keywords, tags, and Google

jimh

Which set of keywords (tags) does Google see: the IPTC keywords in my uploaded files, or the SM "tags"?

There seems to be a problem with SM "tags" and I don't expect it to be fixed soon. Specifically, for some files, SM combines the IPTC keywords into one big tag.
I wish I could just ignore the messed-up SM tags, because I don't need keyword search myself, but if Google sees them when it crawls the page, it won't index the photo correctly.

leftquark

do you have any examples you can share with us? Google is a black box -- they might be downloading the file and looking at the metadata within it, or more likely they just look at the HTML of the page and use whatever tags are there, plus any surrounding text

jimh

I've already been in touch with support, they looked into it told me a bug ticket had been created. That was a week ago. I assume it's ended up on a list of bugs to be fixed... someday.

I've tried numerous keywording tools, and as far as I can see, there's nothing wrong with my keywords - or at least nothing I can fix. Other sites parse them fine. This same problem is mentioned in some old threads here.

I have many images where the IPTC keywords have been combined, here's one:
dhttps://gallery.jimhphoto.com/Minneapolis-photography/i-PLMzB89/A

At this point I'm wondering if I should actually start "fixing" the SM tags themselves, using the SM UI. That's why I'm asking if Google sees them, or if they're just for internal search.

Allen

They all show here. You can easiely pick out what to fix.
https://gallery.jimhphoto.com/keyword

jimh

Thanks Allen.

Allen

I use my page like that to review every so often looking for errors. Did it day before yesterday and found some.
I have an unlisted gallery set to Smart Gallery. Great for bulk changing KW's.
Also open Smart Gallery in Chrome to use the bulk caption/keyword app.
https://dgrin.com/discussion/247592/new-bulk-title-caption-keyword-editor-extension-now-available/p1

jimh

That bulk editor could be helpful too. I'll have to find out what a Smart Gallery is.

Really, it would be so much nicer if SM would just fix their parsing of IPTC as that seems like bedrock functionality for a gallery site. Sorry, I just had to vent a bit of frustration. Now I'll deal with the problem.

Allen

Smart Gallery is just a regular gallery. In the gallery settings click "Smart Rules". Many ways to filter photos.
I use mine as a utility so it's set as unlisted.

jimh

Well it got worse.

I sell prints on a couple of PODs and I've known for a long time that my poor keywording was hurting my sales. I'm in the process of redoing it on all my local files, and updating the sites where I sell - including SM.

Even if I delete my files here and re-upload them, SM keeps all the old keywords (tags) unchanged, it doesn't even look at the new ones. I'm guessing that's because the file names haven't changed. So I'm faced with fixing up every image here by hand, and that's over 600.

Really, at this point, it would be easier to just leave SM and start over somewhere else. Unfortunately I just recently paid up for another year here. So I don't know what I'll do. My time is worth something, too.

leftquark

Jim: Do you have any examples? What tool are you using to keyword your photos? One of the changes I made while still at SmugMug over a year ago was to ensure that "replacing" a photo would re-query the metadata and pull out any new title, caption or keyword. So if you upload new photos it should be pulling the new keywords. SmugMug uses the industry standard "ExifTool" to pull out all the metadata.

(The exception was if you updated the title, caption or keyword on SmugMug first, then we didn't want to overwrite your smgumug made changes, and the new file will have the metadata ignored).

jimh

leftquark, I've found one possible source of the problem. I've used various keywording tools and in the past, one of them (some Adobe product no doubt) put in tags as embedded XMP. Those keywords are obsolete, but it seems SM uses them in preference to the newer IPTC keywords. Other sites that I use just ignore the XMP.

I can't easily strip the XMP from hundreds of files - maybe some sort of command-line invocation of ExifTool might do it. Any way I could get SM to ignore XMP?

But even if I fix that problem, SM still won't parse the individual keywords out of my IPTC.

jimh

Doubt anyone is following this thread, but I don't like to leave things unresolved. I think I've solved this problem - it was a long road. The crux of it seems to be that my files contained redundant keywords because most tools write IPTC but Adobe, in their infinite wisdom, writes XMP. And most programs, and web sites, ignore the XMP so you never know it's there, unless you use something like EXIFTool. SmugMug is an outlier: if XMP is present, it reads those keywords and ignores IPTC. Or, in my case, it sometimes reads them and sometimes does something odd and mashes all the keywords into one tag.

I used a command line for EXIFtool to remove all the XMP data from all my files. Then I deleted all my photos on SM - a gallery at a time - and uploaded again. That seems to have worked.

I think there were actually 2 problems. One is that SM goes with the XMP over the IPTC - that's not a bug, just a feature I didn't want. But there also seems to be a bug because in this situation the parsing seems to often go wrong, resulting in one tag containing all the keywords.

Djm3006

thanks for the info Jim