Yesterday I outlined (in a fair bit of detail) a setup I worked through last Sunday as I attempted to photograph some garden birds feeding, lighting them using three strobes, two of them off-camera. And while I got some shots, they weren’t really what I had hoped for, and so I changed everything. Well not quite everything, but I did, for starters, move to another bird feeder. And I made some adjustments to my lighting which would affect the colour palette of my images. The shot above is the result of the change, and comes with this warning: it is photoshopped… but probably not as much as you might think. Full disclosure, including shots straight out of the camera, below.
The whole purpose of lighting in the way I was lighting was to get me shots out of the camera that were as close to ready-to-publish as possible. And in general, it’s always better to get things right in the camera, even if only because it means less time spent at the computer later.
Having experimented with a lighting setup at the first bird feeder, when I moved to the second I was loathe to change it too much. So the basic setup stayed much the same – two bare SB-600s off the camera, triggered by an on-camera SB-800 which was also adding fill. The power and working distances of the three flashes stayed pretty much the same also, though in hindsight I should have increased the power of the on-axis fill as what post-processing I did have to do (photoshopping aside, as you’ll see) was primarily making up for the lack of fill light that this flash actually provided.
Here is a (blurry) setup shot taken from my vantage point for this setup, which was viewing point 2 in yesterday’s setup diagram:
The change to my lighting setup was two-fold. Firstly, the position of the flashes is different relative to the camera. Where yesterday’s post had one flash at camera left and one at camera right, with the on-axis fill on the camera, this setup has the on-axis fill, a flash at camera right and a flash directly behind (and slightly below, to stay out of the frame) the subject, providing backlight. The idea of this was to separate the subject from the background. Which brings me nicely to the second change to the lighting setup…
For this part of the shoot, I gelled all three flashes with CTO gels, and set the white balance of my camera to Tungsten to match. My subject was to be entirely lit by flash – again here I’m at an exposure that almost zeros out the ambient light falling on the subject – and so the gel and the white balance adjustment would cancel each other out. However my background – in this case the sky – is all ambient light and not lit by flash at all, so that shifts from the grey that it was to blue, which gives me a more colourful palette in my final image. It enhances the feeling that these shots might be taken at dusk rather than in the afternoon because it adds a deeper colour to what otherwise would have been a bland sky.
Indeed, this was key to the rationale in moving from one bird feeder to another half way through the shoot… this bird feeder firstly allowed me to frame against the sky (by virtue of where it was located) and also allowed me to leave some space either side of the bird feeder, so a bird coming from the left or the right would potentially still be in a place in the frame that worked compositionally. Yesterday’s post showed a setup which was really only suited to birds flying in from camera left, and more often than not they actually flew in from camera right.
So with this new setup, and my new vantage point inside the house, I waited for the birds to resume feeding. Initially, every time a bird landed, or got close, I triggered the shutter, which in turn caused the flashes to go off. That resulted, almost invariably, in the bird flying away immediately and staying away for a while. Over time I left them eat for a bit before taking a shot in the hope they might become more used to the sound of the shutter and the triggering of the flashes. Nine times out of ten they still flew away, but there was one (at least I think it was consistently the same one) who was happy for me to take 5, 6, even 7 shots as he fed and didn’t scare like the others did. I came up with the theory that it wasn’t the flashes, but the sound of the shutter that was scaring them off, and that this guy might have been deaf. I don’t know if there is such a thing as a deaf bird, but if there is, I reckon he was one of them. Or maybe he was just more brave or more stupid than the rest!
Anyway, eventually I started capturing shots with more than one bird feeding at a time. Firstly I got this, shown to you here straight out of the camera with no post-processing whatsoever:
(As an aside, you can probably see what I mean about that on-axis fill being set too low). I think that guy at bottom left is my deaf friend, because he stayed there for the next two shots. On the third shot after this, he left and two more arrived to give me this (again, straight out of the camera):
When reviewing these shots later that afternoon I thought that these two would merge nicely, which brings me to the Photoshop trickery I mentioned at the outset and is how I got to that shot at the top of the post. But the only trickery going on is, as you can hopefully see, a pretty straight-forward merge, and a tiny bit of (rushed) cloning that needed to be done to clean up the edges post-merge. Everything else about the image is straight post-processing, primarily playing with colour and saturation, and also trying to compensate for the shortage of fill light.
The dilemma for me with that opening shot is that I think it looks a bit fake. In particular I can see how someone might think that the birds were “dropped in” – especially the guy at top right, and to some extent my deaf friend at bottom left. But as you can see from the unprocessed shots, I haven’t really cheated as much as it might be thought. But I bet if I enter that shot into next month’s competition in the Dublin Camera Club I’ll be accused of much more photoshopping than I’m guilty of.
Which is why I’ll probably enter the following shot instead… it doesn’t have four subjects fighting for your attention for starters, just the one. And I’ve cropped it tighter than it was out of the camera to make the framing better. In general for this shoot I found it better to frame perhaps wider than I might have liked to allow more options for where birds might land. The post-processing is much as for the shot up top, but with no photoshop trickery:
To wrap up, I thought I’d respond here to a couple of points raised by a reader, David, who commented on yesterday’s post. First he asked:
“I was wondering, do you know what the duration is for the flash burst? I tried something myself a while back and found that using a smallish aperture (around f8) required the flash to use a high output and the bird (blue tit) blurred because of this. They move incredibly quickly. Even with yours, you can see the blur of the wings, even though the primary light source is the flash.”
He raises a point I meant to raise yesterday. Yes, there is movement in all these shots, and when I said I didn’t get what I had hoped for from this photoshoot, that’s one of the primary reasons – I hoped to freeze them in mid-air, but I think I was at the wrong time of day for that. Ideally I’d have been able to use my flashes in normal (non-high-speed- sync) mode at a low power (say, 1/16th power, which means flash durations of perhaps 1/20,000s or so). In that situation, with the flash as the only light source, that flash duration becomes the effective shutter speed. And that’s a very fast shutter speed. But that would require almost no ambient light to start with, or at least very little so that I didn’t have to close down my aperture too much and could keep my shutter speed slower than 1/250s.
In my case, the shutter speed became the dominant factor in the freeze time, because as I understand it the way high speed sync works is that the flash starts to output light just before the shutter opens, and sends out what is almost continuous light until just after it closes. So instead of freezing movement at 1/20,000s, I was working 20-25 times slower than that, and not sufficiently fast to stop the motion blur.
The other qustion David asks is this:
“When I did my ones, the camera was about 12 feet away. You must have been very close if you got that with the lens at 48mm on a D700, or have you cropped the photos?”
Yesterday’s shots were cropped ever-so-slightly – today’s aren’t (apart from the last one, and you can see how that is cropped by comparing it to the other shots), so yes I had the camera very close. In the first setup it was 4-5 feet away from the bird feeder. In the second, as you can kind of tell from the setup photo above, it was perhaps 3 feet away. Close enough that quite a few times a bird landed on it before landing on the feeder.
Thankfully though none left a reminder of their visit to the camera!
Recent Comments