Comments on: More Fun With Gradients http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/ On Mac OS X programming Sat, 07 Jan 2012 04:23:42 -0500 http://wordpress.org/?v=2.8.5 hourly 1 By: mr_noodle http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/comment-page-1/#comment-23 mr_noodle Sat, 14 Oct 2006 15:03:36 +0000 http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/#comment-23 Thanks for the comments. First off, I want to point out that my category is just a wrapper around CoreGraphics shading functions and it (CG) should do the proper calculations with regards to device resolution and such. My category is oriented more towards gradients in the UI. This should be apparent by the fact that I convert everything to the device RGB colorspace and that I also have an optimization for the cardinal directions. Looking at CTGradient code, it seems it is much more complete in terms of control over the actual gradient; I only implemented linear axial gradient between two colors in RGB space. The focus of my little exercise had to do more with calculating the start and end points of the gradient where it would fit it to a shape at a certain angle. The CG functions for shading require you to specify the start and end point and I also provided a method to specify those for more control over the geometric extent of the gradient. Overall, my method is more of a convenience function covering a common case than the end-and-be-all of gradient methods. Incorporating my code into CTGradient may be something I could bring up with the author since it seems like it is complimentary with his stuff. Thanks for the comments. First off, I want to point out that my category is just a wrapper around CoreGraphics shading functions and it (CG) should do the proper calculations with regards to device resolution and such.

My category is oriented more towards gradients in the UI. This should be apparent by the fact that I convert everything to the device RGB colorspace and that I also have an optimization for the cardinal directions. Looking at CTGradient code, it seems it is much more complete in terms of control over the actual gradient; I only implemented linear axial gradient between two colors in RGB space. The focus of my little exercise had to do more with calculating the start and end points of the gradient where it would fit it to a shape at a certain angle. The CG functions for shading require you to specify the start and end point and I also provided a method to specify those for more control over the geometric extent of the gradient. Overall, my method is more of a convenience function covering a common case than the end-and-be-all of gradient methods.

Incorporating my code into CTGradient may be something I could bring up with the author since it seems like it is complimentary with his stuff.

]]> By: Chas http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/comment-page-1/#comment-22 Chas Sat, 14 Oct 2006 07:04:35 +0000 http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/#comment-22 Ah, your article brings back memories of the olden days when I used to work at a graphics bureau, one of the continual problems was fixing other designers' gradients in Adobe Illustrator 88 and Aldus Freehand. Gradients were defined by dragging out a rubber-band line, you defined the start and end points, but really what you were doing was creating a vector (a math vector as in a direction and magnitude, not as in vector graphics). We used to have a little Mac desk accessory app to help calculate gradients between two CMYK colors, and to tell us how many inches we could span a gradient on the imagesetter film at a certain linescreen without causing visible banding. It was very tricky to get it right, and mistakes meant you blew 4 sheets of film which could easily be over a hundred bucks. But getting to the point.. I haven't run your code, but from your description of bounding boxes, it appears that you have solved one instance of applying a gradient to a shape, but you haven't solved the general case. The vector that defines a gradient does not necessarily coincide with the bounding box. A gradient fill may not span the entire object, i.e. you might have an object with a large region of 20% gray on the left, a small 20 -> 50% gradient across the middle, and a large region of 50% gray on the right. It seems to me that that this would be easily defined with a vector but difficult to define with a bounding box. Ah, your article brings back memories of the olden days when I used to work at a graphics bureau, one of the continual problems was fixing other designers’ gradients in Adobe Illustrator 88 and Aldus Freehand. Gradients were defined by dragging out a rubber-band line, you defined the start and end points, but really what you were doing was creating a vector (a math vector as in a direction and magnitude, not as in vector graphics). We used to have a little Mac desk accessory app to help calculate gradients between two CMYK colors, and to tell us how many inches we could span a gradient on the imagesetter film at a certain linescreen without causing visible banding. It was very tricky to get it right, and mistakes meant you blew 4 sheets of film which could easily be over a hundred bucks.
But getting to the point.. I haven’t run your code, but from your description of bounding boxes, it appears that you have solved one instance of applying a gradient to a shape, but you haven’t solved the general case. The vector that defines a gradient does not necessarily coincide with the bounding box. A gradient fill may not span the entire object, i.e. you might have an object with a large region of 20% gray on the left, a small 20 -> 50% gradient across the middle, and a large region of 50% gray on the right. It seems to me that that this would be easily defined with a vector but difficult to define with a bounding box.

]]> By: charles http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/comment-page-1/#comment-21 charles Sat, 14 Oct 2006 01:20:09 +0000 http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/#comment-21 That's a good one! That’s a good one!

]]> By: Hioto http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/comment-page-1/#comment-20 Hioto Fri, 13 Oct 2006 14:22:05 +0000 http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/#comment-20 Yet another attempt to get a handle on gradients is CTGradient! http://blog.oofn.net/2006/01/15/gradients-in-cocoa/ Yet another attempt to get a handle on gradients is CTGradient!
http://blog.oofn.net/2006/01/15/gradients-in-cocoa/

]]> By: Hjalti http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/comment-page-1/#comment-16 Hjalti Thu, 12 Oct 2006 16:50:31 +0000 http://www.noodlesoft.com/blog/2006/10/11/more-fun-with-gradients/#comment-16 Thanks! Thanks!

]]>