For those who don't know what I'm talking about...
IPTC metatags allow you to embed textual information into your jpeg files. You can embed right into your image files the title, description, keywords and even categories in some cases.
For those who know, but never used IPTC tags...
Specially if you upload to more than one agency, embeding your photo title, description and keywords into jpeg image files is most useful and a huge time saver. Most websites will read it pronto and submission forms suddenly are filled
without need for any extra work but a button click (and eventually category selection)
For those who use it but things don't work out very well...
... and this is the main theme of this article because it also happens to me and it's a little... annoying.
First, the IPTC fields you must use are "Title" (obvious), "Caption/Description" (not that obvious because you can find more than one that can fit, but only one works) and "Keywords". This one is also obvious and you should create one keyword field for each keyword.
Well, if you stuff some keywords on the same field separated by "," it also works when all you wan't is to fill your submission forms. Since I don't have anyother use for IPTC I do it that way and end up with only 3 ou 4 keyword fields for 10 to 20 keywords. But you should not do it that way, are we clear? ;)
Now, the problem comes when you do it all correctly and still it doesn't work!
I usually work my pics in photoshop and then add IPTC tags using another little program, photoME, if I'm not mistaken.
"Normal" single files taken right out of the raw file into PS, retouched and saved as jpeg usualy represent no problem.
If I need to do some work that requires creating new layers, things start getting messy when using IPTC.
And then comes panoramas. I set the software to create panoramas as to always create separate layers for each source image so that I can correct potencial stitching errors. And almost always IPTC tags would not work when submiting final images!
I asked some websites about it, including DT, and the answer was always something like "sometimes our engine is not capable of reading IPTC". Not the perfect answer.
And then, by accident I found the reason, and it is within photoshop. You absolutely need to flatten your image before saving the final jpeg.
My PS workflow has two final steps that are converting colorspace and changing to 8bits. This clears the file to only one layer also, but is not the same as flattening.
If I save panoramas, or other several layers image, (adjustment layers don't seem to count) as jpeg without flattening, even if all IPTC data is added correctly, most webagencies won't be able to read that IPTC.
I suspect the problem is related with the descriptio field because there's one agency that don't use that field and I get more functional IPTC data over there.
But the trick for me is, flatten the image before saving the final jpeg.
I hope you find this useful, I upload many panoramas and I've wasted too much time because of this!