Thursday October 16
 
16:47
 
RAW

Recently I’ve started to get into the meat of what it takes to create a good digital photograph - not just depth of field and framing, but resolved detail, removing lens distortion and chromatic aberration, and affecting the whole quality of the image.

At first I was focused very much on the camera and lenses, but this weekend, I started to see just how much difference shooting RAW and using decent software can make

Capture NX 2 can automatically remove chromatic aberration. It can allow exposure compensation based on raw data (which can recover shadows or highlights, something which can’t be recovered from JPEGs). It allows D-Lighting at settings far more effectual that the crippled on/off of the D60. It allows white balance to be changed.

DxO lacks Nikon D-Lighting, however as well as automatic chromatic aberration correction, it can also automatically remove lens distortion. Look at this test from my 18-200mm and you can see the power - I did nothing except import the image and tell DxO to process it (excuse the colour and shadows - this was a distortion and sharpness test only).

Before:

After:

I always avoided shooting RAW before mainly because I expected the size of the files to be around 5-10 times that of the JPEGs, but they are not. Nikon compressed RAW files are only double the size of the JPEGs.

I also avoided shooting RAW because of the hassle to my workflow - RAW isn’t immediately viewable everywhere like JPEG, needs converting to be uploaded somewhere, etc. Now I’m using Lightroom and flickr, that’s of no concern - Lightroom handles RAW seamlessly and even converts to JPEG automatically for uploading to flickr.

Why would I ever shoot JPEG again? For the price of double file size I can rescue those images where auto white balance screwed up, where highlights were blown, where I chose the wrong exposure - and virtually no penalty.

These days, there’s a lot more about photography that I’m “getting”. Not just an understanding that choosing this aperture does that, etc, but a real feel for how things hang together. My photos might still be crap, but at least I understand why they are crap.

One of my last photos taken as JPEG:

Thursday October 9
 
16:51
 
DOF, Background Isolation, Diffraction, And Sharpness

When I first bought my DSLR, one of the things I noticed was the difference in depth of field from my previous cameras, even at the same aperture. This article puts a technical explanation to that feeling, explaining that DOF is not just related to f-value, but to other factors such as sensor size. It gives this example:

if one used a 50 mm f/1.4 lens on a full frame sensor, this would produce a depth of field so shallow it would require an aperture of 0.9 on a camera with a 1.6X crop factor - not possible with consumer lenses.

On my old F707, an aperture of 0.4 would be required.

Working backwards, F2 on my F707 is equivalent to F5.3 for a Nikon crop (1.5x) and F8 in 35mm terms.

F2.8 on a DX Nikon is equivalent DOF to F4.2 on a 35mm camera.

All these examples assume the 35 mm equivalent focal length is the same in each case (ie, keeping the perspective and size of the image the same in each case).

Wikipedia has a similar example:

a point-and-shoot digital camera with a 1/1.8″ sensor (7.18 mm × 5.32 mm) at a normal focal length and f/2.8 has the same DOF as a 35 mm camera with a normal lens at f/13

This has implications for me since I’m thinking to buy a full frame camera. It means that if, say, I take a photo of the same thing with a DX and full frame camera at 50mm (in 35mm terms) and keep the aperture and framing the same, than I’ll have greater DOF on the full frame camera. On the one hand that’s good because it give better creative effects - on the other hand it makes it more difficult to ensure everything is in focus. 

This article explains (near the end) why zoom makes a difference to DOF: a longer focal length will flatten perspective. While the DOF is actually the same, the perspective changes and therefore it looks like the DOF is different.

Luminous Landscape puts this into pictures here. What you can see is that the tower in the background is essentially out of focus by the same amount at each focal length - but the perspectives are different so you get greater subject isolation.

This is something I noticed a long time ago: if I keep the aperture and size of the object to be the same but take one photo standing close and one standing further away (using zoom to match object size), then I can achieve better background blurring. It’s often assumed that one needs to use a larger aperture to get background isolation and blurring, but that’s not the only way to do it. You can get great background isolation at F5.6 if you use a longer focal length.

While I’m on the subject of DOF, this article has good explanations of Circle of Confusion and DOF range (1/3rd in front of the point focused on to 2/3rds behind it).

Another thing I’m interested in is sharpness. I’ve long known that lenses are rarely at their most sharp wide open - eg, an F1.4 lens will be less sharp at F1.4 than F4 - and that most lenses have a sweet point, usually around F8. I’d also noticed that when I pushed past this to a smaller aperture, the quality wasn’t so good - but I didn’t know why.

This article explains the reason: diffraction at smaller apertures. The “What it looks like” example illustrates this better than any words can. For my Nikon, it looks like upto F8 is safe, and from F11 it starts to become diffraction limited. Of course other factors such as lens quality affect sharpness, but it’s good to have this knowledge.

There is only one thing that I’m not yet clear about: at F1.4 on my Sigma I notice a kind of haze to the images, especially in brighter situations.

One theory I have is that the haze is caused by fast shutter speed. At 1/2000s the nearest choices are 1/1600s and 1/2500s - steps of 25%, which is quite a bit of difference.  There could also potentially be inaccuracies at fast shutter speeds - my Nikon isn’t a pro camera, after all.

Another theory is stray light - non-image forming light or veiling flare or even sagittal coma flare (whatever that is). I’ve noticed the situation improves if I use a lens hood, but it doesn’t go away.

Of course I can test these theories and try to work out how to improve the situation myself - and I will - but I’d still like to the technical reasons behind the haze.

Tuesday October 7
 
20:31
 
Bokeh

Interesting webpage which shows that bokeh can be more than just blurring backgrounds - cover the lens with a black card that has a shape cut out the middle and see what you get.