On shooting a bunch of people and automating the process - long post ahead

Just to change pace from all the rumors, here's some personal experience some of you might find interesting.

Two days ago I did a favor for an old friend. I was supposed to shoot approximately 150 people in two poses plus standardized headshot, then process from RAW, caption and print the 30x20cm photos as well as generate a "book" that contains, among other things, a two page spread for every person (left page full photo and caption, right page smaller headshot and predefined text). The catch is that it all had to be done in roughly 28 hours (from start of shooting to giving out books). So, here comes a small story of automation and streamlining the process, as well as some obervations on use of affordable consumer level equipment for, what should be, a professional gig.

Eqipment used: Samsung NX1000 and lenses (30, 45, 50-200), Yongnuo 560IV, Samsung SEF42, Yongnuo RF603II, generic equipment (stands, softboxes, flags, reflectors, backdrop, etc).

Lighting setup: nothing fancy, generic 2 lights and reflector (cross) but tuned to be usable for both poses and headshot. All manual, TTL not used (SEF42 in RF603II).

How to identify 150 people quickly and precisely? Not easy, they are from three different departments, some of them work abroad, etc. While we're at it, we should also process RAW files, generate JPEGs, prepare final document for printing and everything inbetween.

Open source to the rescue

Everybody received a small ID card for the meeting with QR code big enough to be read from photo but small enough to be easily cropped away (there was a small stand for card just at the edge of the frame, it had to be cut either way to make room for caption of the photo). So, I made a small python script that processes the JPEG (photos were taken in RAW+JPEG) and finds and reads QR code. Then it uses OpenCV and DLIB libraries to find the position of the face and eyes in the photo. Based on those information it generates XMP sidecar file for darktable (with previously defined processing stack, just adapting few parameters for every image). Darktable was called in batch mode (darktable-cli) to process the images and generate final JPEGs (two processed from original RAWs, headshot is cropped out from second photo - 20 MP is plenty to work with). This ensured that every face in every photo will be positioned perfectly (as there was a collage image to be made in the end with all headshots per department - this was kinda important for me - I'm slightly OCD).

Generating output files for book - Open source to the rescue - again

Everything was done in advance in LaTeX so we just waited for final photos with right filenames to be inserted and final PDF to be generated. Everything took very little real time as I was switching SD card from camera to laptop every 25 or so people. Scripts than took files off the card and processed them in less time it took me to shoot another batch of photos. Final batch took around 1 hour but that included generating book PDF as well. Files were uploaded to FTP server for printing company as they were generated (we had very fast connection). I guess tethered shooting would be even better/easier, but no such luck with NX1000. That said, this approach worked flawlessly.

Printing guys came trough and made everything in time for closing of the meeting on the second day.

General observations:
- This quite affordable equipment was more than enough. Originally thought about using D300 but extra resolution enabled me to crop a lot of a photo out and still have very decent print.
- Preparation is everything - people organizing and guiding people for shooting did a great job. Over 300 shots (450 if you count the headshot) in a bit over 2 hours - that's less than a minute per person. Preparation also made the complete process from shooting to printing possible. Some three years ago we did a similar gig (not such hurry) but for 800 people, two shooters, took us over 10 hours each.
- I have no idea how CIPA conducts battery tests, but NX1000 spent less than one battery for over 600 shots (I also shot some of the event). It should have drained two batteries. I had two spares just in case.
- No flash spent the batteries (and SEF42 shot on every photo - half with TTL so not full power). Also had bunch of spares but did not need them. All flashes, including triggers worked without a hitch.
- SEF42 is tricky for events as it has no swivel head - I used TTL flash extension cord to position the flash the way I wanted (a bit complicated handling). Highly reflective or specular items can throw off the TTL metering when using matrix or center weighted - it's nice to have some latitude in post processing the RAWs
- NX1000 is not an ideal choice for shooting a lot of RAW files quickly (like events) - Processing message bit me couple of times - saved by having people skills, be nice to people and they will be nice to you
- I'm somewhat surprised by 50-200mm at around 70-130mm - it's actually very very good - good enough for me. Used other lenses for event afterwards.

So, with a bit of organization and preparation, you can work quite fast when needed

Very interesting and thanks for sharing! Fun story, must be a great friend to do this for him. Fun to stretch your skills from time to time. In this case the speed of processing and organization skills.

Wow! I wish there were posts like this even once a month. Even though I don't need to photograph this many people, some day I might and now I have big head start on getting it done because of you! I love a post with useful information like this. Thank you so much!!

Thank you for your kind words

Thank you, yes he is a very good friend indeed.