Hi Friends,
I've prepared 2 screen shots to show my problem. I have 2 separate, but except for their shape, identical paths. Applying the same amount of Blur to each one, one is much more blurred than the other. Here's the 1st without blur. The blue arrows point to the 2 paths (light brown color....the color between the dark brown and the gold):


And here is what it looks like after I apply 1.0 blur to each one:


Why does the path on the right appear to be blurred much more that the path on the left?

For some reason, I'm thinking that I should know the answer to this, but I'm drawing a blank. The path on the right is much bigger, extending 4 or 5 times the length that is shown outside the border of the screen shot. While the one on the left ends just outside the border of the screen shot. But I don' think that should have anything to do with it....right?

Does anyone have any ideas about this?

Thanks for your help

Dru

blur radius changes in relation to the size of the object.
a) set the blur for each object separately, not when having both paths selected
b) In Inkscape 0.47(pre) to get the same amount of blurring, use a filter effect (Gaussian blur) and set the equal amount of blur in the filter editor.
c) Dru's idea to combine the paths to one seems an interesting workaround, I haven't tried it myself though - yet ;-)

blur radius changes in relation to the size of the object.

omg! Why? What is the reason Inkscape was made to do so? To me, it doesn't make sense. To me, the blur radius for 1.0 should be the same where ever it is set at 1.0....in fact, I thought the blur setting was exactly that, the measure of the blur radius.

I did set the blur on the 2 paths one at a time (not with both paths selected). But, when I select them both before setting the blur, the blur is consistent for both. Don't have 0.47pre4, so can't try your 2nd suggestion, ~suv.

And yes, yes, druban and ~suv, combining the paths also causes the blur to be consistent for the 2 paths. And in my image, it could be reasonable to have all the paths with this same fill color combined. Actually they were originally drawn as the same path, but got broken apart when I needed to use Break apart in another area than what is shown in the screen shot.

But anyway, why is the blur radius different for different sized objects? And why does selecting them both before setting the blur have the same effect as combining them -- ie-it causes the blur to be consistent for both? Something about that doesn't make sense either????

brynn wrote:But anyway, why is the blur radius different for different sized objects? And why does selecting them both before setting the blur have the same effect as combining them -- ie-it causes the blur to be consistent for both? Something about that doesn't make sense either????

I'm sorry I have no in-depth answers to your questions - I happened to stumble over this issue recently in a conversation on irc and remembered a bug I once read but didn't understand. There might be more answers buried in earlier release notes or in the mailing lists archives, but I haven't researched this question any further yet.

3

I reacted on this too as it seems quite counter-intuitive.. then again there's no unit tied to the blur -- the number could basically be a percentage of the objects size for all I know..
It could be argued that the predefined unit should be the one set in the document properties though.. All in all, I'd really wish blurs were more consistent too. Also, I can't find this Gaussian blur you're talking about..

edit: I just did a test. It seems a blur value of 41.666~ (the decimals are truncated) quadruples the area of any object. Twice the value, or 83.333~, doubles the effect again. This seems true for any object size and unit type in 0.47pre4

Mar wrote:Also, I can't find this Gaussian blur you're talking about..

Gaussian Blur:

1. open 'Filters > Filters Editor…'
2. create 'New' filter
3. Add Effect: 'Gaussian Blur'
4. select object and activate the just created filter for the object
5. adjust the amount of blur with the sliders in the 'Effect Parameters'

Inkscape » Filter Effects—Custom
<http://tavmjong.free.fr/INKSCAPE/MANUAL/html/Filters.html>
Inkscape » Filter Effects—Custom » Filter Effects Dialog
<http://tavmjong.free.fr/INKSCAPE/MANUAL/html/Filters-Dialog.html>
Inkscape » Filter Effects—Custom » Pixel Manipulation Filter Primitives - Gaussian Blur
<http://tavmjong.free.fr/INKSCAPE/MANUAL/html/Filters-Pixel.html#Filters-GaussianBlur>

Druban, I'm not sure if I follow your explanation . If I do, then I would disagree. I do understand what you mean about Inkscape using pixels as a measurement in certain, maybe even many situations. But I don't think that screen size has anything to do with it. Any kind of action potentially related to or based on screen size would result in images made on one screen size appearing distorted on other sizes. I also have not experienced any problems with blur related to magnification or zoom. In as far as I understand what you mean by this:

So if you applied the blur to each object separately with the object filling up the same amount of screen each time, then the bigger object, being magnified less, would receive less blur.

I tried it. But changing the zoom before applying the blur had no effect on the blur.

The problem I'm having is apparently strictly the relative sizes of the paths/objects. And it does appear, as ~suv suggested, that the blur radius is related to the size of the object/path. I just want to understand why it was made this way, if possible.

edit: I just did a test. It seems a blur value of 41.666~ (the decimals are truncated) quadruples the area of any object. Twice the value, or 83.333~, doubles the effect again. This seems true for any object size and unit type in 0.47pre4

Yes Mar, that makes sense to me. I learned early on working with Inkscape, that blurring the fill of a path/object can change the size of the path/object. It's how I came to learn how to use Clip

But if you would try your experiment on an object half the size (and/or twice the size of) your original object, and you use the same blur values for them all, you will clearly demonstrate my problem (or more like my question). You will notice that the blur of the smaller object is much less than the blur on the larger one.

Anyway ~suv
[quoteThere might be more answers buried in earlier release notes or in the mailing lists archives, but I haven't researched this question any further yet.][/quote]
Please don't go to any trouble on my account. I just want to understand in the most basic way possible, why the blur radius is related to the object/path's size. I haven't really undersood what the Question/Answer section of Launchpad is intended for, but do you think this would be an appropriate question to ask there?

Thanks for everyone's comments on this issue. I really appreciate your input

I think Druban's right. When I get home I'll post a few images to help explain. You can test it yourself. With zoom at 50% draw a path and duplicate it twice. Select one path and double its size. Apply a blur 1.0 to the larger path. Select the original path and apply blur 1.0. Finally select the last path and zoom to 100% and apply blur 1.0. The blur of the large path at 50% zoom and the small path at 100% zoom should be the same, and both should be different from the small path at 50% zoom.

I don't know llog. I followed your instructions, and here's my results (oops, sorry it's so big ):

The 3rd duplicate, which is the one on top, is where I zoomed to 200%, after I couldn't see any difference at 100%. But I still don't see any obvious difference. Note that I had to use 10.0 blur to be able to see it well enough to compare.

Just for clarity, here are my steps (which except the the 3rd dup, are llog's)
1 - create object
2 - duplicate 3 times (original is bottom left, zoom 50% -- 1st dup in the middle, the large one, zoom 50% -- far right is 2nd dup, zoom to 100% -- top left is 3rd dup, zoom to 200%)
3 - double size of 1st duplicate
4 - with zoom at 50%, apply 10.0 blur to bottom left and middle (large)
5 - zoom to 100%, apply 10.0 blur to far right object
6 - zoom to 200%, apply 10.0 blur to top left object

Did I do it wrong, or do I not understand?

No new information from me at the moment, only throwing in that blur is independent of any zoom levels - it is part of the attributes of an SVG object (whereas cursor movements, calligraphic stroke widths or the paint bucket tool can be configured to depend on zoom levels. This is unrelated to filter effects or blur AFAIU).

brynn wrote:I haven't really undersood what the Question/Answer section of Launchpad is intended for, but do you think this would be an appropriate question to ask there?

You could try - but I could imagine that a message in the inkscape-user mailing list might result in a broader discussion than in the answer section (that part of launchpad works well for basic support questions, but recently there haven't been many developers engaged in answering - all busy getting 0.47 out ;-).

OTOH the section in the manual about the Gaussian Blur seems quite extensive - I'd probably start from there to get a better understanding about blur, filter effects in general and their relation (in Inkscape and in the SVG specification) to the size of the object.

hmm. You're right Bryn. It doesn't work the way I thought. Resizing an object with blur does change the amount of blur though. I guess that's what I was thinking. Sorry I'm not much help here.

Mar, from the manual

For practical reasons, the limit of an object's blur is two times the blur radius outside the bounding box at maximum blur.

I believe this is what you discovered with your experiment!

And here's part of the answer, from the manual:

Tip
Blurs created through the Fill and Stroke dialog depend on the size of the blurred object. To get the exact same amount of blur on different size objects, you can either use the Edit → Paste Style (Shift+Ctrl+V) command (if all the attributes are to be the same) or use the Filter Effects dialog to set the blur radius (standard deviation) to the same values.

It confirms that the size of the object does affect the blur radius. But it doesn't explain why it was made so. I''ll be reading on.....