Lightroom 4 – A Mixed Blessing?

Adobe has just released the latest version of their Swiss Army knife for digital photography: Lightroom 4.

Since there’s a free trial available (which lasts for 30 days), I thought I’d download it and give it a go.

Lightroom is a Digital Asset Management (DAM) tool for digital photographs. That is to say, it covers all aspects of dealing with digital photos, such as acquisition of the photos from the camera, selecting the ones to keep, editing for the final versions, organisation of photo collections, and publishing.

At the moment, I use IDimager as my DAM tool, so I was interested to see how this latest version of Lightroom would compare. In some respects, the two tools are fairly similar, but there are also some substantial differences. (Note: IDimager is no longer available. Its successor is Photo Supreme, which I am now using)

One area that I found to be similar is the acquisition process – getting your photos off a camera’s memory card and into your PC environment. Both products allow flexible renaming of your files, and applying metadata templates to the resulting files as part of the acquisition process. So far, so good.

Once the files are in the tools’ workspaces then the main work of selecting the photos you want to keep and adding metadata to help organise the collection can begin in earnest. Again, both Lightroom and IDimager have similar features. For example, you can rapidly compare photos side-by-side in a virtual “light table” to aid in selection of those images that you want to keep.

However, I quickly ran into a couple of issues with Lightroom’s handling of photo metadata that, for me, are quite serious.

First, some background. Both Lightroom and IDimager use the concept of a Catalogue to hold a list of keywords that are used to organise your photo collection (or collections). While this list of keywords can be just a simple list, both products support, and encourage, the use of a keyword hierarchy for ease of use and flexibility.

As I described a while back, the keyword hierarchy I use has been built up from a number of sources:

I’ve ended up with a structure that has the following items at the top level of the hierarchy:

  • Activities
  • Events
  • Nature
  • Objects
  • People
  • Places
  • Science
  • Styles

Each of these splits down into further categories as necessary as you go down the levels. For example, Activities splits into

  • Disciplines
  • Hobbies
  • Physical and mental activities
  • Processes and techniques

So then a photo of a tennis match would have the structured keyword string of Activities/physical and mental activities/games/sports/ball game/tennisassigned to it.

You’ll notice that I’m using the “/” character to separate the various levels contained in a keyword. The choice of the separation character is arbitrary, some applications use the period (“.”) or the pipe (“|”) character , since there is no industry standard at the moment. A standard for handling keyword hierarchies in image metadata has been proposed (by the Metadata Working Group), but as far as I am aware, there is no product on the market that implements it as yet.

I chose the “/” character because Microsoft’s Windows Live Photo Galleryuses it as the separator to structure the keyword hierarchy (Microsoft calls keywords “Tags” in Windows Live Photo Gallery).

For example, here’s a screenshot of Windows Live Photo Gallery with a thumbnail photo of a church in the Isle of Man shown being selected (the light blue frame around the image).

Church2

You can see, on the right of the screenshot, that the photo has three keywords (tags) associated with it: architecture, Baldwin, church. These are actually all structured keywords stored in the photo’s metadata as:

  • Styles/design/architecture
  • Places/Europe/Isle of Man/Middle Sheading/Baldwin
  • Objects/built environment/buildings/ceremonial buildings/religious buildings/church

You can see part of the “Places” hierarchy being shown on the left of the screenshot, with the “Baldwin” tag being highlighted.

Because Windows Live Photo Gallery is easy to use for other family members, I’ve adopted this method of implementing a keyword hierarchy, i.e. using the “/” separator, in my main DAM tool: IDimager. Here’s a screenshot of IDimager showing the same photo:

Church1

You can see on the left that IDimager uses the same keyword hierarchy. And on the right of the screenshot, you see the keyword strings that are being stored in the photo’s metadata.

Now, the reason that both IDimager and Windows Live Photo Gallery have the same keyword hierarchy is that both tools are constructing it from the keywords stored in the photo metadata. And because they understand that the “/” is the separator character, they build up the same keyword structure on the fly as they read the photos in my collection. IDimager is the more flexible of the two applications, since you can define different separator characters if necessary. WLPG is fixed, and only understands the “/” character.

So, what happens with Lightroom 4?

Well, at first I thought everything was going to play nicely together. Just as with IDimager, Lightroom 4 has an option to choose the separator character that you want to use when reading the keywords in your photo metadata:

LR4 002

I’ve chosen “/” as the separator character – the same as for IDimager and WLPG.

Sure enough, when I imported my photo collection into Lightroom 4, the keyword hierarchy got reconstructed to match the ones in IDimager and WLPG (click on the screenshot to see it full-size in a new window):

Church3

But then things started to go wrong.

First, I discovered that although this process of recognising the separator works when importing photos into the Lightroom 4 Catalogue, it doesn’t work when reading metadata from individual photos – even though Lightroom claims it does (see the text in the “Preferences” screenshot above).

I added the keyword “Christmas” to a photo using IDimager. This is a structured keyword, so the keyword string that was written to the photo’s metadata was actually: Events/holidays/Christmas. Lightroom 4 correctly saw that the photo had had its metadata altered, but when I used Lightroom to read in the photo metadata, instead of adding the photo to the “Christmas” keyword in the existing hierarchy in its Catalogue, it created a brand-new single-level keyword string: Events/holidays/Christmas – it did not treat the “/” character as a level separator.

Now, this, I think is a simple bug, and has been reported as such. However, much to my dismay, I discovered I was not out of the woods yet.

Up to now, I use IDimager to do all my keyword work. When a keyword in the IDimager Catalogue is assigned to a photo, IDimager will write out the structured keyword string into the photo, and WLPG will then pick up the change and modify its own Catalogue of tags automatically.

As a test, let’s use that photo of the church in Baldwin in the Isle of Man. Here you can see the current keywords assigned to it: architecture, Baldwin, church, as seen in Lightroom 4.

LR4 003

Remember, these are all structured keywords with a hierarchy. Lightroom, like WLPG, is just showing the lowest level of each keyword. Using IDimager, I can look at both the actual photo metadata (showing the full keyword strings), highlighted in red, as well as the Catalogue keywords/labels, highlighted in green:

Church4

Now, let’s use Lightroom 4 to add a keyword to the photo, and then get Lightroom 4 to write out the changed metadata into the photo. Here, I’ve added the keyword wall:

LR4 004

Looking at the photo in IDimager, what do I see:

Church5

Disaster! Lightroom has not written out a structured keyword string, but a series of individual keywords separated by commas. While IDimager has been able to sort the wheat from the chaff (the labels, highlighted in green, show that it knows that the structured keyword wall has been added), Windows Live Photo Gallery is not so clever.

Church6

It is now showing all the individual keywords assigned to the image, and worse, it has created new, false, levels in the keyword hierarchy shown on the left in the screenshot. For example, wall is now shown as a top-level keyword.

So Lightroom 4 will read structured keyword strings using the “/” character as a separator from photo metadata, but, unlike IDimager, it will not write out structured keyword strings to photo metadata. Instead it writes single level keywords and additional, Adobe-proprietary, metadata to describe the hierarchy. This is, apparently, expected behaviour.

Well, it may be expected, but it’s pretty much useless to me if I want to keep WLPG as the easy to use browser for others in the family. If I use Lightroom 4 to do any metadata work, it will destroy the keyword structure that I use as far as other programs are concerned.

So where does this leave me, as far as the trial of Lightroom 4 is concerned?

I have to say that the Lightroom tools for editing and developing photos (especially those in RAW format) are far in advance of anything that IDimager or WLPG possess. So while I could continue to use IDimager for metadata work, I could supplement that with the image adjustment tools of Lightroom 4. Frankly, the other modules of Lightroom (Map, Book, Print and Web) are of little use to me; my other tools give me all that I require in those areas. So, are the image adjustment tools of Lightroom alone worth an investment of 130 Euros to me? If I were heavily into manipulating my images using the RAW format, then, yes, very probably. But at the moment? To be honest, I’m not sure.

About Geoff Coupe

I'm a British citizen, although I have lived and worked in the Netherlands since 1983. I came here on a three year assignment, but fell in love with the country, and one Dutchman in particular, and so have stayed here ever since. On the 13th December 2006 I also became a Dutch citizen.
This entry was posted in Computers and Internet, Photography. Bookmark the permalink.

17 Responses to Lightroom 4 – A Mixed Blessing?

  1. Steven Grimm says:

    FYI (or for the benefit of anyone else reading this later), I replied to the Adobe forum thread with a description of how to use EXIFTool to convert Adobe’s hierarchy format to the form you’re looking for. Of course, whether having to run a postprocessing step is too much hassle is not for me to say, but it’s at least possible.

  2. JL says:

    Rather calm of you, I’d say, saying you’re ‘not sure’ if it’s worth using. Since I’m still repairing the EXIF damage from WLPG a year and a half later, the mere thought of another software program that messes around with my metadata … Good Lord!

    • Geoff Coupe says:

      Yes, I know what you mean. Reading the Adobe forum, I get the impression that most of the people there are photographers who care passionately about the images themselves, and pushing the artistic envelope using Ligthroom’s image processing tools. There are fewer on the other side of the fence, those who care passionately about the information depicted by the photo…

  3. jlbeeken says:

    Being as this is (was) a reputable company, I expected they’d have it together for both sides. The metadata setup in Adobe Photoshop Elements is horrible as well and I thought it might be worth trying a higher end product. Obviously not and thank-you for your review of it. I can just imagine how much work you’ve put into your controlled vocabulary. I’ve spent more hours than I’d care to remember annotating my photos so I can find them easily and I hardly think it’s a non-issue.

  4. EddyKilowatt says:

    I had a bad run-in with metadata handling in Elements Organizer as well, lost a couple years’ worth of keywords moving to WLPG due to Adobe not following standards, swore off Adobe for a couple years as a result. I’ve just been tiptoeing back in with Lightroom, figured it was supposed to be the pro-grade tool so should be robust for an amateur like me. I don’t use keyword hierarchies (yet, though I can see the benefits), so I haven’t had much trouble, but it’s disappointing that Adobe is still writing important information in proprietary formats. I’ll be thinking pretty carefully about upgrading to LR4, and looking at other options more carefully.

  5. elisabethbucci says:

    Thanks so much for this blog entry! I evaluated LR 3.5(?) about two months ago. Like you I am a big fan of hierarchical tagging (my system is similar to yours ;) ) and of WLPG. For this reason when I discovered that LR does not write hierarchical tags (my tests and conclusions were the same as yours), I had to dump LR. (I was so relieved that I am the only one obsessed with hierarchical tags: I was thrilled when I came across your 2009/09/03 post regarding hierarchical tagging and WLPG!). All this to say that when LR4 came out, I was curious to know if they solved the problem of writing hierarchical tags…thanks to you I won’t even bother downloading the trial version.
    Thanks again, love your blog!

  6. Randolf says:

    Well, the whole keyword business is a big historically grown mess. In the beginnig, there were simple keywords, only. Adobe introduced then hierarchical keywords — and in order to not mess up flat keywords with hierarchical keywords, they kept both separated from each other. So, when you look in your XMP file (which is an open Adobe file format, like TIFF, DNG, PDF) you will see your keywords written once as flat separate keywords and once in a hierarchical form. But not all programs evaluate both versions, give preference to the one and ignore the other.

    Regarding the “/” as the keyword separator, when you read carefully the text beside the checkmark, you will see that it says “when importing”. I.e. LR is willing to recognize it as such a separator when reading third-party programs. But nothing is said about exporting. You’re right that there is no formal industry standard. You may love it or hate it: But considering the fact, that the XMP files are an Adobe file format and that currently I don’t see any other company in photo software business which has the potential to survive the next 30 years, I would strongly recommend to follow the Adobe way as the standard, ie. use the “|” as a separator, and forget what others do, if you don’t want to mess up your image collection completely.

    Clearly, you can only transfer your metadata between WLPG, IdImager, LR, Media Pro etc, because they recognize more or less (!) Adobe’s XMP files. And don’t blame Adobe, that Microsoft doesn’t follow Adobe’s standard in Adobe’s file formats and uses “/” instead of “|”. Next time, someone else writes a program which uses an “*” as separator, only — then you can’t blame neither Mircosoft nor Adobe that they doen’t write out the metadata with * as a separator to be compatible with your new favourite software.

    So, my advice: For maximum compatibility and future-proof, stick to the way the owner of the file format has defined it and ignore programs which doesn’t follow that definition.

    • Geoff Coupe says:

      Randolf, thanks for your comments, but I think you are mistaken in some respects. It is true that Adobe first came up with XMP, but now it is an ISO standard, i.e. a full industry standard, and no longer controlled by Abobe. It is also true that Adobe established the use of the “|” character as the separator for keyword hierarchies. However, there is a proposed XMP standard for keyword hierarchies that has been documented by a cross-industry group, the Metadata Working Group. That group has both Adobe and Microsoft as members. The proposed standard does not use the “|” character (or indeed the “/” character), since that was a simplistic and limited way of achieving the goal of keyword hierarchies. The proposed standard is much more in keeping with the XMP way of doing things.

      I’m hoping that both Microsoft and Adobe will put their money where their mouths are and implement the proposed standard in their products.

  7. Pingback: Hierarchical Keywords in Lightroom | Be Careful! | Daminion Blog | DAM for Small Teams

  8. Hi Geoff,

    This is Murat from Daminion Software. I’m sure you remember our PicaJet Photo Manager which now has a new lease of life and is being actively developed under the new name of Daminion.

    I was just starting to write a blog article about the issue of hierarchical keywords in Lightroom when I came across your post:

    http://blog.daminion.net/upcoming-posts/hierarchical-keywords-in-lightroom-be-careful/

    The root of the problem you describe above is that the LR team has added its own metadata scheme for storing hierarchical keywords (as you know it stores as lr:HierarchicalSubject) and also stores flattened keywords in the Dublin Core Scheme (dc:subject).

    According to MWG, the hierarchical path elements of keywords MUST be flattened. This means that each hierarchy node must be stored as a separate keyword entry in XMP dc:subject (a very dubious decision IMO).

    So Lightroom will store a hierarchical keyword in two different places: with a preserved hierarchy and with a flattened hierarchy, and there is no way to determine which location is valid; especially if third-party software that aware of LR’s Scheme updates “dc:subject”. This breaks the compatibility of LR with other software programs.

    Because LR is such a widely adopted program we’ve developed a temporary solution. We’ve set the priority of “lr:HierarchicalSubject” higher than “dc:subject” in Daminion’s Keyword mapping list, but this is simply because LR is a widely used program.

    • Geoff Coupe says:

      Hi Murat,
      Thanks for that explanation of LR’s behavior. That’s an interesting point that you make about the fact that the MWG guidance states that hierarchical keywords MUST be flattened by a Changer application. I hadn’t noticed that before.

      Like you, I am far from convinced that it is a good decision, and I wonder which came first, the MWG decision or the fact that LR does it…

      • MWG decision is known: a new complex way to store hierarchical tags and as you said above there is no product on the market that implements it.

        The probability of ignoring MWG recommendation by LR team is higher, although they are on mountain top and can cares about compatibility between LR’s versions only.

        • Geoff Coupe says:

          What I was hinting at is that Adobe are founding members of MWG, so if LR already had this flattening of keywords by adding them in as single-level keywords to dc:subject, then I wonder if Adobe were behind the putting in of this rule into the MWG guidance…

  9. Fantastic keyword structure you have Geoff!!! Something I must learn to implement!! Ive only been shooting for a couple years now and would love your criticism! – https://www.facebook.com/MJAcreative

    • Geoff Coupe says:

      Er, thanks. Matthew. ‘Tis a small thing, but (almost) mine own. But seriously, I think that metadata is important – especially for professional photographers.

      Criticism? I wouldn’t presume – I’m only an amateur. Good luck!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s