2.46 Use Translate in Video Sequencer - how the size works?

I am trying to use “Use Translate” in Video Sequencer, to position some image over the video strip. It mostly works (= I can change the image position) but…

… but for some reason after I enable “Use Translate”, my image is significantly scaled up (it is immediate: I see small image, click Use Translate, it turns big, click Use Translate to disable, it turns back small).

Is it possible to disable this effect somehow?

I used 1024x768 video clip and 1024x768 image (mostly alpha, with small true image).

PS In general: does there exist any documentation of new sequencer tools?

I spotted something more: the effect disappears if I do both:

  • set the output size to the size identical to the source
  • set the preview size to 100%

If any of those is not fullfilled (either I render 800x600 video from 1024x768 strip and images, or I preview in 50%) the image is enlarged.

It looks like in “Use Translate” mode for some reason the image is not scaled down in sync with everything.

Is it a bug? Should I report it somewhere?

docs are in the wiki, and on release notes page.

Dealing with different sized images and different sized outputs is tricky. Think like a pixel. If you have a mismatch in the VSE, the VSE does try to auto-scale the image to fit. Once you click Translate, you are telling Blender you want to manually re-scale, hence the jump in size.To manually re-scale, use Transform.