selective motion blur after many attempts

[bluphoton]

After a thinking about it for somedays, i got an opportunity to sit next to my riding friend in my car. And try to selectively blur a single section of it. After sufficient trials, this "Image one" picture demonstrates blurring to the left of the subject and right also front and behind. Haven't edited these pictures. I am trying to elucidate it all in numbers, will post it shortly. :)

http://i65.tinypic.com/2lcxmrp.jpg Image one = f/9 1/50 seconds ISO 100. 18 mm Canon 600D. The car was about 80 kmph

http://i63.tinypic.com/10yqvd0.jpg Image two = f/9 1/50 seconds ISO 100. 18 mm Canon 600D. The car was about 80 kmph

I am student of physics and these kind of experiments give me joy, would be happy if some one have sources to such experiments and explanations. Please post them. Sorry If i had to post in in a more appropriate thread, newbie, this is my first post. :D

[Manfred M]

First of all, welcome to CiC.

Interesting perspective, Tikak. In most cases of this type of work, the subject is moving and the camera is stationary. this is of course a complete reversal of that I I find that it does give you a rather unusual result. I've also seen some work (in a gallery) shot from the car where there was a long exposure and the image was just a blur and these were also quite effective from an artistic standpoint.

As for the physics, I'm not sure where you are coming from? What you are doing is experimental photography and I don't know of too much out there specifically on this topic. I would suggest that the images would look better if you tried taking them in more interesting light (just around sunrise or sunset, during "golden hour" or "blue hour".

There are some interesting techniques where flash is combined with motion (this is often referred to as "dragging the shutter"), where you get the flash to freeze part of the image and allow motion blur to be recorded from the ambient light hitting the subject. Another take on this type of photography is to zoom in or zoom out while the shutter is open for up to a few seconds.

Happy experimenting!

[bluphoton]

Hello Manfredm,

Thank you. I am excited to try something with flash as you mentioned now. Thanks again.

[Manfred M]

Here is an example of the zoom technique I wrote about:






This shot uses flash and ambient light with a slow shutter speed.

[Shadowman]

Tilak,

Nice examples, has a bit of a 3D effect as well. I'm wondering if yours and Manfred examples are more suiting for electronic media rather than still images, they work as shown but could have additional possibilities.

[Manfred M]

John - I have no idea why you would say that. I have seen examples of this type of work for sale in prints at galleries.

[bluphoton]

Hi,
Thanks john. Manfred, in that car example.. so essentially the frequency of the flash, is like another localized shutter speed for the car itself ! wow so cool ! an image that sort of has two shutter speeds, crazy ! [i know not really, but can safely say that]

Thank you.

[bluphoton]

Argument:

1. Since it was a passing click i do not know how far the subject was = s
2. Or its real dimensions
3. I have assumed that the camera was rotated in retrograde in a uniform circular motion
4. Assumed that the subjects were in motion relative to the camera, although it was the other way that the shot was developed
5. Any non-circular motion of the camera, [in the direction parallel to the motion] can be subtracted from the velocity of the car. [in perpendicular] will give rise to a curve for the movement of the subject, causing inaccuracy in the angular velocity calculation.
6. Have used a Canon 600D/T3i and 18-55 lens, @ 1/50, f/9 and ISO 100
7. The camera is known to have circle of confusion = 0.019 [refer:http://www.dofmaster.com/digital_coc.html]
8. Hyperfocal distance at 18 mm and f/9 can be calculated to be = 1.9127 m
9. Since the picture demonstrates that there are a layer of off-focus objects behind the focused ones, we can safely assume that s < 1.9127 m
10.Depth of field; D= (2s - 36.34)[t(36+t-s)(t+s)], t = 1876.73 [refer: http://www.dofmaster.com/equations.html]
11.Assuming the car was at 80kmph, the 1/50 seconds of exposure = 44.44 cm of trail [this is really surprising, to have got a still at a region despite such speeds]
12.Circular rotation of the camera must have happened at an angular velocity= 100 arcsin (1/45s) [refer: https://en.wikipedia.org/wiki/Circular_segment]

Have tried to depict it all using this diagram, but i must admit, it is way out of scale :P Needed the value of 's' very crucially any ways to approximate it ? or Must think of a controlled experiment to analyze this photography better ! when i get the time and support for it, I welcome more technical insights, corrections, resource-sharing and discussions. Thank you.

[Manfred M]

You are definitely overthinking this.

The one comment I would make is that shooting from a car that is traveling at 80km/hr is certainly going to be much more difficult than doing this from a slower moving vehicle. I got back from India yesterday (Bangalore, in fact) and found that there were very few roads that we drove that fast on, or for that matter would have wanted to drive that fast on.

You'd have an identical effect driving at 40 km/hr and shooting at 1/25th sec, 20 km/hr at 1/12th sec, etc. Slower speeds will mean more accurate targeting of the subject. Both of the shots I showed in this thread were taken with the camera on a tripod. I have a car window mount that I can use to shoot from a moving vehicle (I've only done this with video so far, but do want to try things with a camera at some point).

Also be careful when looking at the circle of confusion and hyperfocal distances as both assume a particular degree of enlargement for the image (usually A4 size). Frankly, with this type of shot, I wouldn't get too worried about loss of sharpness from diffraction; you are going to get a more pronounced effect on sharpness due to motion blur.

[Shadowman]

John - I have no idea why you would say that. I have seen examples of this type of work for sale in prints at galleries.

Manfred,

As I said it works, but the 3D effect could be enhanced if to virtual levels with a bit of digital movement.

[TonyW]

Hi Tilak, I like your second picture - to my mind it works very nicely. It is a little bit like a standard panning shot except that the background is moving at the same speed relative to the camera as the main subject, which I take to be the tree in this instance.

I would have thought that it would be more convenient in your analysis to make the tree the central fixed point of your coordinate system since that is the fixed point in your image. Then during the 1/50 of a second of your shot, the camera rotates anti-clockwise, say, about that point and the other points of interest rotate by the same angle, given that the camera is always pointing towards the tree. The hill in the far distance then moves a large amount and the points near the tree by a smaller amount, as seen in your image.

[george013]

Tilak,

Your photos are the result of a 3-way movement during exposure.
1) the car is driving, the x-ax
2) you keep while driving the main subject on the same place in your view finder, the y-ax
3) the camera was rotated a little doing that movement, the z-ax
The effect can be declared with some gonio during exposure.

I tried to understand what you wrote using the diagram, but I don't see the connection between your statements in relevance to the pictue. Individual they might be true, I don't get the overall idea you want to tell.

The depth of field is not a difference in distance from your camera as drawn in your diagram. It's the diffence in distance between 2 planes parallel to your sensor, something like the lines b and d.

I do like the effect in the pictures.

George

[dem]

Hi Tilak. When you draw depth of field, you need to use parallel lines not concentric circles. If you draw a plane perpendicular to the camera-subject line, everything on this plane will have the same sharpness. Using the camera-subject distance as a radius of a circle is not a good idea.

In practice the focus plane (note it is called "plane" not "sphere" and it goes through the subject, do not mix it with "focal" which is "inside the lens") is not a perfect plane due to the lens not being ideal but it is still has a very small curvature.

The most likely reason why the tree trunk is sharp and the brunches are blurred is the rotation of the camera. Did you turn the camera to keep the focus on the tree trunk?

In statement #9 you seem to assume that the blur is related to shallow DOF which leads to a conclusion that the camera-subject distance was less than 1.9 meters. I bet it was more than that and the blur you see is entirely to do with motion.