Bizarre result after deleting a KF

terrypin wrote on 11/27/2013, 9:38 AM

This is an obscure one. I set several KFs on an image to give a smooth zoom, with pauses at start and end. As sometimes happens an extra KF got created somehow, just to the right of the last one, with almost identical values. So I simply deleted it. But that completely changed all the previous KF values!

I lengthened the image in case there were hidden KFs on the right, but there was nothing.

Here are a few screenshots in case the above needs clarification:

Or for better quality:https://dl.dropboxusercontent.com/u/4019461/MEP2014-KF1.jpg

Or for better quality: https://dl.dropboxusercontent.com/u/4019461/MEP2014-KF2.jpg

Or for better quality: https://dl.dropboxusercontent.com/u/4019461/MEP2014-KF3.jpg

Anyone have any idea what might be going on there? It remained the same after a restart of MEP 2014 so it's not just spurious behaviour.

--
Terry, East Grinstead, UK

Comments

cpc000cpc wrote on 11/27/2013, 3:55 PM

Terry,

Started with two thoughts: First check if you have the option to link the keyframes to the length of the clip selected (or not) and try the other way. Second, don't worry about the extra KF.

Then I did an overlay of your images:

And my third thought is that there isn't any difference AFAIKS other than the last FK being eliminated. The rest are in the same positions?

ADDED: Ah... I was only looking at the positions of the KFs even though you clearly were concerned about their values.

Was very confused,

Carl

terrypin wrote on 11/27/2013, 4:30 PM

Carl,

All the values of the remaining KFs changed, as shown in the third screenshot.

I'll get back on the case in the morning.

Terry, East Grinstead, UK

johnebaker wrote on 11/27/2013, 4:31 PM

Hi Terry

When setting the zoom, are you using the slider?

If so I have found that when you intially click on it to activate the keyframe editor, then move the timeline cursor so you can drop a keyframe onto the object, it does create a keyframe at the initial position of the cursor.

This appears to be caused by a slight movement of the mouse while clicking the slider, hence the value close to the one you deliberately set and the change of all the other keyframe values when you delete it because it is the initial keyframe which appears to set the initial state of the object.

You will also find this can happen - and has done in every version from MEP 15 - when you use other effects with keyframes.

I use a tracker ball rather then a standard mouse so I can let go of the ball during clicking operations without moving anything.

HTH

John

terrypin wrote on 11/28/2013, 1:19 AM

Hi John,

Thanks, I'll see if I can reproduce that spurious creation. What I'm having trouble gettng my mind around is how deleting that last KF in my example can change the values (x, y, etc) assigned to the earlier KFs.

--
Terry, East Grinstead, UK

johnebaker wrote on 11/28/2013, 2:02 PM

Hi Terry

. . . how deleting that last KF in my example can change the values (x, y, etc) assigned to the earlier KFs . . .

A hypothesis for experimental proving/disproving:

The first keyframe size/position set is the reference point wrt to the screen for all subsequent KF's.

If the KF size/positions are relative to the reference point then deleting the first KF set would make the second KF set become the reference point , and the values of all subsequent KF's would change and so on if you delete the second KF set.

Rational for the hypothesis is that relative positioning is easier to implement for, for example, down sizing from HD to DVD to keep everything as close to the original HD as possible and the warning message that objects may not appear exactly in relation to each other when changing from eg 1920 * 1080 to 720 * 576 keeping the same AR.

John

terrypin wrote on 11/29/2013, 2:25 AM

Hi John,

Many thanks. I've put this on the back burner until I finish my current project, which has been progressing much slower than it needed to!

Then i'm going to try to overcome the main obstacle with this sort of issue, which is getting it consistently reproducible. I did try briefly but failed.

This was basically just a 'delayed pan' from left to right and the first KF was simply the leftmost position of a panoramic JPG. It panned smoothly as required - until that last KF was deleted, then went haywire as you see from my third screenshot.

Will report back here later. (May be in 2014 as I'll be away from MEP from 12th - 29th December, surviving with iPad communications in USA!)

--
Terry, East Grinstead, UK