ACDSee Database Primer

Are you confused with cataloging, embedding, ITPC vs. ACDSee keywords, converting a database and more? I’m going to try to take away some of the mystery. Given I’m a database-kinda-guy, it may be difficult to keep it simple, but I’ll try.

This will be explanation by example, and weighted towards understanding what happens when you need to get going on a new version of ACDSee. Spoiler alert: When you have a new version installed, Go to: Tools | Database | Convert Database… this actually converts the old database and puts it into the empty database that is running in your new version. Not the most intuitive naming in the world, but that’s what it does. That’s getting ahead myself, but if that’s all you came here for, you’re welcome.

Overview

l started in ACDSee 2023 with a new, empty database and created some folders with test images in them. Then I added metadata to some images and processed each folder with specific database tools. This article starts at this point and shows you what I started with.

Then in ACDSee 2024 I switched to a new empty database and show you how the data changes in the database: initially, after cataloging, and after doing a Database Convert on the ACDSee 2023 database.

Along the way, I tried to dig a bit deeper where that is useful, and point out places where what you might expect may not be what happens. I should let you know that in the video I made before this article and in re-creating the environment for this writing, I found it impossible to get from beginning to end in “one take” – I had to go back and redo some process because of something I’d missed. (For instance, one of the folders is intended to contain images that did not have the data embedded. However, while it started that way, it did get embedded because my options in ACDSee 2024 specify embedding on close-down. So the next day I did not get the result I expected, and I had to go back and redo some screenshots. If you notice some continuity errors, that’s why. If you find any errors more serious that that, please let me know so I can address them.

Demonstration set-up

Rather than trying to demonstrate how things work with a database with 40K items, I created a new database in ACDSee 2023. Then I set up a test folder with four subfolders, each with the same six photos in them. (Well, one is a bit different, but we’ll get to that.) This means that the database only “knows about” these 26 images. The other forty thousand images are still on my disk, of course, and another database has all of the cataloging and everything else, but as far as the test database is concerned these are the only images that exist. For instance, when I list categories, only the standard ACDSee categories and the ones I set up in this test show.

Here’s what the test images look like:
(If you click on the image (or, come to that, any image) you can see it larger.)

Both the names of the files and the sub-folders indicate how they have been processed. One folder (red thumbnail) has no metadata at all, all the others have the bird, cat, rabbit and and moon with appropriate categories, and bird, cat and rabbit keywords. One folder (blue) has nothing embedded, the yellow and green thumbnail folders has everything embedded. Except for the red folder, every image has a label attached, green, blue and yellow. The “Labels” metadata is used to easily identify which folder an image is coming from – for instance all the images in the “Other” folder are labelled yellow, and a thumbnail is included which consists of an essentially yellow image.

The “Other” folder has some edge cases:

• The moon image has an IPTC keyword (“MoonIPTC”) added
• There is a RAW image of a frog in this folder only
• The frog has an ACDSee keyword (“Frog”) added
• The frog has an IPTC keyword (“FrogIPTC”) added
• The Folder thumbnail has a category of “Various” and a keyword “Fall” plus an IPTC Description, Keyword (“TestIPTC”), Category “Human Interest”, has a rating of 2, is tagged, and has an ACDSee caption and an ACDSee Notes entry.
• The sidecar file (extension “xmp”) created for the metadata belonging to the frog image was purposely deleted outside of ACDSee
• The rabbit image was purposely deleted outside of ACDSee
• All of this is embedded.

What does “embedding metadata” mean in ACDSee?

Metadata means “data about data”. ACDSee allows you to add extra data about your images. The image is the data; things like the rating, keywords, categories, labels, and even a list of actions you took to change the appearance of your image are metadata: data concerning your image.

The confusing thing about all of this is how and where it is kept, and which pieces of data are kept in a proprietary format and which are kept in a place and form where any program can find and use it. The nice thing is that if you only use ACDSee then you don’t need to care, it just takes care of it. The other simple concept is that (so far as I am aware) ALL of the metadata created in ACDSee is kept in the ACDSee database. Data in a database is very easy (and fast) to search, sort and select. Imagine if you wanted to see all of your images which had a rating of “4”. Without a database, any program would have to read every image file (and many of these would be megabyte files) before it could start selecting the 4’s.

So we have a database that keeps all the metadata and also a thumbnail which is simpler and quicker to display. But what happens if the database (which you didn’t get around to backing up yet) gets damaged, accidentally deleted, encrypted by malware, or lost in a move to another PC? Many people (including me) choose to play safe and have ACDSee periodically check its database to see which metadata is only in the database, and make a copy of it in or near the image file. ACDSee calls this embedding.

You need to turn on automatic embedding, or manually embed metadata for selected files only, which is what I did for this tutorial; you can see the steps in the screen shots below, starting at Tools | Metadata | Embed ACDSee metadata. Or you can set it up to do so with varying degrees of automation (see Tools | Options | Embed ACDSee Metadata). ACDSee 2024 makes this particularly painless.

Note that it is a proprietary format so other image editors or photo managers cannot read most of it.

If the database is not available, a new database can be created from the embedded data. Or if you have a backup database which is not up to date, embedded data can make up for it.

Where the image file format is not designed to contain optional metadata (for instance RAW files), ACDSee saves metadata (and develop “instructions”) in what they call a sidecar file with an xmp extension. We’ll see in the next section that programs that can find keywords stored in IPTC-standard format can display this information, but cannot read the same information when it’s in an xmp file.

By the way, ACDSee hides all the xmp files by default, so you never see them unless you take a peek with a file explorer.

What happens when you delete a sidecar (xmp) file?

The screen shot above was created before deleting the xmp file. It shows that while the IPTC keyword shows up for the moon (a jpg file), nothing appears for the moon image, because that information is stored in proprietary format within the sidecar file, or xmp. Deleting this xmp file (NOT recommended) will be instructive when we see how ACDSee deals with it. I’ll come back to this when I get to moving the data to a new ACDSee version.

What happens when you delete an image (outside of ACDSee)

We can see what ACDSee does about a deleted image file without waiting to transfer the database information to a new version. I deleted the rabbit file from the “Other” Folder, using an external file manager. So the database “thinks” it’s still there, and when I did a search for “rabbit” it presented four images, because there is still a thumbnail in the database. However, when I double-click on it, it needs the real image file to show a full-size image and it’s not there. This is what I got:

You can see details of all of this set-up work in the screenshots that I’ve included in the gallery below (if you want to). One of the things I show below is loading every kind of metadata I could think of on the Yellow thumbnail, so I could see the results. Remember that you can see any of these full size by clicking on it.

OK, wow, that’s quite a lot of set-up. I’m hoping you may find some new ideas in looking at these – for instance, I hid various parts of the interface to get a more concise screen shot, and grouped the images by folder name to get all of the images to show in an organized fashion.

Bringing database information into another version or instance of ACDSee

This is the stage you’d be at if you had just installed a new version of ACDSee – or the same version on a new computer. I didn’t want to re-install on my PC, but opening an empty database (File | Database | New…) simulates the initial empty database which accompanies a new install very well. Here’s what the test folders look like before doing anything special, just looking at them in Manage mode.

Just looking at a folder in manage mode loads some metadata

You can see that labels and ratings already show up. There’s another screen shot (see the gallery below if you’re interested) showing no keywords or categories are associated with any of these files yet. Looking at the Yellow thumbnail in the Other folder, we can see that IPTC metadata is also present, which makes sense because we know it’s present within jpg files. Just browsing to the folder has added that info to the database.

The frog didn’t fare as well – at this point the database knows very little about this image, because I deleted its xmp file.

Before we move on, I’m adding an IPTC keyword “New IPTC” to the Yellow thumbnail, to see what happens later when we bring in the 2023 database. Spoiler alert: it gets overwritten by the data for the same image.

Cataloging the test folders

Now I’m going to select the top folder in our test and select Tools | Database | Catalog Files… (I thought I had screenshots of this but they aren’t there, sorry. But it’s not complicated.)

In the embedded folder, categories and keywords are now showing up.

Looking at the Not Embedded (Blue label) files, I do see my face being recognized, but I realize now that I should have given myself a different name in this test, because it’s not clear whether that information came from the file itself or from facial recognition and ACDSee somehow getting the information from all the other “me” photos on the disk with FR embedded. I suspect the former, but I could be wrong. Bad test design. You can see that the database “knows about” more categories, but none are assigned in this folder.

In the Other folder, which did have data embedded, the frog still has no connections to categories or keywords. Catalog looks at embedded data in the file, and in this case, the file with the data (an .xmp file) was deleted, so the frog is still data-less, except that AI keywords were assigned.

So if all your files have up-to-date embedded data, you can re-create your database.

Other menu options under Tools | Database

Before we get to Convert Database, let’s take a quick look at other options which one might be tempted to try:

• Import – this command only works to bring in an existing database created on the same version of ACDSee. Use this for setting up a new computer, or images from another party who has the same version of ACDSee
• Restore Database – this is to restore a database from a backup

Convert Database

The wizard (Tools | Database | Convert Database… ) suggests a database to convert, but does not provide a list of all possible databases. For instance, it did not show the full database that I still have for 2023. I have screen shots of all the steps in the gallery below. For my trivial test database, conversion was practically instantaneous, but my full database took a while – not multiple hours, but longer than a coffee break. Much of this time appears to be taken with optimizing the new database.

After running this, the “Not embedded” folder now has keywords and categories. The frog finally has the keyword “frog”, and its yellow label. This really is the way to go.

Convert Database replaces metadata for the same file

Somewhere in ACDSee I’m pretty sure I read that “Convert Database” will replace metadata in the new database if it has metadata in the converted database for that same file. Remember the “New IPTC” keyword I added to the Yellow thumbnail? After running Convert Database, it was no longer there.

Screen shots on 2024 version

Upgrading from ACDSee 2023? Convert Database is the answer

Soon after installing a new version of ACDSee, I’d recommend running Convert Database. I’d love to see some code in ACDSee which, right after installation, looks for a previous database and asks if you want to use it. And then starts Convert Database for you> Until we get that, do it yourself before you start trying out nifty new features, because, depending on what you do, you might lose all or some of that new-version work when you do get around to converting.

You should also decide if you want AI keywords, because if the options are not turned off, (Tools | Options | AI keywords) you will find it chugging through all your photos. It’s remarkably fast, but may not be what you want it to do right then.

Moving the database to another location

The default location for an ACDSee database is in C:\Users\xx\AppData\Local\ACD Systems\Catalogs (where xx is your user name). I have two reasons why I’d prefer not to leave it there:

• I have a system drive and a data drive (among others) and the system drive is an SSD or solid state drive. Appdata is always on the system drive, and I’d prefer to keep user data elsewhere. I don’t want to run out of space on my relatively small SSD just because I have too many thumbnails.
• If I ever need to re-install Windows, I’d expect to have to re-install all my applications, but I do not expect to lose my data. I’d like to think that I can physically remove my D: drive put it in another PC and (providing the same applications are installed) have all my data available.

So my workflow for installing a new version of ACDSee includes creating a new empty ACDSee database on my D: drive as one of the early steps, then when I convert the data from a previous version, the new database is where I want it, not hiding in the depths of the Windows operating system. I also delete the database ACDSee created in Appdata, so there’s no confusion later on.

Video version on YouTube

Before I wrote this article, I recorded a similar process, but with less detail. That video is available below:

Secret Bonus: Early thoughts on AI Keywords in ACDSee 2024

• It doesn’t take long to process, although it might be a little rough on your HDD – I could hear it working.
• It gets keywords right most of the time – probably over 90%
• However it looks at everything on the image not just the main subject. So a car in the background registers.
• Future enhancements might included cross referencing with face recognition and omit “Car” (for instance) if a recognized face was also there.
• I haven’t found a way (and I haven’t really tried hard either) to turn off some keywords. For instance, I don’t care about fashion, clothing, shoes, jewelry etc.
• So the effect is that each image gets far more keywords than I would choose to include. For example: Animal | Bird | Duck and Nature | Body of water. Five keywords and “Duck” would do it for me. Except that it was a cormorant. (That last was a cheap shot, sorry. I don’t expect it to correctly identify every species, but maybe they should have gone with a more generic “Swimming bird”.)
• Stock photographers might find this speeds their workflow considerably, as they presumably want every possible keyword.
• I haven’t found much use for it, but it might come in useful occasionally. I’d want to see “AI keyword” in the select images menu for it to be really useful. When I’m processing a lot of new photos, my last step is putting similar subjects in the same folder, and that would be useful.
• For me, I’ve used exclusively categories, no keywords, so I’d have to do an extensive conversion, or collect all the “ducks” and add my preferred category.