the Sangfroid library.

Example#

Let's take the first two lines of How Can I Sing Like a Girl? by They Might be Giants:

TMBG performing the song

The subtitle file used to caption this gif is in SubRip format (.srt) . It looks like this:

1
00:00:00,000 --> 00:00:02,600
How

2
00:00:02,700 --> 00:00:06,000
can I sing like a girl

3
00:00:07,000 --> 00:00:10,500
and not be stigmatised

4
00:00:11,000 --> 00:00:14,000
by the rest of the world?

Since each line has a half-bar pause, we have split them both into two parts, giving four lines-- each with a start time and an end time.

In practice, it's fine to keep the Bywords subtitle file separate from the one you use for captioning.

Now, we'll combine it with this animation:

The basic animation.

All text is always visible, except for the ampersand, which we'll discuss below. The first line pivots away at the appropriate place. This, and the ampersand, are the only animated parts.

Here are the layers of that animation:

The layers pane in Synfig Studio. Four groups (highlighted) have names beginning

Bywords will matches the subtitle lines to these groups based on the text they contain. The names must begin with "bywords", but the rest is arbitrary; the name may simply be "bywords".

Now, suppose we type 

bywords sing-like-a-girl.sif

(We don't have to specify the .srt name, as long as it's in the same directory with the same basename. Otherwise, you can specify it as an extra parameter.)

The .sif file will be modified in place, giving:

The words of the song, appearing at the times given in the subtitle file.

Everything else is kept the same. If you don't trust it to work, make a copy of the .sif first-- although you should be backing everything up anyway.

All the lines, except the first, contain multiple words. The pauses between these are weighted by the number of letters in each word. This is because longer words take longer to say. It's only approximate, English spelling being what it is.

If instead we type 

bywords sing-like-a-girl.sif --letter

then we get:

Letters appearing one by one, instead of words.

Text justification#

Make sure that the "Orientation" field of each text layer is set to (0.0, 0.0), which means it's measured from the top leftmost point of the layer:

Synfig Studio, with the Orientation field of a text layer highlighted.

You should do this because Bywords animates the "Text" field of the layer, making the layer wider as it goes. So if the orientation isn't calculated from top left, the text will slide across the screen. Here's an outtake, demonstrating the problem:

The words "how" and "girl" are centred, causing them to expand from the middle of the canvas.

This doesn't apply to single-word layers in word-at-a-time mode (or single-letter layers with --letter).

You might find it useful to centre the text while you're laying the page out. Then duplicate the layer, and set the "Orientation" layer of the new layer to (0.0, 0.0). Drag the new layer until it's exactly over the old layer, and then delete the old layer.

Animating the ampersand#

The line

and not be stigmatised

is an example of displaying text which isn't directly controlled by Bywords. In the animation, it's shown with the word "and" replaced by an ampersand:

& not be stigmatised

It corresponds to the group "bywords and not". This group does contain the word "and"-- otherwise Bywords wouldn't match the group-- but the word is set to invisible.

After we've run Bywords once, we can look up where the invisible "and" is set to appear:

Synfig Studio, with the "and" layer selected. The time cursor is on the waypoint which brings in the word "and".

...and add a new waypoint manually, at the same position.

Now we have the "&" layer selected.