Status: 

OS version: 

  • Mac OS 10.11

Steps causing the bug to occur: 

  1. Create a text layer with revealing characters (cut text increasing)
  2. Align it for example left to the window
  3. Notice how as characters are revealed, the layer moves (see joined video & composition in the .zip)

Could this have anything to do with pixel alignement ? This video was captured on a non-retina screen.

Which lead me to question 2, could or should I create a feature request (would it make sense) to optionally enable / disable pixel alignment for "Make Text Layer"?
Because for scrolling very big text sizes, Jaymie's fix to use Make Text Image instead, result in huge image sizes (can even outsize the max image width) (see topic Simple text scrolling is not consistent or smooth)

I wonder though for what hardware range / year this request would be useful for though, since Kewl on that topic noticed the composition I shared there ran smoothly on his hardware

Compositions: 

AttachmentSize
Package icon Aligned Text Moving.zip482.94 KB

Comments

I tried your composition with

Kewl's picture
Submitted by

I tried your composition with Vuo 2 and I'm seeing the horizontal displacement, but to a lesser degree than in your video. I have a Retina screen. The higher pixel density must explain why the horizontal shift is more subtle.

MacBook Pro (Retina, 15-inch, Late 2013), NVIDIA GeForce GT 750M 2 Go, Intel Iris Pro 1536 Mo, macOS 10.14.6.

[Kewl] I tried your

jstrecker's picture
Submitted by
Status:
Waiting for review by Vuo Support Team
»
Accepted

[Kewl] I tried your composition with Vuo 2 and I'm seeing the horizontal displacement, but to a lesser degree than in your video

Same here.

[Bodysoulspirit] could or should I create a feature request

No need. This is something we can/should fix within the text rendering code.

[Bodysoulspirit] Because for scrolling very big text sizes, Jaymie's fix to use Make Text Image instead, result in huge image sizes

In the long run, the best solution would be a 2D vector paths node set (similar to Kineme2D).

No need. This is something we

Bodysoulspirit's picture
Submitted by

No need. This is something we can/should fix within the text rendering code.

Great ;) But would a pixel alignement option, like on the image feedback node make sense for make text nodes though ?

In the long run, the best solution would be a 2D vector paths node set (similar to Kineme2D)

I see. I've been wondering to ask if a feature request for pure text nodes would make sense (render text to window). I don't know the technical details behind the whole thing, but I mean not text on an image, layer or 3D object, but text, rendered with some text "engine", like in a text editor, that you'd even be able to select & copy (maybe related to the "text input" sub-part of the User interface (UI) node set request, but that would probably require to be able to connect several different data types to the same Render X to X nodes (otherwise you'd have to convert it anyway)). But since text are vectors, maybe 2D vector paths support is somehow / is the same ;)