Jumping_exposure.
I took some higher resolution sequential jpeg image to York University the other day where one of the professors loaded them on to a computer in order to render them with a high definition codec.

While watching the playback (a while later) he noticed the changing dark noise in a girls shirt as she ran down the street and commented on the noise in the shadows.

Something didn’t seem right to me (and I had not paid that much attention to the sequence before). The shading seemed to differ too much from frame to frame to be caused by any shadows. The changing tone can be seen in the jpegs below.

Yesterday, being curious to the point where everything else comes to a halt. I isolated this sequence in the raw scan files and processed them a couple of ways.

reference_based_level - this sub gallery shows the sequence using a fixed reference as the exposure key for the scanned files,

image_based_level – this sub gallery shows the sequence deriving the output level from the image contents.

Both means of processing the images show that it was not shadows or processing that cause the tone of the images to change. As the cameraman ran with the girl the exposure-mechanism in the camera responded to the bounce by opening and shutting. You will notice frame 8mm_Nath001_01_002512.bmp is almost black whereas frame 8mm_Nath001_01_002514.jpg is bleached out.

In each of the frames you can see parts of the previous and following frame to verify the exposure is changing rapidly.

I learned something new today.