Watermark question

[Bernard]

This has been discussed several times in this forum, but I could not find the answer I am looking for.

How do I add a watermark to several images in an automated workflow, when all images have a different size ?
I have read the text effects white paper, which describes usage of a mask, but afaik a mask must have the same size as the base image.

I would like to be able to add a watermark to several 'standard' relative positions, like center, bottom right, bottom center, bottom left, and intermediates ie between center and bottom center....

Thanks,
Bernard

[den]

PWP5's workflow 'Text' widget seems to do most of what you describe without the use of a mask, i.e., ...scale and locate semi-transparent texts with or without drop shadows surrounded by semi-transparent contrasting background text boxes for differing image overall pixel dimensions and orientations [portrait or landscape]... but you will have to re-group your images into sets for the desired "'standard' relative" watermark/signature positions... at least that is what my brief experiments seem to indicate.

[Charles2]

Take a look at the free program Fast Stone, specifically Tools | Batch Convert | Advanced Options | Watermark , including choice of position.

[ksinkel]

A number of features in the text widget are designed to make text positioning insensitive to image size and shape. You can specify any of the corners and the top and bottom centers as the reference point for positioning the text. Choose the point closest to your text as the reference. You can also choose to scale your text to the image. By using these features you can make your text look very similar regardless of image size.

As far as mask size is concerned, during execution the workflow automatically resizes the mask to the size of each image. Thus a single mask will work with images of mixed sizes.

Kiril

[Dieter Mayr]

As far as mask size is concerned, during execution the workflow automatically resizes the mask to the size of each image. Thus a single mask will work with images of mixed sizes.

Does this only work for images of the same w/h ratio ?
And if not, how images with different w/h ratios are handled with respect of positioning the mask ?

[tomczak]

Be aware that FastStone Viewer can read 48bit TIFFs but saves them as 24bit ones. Also, batch processing in FS will strip the ICC profiles.

[ksinkel]

Does this only work for images of the same w/h ratio ?
And if not, how images with different w/h ratios are handled with respect of positioning the mask ?

When you create a workflow, you use a 'training' image to set all your parameters. The mask used at that time must be the same size as the training image.

During execution, this original mask is resized to the exact dimensions of each image, if necessary. Once it is resized, there is no problem in positioning it -- it is pixel for pixel the size of the image. Of course the proportions of the mask will change if the proportions of an image are different from those of the training image.

[den]

Benard, Kiril, DieterM...

(1) I developed two general purpose workflows, landscape4x6r and portrait6x4r, for placing a watermark in the top/bottom left, center, and top/bottom right 6th grid crossing image areas of two, 2304x3456 and 3456x2304, training images.

(2) These general purpose workflows work well within certain limitages of scaling and proportions from their training image design. I found it necessary two have two workflows as the scaling and proportions between landscape and portrait orientations varied too widely from each other to have preference appearances and were apparantly outside programing considerations as well.

(3) When applying these general purpose workflows to a large group of images, they worked best if the images were seperated into landscape or portrait sets... then each set processed with the appropriate workflow with its training image as the first image of the workflow's selected images to be processed.

The processed landscape training image:

watermark1-den-1_LO-2.jpg

If there is further interest, I can make the workflows and trainning images available for use and experimentation. The workflow specifics [text, opacity, fonts, etc] can be changed to your own preferences.

[Bernard]

Kiril, Dieter,Den,....
Thanks everybody for these ideas; just coming back from Paris, I will need a bit of time to experiment .
I am buiding a web site, and want to display my images in large dimensions, therefore the need for a good watermark workflow.
Bernard

[den]

Benard... ...Others

While experimenting with the Text widget, I discovered that a double widget application in the watermark workflow for a given text location: ...where the first widget is set to "Bold" with a tone/color compliment to the preference font tone/color, followed by the second widget with the "Bold" off and the preference font tone/color... ...will provide a more interesting watermark text-font combination... than a single widget application...

An emphasized font to image content ratio for illustration purposes:

font_outline.jpg

---Edit1---
In the illustration above, the 75% transparent black Mistral font with drop shadow [black: A=default; O=10; T=50; S=12.5] is underlayed with a 75% tranparent Bold white Mistral font with its drop shadow [black: A=default; O=10; T=50; S=12.5]. Similar results can be achieved with other fonts. Mistral happens to be my chosen 'marquee' for "IMAGEs by DEN".