1. Carl Seibert says:

    Breaking news: It turns out you CAN do a NOT search in Lightroom!

    A viewer on my YouTube channel left a comment wherein he explained that you can use a Smart Collection as a workaround for the missing Filter or Find functionality. And here is my recap of his comment:

    *Go to Collections and make a new smart collection. You can reuse it, so you can name it whatever you want. I called mine “AA_NOT_SEARCH”.

    Smart collections use the same filters that you have in the Filter or Find interfaces, but there’s a little plus sign at the right that you click to add additional filter statements.

    *Just click the plus and add a second line.

    *Choose “Match all of the follow rules” in the pulldown at the top.

    *Use the pulldowns to make your first line be your main search, like “Caption contains words [some search terms]” (Note that “Contains” puts an OR operator between search terms, like “Contains Any Of”. “Contains Words” or “Contains All” are probably what you are looking for.)

    *Put your NOT statement in the second line. ” Keywords doesn’t contain [some keyword]”.

    *Click ‘Save’. And you’re good to go.

    Note that both the Caption and Keywords fields in the first column are under the ‘Other Metadata’ flyout.

    *To reuse the Smart Collection, right-click on it and choose ‘Edit Smart Collection’.

  2. Todd Poole says:

    Hi Carl,

    As a subscriber to David Riecks controlled vocabulary, do you suggest adding/appending my newly created and thought out keywords to that database or keeping them separate? I shoot high school and college sports in the Burlington Vermont (and outlining) area. I would also include sports portraits, team photography and event photography as a description of what I shoot.
    Davids’ list only contains a half dozen cities/towns in Vermont (oops, now just remembering to use the location fields, not keywords, for locations). I do think that School names, team names are important metadata for my use and others. So would those keywords and appropriate child keywords be added to the list?

    • Carl Seibert says:

      Hi Todd,

      What I would do (your mileage may vary) is start my hierarchy, or hierarchies, at the top level in the left column, separate from your existing ControlledVocabulary one(s). I would backup (Save) everything, then Merge in my own file(s), then make final edits with the interface in the Structured Keywords panel. Or, you could build your hierarchies from scratch, right there in the panel. Then backup again. (I’m assuming for some reason you’re using Photo Mechanic here. If you’re using Lightroom, you’ll need to translate as we go along. But the principles are the same.)

      I would think the advantages of doing it that way would be that it’s way easier to build a keyword file if it’s to start at the top level, rather than to make it fit into the middle of an existing hierarchy, and I would think navigation would be easier with fewer levels. (As you can see in the illustrations in this post, I did actually build my own this way.)

      If you use the Find function, it doesn’t matter where the keyword you seek might be; you get to choose in the Find dialog. But if you navigate to the keyword, then whatever path makes the most sense for you is what you want to build. (You might start at “`MyKeywords”, or you might have one hierarchy for high schools and another for colleges, or one for golf and one for football, for example.)

      In your case, place names are commingled with institutional names. “Burlington Central High School” (I made that up) is certainly keyword material. But they might play a game in, say, the CapitalOne Arena in Washington, D.C. In which case, you have a place name that isn’t the location of the photo. So, you’re probably stuck with places in your keywords, one way or the other. Depending on software, “Burlington” might or might not return a photo with that keyword. If you want to be sure that “Burlington” does hit that photo, you would do “Burlington, Central, High, School,”

      (Of course, school names (hopefully conforming to style) and playing venues will probably be in your captions, too.)

      Now that I’ve brought it up, there’s the matter of “High School”. It’s a style issue, for sure. Your style is probably “High School”, but if there are a few “H.S.” entries, a synonym would help. And “high school” is a category of pictures and would most certainly be denoted in the keywords. But “High School” may or may not be part of the name of the school. “Acme Prep”, say. So, “high school” should probably be a keyword of its own. Same for “college”.

      So many fiddly decisions go into a keywording plan.

      By the way, if you are using Photo Mechanic, the location fields allow you to chose from a list. I built mine from Census data. (South Florida has about 100 municipalities. Nobody knows the correct names of all of them.)

      I hope this helps.


      • Todd Poole says:

        Thanks Carl! And yes, I am using PM for a few months now. LR and PS are my down stream editors (in my work flow) respectively. Using PM for culling and exclusive meta.data.adder….

  3. Cami says:

    Referring XnView, how many keywords / max. characters can be added to a image, not affecting, without a performance penalty?
    Thank you.

    • Carl Seibert says:

      In the old IIM standard (and I can’t lay my hands on my copy right this second) Each individual keyword was limited to 64 characters. I don’t recall if there was a limit to the total size of the keyword field. In the XMP, there are no limits.

      If by performance you mean search/Digital Asset Management performance, I don’t think it really matters for XnView, because XnView doesn’t really have much in the way of search or DAM functionality.

      In terms of XnView writing metadata, considering how tiny metadata is comparing to the image itself, I can’t imagine noticing a difference one way or the other.

      If you mean performance as in page load times on the web, each character is a byte, basically. (well, two, if you consider that most fields, like Keywords, appear in both the IIM and XMP.) For most images, the IIM and XMP data blocks are four to eight KB or less, combined, unless an absolutely outrageous amount of metadata is there. The Exif can be a little bloated if it contains over-large previews. Altogether, we’re talking about only a few milliseconds of load time with most normal bandwidth, something on the order of well under a tenth of a second, usually.

  4. Keith Bamber says:

    I just wanted to say how valuable I found your blog and video. (only the first of the latter sofar). I am only an enthusiast but it is by far the best introduction I have read in many years that I have studied and practiced photography. Thank you.

  5. I loved your article, very succinct, well thought out, and cleverly puned. Like keywords. Hmm, imagine that?

    I am just getting started on stock photography, and have had several images accepted by Adobe stock. A real greenhorn. I have been researching for weeks about keywords, and this is by far the most detailed explanation of how to (and how NOT to) use keywords. You went beyond the rinse and repeat tips and tricks of the common rabble and presented real information with associated thoughts and rationales.

    Thank you.

  6. Maritza says:

    As always, thanks. After months, or maybe years, struggling with the best way to tag my pictures -well, not “my” pictures, but pictures that I collect, just because they’re pretty and I like to look at them. Probably what you’d call “meaningless stock pictures” but then I’m no photographer, but I enjoy beauty- I had already discovered some of these tips by myself, but there are many that would have taken me many a month more.

    It never ceases to amaze me how you manage to make your posts funny and a really entertaining reading, while writing about metadata, no less.

    I know this blog is meant for photographers, not to the likes of me, but I’ve found very useful information here to help me with my hobby, and I’m very grateful for that.

  7. Robert Hendricks says:


    I am using Photo Mechanic Plus for managing my metadata for a collection of about 35,000 images. I find your discussion of Keywording most helpful, but have a few questions with regard to PM+.

    With the exception of the Keyword entry, all of the data entry pull-down menus for the searchable IPTC fields are flat (or term) files; none support hierarchical lists. Of the searchable data entry fields (except Keywords), only Title/Object and Headline are logical entries for a description of the subject of the image. The Description/Caption is not searchable. Thus, Keywords appears to be the only choice for a structured, hierarchical keyword subject list.

    Let’s say you want to enter information about flowers and horticulture in a structured list as in Keywords. Now say you also have images that depict animals, you use a different hierarchical list, etc., etc. When you look at the Browse menu in the new cataloging system, all of the entries from these different sources are sorted alphabetically starting with the first character in a multiword string unless you insert the entire path in the hierarchy.

    This may be OK if you are preparing a list of keywords for an agency or for others who may look at and search for your (really good) images, but if the goal is have keywords that are strictly for your own convenience is searching for images in your collection, this gets to be more than cumbersome.

    What to do?

    The Title/Object name might be a good place for the descriptor of the subject, but you would have to build a flat term list. But still, your animals and your flowers (to pick only a couple of subjects) get mixed together. Of course, you could always use a prefix descriptor to sort them out, such as animals>mammals>bears and plants>genus>species. A controlled vocabulary would be critical in keeping this all straight.

    What am I missing in this discussion? My goal is for my own use. I am not a pro, and I do not sell images to agencies, but I do produce TV shows and/or slide shows for public audiences and sometimes my presentations get put on someone else web site.

    You comments would be most helpful.

    Many thanks,

    Bob Hendricks

    • Carl Seibert says:

      Check out my review/How-To for Photo Mechanic Plus. It’s one of the more recent posts, so it shouldn’t be too hard to find. Or just use the search box to search for “photo mechanic Plus”.

      The Caption most certainly is searchable. It’s the most important field. It’s where the specific description of what’s in and what’s happening in the picture goes. (In some narrow areas, like stock photos -which often aren’t “real” to begin with – the keywords can be the more important field.) In PM+, next to the search box, in the “star” pulldown, there’s a list of all the searchable fields and what they’re called in PM+ (Which is very handy if you want to do fancy searches against specific fields.)

      Keywords are flat by nature. Once an asset is at the point of searching for it, there is no such thing as a hierarchical keyword. There’s just a bunch of potential search terms in a bucket. It doesn’t matter how they got there. If you search and hit one (or three, or however many), that asset comes up in the return. The whole hierarchical keywords thing is only applicable to the process of assigning the keywords in the first place. Photo Mechanic has a dandy tool for hierarchical keywording. (Structured Keywords, they call it.) Adobe muddied the water by coming up with an overly complicated (and dangerous, IMHO) scheme to allow for hierarchical edits of keywords after the fact. But that’s only for editing. Once you get to the search phase, Adobe’s keywords are just as flat as everybody else’s. (And FWIW, Photo Mechanic has a method for editing already assigned keywords that, in my opinion, is safer because it makes you work harder to mess things up.

      If you do have a case where a user might need to see a taxonomical string, that needs to go in the caption. It’s bad practice to assume the human beings will ever open and root through the keywords field. Check out my post on the Statue of Liberty postage stamp debacle for an illustration of why this is so.

      Title/Object Name and Headline are interesting fields because they are used in very different ways by different organizations. I have written about them before. The standard allows sufficient leeway in the definitions of those fields that this is OK. Most photographers are perfectly safe not bothering with them. If you contribute to a certain organization and they want you to use those fields in some certain way, they will certainly speak up. For most photographers, the caption contains – or should contain – any information that might be headline or object name-worthy.

      BTW, You can configure the per-field pulldowns in the IPTC Editor (and Template Editor) to work as controlled vocabularies by going to Edit>Settings>Auto Complete and ticking “…. only allow completions from the field’s own list”. Once you have more than a few terms for a given field, the pulldown functionality gets a little wonky. But I do this for Keywords, for when I don’t want to drag out the big Structured Keywords tool.

      I keep threatening to write a stand-alone post on how to write a good caption. Until then, there’s a lot on caption writing in my post on Creative Commons, and there’s a very old post on what each IPTC field is for. I’m sure if I re-did that one, it would be better. But it may be worth a look.

      I hope this helps.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.