Zoom: seams to accelerate

[Esben]

Why is that? What am I doing wrong?

I start at 300% and within 20 sec. it is going down to 100% screen fit (2 keypoints). It starts very slowly but speeds up at the end (seams at least). I wish it went even.

Any hints?

Thanx,

Sven

[ronwil]

In a thread started by Lin on 1 May headed "Beta 5 Tutorial, "Beta" of tutoral", your question cropped up then on 2 & 3 May and I quote:

LumenLux on 2 May - "I'm wondering if you have noticed one characteristic of the zoom in and out feature. My observation is that (at least in some instances) a long duration zoom makes more "progress" in the first few seconds than in the last few seconds. That is the zoom rate slows as the duration increases? Maybe this is just my perception but I am curious what you or others or Igor might be able to explain."

Admin on 3 May - "It's only visual effect. Currently Pan/Zoom are linear with constant speed."

ADB on 3 May - "I think the zoom speed is an optical illusion if you start zoomed in and zoom out for example objects initially appear closer and feel like they are moving faster but as they get smaller the visual rate of change is less so the zoom appears slower and vice versa if you start zoomed out then zoom in.I have seen the non-linear zooms in some other programs which speed up initially then slow down and vice versa depending on where you start from, I think Igor will implementing this in future."

I hope this helps.

Ron [uK]

[Esben]

Exactly, I thought so too. However, visual effect or not, Reality is how it seams in our perception, isn't ist ;-) So my question is, is there a way to compensate that by a NON linear Zoom effects?

If not, it is fine as it is, just wanna make sure I didnt miss anything. No rude ask more and more Igor ;-)

In a thread started by Lin on 1 May headed "Beta 5 Tutorial, "Beta" of tutoral", your question cropped up then on 2 & 3 May and I quote:

LumenLux on 2 May - "I'm wondering if you have noticed one characteristic of the zoom in and out feature. My observation is that (at least in some instances) a long duration zoom makes more "progress" in the first few seconds than in the last few seconds. That is the zoom rate slows as the duration increases? Maybe this is just my perception but I am curious what you or others or Igor might be able to explain."

Admin on 3 May - "It's only visual effect. Currently Pan/Zoom are linear with constant speed."

ADB on 3 May - "I think the zoom speed is an optical illusion if you start zoomed in and zoom out for example objects initially appear closer and feel like they are moving faster but as they get smaller the visual rate of change is less so the zoom appears slower and vice versa if you start zoomed out then zoom in.I have seen the non-linear zooms in some other programs which speed up initially then slow down and vice versa depending on where you start from, I think Igor will implementing this in future."

I hope this helps.

Ron [uK]

[thedom]

An other explanation might be that you moved the "center" of the object on one of the keypoint.

That can increase the impression of non regular zoom...

Could you post your project please ?

[Lin Evans]

Exactly, I thought so too. However, visual effect or not, Reality is how it seams in our perception, isn't ist ;-) So my question is, is there a way to compensate that by a NON linear Zoom effects?

If not, it is fine as it is, just wanna make sure I didnt miss anything. No rude ask more and more Igor ;-)

At present the zoom is linear, but I believe Igor will implement a non linear option eventually. To get around the perception of non linear (even though it is absolutely linear at present) you can insert additional keypoints and change the amount of zoom between them keeping the overall percentage identical but varying the percentage which happens between the keypoints.

For example, lets say you wanted to slow down the last three seconds of a twenty second zoom. You insert a keypoint at 17 seconds and change the percentage of zoom at that point to a different value than it shows as a linear zoom. Then the rest of the zoom between the second and last keypoints will happen either faster or slower relative to the value you choose for the intermediary keypoint.

Lin

[thedom]

At present the zoom is linear, but I believe Igor will implement a non linear option eventually.

He definitely will :

We'll add speeds (Linear, Smooth, etc) to Pan/Zoom/Rotate effects.

[Conflow]

Esben

That 'phenomenon' you are experiencing is quite natural to the Human Visual senses and is also natural in the Human Aural senses. All Human senses obey a Logarithmic Law and the opposite Anti-Log Law.

Simple examples are:-

"Cablecar" When descending,the ground seems to be rushing up to meet you and when ascending the oppsite is true, although you have the same velocity in both directions.

"Amplifier"

Amplifier Volume Controls are Logarithmic because the Human Ear obeys a Log-Law what's called 'A-Log'

"Doppler Effect"

Both Visual & Aural senses are responsive to 'doppler-effect'. White-Light coming towards you has a different wavelenght than departing White-Light. Sound is similar,approaching Sounds have a higher pitch than departing Sounds and so on and on...

It's all part of being Human ~ unfortunately to try and correct this Visual phenomenon to a Linear characteristic is electronically very complex in the extreme ~ 90% of the time it simply doesn't work because the 'senses' realise there is something phony about the Linearisation.

The simple way around this:- Don't Zoom so quick, try and keep Zoom's above 400millesecs execution time.

Hope this helps somewhat, and before we go down this road, let Igor finish the Program before we ask for the 'Salt & Pepper'...

Brian.Conflow.

[JohnFeg]

Esben

That 'phenomenon' you are experiencing is quite natural to the Human Visual senses and is also natural in the Human Aural senses. All Human senses obey a Logarithmic Law and the opposite Anti-Log Law.

Brian.Conflow.

I’m creating a particular zoom effect and have come up against the buffers of ignorance! Principally it’s because, after all these years, I cannot remember how to do the maths.

The idea is to use a high definition map (of the Aegean) onto which has been superimposed an even higher definition aerial photograph of the island of Ikaria.

I find that I can zoom in from 100% to 30,000% with good results. For the last part of the zoom, because of definition constraints, the map component is at 0% opacity, it’s place being taken by a “blue sea” background.

Now, as the zoom is linear the impression is that the rate of descent constantly decelerates. I have worked out a couple of (obvious!) points to make changes to the rate of zoom so that the apparent speed of descent is about the same at the beginning and at the end. i.e.:

a) 100 – 700% zoom is achieved in 6.0 seconds

700 -1400% zoom is achieved in 3.0 seconds

c) 1400 – 2100% zoom is achieved in 1.5 seconds

This works, after a fashion, but, there is at each of the points of change an obvious change in the rate of zoom. Therefore, I want to put in more change points to try to smooth things out a bit.

For the life of me, I cannot work out at what time I should set the changes for 200%, 300%, 400% zoom etc.

Can anyone help? Or, tell me if I am going about things in the wrong way?

John

[Lin Evans]

I’m creating a particular zoom effect and have come up against the buffers of ignorance! Principally it’s because, after all these years, I cannot remember how to do the maths.

The idea is to use a high definition map (of the Aegean) onto which has been superimposed an even higher definition aerial photograph of the island of Ikaria.

I find that I can zoom in from 100% to 30,000% with good results. For the last part of the zoom, because of definition constraints, the map component is at 0% opacity, it’s place being taken by a “blue sea” background.

Now, as the zoom is linear the impression is that the rate of descent constantly decelerates. I have worked out a couple of (obvious!) points to make changes to the rate of zoom so that the apparent speed of descent is about the same at the beginning and at the end. i.e.:

a) 100 – 700% zoom is achieved in 6.0 seconds

700 -1400% zoom is achieved in 3.0 seconds

c) 1400 – 2100% zoom is achieved in 1.5 seconds

This works, after a fashion, but, there is at each of the points of change an obvious change in the rate of zoom. Therefore, I want to put in more change points to try to smooth things out a bit.

For the life of me, I cannot work out at what time I should set the changes for 200%, 300%, 400% zoom etc.

Can anyone help? Or, tell me if I am going about things in the wrong way?

John

Hi John,

So you do not end up fighting multiple issues remember that whenever you insert a keypoint, the zoom stays with it even if you subsequently move the keypoint. For a test, take a zoom in from 100% to 400%. Set the overall time with only two keypoints. The first defaults to zero so then set the time to 10 seconds for the entire zoom and observe what happens. You have a rather nice linear zoom. Now insert a keypoint at about 1 second then slide the keypoint to 9 seconds and watch again. Strange?? Yes - the actual zoom percentage stays with the insert point at about 1 second so that you end up with an extremely slow zoom to 9 seconds followed by an extremly fast zoom for the last second.

In order to keep your zoom linear or to compensate for the perception of non-linear zoom you need to know the degree of total zoom as a percentage then divide this total zoom percentage by the number of increments you intend to use to know what "should" be at any particular part.

For example, in the example above you have 400% size increase over 10 seconds. So 400/10 - 40 or 40% increase for each 1 second of time. So at 1 second we would have 140%, at two seconds 180%, three seconds at 220%, etc. So at 9 seconds we would have 360 percent (9x40). This is what you would need presently to calculate how to keep the zoom percentage linear in "reality". Now to compensate for perception.

So to figure out your percentages for each point you must make some estimate at to the "perceived" rate of zoom at the start and end then factor in the percent of change necessary to achieve what "appears" to be linear to compensate for the human perceptual difference. I can't give you exact figues because I don't know how the perception of zoom differs from the reality of zoom, but perhaps you could make an estimate based on "perception" then alter the true linear zoom figures for any given point in some compensatory fashion based on perception to arrive at what appears to be linear to you. Best to make the compensation spread over more rather than fewer increments to minimize the perceptual changes at any given point.

Best regards,

Lin

[Conflow]

I’m creating a particular zoom effect and have come up against the buffers of ignorance! Principally it’s because, after all these years, I cannot remember how to do the maths.

The idea is to use a high definition map (of the Aegean) onto which has been superimposed an even higher definition aerial photograph of the island of Ikaria.

I find that I can zoom in from 100% to 30,000% with good results. For the last part of the zoom, because of definition constraints, the map component is at 0% opacity, it’s place being taken by a “blue sea” background.

Now, as the zoom is linear the impression is that the rate of descent constantly decelerates. I have worked out a couple of (obvious!) points to make changes to the rate of zoom so that the apparent speed of descent is about the same at the beginning and at the end. i.e.:

a) 100 – 700% zoom is achieved in 6.0 seconds

700 -1400% zoom is achieved in 3.0 seconds

c) 1400 – 2100% zoom is achieved in 1.5 seconds

This works, after a fashion, but, there is at each of the points of change an obvious change in the rate of zoom. Therefore, I want to put in more change points to try to smooth things out a bit.

For the life of me, I cannot work out at what time I should set the changes for 200%, 300%, 400% zoom etc.

Can anyone help? Or, tell me if I am going about things in the wrong way?

John

Hi John,

Yes you are quite right in your Mathematical approach and as you are using 'Arial Photography Shots' I would suggest to use the G-Force Curve (Gravity Curve) which starts off slowly at 32ft/sec/sec and accellerates very fast and slows as finite terminal velocity approaches some 120mph. (Old Engineer)

You would need to "fix" a minimum of 7 points on this curve to give effect to those perceived phenomenon experienced by the human body in 'free-fall'. If you use the correct algorithm you will fool the human eye and what an effect that will create...You should try a 'Roller-Coaster' as your next Project, and that would be some Show.

Get on to www.answers.com for your Maths, here's a Link:-

http://www.answers.com/main/

Brian.Conflow.