Whilst waiting for this year to end I have build another little Flash widget that uses the Flickr API: Clockr is a clock that displays the current time by using images from Flickr that show single digits. If you want to contribute more numbers to the pool use the tag “onedigit” but please follow the rules of that group (just upload real photos for example).

If you don’t like a number just click on it so it will switch to another one from the huge pool - as more and more numbers are loaded the available selection gets bigger the longer you wait.

For your post-holiday, pre-holiday or in-holiday enjoyment, here is a desktop image for you of a new tree-type that I rendered over the holidays. Enjoy!

1024 x 768

1280 x 854

The holidays have been a bit… trying… but I’m back on track for the New Year. There’s one more desktop image to go, which I’ll probably post later today or tomorrow.

For your post-holiday, pre-holiday or in-holiday enjoyment, here is a desktop image for you of a new tree-type that I rendered over the holidays. Enjoy!

1024 x 768
1280 x 854

The holidays have been a bit… trying… but I’m back on track for the New Year. There’s one more desktop image to go, which I’ll probably post later today or tomorrow.

I finally had some time to play around with the Flickr API and created Flickeur. This piece retrieves random images from Flickr.com and creates a stream-of-consciousness type of video clip out of them. By adding image layer upon image layer it creates an endless loop which is in constant change. Sometimes this principle generates the strangest stories. Due to the random nature of the piece you may have to be patient a bit until something starts to happen.

I have used Kelvin Luck’s Flashr wrapper which makes all the communcation with the Flickr API very easy. Unfortunately there is a real problem with the way the security sandbox in Flash handles the draw() command: whilst you can load and display images from other servers you cannot draw() them to a bitmapData object. In order for that to work the loaded clip has to issue a System.security.allowDomain(myDomain) which of course only swfs can do - not JPEGS. This means you cannot do any fancy bitmap operations with these images. The only workaround I have found until now is to use a proxy on my own server which fetches the images from flickr. Of course that is not satisfying at all as now all the traffic is adding to my server bill.

No this isn’t another post about The Castle. I just wanted to set the record clear about the K in CMYK.

The often-collaborator/sometimes-student/always-fun Pierre-Erick Lefebvre just asked me to look at his blog, and his new CMYK-controller + Psycho Chicken game. I can’t wait to get my mits on it, but he says something in there about the K in CMYK being “Kholâ€? (whatever), and that’s where the evil-professor split personality kicks in and reminds anyone who will listen and everyone else what that damn K is for in the first place.

The Wikipedia CMYK entry wasn’t so helpful about the origins of the K, but at least it got it right: K stands for “Keyâ€?. So I rooted around my study and found the following passage that works pretty good for me. It’s from Getting it Right in Print, one of those how-to books that occasionally come in handy when working with the printer. From page 31:

Most graphic designers know that the inks used in four-color process printing are cyan, magenta, yellow and black (CMYK). While it is reasonable to abbreviate cyan, magenta and yellow to C, M, and Y, how come black is given the letter K? Here is how I found out…

Towards the end of my very first day working in the printing industry, one of the printers told me that he needed a new “key plate� first thing the next morning. I had no idea what he meant and thought he was joking. When I eventually arrived an hour after him the following day, he was not happy, to say the least.

This was not good. When the presses stop, the shop is no longer printing money. A new key plate, I was rapidly informed, is a plate that prints the color that all the other colors key to: i.e. black. If you think about it, it is obvious. Text and image borders are typically printed in black. Printing them first often makes it easier to position — or “key� — the other colors to the job. So “key� is actually what the “K� stands for.

It is a common misconception that black is assigned the letter “K� because if it was called “B� it could be confused with blue. While plausible, this is not the case.

So there, K stands for “keyâ€?. Now that that’s out of the way, the pedantic academic can go back into hibernation.

I’ve been having a lot of fun playing around with the GD library off and on for the past couple of days. For those that don’t know, GD is the standard image treatment library that comes with most installed distributions of PHP. Ultimately, I’m trying find a decent background image for the abstractmachine homepage, but so far have yet to find anything that catches my fancy. So as usual, when in doubt, make it random!

Coding images on the server probably isn’t the best workflow, but it actually works out fine for me. It also allows me to mashup all my data on the server directly without having to later change anything. My FTP program and BBEdit make the whole process transparent: I just write the code in BBEdit, save, and refresh the page in one of the various browsers to see the results.

GD isn’t Processing, not by a longshot. But it’s actually quite pleasant to play with basic image building blocks, and I do so love programming in PHP. Variables and arrays are so easy, although I still don’t like the whole dollar $ign stuff. I left that world back in 1991. I’m a €uro man now. (ok, bad joke)

Two updates in one day? Well, whilst playing around with the class today I just had a few more ideas how to make it more useful, so I quickly implemented them:

I’ve added a blend() method which allows to blend between two ColorMatrices - this makes color tweening very easy because you just have to tween one value - the amount amount of blending.

Additionally I’ve added a “filter” property which returns a ColorMatrixFilter object initialized with the matrix values - this saves you one step.

Furthermore there’s now a randomize() and a clone() method.

Here is an example which demonstrates the tweening between two ColorMatrices (something that Grant Skinner demonstrated so elegantly with his ColorMatrix class already months ago):

I have added a colorize( rgb, amount ) method to my ColorMatrix class - it allows you to tint bitmaps similar to the “colorize” method of Photoshop. Correct me if I’m wrong here, but to my knowledge this kind of colorizing is not possible with Flash’s normal colorTransform or tint.

Here’s the demo:

view applet

Poker is a little hobby of mine and I’ve been keeping track of hands played (in computerized versions of the game) in a database. Amusingly enough, over the last several years, I’ve collected about 16,000 hands. I’m interested in creating some visualizations with this massive amount of data. . . A first, rather rudimentary pass (built with Processing) is at the link above. It shows Texas Hold’em starting hands and their win rate — green for good, red for bad, the brighter, the better/worse (the values are all normalized so the betting amounts are irrelevant, each hand is valued according to it’s average “unit” win per time played. “s” indicates if the hand is suited, “o” for offsuit) The obvious ones — AA, KK, QQ, AK — are best, of course. What’s interesting to see are the hands that are overplayed — QJ suited, Q10 suited, J10 suited. Oh, they sure do look pretty, don’t they. . . Of course, this takes very little into account, nothing about how many players, limits, position, etc. But it’s a start. . .

BlobDetection_kaleido_fx
A kind of RGB channel sinusoidal disto was added on top of the 3D rendering, giving quite a psychedelic effect to the whole stuff.

BlobDetection_kaleido_fx
A kind of RGB channel sinusoidal disto was added on top of the 3D rendering, giving quite a psychedelic effect to the whole stuff.