Part II – the analysis


Following on from the last post, I’ve been doing a little more technical analysis to convince myself I don’t need a new camera.

First off, noise at high ISO. Here are some crops of the image above, taken at various ISO settings. The full size view, which you can see by clicking on each one, is a 100% crop with no noise reduction.

20141213_0029-3200 20141213_0030-1600 20141213_0031-800

In order, these are ISO 3200, 1600 and 800. By 800 the noise has practically disappeared, but even at 3200 it really isn’t intrusive. To give an idea of how irrelevent it truly is at web sizes, the full frame shown at the top is the ISO 3200 version.

Also note that the ISO 800 image isn’t as sharp as the others. Camera shake, I’m afraid, even at 45mm and 1/125s. And that’s with VR on – really not good enough! It’s a reminder, should I need one, that the biggest limitation I have is not my camera – it’s my technique.

Finally I tried a little experiment with another image to see how it might pan out enlarged.


I took a crop and scaled it to A4, A3 and A2 and printed the result. The A4 crop is actually a slight reduction from full size and in each case I sized to print at 300dpi. It really was a quick and dirty experiment, as you can see from a scan of the resulting print on A4 paper:


It should give you an idea of how big the image would appear printed at each of those sizes. The final one, A2, would be near enough a 24″x16″ print. Looking at the print at arm’s length (and when do you ever view an A2 print much closer than that?), I reckon it’s good enough.

With care, I could probably get an A1 print out of the camera. This shot was handheld, at ISO 400 – if I was shooting landscapes with a view to making massive prints, the camera would be on a tripod at ISO 100 and I’d take a lot more care over the post-processing.

So what would a 24MP full frame camera give me? A bit more leeway, mainly, if I was making large prints – they’d be slightly smaller enlargements and therefore a little more forgiving of less than perfect technique. That’s about it, really.

Given how rarely I ever print anything, there really is no need to change my camera – although I do have to work on my technique! It’s been a useful exercise in that respect.

I don’t need a new camera


I really don’t need a new SLR. I just need to convince myself I don’t.

20141213_0037My current SLR is a Nikon D7000 that I’ve had for nearly three years. At 16MP, it’s been overtaken by the current generation – 24MP is even available in Nikon’s entry level D3200. So it begins – the nagging feeling that I want to upgrade.

The thing is, though, I don’t really need more megapixels. How often do I print to A3 or larger? Almost never, and my first DSLR, which only had a modest 10MP, was perfectly capable of decent prints to that size anyway.

So what else has improved? High ISO performance is, amazingly, better on the current generation of 24MP cameras. I say amazingly, because in general cramming more pixels onto a sensor of a given size ought to increase the risk of noise. Again, though, my current camera performs perfectly adequately. Most of my shots are taken in the ISO 100 – 400 range anyway, and the 7000 is good enough for the rare times I need to venture into the higher range.

There are a few other things that have improved, but the real boost in image quality would come from jumping to a full frame camera. I’d rule out the 36MP D800. Too expensive, and with RAW files taking about 76MB it needs a lot of storage, both on and off camera. The D750 is tempting. At 24MP it’s inline with the current DX generation, but the larger sensor should mean less noise.

20141213_0024Of course, going full frame means new lenses as well – all of my current lenses, with the exception of the 50mm 1.8, are designed for DX cameras and won’t cover full frame. Any change would require major investment. So why do it? Just for the remote possibility that I might want top notch prints at silly sizes? Just in case? It really doesn’t seem worth it.

I’m trying to convince myself… when the urge to splurge on a new toy gets hold of me it takes a lot of willpower to resist it, even when it’s a lot of money.

I’ve been putting my current camera through its paces to try and convince myself I don’t need a new one, and will be posting a more technical run through soon.

Colour Calibration


How does this picture look to you? I can guarantee one thing – it looks different to me, and to anyone else viewing it.

I’m talking specifically about the colours. I’ve viewed this picture on several screens – my computer, my phone, a couple of different tablets and another PC – and the colours were different on each. On the other PC, in particular, the colours were far less warm, and the lower part of the sky was a brighter cyan.

On my Mac, the colours are quite warm with a hint of a magenta cast in the clouds. On the tablets the colours are also quite warm. On the Windows PC, the cast disappeared and the overall colour was colder.

It’s a problem for photographers. A keen photographer may invest in colour calibration equipment to get the colours just right. That’s important when printing, so that soft proofs on screen are as accurate a reflection as possible of the final print. How often do we actually print though? I would imagine I’m like the majority of amateur photographers these days in that my photos are only seen on screen, and rarely printed.

That, then, is the problem with colour photography on the internet. You can tweak the colours all you like to get something that looks right to you, but have no control over how it’s seen by anyone else. No wonder I convert a lot of my photos to black and white!

Experimenting with star trails

20130814_0001-EditWe’ve had a couple of clear nights recently and with the prospect of seeing the Perseid meteor shower I’ve been spending some time in the garden of an evening. I didn’t have much joy – maybe two or three over the course of a couple of nights.

While I was out there, I decided to set the camera up and have a go at star trails. I was mainly aiming to try out the techniques and see what happened. I figured the level of light pollution where I live would rule out getting anything worthwhile, and I was pretty much right.

20130813_0007I started off with a few 30 second exposures, such as the one on the left here, just to get a feel for what I could capture. At that duration, with the aperture wide open, the stars are quite clear but the street light is starting to become obtrusive. The camera has picked up far more stars than were visible with the naked eye.

I tried the stack approach of taking a series of exposures of around five minutes each with a view to combining them in Photoshop. Five minutes is long enough to capture some movement in the stars, and as long as the gap between exposures isn’t too long it should be possible to get continuous trails. Unfortunately none of those experiments quite came off, but I’ve learnt some useful lessons.

Finally I decided to go for a really long exposure – almost 25 minutes (I was aiming for half an hour, but it was cold and I wanted to go to bed!). As expected the sky was getting really bright from the ambient street lighting and the sensor noise was building up. Still, with a little judicious tweaking of levels it was possible to see the stars rotating around the pole star (neither of these work well at thumbnail size, just click for a larger version).

For an experiment it worked out reasonably well, and I now have a better idea of what to do if I can find the combination of a clear night, no moon and darker skies a little further away from civilisation.


The model village effect

156The model village effect is an attempt to make real life scenes look like models. I’ve seen it done in exhibitions and even TV advertising, and thought I’d have a go.

The idea is to replicate the shallow depth of field that you’d get when shooting a small model. Read More…

Night shots – doing it wrong

138139Another trawl through the archives for this post, and some old night shots I took in London a few years ago.

So why the title – what was I doing wrong?

Well, just look at the sky. It’s an expanse of black. I was clearly there far too late.

The best night shots aren’t actually taken at night. Dusk is the best time, when the street lights are on but there’s still some light in the sky. Read More…

Full moon fever

127128It’s a full moon on Friday, and that’s prompted me to dig out some of my old moon photos.

The moon is surprisingly bright. The photo above was taken at 1/160th of a second at f9 (ISO200) – virtually a daylight exposure. It’s also small – this is a considerable crop from the full frame, which was taken at 300mm (450mm equivalent in 35mm terms). Read More…

Compact vs. Smartphone – round 2

102Round two of my smartphone vs. compact comparison, and I’m doing the same as I did in round one, this time pitting my phone against a Canon Powershot SX230. The Canon is what is often termed a travel zoom – a compact with a huge zoom range that’s ideal if you need a small but versatile camera. It came out well in a group test, and while it won’t compete with my DSLR, or even the GF1, it’s much easier to fit into a pocket!

So, first up, here’s the image from the phone (click for the larger version). At web size, it’s perfectly adequate and again shows how handy it is to have a reasonable camera constantly to hand. Read More…

Compact vs. Smartphone – round 1

98Apparently the cheap compact camera is soon to be no more, thanks to the rise of the smartphone.

With that in mind, I thought I’d do a little experiment and pit my phone, an HTC Desire, against a small camera. I say small camera, because I’m cheating. I didn’t use the sort of cheap compact referred to in the article but my Panasonic GF1 which really isn’t what they had in mind, especially with the prime 20mm lens on the front. Still, it is quite compact!

First up, here’s the image from the phone, taken at full resolution and on the finest jpeg setting then resized for web use (click to see the 700px version). It’s pretty good at this size, and for happy snaps it’s perfectly adequate. Read More…

Poor man’s tilt and shift

95Bit of a techie one today! As anyone who’s ever tried to take a picture of a building will know, unless you can shoot from a height or a long way back converging verticals are a problem. As soon as you tilt the camera upwards to get the whole of the building in shot, you get that characteristic “leaning over backwards” look as lines that should be parallel converge. Read More…