preahkumpii

I have found this plugin one of the most useful of all. I have a question/problem with it, though. I understand the concept of the plugin and am able to implement it. However, when I have a picture, apparently when the picture includes transparent pixels, the plugin does not seem to render the mask correctly. Consider my layer toolbox: I have a blue background and a black layer with the text cut out by using the alpha mask earlier. So, where the text is, it is transparent. I applied gaussian blur to the black layer, creating a feathered look/inner glow within the text. I then use the alpha mask (a white background with black text; a png) and expect this: Instead of that I get something like this: Apparently, the pixels in the black layer that have any (or a certain amount of) alpha are just being rendered as without alpha and opaque. This is my conclusion based upon playing with the alpha mask. It seems to render a certain a pixel with a certain percentage of alpha (say less than 50%) as opaque and a certain percentage of alpha (say over 50%) as transparent. I am expecting the mask to keep the pixels that are to be shown exactly as they are on the layer, but that is not happening. I think this is a bug of sorts, or maybe just an oversight? Any input would be appreciated.

Any idea if this plugin will incorporate alpha/transparency into the choice for colors for the gradient bar in future? Adam

The difference between input and output is exactly my problem. I figured out also that placing a dot on the left side of the diagonal line causes what effect you have selected (luminosity/RGB) to get brighter. Placing a dot on the right side of the diagonal line makes the pixels darker. However, a big hole in my understanding is what is the difference between placing a dot on the bottom left of the line and the top left of line (diagonal line, that is) or both? The Levels adjustment is simply an unknown to me. That is, the graphs and all. I did notice one thing. The horizontal and vertical line have values between 0 and 255, which corresponds with the RGB color values respectively that I find in the expanded colors dialog. That has something to do with it. What, I don't know. I know the pair of numbers (I call a coordinate pair, from high school :!: ) have 0-255 respectively and refer to the place on the grid/graph.

I have been fooling around with Levels and Curves a little bit. I have tried to figure out what input and output [channels] refer to. I cannot figure it out. Can anyone give a thorough but not too long explanation of what the difference between input and output is? I understand the channels (rgb). For instance, in the Curves dialog, the help on the website says the horizontal lines represent the intensity of the input and the vertical the output. I can see what it does, but cannot understand enough to use it and know what I am actually doing. Can anyone help? Help on the "Levels Adjustment" dialog would also be of help to me. Thanks.

Once I figured out exactly how to use the alpha mask, it was quite easy. I have not tried the feather plugin yet. Notice the below picture, the edges are smooth by using the alpha mask, which is excellent. The problem mentioned in the previous post is shown in the second picture. It is the inside of the A and at the sharp corners which look abnormal or unnatural. Any ideas? Preahkumpii

david.atwell, I think I found a really good way to do the bevel using the alpha mask plug-in. Previously, the object with the bevel shading/highlighting would be jagged on the edges, like without anti-aliasing. Using the alpha mask, you can keep the anti-aliasing. It looks really sharp that way. The only problem is with sharp turns, holes, or concave parts of objects. This produces a sort of goofy look. For instance, the shading/highlight on the letter "A" looks rather odd (that is, on the hole in the middle). It there any way to make the black/white layer look right after offsetting it from the original object and applying a gaussian blur? Am I being clear? It seems not. The shading/highlights are not extending all the way to the corners of the objects in these cases. Preahkumpii

I have been fooling around trying to manually reproduce the bevels found here: http://www.joesfilters.com/joes_bevel_and_emboss, but I have had some problems. The primary problem is the ability to remove the shaded/highighted layers in a directional sort of way, creating the effect of light. Using the erase tool is not very efficient and is pretty time consuming. Besides, it does not produce very good results. Any ideas as to how to erase the shadow layers so the shadows/highlights look half-way realistic and directional? Preahkumpii

You lost me a bit here. When you gaussian blur the top layer, then select outside the blurred text with low tolerance, a larger version of the text is selected. So, I raised the tolerance to about 68 or so to select a smaller version of the text. I then tried to follow your tutorial, but got a little lost. I ended up going back to what I suggested in my last post. It worked okay. One of the main problems with this method is when using the magic wand, anti-aliasing is lost, creating choppy edges. I don't think this was the case with the first method I used, but only when resizing the shape/text is done. So, that problem remains. Also, the manual labor of erasing the highlights or shadows is somewhat annoying. One good thing out of it is that the gaussian blur selection thing you showed me was neat. I never knew how to select a smaller version of an object. Still looking , Preahkumpii

I tried Pyro's outline object plugin. This did not work because it adds the outline in addition to the shape/text/object. This makes the bevel not a part of the actual object. However, from this I got to thinking of another way to do it. I simply made two new layers after the object layer. With tolerance set to 0, I highlight all but the object. Then, paint bucket fill the selection black on one layer, then white on another. Then, for each layer I ran a gaussian blur to whatever softness I wanted. Then I used the gradient tool set on transparency mode to fade out the black and white to create the beveled look. Last, I select all but the object again, then delete the excess black and white from the layers. This leaves a beveled object. You can even control the "choke" of the bevel by reducing the size of the selection when you select the original object. Then, do the above and it causes the bevel to be thicker. The only problem with this method it that it only works on normal shapes (rectangles, circles, ovals, triangles, etc.) but not on text. With text and odd shapes, you can do the bevel the first way mentioned, which makes the bevel options somewhat limited, but the method using the resize selection technique will not work because text cannot be resized without distortion problems. Any other help would be good. Preahkumpii

This bevel plugin only does bevels for a rectangular shape. I wanted something that would work for any shape, including text. I know there's no plugin to do it. That's why I was looking for a tutorial, and a rather advanced one at that. Preahkumpii

I have searched for a good bevel tutorial that can do any shape (not just a rectangular shape) and have not been able to find one to suit me. I have tried to figure out how to make a good bevel that has a fair amount of flexibility with the blur, size, and other attributes. At this page, http://www.arraich.com/effects1/bevelandemboss.htm the second paragraph gives a little hint about Photoshop's method of making a bevel. I was not able to do it. I can make the crescent-shaped layers that would make the beveled lights and shadows, but when I apply a gaussian blur, it doesn't work well. Can anyone help? Preahkumpii

I am having trouble searching the new forums. For instance, I searched for "alpha masking" after manually finding the forum post in the new forum, and the engine said there were no results. The post was there, the search (I did check search criteria) did not see or find the posts.

From my reading, I found the same as what you said, that the cmyk spectrum simply cannot reproduce exactly what rgb produces, especially in blues. I just thought, seeing the source codes for those converters is available, it may be possible to implement something from them that has made those converters appear successful.

This is mainly directed towared boltbait. I looked on your website and viewed your cmyk plugin and gave it a try. I was going to just suggest something I found on gimp. I have no idea if it is feasible or anything. The website below had a very good cmyk conversion plugin that basically makes the image unchanged as far as appearance is concerned. The website is: http://plasticbugs.com/?p=361 There are two different plugins mentioned there; apparently one is better than the other. Like I said, it may not be usable at all, but I figured it was worth looking at. If anyone else knows more, I would be interested.