Perspective Correction for Zoom

[JEB]

Hi Folk,

I have been playing about with this over some long slow zooms and can't say that I notice that it does anything.

I have used various custom and linear zooms, separated and glued and notice no difference.

What am I missing and perhaps more importantly what should I be looking for??

Any suggestions as to how I could create a before and after illustration would be appreciated.

I am using a relatively high res image of an ornament though for the project I have in mind I would be using a much higher res image.

Regards

John

[stardealer]

Hi John,

When using long slow zooms, the difference is nearly unrecognizable. So to see the difference I would advice to test it with a (smaller) image and fast zooms.

Then you can see that linear starts harsh and goes linear to the end and then stops instantly. The other options start with a smooth acceleration and/or end with smooth deceleration.

Seperated and glued only makes a difference when not using linear. At every not glued section the motion comes to an end (smooth or instantly as selected) and motion then starts again as defined.

Regards, Frank

[stardealer]

Sorry for not responding to perspective correction.

Imagine you are in a cathedral with evenly spaced columns on the side.

Now you look straight forward and walk with linear speed. The perspective changes with every step you take and the columns you pass leave your field of view at even timelaps.

In a picture of a cathedral the perspective is static and so to create the imagination of natural walk (zoom) needs a non linear speeddesign of zooming. Checking "perspective correction for zoom" forces P2E to transform the linear zoom automatically to a non-linear zoom calculated based on starting and ending zoom factors. (Anyone should correct me if I'm wrong).

This effect can best be seen by arranging the same picture side by side with different settings for perspective zoom.

Regards, Frank

[JEB]

Thanks Frank,

Yes, I understand and am aware of linear and custom zoom, it is the Perspective correction I am trying to get my head round.

I think I have an image similar to your suggestion so will dig it out and have a play.

Thanks in anticipation.

John

[sanewcomb]

Yes, I understand and am aware of linear and custom zoom, it is the Perspective correction I am trying to get my head round.

There is actually quite a bit of difference on long zooms between the two algorithms. By long I mean the starting and ending zoom percentages are significantly different. Try 100% to 500% to highlight the difference, which occurs primarily at the very beginning and ending of the zooms.

Almost all programs with a zoom function, including P2E without checking the Perspective Correction box, perform a simple calculation to change the PERCENT of zoom in a linear way. Said another way, in 10% of the time the zoom percent is 10%, 20% of time zoom percent is 20%, etc. While this is simple to program, it is not what happens when you walk through space or move an actual camera on a stage with a constant (linear) speed towards a photo, as Ken Burns did in his documentaries (though he didn't always use linear zooms). This simple algorithm actually gets the real world physics backwards.

With Perspective Correction box checked, P2E uses a more sophisticated algorithm to simulate what it would look like if you were moving a real camera towards a printed photo at a constant speed.

In terms of what you see on the screen

Without Perspective Correction, the closer you get to the image, the slower the rate of approach. The further you get from the image, the faster the rate of retreat. If you want the image to slow down as it zooms in, use this setting. Most useful when fading from one picture to another. Custom zoom works better if you want to come to a stop and display the picture for more time or pan across.

With Perspective Correction, the closer you get to the image, the faster the rate of approach appears (approach is approximately constant, but our perspective makes it seem like it is moving faster). The further you get from the image, the slower the rate of retreat appears. If you want the image to linger as it pulls away from the subject, use this setting.

This difference doesn't carry over that well once you start playing around with the non-linear or custom zoom settings, which effectively add additional calculations over one another. Except I will say I get a more predictable (and thus satisfying to me) result with custom zooms if the check box is checked with my work.

Steve Newcomb

Tucson, AZ USA

[JEB]

Steve, Frank,

Thank you both for your responses. I feel I have got some sort of handle on this now and am due you some sort of feed back.

I took your advice and used a 500% zoom over four 20 second slides.

1. Perspective/Linear

2. No Perspective/Linear

3. No Perspective/Custom/Smooth

4. Perspective/Custom/Smooth

3. would probably be my normal first choice and will, as you suggest Steve, remain so when stopping starting and panning.

1. will now become my preferred starting point for a standard slow zoom as it simulates the characteristics of 3.

4. this was interesting and in this instance quite attractive in that it gave an extremely slow take off and gentle stop.

Horses for courses I suppose.

Thanks again

John

[Lin Evans]

Hi John,

Just a suggestion. Do all your animations, zooms, etc., in linear and be certain you have the essence of your show completed before switching to any non-linear mode. The reason is that when you make changes to keyframes and you have a non-linear function enabled things don't always work as you might expect. Also if you have very sharp images, "smooth" will exacerbate any edge distortions such as normal resize artifacts and cause excessive "shimmer" at the start and end of an animation. This is because there are actually more images created during the slow periods of non-linear smooth start and finish. The way around this is to apply about a 1% gaussian blur to frames or hard, straight line edges selectively to avoid these rather unsightly artifacts. Smooth really works well and is one of my favorites because it avoids the jerky start and finish to a zoom or pan but it does have the side effect of the shimmer of really sharp areas of an image, especially if the zoom is slow.

Best regards,

Lin