Borders and Frames

[Mij]

BORDERS AND FRAMES
There are now a number of different methods by which frames can be added to images in Irfanview, so I thought it might be useful to post some tutorials here to explain them.
I will post 6 guides in succession, and the thread is then open for your ideas, questions or comments.

Irfanview is really a box of tools that you can use for many different purposes and this tutorial tries to explain how to use some of them. So even if you do not want to do the precise task being described, or you know a better tool to use from another application, I hope you can still learn something useful.

Notes.
(a) In my examples I have used 2MP images of size 1600x1200 pixels (or 1200x1600 in portrait format). I chose this size to avoid attachments becoming too large. Just scale border widths up or down to suit bigger or smaller images.
(b) The paragraph numbers refer to an image that is attached at the bottom of the post.
(c) Unless otherwise stated all instructions are for Irafnview v4.36 but will, hopefully, continue to apply to later versions.

[Mij]

The simple border tool

1. For a simple border in a single color use Image > Change canvas size.
2. Use method 1 and set the required border width for left, right, top and bottom sides or use method 2 and set what the total width and height is to be after the border is added.
3. Click Choose alongside canvas color to open the Color dialog. Click on one of the small colored boxes to choose a basic color for your border or choose a custom color by dragging the cursor around on the multicolor palette (hue/saturation) and the vertical bar to the right (luminance). The color selected is shown in the rectangular box below it. You can also specify a color numerically in the boxes to the right if you understand the H,S,L or R,G,B system. When satisfied click OK to close the dialog.
4. Click Apply to original image to preview the result. If you are not satisfied click Undo, change the border dimensions or color and press Apply to original image again. Finally Click OK.
5. Save your result.

Notes:
(a) I chose 60 pixels for border widths all around which is 5% of the width of the shorter image side but you can select any width you want for any side. Some prefer a slightly larger border at the bottom and smaller at the top.
(b) By entering positive numbers for border widths in method 1, the borders are added to the outside edges and you still see the whole image. Negative numbers (e.g. -40) will add the borders to the inside edge. If you want to use this feature you must check the box "If negative values used: add canvas inside" by clicking on it. If you do not, the image will be cropped instead and no border will be seen. You cannot create inside borders with method 2. If you enter a size smaller than the existing one the image will always be cropped to that size.
(c) An inside frame masks part of the image edges. Some of the methods that follow this one are also masks. If you do not want to lose part of your image area in these, you may want to use the Change canvas size tool before you start to add a border that can then be covered over instead.
(d) An easy way to remove a single color border is to click on Edit > Auto crop borders.

[Mij]

The frame tool

You can add more complicated borders by selecting Image > Borders/Frames. From the dialog that opens you can select single or multiple bands of color to form borders and other frame effects. All selections can be edited to change the colors and widths of the bands before applying to the image. Some examples are shown below.

1. Here I have selected the Warm border/frame style.The default settings looked to be what I wanted for both color and width judging by the preview image.
2. However in the result the border is much narrower than expected. This is because the preview border is not scaled to the size of the preview image but is instead shown as it would appear in a full size (1:1) image.
3. The default border widths (3+10+1+0=14 pixels) had to be scaled up by about x4 to get near my preferred border width of 60 pixels. Add up each box to see the full border width (12+40+4+0=56 pixels). The preview now makes the border look far too wide.
4. However the result is what I wanted. Save the image.



5. In this example I have selected the Red gradient for a 3-D frame effect but I will edit the colors to brown to look more like a natural color for a wooden frame. I have clicked on Choose for Border 4 color, the color at the inner edge of the frame. In the color dialog I have selected a strong brown from the multicolor (hue/saturation) palette but left the setting on the right hand bar (Luminance) close to where it was in the defaults.
6. Now setting Border 1, the color of the outer edge of the frame. I want a slightly different brown with a little more yellow in it. I cannot make the hue very different to the inner edge though or colored bands will be apparent, but the hue colors do wrap around horizontally so a hue near to the left edge of the palette for the outer is compatible with the one near the right edge that was chosen for the inner. The inner edge must always be darker than the outer one or the gradient is incorrectly generated. So the luminance is set higher than it was for the inner.
7. The width of the whole frame must be set in border 1. I have set it to 60 pixels which once more looks too wide in the preview. The setting of border 4 has no effect. Press OK
8. Again what I wanted though. Save the image.



9. The topmost selection, Inside fading frame, is rather different and was added in Irfanview v4.36. Try it yourself to see the effect.
10. The default, fading to white is quite a popular effect but you can set any color to fade to. Setting a normal border width of 60 pixels gives too hard an edge
11. but setting 250 pixel width (255 appears to be the maximum possible) produced this quite pleasing effect on a 1600x1200 image.



Notes:
(a) With the Auto-load default... boxes at the bottom of the dialog checked, all values are automatically reset to default when you select a new style. With either or both boxes unchecked you can lock the values in the width and/or color boxes as you change styles so that they are then applied to a different one. This feature can also be used to save width/color settings from one session to the next but remember it is the settings within the dialog box that are saved, not the settings for individual styles.
(b) You can use Edit >Auto crop borders to strip multiple colored borders but you may need to repeat it many times before it is all removed. You can increase how much is removed each time by increasing the Tolerance value for auto cropping in the Browsing/Editing tab of the Properties dialog but you do risk stripping off some of the image as well if you increase it too far. A better solution is to leave the Tolerance at 20 and use the keyboard shortcut. Hold down CTRL and Shift keys and then press and hold down key Y until the border has all gone.

THE COLOR DIALOG
When using this tool there is often a need for saving more custom colors than the limited storage described in (a) can accommodate. It is useful to learn how to use the 16 small custom color boxes at the foot of the Color dialog box for this purpose. The custom color in the Color|Solid preview rectangle is copied to one of these small boxes every time Add to custom colors is clicked. The boxes are populated in order starting from top left. These are saved between sessions so that you can later click on any one of them to transfer that custom color back to Color|Solid and then carry it back to the Border/Frames or other dialog when you click OK.
Each time you open the Color dialog, the start box for populating new custom colors returns by default to top left, so you risk overwriting previous colors that you may want to keep. You can click a box and it will start to populate from there instead. Do it before you create the new custom color though. You will lose your new color if you click later. If you forgot, you can still press the Tab key repeatedly until you see a custom color box highlighted and then move the highlight to your chosen start box using the keyboard arrow keys.

[Mij]

The paint tool

The paint tool is a plugin and the file PAINT.DLL must be installed in the plugins folder for it to work. The tool can be used to paint or draw a mask directly onto the image to form a frame.
In the forum there have often been requests for an oval frame and drawing an ellipse shaped mask over the image using the Paint plug-in provides the simplest way to do this. This is the procedure that I use to accurately position the ellipse whilst minimizing the encroachment of the mask onto the image.

1. Start by adding a fairly wide border around the image in the color chosen for the frame. Around 1/6 of the smaller image side is suitable, so for this 1600x1200 pixel image I added borders of 200 pixels all around.
2. Open the paint Plug-in (F12) and select the Color picker tool. Left click on the border color which sets the (foreground) color, in which the ellipse will be drawn, to be the same as the border chosen in (1).
3. Select the Circle tool and set the width to twice that of the border (in this case 200x2 = 400 pixels). Ensure that Fill is unchecked.
4. Draw the ellipse, starting by pressing down the left mouse button as close as possible to the top left corner, dragging diagonally across the image to a point as close as possible to the bottom right corner and there releasing the mouse key.
5. Close the Paint plug-in, draw a selection box around the oval and crop to the desired size.
6. Save the image.



Notes:
(a) If you choose too small a border in 1, you will get areas of canvas color left showing between the border and the outer edge of the ellipse. You can fill this in using the Brush tool but that is time consuming and best avoided.
(b) If you want a more accurate crop in 5 you can use Method 2 in the Change canvas size tool. In this case entering 1700x1300 (centered) would have achieved much the same result as in 6.

[Mij]

8bf filter tools

The Adobe 8bf filter tool is a plugin and requires the file 8BF_FILTERS.DLL in the plugin folder. This plugin allows filters written in the Adobe 8bf format (for Adobe Photoshop) to also be used in Irfanview.
Bordermania contains 8bf filters that can be used as frames and can be downloaded free from this site http://www.pluginfilters.com/bordermania/index.html
After download, install the filters you want to use in the Adobe8BF folder that you can find in the Plugins folder of Irfanview.
Select Image >Effects >Adobe 8bf filters (or press Ctrl+K). Click on Add 8BF filters and browse to the Adobe8BF folder where you just installed the Bordermania filters. Click OK and you will see all the filters listed in the Adobe 8BF filter dialog.

1.To use the filter open the image to be framed, select Image >Effects >Adobe 8bf filters and choose the filter you want to use.
2.These are the settings I used for the Rings filter
3.and this is the result



4.These settings were used for the Ripple filter
5.and this is the result



6.In the Picture frame filter you have an option to add a colored pattern from a Texture file. I created one by capturing a slice of the color dialog from Irfanview.
7.Applying these settings to the Picture frame filter
8.gave this result



Notes:
(a) You need to register and leave an email address for thepluginsite.com when you first use it, but you do not need to receive their newsletter unless you wish. Uncheck the relevant box and all you will receive from them is the one email to verify the registration.
(b) You may find two files already pre-installed in the Adobe8bf folder, HarrysFilters.8bf and PopArt.8bf
(c) Other 8BF filters do not have to be in that folder. They can be placed anywhere on the computer that has read/write access. Since a whole folder of 8Bfs are loaded together, it is best to divide them into small groups each in a different sub-folder.
(d) You can keep all the Bordermania filters that you download in one folder but not all are useful for creating frames. The ones I consider best are shown in image 1. They correspond to the Bordermania files nvrph4.8bf, nvrph5.8bf, nvrph7.8bf, nvrph8.8bf, nvrph9.8bf, nvrbm18.8bf and nvrbm19.8bf
I suggest that you move the others to a different folder and keep them there. Some you may find useful for other purposes.

[Mij]

Overlay tool

Masks can be created from image shapes downloaded from the internet. A good set is Harry's edges available here. http://thepluginsite.com/products/harrysedges/
Download them and put them into a read/write folder anywhere on the computer.

To use the masks in Irfanview you must resize them to match the size and shape of the image you want to apply them to and to save as a PNG file using transparency options. The PNG file is then applied to the image as an overlay.

1.To show the transparency of a PNG file used as an overlay the option "Show PNG/TIF/TGA/DDS alpha/transparent color)" in the Viewing tab of Properties/settings must be set. It is wise to check that has been done before you start.



2.Open the image file to be masked. Find the folder and choose the Harry's Edge to use as the mask. I double clicked on Star9.png and opened a new copy of Irfanview in another window on top (I have PNG files associated with Irfanview).
3.Resize the mask to equal the dimensions of the image to be masked. The image dimensions can be read below it in the Status bar.
4.Save the resized mask with "Save transparent color" selected in the PNG Save options.
5.When asked to select the transparent color click on the white area in the center.
6.Close the second copy of Irfanview showing the mask. In the first copy, showing the image, select Edit > Insert overlay/watermark image.
7.The saved PNG file has been selected as the Overlay/watermark image and the Center option set. Click Preview to see the result and if it looks good click OK.
8.Save the result.



The Star9 mask used above is just Black and white. The white area became completely transparent and the black completely opaque. Some of the masks also contain shades of gray which allow some areas to be only partially transparent. In this example we will use one of these, Twirl14.png.

9 Open another image in Irfanview, then open Twirl14.png in another copy just as in (2) above.
Then resize the mask to 1600x1200 pixels to match the image dimensions as in (3).
10 For this example we will have a colored mask. Open Image > Color corrections and move the Red, Blue or Green Color balance sliders to the right to get the mask the color you want.
11 This time when you save as in (4), also check Save Transparency as Alpha channel in the PNG save options. When asked to select the transparent color click on the white area as in (5) to make that fully transparent.
12 The overlay is added just as before in (6) and (7) and the result saved.



Notes:

(a) You can also use masks bundled with other software applications such as those in the masks folder of Faststone image viewer.
(b) The purpose of opening the image in a window first, and then processing the mask in a second copy of Irfanview opened on top of it, is primarily so that the image dimesions can be read and used to resize the mask. I find it helps the workflow too. If you find it confusing either close or minimize the first copy after noting down the image dimensions.
(c) Unless you want to reuse the mask after it has been colored and resized it need only be stored temporarily until it has been used as an overlay. I save all temporary overlay files as mask.png in a Temp folder. That avoids the need to change the name and location in the overlay dialog every time.

[Mij]

Combining the tools

The oval vignette is the most complex effect I have tried and uses a number of tools. If you do not know what a vignette is look at image 12 first. Then you can see where I am going.

1. You can find a number of gradient images on the web which fade from white to black with all the shades of gray in between. This one is a radial gradient with a tiny white circle in the centre fading outwards to fully black at the corners. Open it in Irfanview.
2. Select Image > color corrections dialog and move the contrast control right to maximum. This converts all grays to either black or white. By adjusting the brightness slider you can make the white circle at the center any diameter you choose. Adjust so that the white circle almost fills the screen.
3. Now reduce the contrast a little to create a gradient from white to black instead of a hard edge. The white will be the fully transparent image area and the gradient will become the vignette region where transparency decreases gradually towards the corners.
4. Add some color to the mask by moving one or more of the Red, Green, Blue sliders to the right. In this case I have just moved the Blue slider.
5. Click OK to close the Color corrections dialog.
6. Now resize the mask to match the size of the image you want to apply it to. In this case it is 1600x1200 again.
7. Save the image as mask.png in the Temp folder with both Save Transparent color and Save Transparency as Alpha channel set in the PNG save options.
8. Click in the white center region when asked to set the transparent color.
9. We now have a blue oval vignette mask. If you wanted to apply it to a number of different photos you can save the mask at this stage in a permanent location. You could then use it in batch operations. For this example we will just apply it to one.



10. Open the image to be masked.
11. Open Edit > Insert overlay/watermark image and set as in previous examples.
12. Click OK and save the final result.



FADE TO WHITE
This is also a vignette but with a white surround. A white surround would leave
the alpha layer transparent at the edges, however, so we need to use a workaround.

13. Repeat steps 1 to 8 but omitting step 4. We have saved a black oval vignette mask this time.
14. Open the image to be masked.
15. Select Image > Negative > All channels
16. Add the overlay from 13 as you did in 11.
17. Select Image > Negative > All channels again
18. Click OK and save the final result.

[theDUKE]

CUDOS to you! Great advice and tips for those wanting to enhance photos without the need of Expensive Photo Software!

BORDERS AND FRAMES
There are now a number of different methods by which frames can be added to images in Irfanview, so I thought it might be useful to post some tutorials here to explain them.
I will post 6 guides in succession, and the thread is then open for your ideas, questions or comments.

Irfanview is really a box of tools that you can use for many different purposes and this tutorial tries to explain how to use some of them. So even if you do not want to do the precise task being described, or you know a better tool to use from another application, I hope you can still learn something useful.

Notes.
(a) In my examples I have used 2MP images of size 1600x1200 pixels (or 1200x1600 in portrait format). I chose this size to avoid attachments becoming too large. Just scale border widths up or down to suit bigger or smaller images.
(b) The paragraph numbers refer to an image that is attached at the bottom of the post.
(c) Unless otherwise stated all instructions are for Irafnview v4.36 but will, hopefully, continue to apply to later versions.

[jewe.org]

I hope I'm not intruding on your thread, Mij. I thought I'd contribute a few border and frame effects for the IrfanView Filter Sandbox plug-in. You can find this plug-in in IrfanView under Image -> Effects -> Filter Sandbox, if you have the IrfanView Plug-ins bundle installed.

Sepia Frame

This effect creates a fuzzy border around the image. Despite the word "Sepia" in the name, the actual tint color can be chosen in the effect's parameters. The effect works by drawing a tinted version of the original image back onto the image, using a circular, anti-aliased brush.

[ATTACH=CONFIG]4110[/ATTACH] [ATTACH=CONFIG]4111[/ATTACH]

Brush Size: Specifies the size of the brush to use. Depending on the size of your image, you may need to adapt this so the border looks right. If you set the step-density too high, there may be visible gaps between the brush strokes. Increase the brush size or decrease the step-density to avoid this.

Step Density: Specifies how close to each other brush strokes will occur. If you set this very low, this will result in a more regular and less fuzzy frame around the image. But it will also take more time to render. If you set this very high, this may result in visible individual dots or lines.

Tint Color: Specifies the tint color to use. The default is a faint sepia color.

Random Seed If you don't like the distribution of the random strokes, you can experiment with different random seeds to see if it yields better results. When using the same seed, the random function will always produce the same result.

You can also apply this effect more than once for some interesting accumulative effects. The next image was made by first using a dark blue tint and a step density of 50 pixels, resulting in single dark blue spots. On top of that the effect was applied with light blue tint and a step density of 10 pixels.

[ATTACH=CONFIG]4112[/ATTACH]

[jewe.org]

Radial Brighten (Vignette)

Next up, this effect allows to darken an image around it's borders and edges and brighten it's center. The effect works by generating a radial gradient of the specified shape and multiplying that with the original image. The result of that is then added to the original image.

[ATTACH=CONFIG]4113[/ATTACH] [ATTACH=CONFIG]4114[/ATTACH] [ATTACH=CONFIG]4115[/ATTACH]

Scale: This determines the shape and size of the frame. Smaller values result in a more square-shaped mask. The larger the scale factor is, the more circular and wider the frame gets. This can be best demonstrated by applying the effect on a white background. These are the scale settings 0.5, 1.0 and 2.0:

[ATTACH=CONFIG]4116[/ATTACH] [ATTACH=CONFIG]4117[/ATTACH] [ATTACH=CONFIG]4118[/ATTACH]

Base Intensity: Specifies the brightness of everything that is not multiplied by the radial mask. If you set this to zero, this will result in the borders to become pitch black. If you only want to darken the borders, slightly increase this value.

Radial Intensity: Specifies the brightness of the masked image. If the vignette effect is too strong for your taste, you can reduce this value to suit your needs.

Ideally, the sum of base intensity and radial intensity should not exceed 1.0. Otherwise you risk white-saturation of the center of the image. However, this effect may also be desired. Also, if your image is rather dark, it can be a good idea to intentionally boost the center brightness a little.

[jewe.org]

Image Cook

This effect is part of IrfanView Filter Sandbox 1.3, which will be released with the next IrfanView version. It's purpose is to make an image look old and worn. The effect is composed of three effects: Color and Brightness, Noise and Dust, and Scratches and Speckles. We'll focus only on Color and Brightness for now, by setting the intensity of the other two to 0%.

Color and Brightness lets you choose saturation and brightness for the center of the image and it's borders separately.

[ATTACH=CONFIG]4119[/ATTACH] [ATTACH=CONFIG]4120[/ATTACH] [ATTACH=CONFIG]4121[/ATTACH]

Frame Size: Specifies the shape and width of the frame. The effect uses the same radial gradient as the "Radial Brighten" effect discussed above, so this parameter works in the same way. Lower values make the frame thinner and more square shaped. Higher values make the frame thicker and more circular.

Center Brightness: Specifies the brightness of the center of the image. A value of 100% means no change, lower values darken and higher values brighten the center.
Center Saturation: Specifies the color saturation of the center. Again, 100% means no change. Lower values desaturate, higher values saturate it.
Frame Brightness: Specifies the brightness for the frame. Same function as center brightness, only for the frame.
Frame Saturation: Specifies the color saturation of the frame. Same function as center saturation, only for the frame.

If you reduce the saturation around the borders of the image and add some dust and scratches, you can make any image look old and worn:

[ATTACH=CONFIG]4122[/ATTACH]

[Mij]

I hope I'm not intruding on your thread, Mij. I thought I'd contribute a few border and frame effects for the IrfanView Filter Sandbox plug-in.

Not at all. I am very pleased to see your contributions here. It seems to me that there are many valuable tools available in Irfanview that are ignored because very few people know how to use them. So I hope that your tutorials will persuade some of them to give the Sandbox plugin a try.

[rkghadai]

Masks can be created from image shapes downloaded from the internet. A good set is Harry's edges available here. http://thepluginsite.com/products/harrysedges/
Download them and put them into a read/write folder anywhere on the computer.

To use the masks in Irfanview you must resize them to match the size and shape of the image you want to apply them to and to save as a PNG file using transparency options. The PNG file is then applied to the image as an overlay.

1.To show the transparency of a PNG file used as an overlay the option "Show PNG/TIF/TGA/DDS alpha/transparent color)" in the Viewing tab of Properties/settings must be set. It is wise to check that has been done before you start.

[ATTACH=CONFIG]3656[/ATTACH]

2.Open the image file to be masked. Find the folder and choose the Harry's Edge to use as the mask. I double clicked on Star9.png and opened a new copy of Irfanview in another window on top (I have PNG files associated with Irfanview).
3.Resize the mask to equal the dimensions of the image to be masked. The image dimensions can be read below it in the Status bar.
4.Save the resized mask with "Save transparent color" selected in the PNG Save options.
5.When asked to select the transparent color click on the white area in the center.
6.Close the second copy of Irfanview showing the mask. In the first copy, showing the image, select Edit > Insert overlay/watermark image.
7.The saved PNG file has been selected as the Overlay/watermark image and the Center option set. Click Preview to see the result and if it looks good click OK.
8.Save the result.

[ATTACH=CONFIG]3657[/ATTACH] [ATTACH=CONFIG]3658[/ATTACH] [ATTACH=CONFIG]3659[/ATTACH] [ATTACH=CONFIG]3660[/ATTACH] [ATTACH=CONFIG]3661[/ATTACH] [ATTACH=CONFIG]3662[/ATTACH] [ATTACH=CONFIG]3663[/ATTACH]

The Star9 mask used above is just Black and white. The white area became completely transparent and the black completely opaque. Some of the masks also contain shades of gray which allow some areas to be only partially transparent. In this example we will use one of these, Twirl14.png.

9 Open another image in Irfanview, then open Twirl14.png in another copy just as in (2) above.
Then resize the mask to 1600x1200 pixels to match the image dimensions as in (3).
10 For this example we will have a colored mask. Open Image > Color corrections and move the Red, Blue or Green Color balance sliders to the right to get the mask the color you want.
11 This time when you save as in (4), also check Save Transparency as Alpha channel in the PNG save options. When asked to select the transparent color click on the white area as in (5) to make that fully transparent.
12 The overlay is added just as before in (6) and (7) and the result saved.

[ATTACH=CONFIG]3664[/ATTACH] [ATTACH=CONFIG]3665[/ATTACH] [ATTACH=CONFIG]3666[/ATTACH] [ATTACH=CONFIG]3667[/ATTACH]

Notes:

(a) You can also use masks bundled with other software applications such as those in the masks folder of Faststone image viewer.
(b) The purpose of opening the image in a window first, and then processing the mask in a second copy of Irfanview opened on top of it, is primarily so that the image dimesions can be read and used to resize the mask. I find it helps the workflow too. If you find it confusing either close or minimize the first copy after noting down the image dimensions.
(c) Unless you want to reuse the mask after it has been colored and resized it need only be stored temporarily until it has been used as an overlay. I save all temporary overlay files as mask.png in a Temp folder. That avoids the need to change the name and location in the overlay dialog every time.

First of all i learned a lot from this thread, so big thanks. Clearly the "transparency as alpha channel" is for gradual transparency change. Transparency of a pixel depends on how 'close' is its color to the color i choose as transparent.
Now, this works as expected for a mask with black to white transition as in the example u hv given, i.e. if i choose white(rgb=255,255,255) as transparent color, gradual shades of black will be less transparent; fully black(rgb=0,0,0) at the border will be fully opaque. In second example with white to blue transition, transparency calculation is not so straightforward; the full blue part is not completely opaque. Similar is what i see for any other random color combination: i choose color1(r1,g1,b1) as transparent color, then i get partial transparency in some unwanted region, apparently depending how much of (r1,g1,b1) is in those pixels.

Can i know the formula irfanview uses for transparency calculation? What is transparency of a pixel (r2,g2,b2) if transparent color is (r1,g1,b1)?

[Mij]

Transparency of a pixel depends on how 'close' is its color to the color i choose as transparent.
Now, this works as expected for a mask with black to white transition as in the example u hv given, i.e. if i choose white(rgb=255,255,255) as transparent color, gradual shades of black will be less transparent; fully black(rgb=0,0,0) at the border will be fully opaque. In second example with white to blue transition, transparency calculation is not so straightforward; the full blue part is not completely opaque. Similar is what i see for any other random color combination: i choose color1(r1,g1,b1) as transparent color, then i get partial transparency in some unwanted region, apparently depending how much of (r1,g1,b1) is in those pixels.

Can i know the formula irfanview uses for transparency calculation? What is transparency of a pixel (r2,g2,b2) if transparent color is (r1,g1,b1)?

I do not know the answer to what you ask. I would probably have created the alpha channel using functions already available in Irfanview such as "Convert to Grayscale" then "Negative > All channels" but it does not seem to be quite that. (In case you do not know, you can extract the Alpha layer for comparison by opening the image you saved and using "Show channel > Alpha" from the Image menu).

I have never found any use for this alpha layer creation feature other than the one I put in this thread, so the exact way in which transparency is calculated does not seem too important. Personally I think that being able to import an alpha channel that you have created yourself would be a much more useful feature.

[rkghadai]

I do not know the answer to what you ask. I would probably have created the alpha channel using functions already available in Irfanview such as "Convert to Grayscale" then "Negative > All channels" but it does not seem to be quite that. (In case you do not know, you can extract the Alpha layer for comparison by opening the image you saved and using "Show channel > Alpha" from the Image menu).

I have never found any use for this alpha layer creation feature other than the one I put in this thread, so the exact way in which transparency is calculated does not seem too important. Personally I think that being able to import an alpha channel that you have created yourself would be a much more useful feature.

as i said my problem is i am getting (partial) transparency in regions where i dont want. Knowing the formula of alpha transparency would help me make educated guess of which colour to choose as transparent colour.

What i am trying is to do editing on non-rectangular selection in irfanview. [I know better programs exist for that, but nothing better if such a fast program like irfanview can do that.] So i create the selection in paint or irfanpaint, edit the selection as desired, make a mask with the rest of the image (outside the selection) transparent, add this mask as "insert overlay image.." on top of original image.
What i want: edited selection part comes bcoz of fully opaque part of mask; rest of original image comes bcoz of full transparency.
Using transparency mask without alpha channel gives a hard edge at the selection boundary; so i use transparency mask with alpha channel, so that the transition is smooth.
What i get: edited selection part comes but with some of the underneath original image visible, indicating partial transparency there; rest of original image comes bcoz of full transparency.

So i gather, the issue is which colour to choose as transparent colour that would result in near full opaqueness in a desired region.