Tuesday, 15 March 2011

My post-processing workflow

I thought it was about time I wrote down my post-processing workflow. This is just how I work just now and it's always subject to refinement.

For anything I care about, I shoot raw - I'd rather have the extra latitude in processing the images and I'm never shooting where I need the shots straight from the camera. I've never actually seen any real reason to use sRAW since the space saving is minimal.

I use Lightroom 3 to manage all my images. I only really use two catalogues - "General" and "Modelshoots" (though I'll likely start a new "Portraiture" catalogue in the near future). Minimising the number of catalogues makes it easy to find everything and manage my online galleries.

Just to give perspective, a typical shoot for me can be anything from one hundred to six hundred images depending on how many sets/looks/locations we work.

You might even notice that I have a fear of commitment - I generally don't like to throw something away or make something permanent. I'm getting better about throwing stuff away though ;)

As I shoot:

If I'm indoors, I'll use the an Eye-Fi card to transmit small JPEGs to Lightroom on the laptop as we shoot (and load the raw images later). The Eye-Fi card will transmit raw files, but that's just too slow, so I configure to shoot raw with a small JPEG written to the Eye-Fi card.

I've also got a Vodafone Mobile Wi-Fi device which allows me to use this anywhere I can actually use the laptop.

If you're prepared to work with a cable connected to a computer, Lightroom 3 will do this for a lot of modern cameras that don't have an affordable wireless option - I'd suggest some gaffer tape to make it harder to inadvertently disconnect the cable though.

Raw file import:

I import the raw images using a card reader. If I'm paying attention I will hopefully pick an appropriate meta data preset when I import the raw files - this consistently sets the authorship and copyright IPTC EXIF. If it's a location that I shoot at regularly then I'll create a preset - otherwise I'll just use a base authorship/copyright preset and update the location information later.

Quick review:

A quick review pass of the images lets me discard any frames that are obviously not worth keeping. I generally only discard the real duds, but I'm trying to get a bit more ruthless - I just can't fit enough internal storage to my laptop to accommodate more images.

Batch adjustments:

I'll enable lens correction over all the images (I really should set this as a default import setting). I also set a consistent white balance for each set (the variability of the Lightroom auto white balance is my one main gripe about Lightroom).


If I'm planning to provide a quick online proof gallery I'll generate the watermarked proof JPEGs at this stage (or maybe earlier - I'm not overly concerned about the proof images and if I'm in a hurry I'll often make the initial batch adjustments after I've generated proofs). Since it's currently only possible to associate an image in Lightroom with a single copy on Smugmug I now perform this step manually as an export/upload rather than use the Smugmug export plugin.

I'll normally have a couple of passes through the images rating and selecting and also incorporate any choices from the models and MUAs. This may also result in a few more discarded frames.

I've tried using collections to group the selected images, but keep going back to coloured labels and ratings.


Around about this point I also add keywords and image title/captions and rename the files to something that works better with search engines than the original filename. I don't personally care about the filenames since I use Lightroom to find the files, but if it helps with SEO on Smugmug then the couple of minutes it takes is worth it.

Image edits:

Lens correction and white balance should have been set for the whole shoot by now.

If the image needs a serious corrective adjustment (e.g. highlight recovery), then I will make the appropriate changes now within Lightroom - if the changes are just a subtle refinement then I'm likely to leave them until later.

At this point, I avoid making any other edits in Lightroom before I retouch in Photoshop.

I usually do most of my image retouching in Photoshop (using "Edit in Photoshop" from Lightroom) since I find it so much more powerful and simpler than messing with the circular clone/heal tool in Lightroom. Photoshop also gives me liquify when it's needed.

For studio work, I've actually accepted that I can now take shots that I previously would not have considered (at the cost of edit time in Photoshop). Accepting that the background can be extended or that a stand can be removed allows me to take shots that I would not have attempted before.

Of course, if the image doesn't need much retouching then I'll perform the edits in the develop module in Lightroom.

Even if I plan to crop/rotate the image (and I almost always need to straighten my images) I avoid doing this before I edit it in Photoshop. I almost never crop in Photoshop (I will often extend the image canvas though to allow me to crop/rotate later in Lightroom). This has the advantage that the decision of the crop isn't set in stone (a big advantage of generating the final output from Lightroom).

If I think they are needed, I will make selective tonal adjustments (e.g. dodge/burn using the Totally Rad Photoshop yin/yang action), but I don't often adjust the whole image in Photoshop.

I own copies of the Totally Rad Photoshop actions and the OnOne Photosuite - these have seen a little more action recently - if I've used these in Photoshop I'm less likely to apply styling presets/effects in Lightroom afterwards.

I've also taken to using Faye Yerbury's recipe for a black and white Photoshop conversion - this is simply the Totally Rad "Boring Old B&W" with a little added warmth (I'd like to be able to replicate this in Lightroom rather than commit myself to a final B&W image without creating a second copy).

I also avoid flattening my Photoshop layers - I like to be able to undo almost everything (and this also has the advantage that I can actually see what I did).

Normally at this stage, I'll return the retouched image to Lightroom and finish the final adjustments there. This gives me a base retouched image which I can then manipulate to give me my desired final image (with the advantage that if I edit a sequence of shots it is very easy to keep the treatment consistent - that's undoubtedly possible in PS, but I don't know how!).

If a Lightroom adjustment highlights something that wasn't originally obvious in Photoshop, I'll
just drop back into Photoshop and fix it (Edit in Photoshop, pick edit original in Lightroom - just turn off the save metadata to PSD/TIFF in the catalogue settings for this to work properly). Lightroom subsequently re-renders all it's changes on top of the updated image.

This is when I add sharpening/noise reduction and play with the final tonal adjustments (and occasionally some creative presets) - I'll try to leave any crop or straightening until the very end.


All my output - web, JPEG and print is handled from Lightroom.

Updates to my web galleries on Smugmug and Flickr are handled using Jeffrey Friedl's plugins (not the stock Lightroom 3 export plugins) - these make it easy to maintain the uploaded images (any modifications just flag the need to republish the images).

I distribute JPEGs via a disk publish service which is set to use shared folders on Dropbox - so much easier than keeping track of which images I've emailed. I also use another publish service to export images to my iPhone (this is also stored on Dropbox to allow it to be shared between my laptop and desktop systems).

Unfortunately, Facebook doesn't allow any programatic upload of images to business pages (so I can't make much use of the Lightroom Facebook plugin) - I just manually upload a subset of images from the shared Dropbox folder.

Pruning and housekeeping:

I must admit I find it hard to discard images, but given disk size constraints on laptops I occasionally go through and prune images from old shoots - this is a losing battle though since I would likely have to delete six raw images to make space for an unmodified TIFF file from Photoshop.

My Lightroom database is also getting a bit unwieldy (currently >2.5GB for 22000 images) - just watch it grow as you use the local adjustment brush. Fixing this is going to be a lot harder - one option is add snapshots and to purge some of the edit history, but again I run into my commitment issues since this would never allow me to go back.


I maintain multiple online copies of my Lightroom catalogues using rsync (I'm running on a Mac now). These are just full copies of everything, but rsync is smart enough to only replicate changes (thought it's not smart about moves and renames).

I tried an experiment with Virgin's V-Stuff online backup service, but in my professional IT capacity I will not recommend it to anyone who cares about their data. It regularly skips files without explanation, I've seen the reported uploaded size lose gigabytes(!) and even after months I haven't completed my first backup!

1 comment:

  1. That's a good (and interesting) read. I'll try see what I can find on the work intranet about V-stuff, and see what it's limits/issues are.