Skip navigation
Frank Vollono
Currently Being Moderated

How would you create this effect?

Feb 22, 2013 7:13 AM

I am curious if there is a fairly straight-foward away of creating the effect seen in this image. I am talking about the pixelated effect made up of many little red boxes of various hues at the top part of the image.


I was thinking of zooming in real close to an image in photoshop and doing a screen capture, but I'd rather have a vector image that can be easily scaled in size and easy to change the hue to different colors.


Something like this able to be done in illustrator without building it a box at a time?




How do I do this effect.jpg

  • Currently Being Moderated
    Feb 22, 2013 7:47 AM   in reply to Frank Vollono

    If your objective is to keep it vectors, I'm afraid you're stuck with individual boxes to some extent, however you don't have to draw each one individually. (Applying their color fills to effectively emulate diffusion may be a more tedious matter.)


    You can make a "pixel grid" of vector squares in a few easy steps using a rectangle the size of your artboard and slicing it up by constructing a grid of paths, then using the Pathfinder Divide function. Making the grid can be done a few different ways. The grid tool comes to mind, or you can use the Blend feature. This tutorial includes a way to do it:


    Mark as:
  • Currently Being Moderated
    Feb 22, 2013 8:03 AM   in reply to Frank Vollono

    Select the image & try creating a mosaic by choosing Object > Create Object Mosaic...experiment with the options in the dialog.

    Mark as:
  • Currently Being Moderated
    Feb 22, 2013 8:33 AM   in reply to [Ravi Kiran]

    Ah, now that's interesting! Thank you, Ravi.


    To the OP: Experimenting with Ravi's suggested method, (I'd never used the feature before), I found that the command Object > Create Object Mosaic, if I understand it correctly, only works on raster objects. So you could apply a gradient fill to an object, then choose Object > Rasterize before applying the Mosaic command. To my delight, the resulting "Mosaic" is constructed of vectors, separate from the original raster image. This is definitley the way to go...

    Mark as:
  • Currently Being Moderated
    Feb 22, 2013 10:38 AM   in reply to John Mensinger

    A solution has already been posted, but for discussion purposes only, I'd like to suggest an alternative method that brings you to the same end result:


    Create a wide rectangular shape and fill it with a gradient. Go to Object>Path>Split Into Grid. The pop up dialogue is, I think, a little clearer than the mosaic approach, has the added benefit of the preview which will show you what your grid will look like as you enter values, and doesn't require round-tripping through raster. What I don't know is if Illustrator has a way to randomize the color in the newly formed grid (via either approach) but I suspect a talented scripter could do it.


    Michael Riordan

    Adobe Certified Instructor

    Digital DesignLab

    Mark as:
  • Currently Being Moderated
    Feb 22, 2013 6:57 PM   in reply to Frank Vollono

    Create a wide rectangular shape and fill it with a gradient. Go to Object>Path>Split Into Grid.

    That will result in a bunch of rectangular paths, each having a grad fill, not with each having a solid fill, emulating pixels.


    I'd rather have a vector image that can be easily scaled in size and easy to change the hue to different colors.

    In a design like that, how is a raster image any less scaleable than a vector graphic? Can you tell me which of these obviously scaleable graphics is a raster image and which is a bunch of vector rectangles?:



    Now click the link to a PDF below. Zoom in on it and tell me whether it's a vector or raster graphic:


    Raster Or Vector?



    Mark as:
  • Currently Being Moderated
    Feb 24, 2013 10:15 AM   in reply to Frank Vollono

    Well I'd say that PDF is a vector.

    The PDF contains nothing other than a raster image. It contains no vector paths.


    (By the way: "A vector" is a single expression of a direction. Calling a vector-based illustration "a vector" is like calling a raster-based image "a pixel.")


    I zoomed in as close as I could and didn't see any pixels.

    All you could see is pixels. That's all that's there.


    If it's raster I'd be interested in knowing what resolution...

    Its resolution is 1. It's a single-pixel raster image (as given away in its title).


    I've never experienced a raster image that didn't get more pixelated the larger you scaled it up.

    Yes, you have. That's exactly what I demonstrated in post 5 and in the PDF.


    The operative word in your comment is more pixelated. That misconception is the point of the demonstration. Your image is already pixelated, and that's exactly what you want. It won't be "more" pixelated whatsoever, no matter how large you scale it.


    At the heart of the [quite common] misconception is this: What you are calling "resolution" is not. True resolution is the measure of the amount of information contained. A raster image contains only a fixed amount of information. PPI is merely a scale factor. It's merely an expression of information density, not quantity. As you can see, that PDF is every bit as "resolution independent" (i.e.; scaleable) as a vector square.


    The same priciple applies to your raster image. Your whole intent in your posted example is to display pixelation. So if the very thing you're trying to depict is the pixelation of, say, a 10 pixel x 20 pixel raster image, then all you need in order to depict that at any scale is a 10 pixel by 20 pixel raster image. At a scale of 10 PPI, you've got a perfect depiction of 10 square Pixels Per Inch. At a scale of 200%, you've got a perfect depiction of 5 square Pixels Per Inch (i.e.; 10 Pixels Per 2 Inches). You can scale it to the size of the moon and you'll have a perfect depiction of 10 Pixels Per Moon. It's just as scaleable as a 10 x 20 array of vector squares.


    In other words: All the actual resolution of your raster image (all its information—all its pixels, in all their squareness) is already visible. Reducing its information density (by scaling it larger) will not reveal more of its actual information (the square shape of its pixels—the information that one usually wants to conceal or disguize about raster images, but not in this case).


    The whole purpose of making sure your raster images have "sufficient resolution" in desktop publishing is to disguise the squareness of their pixels (what is commonly called "pixelation"). But that is exactly what you are intending to depict in this case. And you want to be able to depict that with equal clarity at any scale. There's no need to re-draw it as vector paths in order to do that. In this situation, you are not trying to avoid "pixelation," you are trying to show it, which is exactly what a raster image does by its nature.


    Think of it this way: If all you are trying to draw is 200 squares, it can easily be argued that a raster image is more "data efficient" than a file containing 200 square vector paths. For each vector path, you have to include four coordinate pairs (800 anchor points). For a raster image, all you need is 200 color values.


    This is not just a symantic or academic or pompous know-it-all point of argument. It's a point widely applicable to many common situations, and something every graphics professional should understand. For example, consider the software instructions author who pointlessly frets over the misunderstood "requirement" for pixel density when his illustrations necessarily consist of screenshots that are, after all, supposed to accurately and clearly depict what the user sees on his screen. Or the fact that you see single-pixel raster images scaled to the full width of the page, or dimensions of a table cell or large section of a whole background on web pages everyday. Or the widespread misconception that intentionally fuzzy raster effects like soft drop shapows "require" 300 PPI.



    Mark as:
  • Currently Being Moderated
    Feb 25, 2013 2:41 AM   in reply to JETalmage

    i'm going to start specifying things in Pixels Per Moon from now on. i hope my colleagues are all versed in standard form.

    Mark as:
  • Currently Being Moderated
    Feb 25, 2013 5:47 AM   in reply to dougofakkad

    i'm going to start specifying things in Pixels Per Moon from now on.

    Yes. Also be sure your colleagues are clear on CPM—Colors Per Mile. Never know when someone's going to ask how many colors it is from New York to Atlanta.





    Mark as:

More Like This

  • Retrieving data ...

Bookmarked By (0)

Answers + Points = Status

  • 10 points awarded for Correct Answers
  • 5 points awarded for Helpful Answers
  • 10,000+ points
  • 1,001-10,000 points
  • 501-1,000 points
  • 5-500 points