Austin’s Odds

Man with Bottle

Man with Bottle

A few years ago I was given an old suitcase, full of old photographs. The suitcase had come from my wife's grandmother's flat, when it was being cleared out after her death, and it was given to me as I am the designated family historian and keeper of old photos. There were hundreds of them in the suitcase, along with a couple of books and some holy pictures. Do you remember holy pictures? Well, the great thing about these holy pictures was that they had dedications on the back, with dates and "love and kisses from Alice".

The photos were a different matter. In all, I think only two had any form of identification, and all the people who might have known those pictured were themselves dead. What is the poor designated family historian to do with them? Sensibly, the only thing is to bin them, but as I am also the designated keeper of old photos, I have put them all carefully back in the box, for some future keeper of old photos to worry about.

Nowadays of course we have thousands of photos – I average 2,000 a year, after deleting rubbish and duplicates – and few of them get printed, so there can be no writing on the back. But they are digital! Your camera will have recorded the date and time that you pressed the shutter, along with all the camera settings that you used, and some cameras will even have recorded the GPS co-ordinates. So all you have to do, to be able to answer "Who is that?" as your memory fades, is use your software to add a caption, which will be stored in the EXIF data with all the other details. You owe it to posterity to do it.

Meter Settings

Meter Settings

My camera, in common with most other SLRs, has a built-in light meter. It can be set to take readings from just the centre of the scene (spot), mostly around the centre (centre-weighted) or from the whole scene (multi-segment).

Last week, when shooting at a school play on 1916, I set metering to spot. I find this best in such situations, as then the exposure will be taken from the face of the subject without being unduly influenced by the background, which is generally much darker.

Two days later I was at the St. Patrick's day parade in Carlow town, and took a few dozen shots. When I looked at them at home, many were over-exposed, as I had forgotten to set metering back to centre-weighted or multi-segment. When the spot that was metered from was relatively dark compared to the rest of the scene the brighter spots were over-exposed, and I didn't check the histogram on the camera's view screen, as I should have, after any of the shots.

In the shot of the band, metering would have been from the dark banner, leaving areas such as the van at the back and the white coat on the right clipped.

Presentation Band

Presentation Band

Histogram

Histogram

The "sensor' in a digital camera is actually an array of millions of tiny light-sensitive sensors that gather the light coming through the lens while the shutter is open. Each of these tiny sensors produces a "picture cell" or pixel – a dot of colour – in the resulting image. The image from a typical APS-C sensor has about 3200 rows of 4900 pixels, and so about 15.5 million pixels or mega-pixels.

 

A computer screen is also an array of pixels. If I had a computer with a screen with 3200 rows of 4900 pixels I could transfer the image from camera to computer, and when I displayed it it would just fill the screen. However, most screens are a lot smaller than this. Mine, for example, has only 1050 rows of 1680 pixels, so when I ask for the image to be shown it has to scale it down to about 30% to fit on the screen. Each screen pixel in the display will be an average of the pixels from around the equivalent position in the image. The software allows me to zoom in, so at 100% each screen pixel is the same as the equivalent image pixel, but then I can only see part of the picture and must scroll to see other parts of the picture. I can of course also zoom out, so that the image takes up only part of the screen. Either way, the display will usually show the amount of scaling as a percentage.

 

Note that I have made no mention of dots per inch, or resolution, as this has no bearing on how the image is displayed on the screen. Some cameras set it to 300, others to 240 or 72, but whatever it is set to, it will not affect the on-screen display.

 

Now let's consider printing, where resolution does matter. Typical ink-jet printers space their tiny dots of ink at about 300 to the inch – hence dots per inch, or dpi. The long side of my image is 4900 pixels, so to print it at 300 dpi would need paper 4900/300, about 16", on the long side. But my camera tells the computer that the image dpi is only 72, and printing at 72 dpi would need paper 68" on the long side. Printing with the dots so far separated would give a very poor picture, so, before printing, I need to change the dpi, which software like Photoshop (and many others) allows me to do. When Photoshop, and similar programs, change the dpi, they need to know whether to keep the same number of pixels (smaller print), or the same print size (more pixels), and you tell it this with the "resample image" check box. If you bring up the image size dialog, and change the 72 to 300, and leave the resample image box ticked, Photoshop will take your 4900×3200 pixels and insert extra pixels to make it 20,500×13,600 – that's a lot of extra pixels, and the resulting file will be enormous. Untick that box, and the pixel count stays the same but the document size will go down from 68" to 16" (and corresponding changes for the short side).

 

The 16" is too long for printing on 11.7" long A4 paper, so I would need to scale the image down a bit, to about 3500 pixels. Just as when displaying the image on screen, the computer will do some averaging to get my 4900 down to 3500. If I want to print at the larger A3 size, on the other hand, I haven't enough pixels, so the computer will insert extra to bring the long side up to 5300 pixels. In practice, I only do this if printing elsewhere, as the computer will automatically scale the image before printing on my own printer. The critical thing here is first to set the resolution and then, in a separate operation, scale the picture if necessary.

 

Considerations when sizing an image for display on the internet – the CPS gallery, for example, or facebook, or just for emailing to a friend – are somewhat different. Most people that view the image will do so on a monitor that is 1200 pixels or less on the long side, and often the picture area on screen will be a good bit smaller than this. For such purposes I would normally leave the dpi setting unchanged – it is not going to affect how it is displayed – and scale the picture down to 1200 pixels long. When saving the image I would use a JPEG quality of maybe 70%, but JPEG settings are a topic for another note. Someone viewing the image won't be able to make a great print from it, but if they want a print they should ask for a print-quality copy. If I need to put a file up for printing I will change the dpi, do any necessary sizing, save at maximum quality JPEG, store the file on Dropbox and send a link to the person who needs it.

Elmo was recently told that he could not photograph certain buildings in the IFSC in Dublin. Basically, if you are standing in a public place, you can photograph anything or anyone you see. There are links from the Educational page (External Links) to pages on Irish and UK Photographers' Rights – they are not quite the same. You might also like to look at the video on this page.

Did you enjoy Carlow Arts Week? I certainly did, and the buzz was great on Dublin Street – particularly around our pop-up shop. We sold far more than I had expected, and the comments from the visitors was overwhelmingly positive. The shop manager (that was me!) would like to thank everybody that helped in any way for making it such a success.

Time to start planning for next year!

IMGT8543