Perspective adjustment on zoom out

[LumenLux]

I don't think this is a bug. I'm not sure if it is the way it should be. I know I don't fully understand the perspective adjustment for zoom in the O&A window in V5. But here is an observation. The attached project has one slide. The photo zooms out and then repeats per the "repeat show" option.

1.If the "perspective correction for zoom" box is checked, the zoom-out goes all the way out, as intended. The next cycle of the zoom starts while the zoomed out image is at it's widest intended angle.

2. If the "perspective correction for zoom" box is not checked, the zoom-out cycle begins again before the first zoom-out reaches it's widest angle.

You can notice the difference clearly by a person on the left that appears in case (1) above, but never appears in case (2)above. With the O&A window open, if I use preview there, I can see that in the specified time for the slide, the zoom-out does not get as far "back" during the allotted time for the slide. Is this the way it "should be"? I guess my concern is, that if we use the check box to turn "perspective correction" on or off - we are also changing the amount of zoom if we leave the slide duration the same. If instead, we change the slide duration to result in identical focal lengths at zoom conclusions, are we really changing the "perspective" or not? Or - can someone make this more understandable?

OceanPerspect.zip

[sanewcomb]

Hi LumenLux,

The behavior you see is a consequence of placing the last keypoint outside of the slide's time duration.

1.If the "perspective correction for zoom" box is checked, the zoom-out goes all the way out, as intended. The next cycle of the zoom starts while the zoomed out image is at it's widest intended angle.

Actually, the full zoom out is 112% and would occur if the slide duration was set to the same as the last keypoint, 31584 msec. What you see is a zoom at 132% which occurs at a time of 25000 msec.

2. If the "perspective correction for zoom" box is not checked, the zoom-out cycle begins again before the first zoom-out reaches it's widest angle.

In this case, the zoom at the slide's end is 182%.

The reason is because the function used to interpolate the zoom from start value to end value is different for each case. So when you place a new keypoint (in this case it is the end of the slide, rather than an actual keypoint) in between the start and end the zoom% value will be different. So you see a different zoom% when the slide ends.

Looking at this I discovered a potential problem, or rather something to be aware of when changing an object's zoom function and using more than two keypoints (start and end). You may not end up with a smooth zoom if you add a 3rd,4th, etc keypoints to an object AND change the Perspective Correction for Zoom checkbox back and forth while doing this. The reason is the interpolation for the zoom is only done between two existing keypoints when you add a 3rd keypoint. It can be avoided if you remember to add additional keypoints in a consistant way. For example, always add new keypoints with the Perspective checked off.

I guess my concern is, that if we use the check box to turn "perspective correction" on or off - we are also changing the amount of zoom if we leave the slide duration the same.

Only if you place keypoints outside of the slide's duration. In general I would avoid doing this. Since you only see the slide and effects on it during its time duration, placing keypoints outside of this time creates the potential for unexpected or uninteded results. This applies to effects other than zoom as well.

Steve Newcomb

Tucson, AZ USA