So, I have read most of the FAQ but have a few questions before loading the photos and starting with photostructure. The photos are on a range of devices (iPhone, USB drives, folders on computer h/d, etc.), there is no naming structure, a lot of duplicates will exist and some photos (those that are scanned from prints) will have incorrect meta-data. Is it best, in this situation, to just load all the photos into a single folder or is there benefit in categorizing the photos into sub-folders; e.g., by year / album? And how do I deal with the fact that there might be photos from different devices with the same name?
There are ways around this, and future versions will make this fixable from within PhotoStructure, but for now, if you use a tool like ExifTool to clear incorrect metadata records, and place the assets into somewhat-correct year-month or year-month-day subdirectories, PhotoStructure will figure it out.
If you have the free disk space, I’d just set everything to automatic (auto organization and auto scanning), and let 'er rip.
When you enable “automatic organization,” PhotoStructure copies unique files into date-stamped subdirectories, which means you really only need to worry about photos from different devices with the same name taken on the same day, but know that PhotoStructure handles this case by appending a “-1” suffix to the name of the file. As an example, if you had several files taken today, all named IMG_0001.JPG, they’d get copied into $your_photostructure_library/2020/2020-12-28/. The first file imported would be named IMG_0001.JPG, the second IMG_0001-1.JPG, and so on.
Your PhotoStructure library database retains the source of each asset file, so within the asset information panel (click on an asset thumbnail, and tap the “i”) you’ll see the original folder and filename.
Much appreciated. I’ve run the Photostructure process with the results outputted to a mapped drive that corresponds to a directory on my (Synology) NAS. The process appears to have worked with images organized by year / month / day / image_#. One, significant, issue a bunch of photos taken across different years have been placed into the 2000-01-01 folder because the camera date | time wasn’t set so all photos defaulted to this date. Similarly, as expected, photos that have come from scans are listed in the year / month / day of the scan as opposed to the year / month / day that the photo was taken. Any suggestions or tools on how to address this situation?
Brilliant. So, to be clear, I use EXIF to correct the date of the photos that were output from the initial Photostructure process and then re-run the process with the original output folder used as the input?
Note that you need to provide “YYYY-MM-DD HH:MM:SS”). Note that ExifTool supports “0” for months and days, but I just found this out, so I need to change PhotoStructure to support unset months and days. Until then, just use the first of the month if you don’t know the day, or January if you don’t know the month.
If you’d rather just put images into yyyy-mm directories, use this command to delete the date tags from your files (-AllDates followed by an equal sign, then a space.)
Say I’ve already imported a bunch of files using “automatic organization” and they went into the wrong folder. If I modify the date of a photo then rebuild the library, will PhotoStructure move its copy of the photo into the right folder? I just discovered that all my wedding photos have an EXIF date in 2014 even though I got married in 2018… I guess the photographers didn’t have their camera date set correctly
Not with v0.9.1: once a file is copied into your library, it doesn’t get moved around.
People commonly rename directories to include an event name, and I didn’t want to undo that work.
However, perhaps there should be a setting that says “it’s ok to move files found in datestamped directories that don’t match the captured-at time”?
(Feel free to brainstorm nicer/more robust heuristics, and to add this as a feature request so others can vote on it: I’d vote for it!)
