slow sText node - 18x slower than Text+

I'm working on some Fusion compositions that display the text of entire long chapters of a book. Hence, typically I could insert about 25,000 to 35,000 characters-including-spaces into a Text+ node.

In Davinci 18.6.6 running on a Mac Mini M1 16GB RAM, the Text+ node can readily handle this.

For example, using 18.6.6 if I change the font size in the Text+ node, I can see the change in font size in the viewer in about less than a second.

Whereas, Davinci Beta 19.b1 running on a Mac Studio Ultra 64GB RAM, the sText node is very slow. For example, using the same 25,000 characters in the beta sText node, I have to wait about 8 seconds before the sRender node turns green to indicate it is processing. Then it takes a further 10 seconds for the change in font size to be displayed in the viewer. That's a total of 18 seconds.

I tested the beta sText again, and it was 11 seconds.

I tested the beta sText yet again, and it was 18 seconds.

Compared to about 1 second for the 18.6.6 Text+ node.

(You might ask: why don't I just keep using the earlier Text+ node? That's because I want to take advantage of the new sText where the text is not degraded when increased in size.)

Update:

I'm dual booting my Mac Ultra M1, 64GB RAM.

So I compared the same computer running 18.6.6. versus 19b1.

18.6.6, the Text+ node reacts in real time to changes in font size for text that has 25,000 characters-and-spaces.

19beta1, the sText node takes around 18 seconds to do the same font size change.

This might be a mac issue. I just copied over 29,000 characters and even though I see that sText is a little slower, it's about one second slower.

I still don't quite understand why you wouldn't use the Text+ node and just animate the size in there.
Text+ is also vector based, so if you resize there you won't lose any resolution either.

I still don't quite understand why you wouldn't use the Text+ node and just animate the size in there. Text+ is also vector based, so if you resize there you won't lose any resolution either.

Because the animation simulates zooming in and out of a page of a book - zooming in to show the actual words, plus markings and highlighter colors.

Hence, with the sShapes method, the background paper, the text, the markings, the highlights, all maintain their spatial relationships when the zooming is done prior to the sRender node.

If I were to use the Text+, it's much harder to maintain the spatial orientation of the words on the 'paper', and other annotations. For example, if I simulate zooming into the page, by increasing font size, I have to find a way to synchronise the white background increasing in size to maintain the same ratio of the page-edge-margin.

Whereas, it's really easy using the sText/sBoolean method (as per my other questions on using the sBoolean, which you responded to on the other thread). Except that the sText node is slow on my macOS.