Thing-a-day.com

This was inspired by a short film created by Andre Silva. Watch the great short film here.

And its a line from a T.S. Elliot poem.

I’m in shop biz mode. This is part of my MySpace marketing images.

via mefi, an interesting and simple to make optical illusion.

Instructions: stare at the black dot, don’t look elsewhere. After 20 seconds the image will shift. If you don’t shift your gaze it will appear to be in colour, but as soon as you look somewhere else it goes back to black and white.

The instructions are very straightforward.

1. open the original colour image in photoshop and create a duplicate layer
2. select the bottom layer and do image –> adjustment –> desaturate
3. pick a 50% grey from the colour palette
4. select the top layer and do edit –> fill; use foreground colour, mode = luminosity; opacity = 100%
5. still on the top layer image –> adjustment –> invert
6. make a new layer and draw a black dot in the middle
7. save as jpg
8. delete bottom layer and save as a second jpg
9. either create an animated gif or follow instructions to make a mouseover effect

I did an even more fun standalone page where you can control how long you look at the negative image, then you mouseover to flip to the actual image.

Alright, so I started pixel-averaging on the digital camera noise signature sample material, but I’m pretty sure that my code is wrong. Actually, I am dead certain that my code is wrong. For some reason I can only sum the pixels, but not divide them to get an average. Any which way, I’ll continue trying to figure that out. Have I ever mentioned that I hate C++?

So, in short, these pictures really have very little to do with digital camera noise signature detection, but I thought they looked cool and I needed something for the day, so I posted them.

embedded by Embedded Video

Alright, with a little help from my friend I was able to get starting in converting the image sequence of black images into high-contrast grayscale images. Now starts the daunting task of figuring out how to figure out what the digital camera noise signature is. And once that I think I have done this, I then need determine a way to prove it. I have a few ideas of how to start, but I still need to give it a little more thought.

If you watch the animation above long enough you can start to see some general patterns start to emerge (or you would if the compression didn’t suck). Could this pattern simply be the noise signature? Also, the video above may be too low-res to notice, but there are a couple colorful pixels almost stationary towards the bottom right hand corner of the screen. At first I thought this might be a blemish on the camera, but I feel as though a blemish would block light and make those pixels black, not bright and colorful (the brightness should be an indication of light and the color indicates a noticeable variation in the RGB values… which logically should be an indication of light). Anyway, what could that be about?