A most compelling evening (well if you are a camera club member) had courtesy of Keynsham Photographic Centre. Simon and Neil took us through papers, sizes and types of processes KPC offer and their knowledge and enthusiasm went a long way to explaining why KPC is a destination of choice of so many serious photographers here-abouts. Around 2 in 3 of those present were already users but we all had something new to learn about the services they offer, which are quite extensive. I have always found them to be helpful and friendly and I am happy to pass on a recommendation to use them.
Why print? There is an emotional and a logical answer to this. The emotional one is that we do seem to react differently to a print than to a projected image or one on a phone or other screen. This is certainly the case with the written word. With pictures it might be an age thing, with those of us of a certain maturity having an emotional tie to what we started with, the family albums and so on, but as these are family history it is not uncertain that these emotions are passed on. It might be the way we view what is art. Even slides are, these days, often rendered digital. Prints have a longevity that digital files cannot match.
The logical one is more existential. Our photographs aren’t photographs until they are printed. They are computer files that are capable of being rendered as images given the right, seldom very cheap, equipment. To that extent they are, at best, semi permanent records. They exist not as pictures but as 1’s and 0’s and that makes them vulnerable to damage, loss and redundancy in many different ways.
One of the questions that came up in discussions was the notion of the minimum resolution (pixels per inch or PPI) as a ratio to print size that is compatible with a reasonable quality print. In order to look at this we need to look at and differentiate between two measurements that the computer file we turn into a picture holds – and which can be reset via (photo editing) software. The first is Pixels Per Inch and the second is Dots Per Inch, or as we will refer to them PPI and DPI.
There is a previous post dedicated to this and a good start if your interest is more than passing. Essentially the PPI is a measure of the resolution of the computer file when shown on a screen and the DPI a measure of the resolution that a printer will reproduce the computer file as a digital print. They are not the same thing, they are not interchangeable, the connection between them is that you cannot use DPI to overcome low resolution in the PPI. That might not seem like much of a connection but it is a common error, at least it is a common misunderstanding.
Just as one is not strictly reliant on another there is an optimum setting for both, but it is not a single magic number or formula. This is because different manufacturers of camera sensors and printers construct their wares in slightly different ways to slightly different priorities. The only rock solid constant you can apply is garbage in garbage out aka GIGO. What doesn’t help is that there are always going to be subjective elements to the ideas of output and acceptable, but we can set these aside.
A slight diversion, but one that is worth taking, is the purpose for which we are taking our images in the first place. The digital age has brought with it a number of outlets that, even 20 years ago, were pipedreams or even unimaginable to most photographers. Film cameras would enjoy their predominance for another 10 years in the “serious” and professional markets. A decent living could still be made from stock photography. These days there are far more outlets and they demand different things of the data files they publish. What remains is the fact that proper prior planning prevents poor photography. Start with the end in mind and choose formats etc accordingly.
The argument for recording images in camera RAW is that all the information is left in. The argument for JPEG is its universality and space saving compression. The fact is you can shoot in both and convert your final image from RAW to JPEG or any other format (has to be some other format as you cannot save in RAW). In your final version of the image is the information that other devices are going to use to present your image, including colour space, PPI and a whole lot else. The point here is that you are going to have to make, or let programme defaults make, these decisions to determine the look of your output (as best you can).
So, long way round to the answer we were looking for but the fact remains, that for all the reasons above, there simply is no answer to the question of a ratio of file resolution to print size because Pixels are not the same size on all cameras – 24 mega pixels on a 35mm sized sensor take up more space than 24 mega pixels on an APSC sensor. There are approximations, guidelines and above all, experience. Let’s look at rules of thumb, but first a caveat.
Pixel peaking, the urge to zoom to the maximum and suck the teeth at the lack of relative clarity, is neither helpful nor useful, outside of certain photographic publications desperate for some comparators in a world where you really have to go some to find a “bad” lens or sensor. It’s the aggregate of the pixels used that we view and a badly composed photograph is a badly composed photograph regardless of the amount of money you have spent or your equipment’s DXO rating. We are here looking at making a print of what you have got, regardless of your motive to print it.
Lets first talk of PPI. Smaller sized prints, say up to 10 x 8 (ish) we can get a reasonable print at 125 ppi. That is to say at arm’s length it’s going to look OK. Press your nose against it or use a magnifying glass it is going to look pretty rubbish. I have to say that such a view point rather spoils the whole point as, again you are looking at the aggregate of the pixels after all, or as I like to call it, “The picture”. This is a fall-back position only, If your default is 72ppi (i.e. it’s designed to show up on your computer or phone screen) then you are not going to get much above a 6 x 4, maybe 7 x 5 at a pinch, out of it (feel free to prove me wrong because you are only going to find out by using your combination of equipment and the printer you choose). The best lies between 200 and 300 ppi, by more or less common agreement. KPC ask for 305 PPI. Basically of you use the 300 end of the range you are likely to be close to the optimum for most commercial printers at standard sizes. Refer to your supplier for the necessary information. To change images PPI we use editing software like Lightroom, Gimp or Affinity.
But we were looking for a rule of thumb. Any minimum figure is a result of the combination of equipment particulars and specifications and materials used.
So, take the longest edge of the image (measured in pixels) and divide by the longest edge of the desired print size (measured in inches). If an image measures 3,840 x 5,760 pixels and you want an 8 x 10 inch print. 5,760 pixels ÷ 10 inches = 576 ppi. That’s more than enough resolution. If you want make a 30 x 20 inch poster out of that image, you’d have a resolution of 192 ppi (5,760 ÷ 30), which isn’t high enough for optimum, if we are looking at least at 200 PPI. This is a minimum remember, so adjust accordingly.
Alternatively, and my personal method, size the image in inches at a ratio of 1:1. So an 18 x 12 print is sized as an 18 x 12 image in my editing software (Gimp). The resolution is set at 305 by 305 dots per inch as I use KPC, and the image, in pixels, is 5490 x 4118. The down side is the file size, the upside is its going to fit without faffing around.
So http://www.keynshamphoto.co.uk/ and start printing!
Aaaaaand we’re back. No wait, did that one last year. Try this. Hello, again. The first of season 2016 – 2017 we did out annual what-I-did-on-my-summer-holidays session. No, wait, it was way better than that, at least after I’d done my stint. What he we had was a good variety of people, landscapes, details, street, and what else taken over the break by the membership. Even those club shoots we did, where there are a lot of views of the same thing, showed a refreshing variety of angles and interpretations of some very stimulating evenings shooting. Quite energising, as it is always a range even when taken of the same subject. Angle, crop, the way the light falls all make up some subtle differences and that is before you get to the mechanics of the way different firms sensors render colours, or, indeed whether the photographer has chosen black and white.
We followed that with a Q&A in week two, with varying degrees of success and it is this that I want to follow up but because at least two of these answers is about the length of an average post in itself, then I am going to split these over approximately the next month in association with content from that week.
What difference does it make when I change the DPI and its relation to pixels?
The simplest answer to this is none whatsoever, but simple answers don’t always mean helpful answers. It can cause confusion, not least because we are seeing the term PPI banded about more frequently of late. Add to that the fact that you can set both pixel count and DPI often on the same menu and club photographers most often come across these when resizing for competitions and voila we have a strong correlation by association. They are not, however, interchangeable in any meaningful way.
Let us start with DPI or Dots Per Inch. This term has been around longer than the common notion of pixels and has been associated with different aspects of image quality over the years. DPI has come to mean more than it actually does in this context and can be the source of confusion. In fact DPI refers to prints only. It is not directly related to image quality at the core, that is to say on the sensor. Your printer creates the final image through spraying jets of ink – it’s sort of in the name – on to a surface. The resulting dots are engineered to be thinner individually, than a human hair. It is a sort of the ultimate in Pointillism and is philosophically absolutely nothing to do with that school of art and so isn’t, if you see what I mean. No? Move on.
What it practically has in common with pointillism is the capacity of the human optical and processing system (that is your eyes and your brain) to create a fuller tonal range from the arrangement of dots of colours of far less actual colours being applied. Hence RGBY (Red Green Blue Yellow) in one cartridge and black in another on your basic home printer can give you a fair rendition of a much, much wider tonal range.
So, why do we even have the dots per inch option built into our editing suites print menus? The answer is tied up with that tonal range and the viewing distance. Another confusing label attached to this is “Professional”. “Professional” prints are printed at around 300 DPI (depends upon the printer system being used). It’s to do with quality, we are told. Yes it is, very much. Doesn’t explain why your average billboard poster is printed at around 6 DPI and they are professional too. The key to understanding that is the viewing distance. Billboards are, most often, not viewed at arm’s length. If you have to stand back further to get the whole image in then you need less dots. If you have ever bought printer ink then you know that it is, per litre, one of the most expensive things on Earth to buy. Someone, I forget who, calculated that to fill your average family car petrol tank with the stuff would cost you around £90,000 at the pump. And then it wouldn’t go. So that’s £90k + a garage bill + some very awkward questions to answer from your other half and the bank.
It should be noted that this is all about the final quality of the print. If the quality isn’t in the file you are trying to print in the first place then daubing more paint onto your canvas just (expensively) glorifies the mess you have started with. Which brings us to pixels.
Pixel is a compound word made up from the term Picture Element. It is not purely a photographic term by any means. It is the smallest programmable element in an electronic visual display. It is a single point in that display. Your TV works on pixels. Your computer screen works on pixels and yes, your digital camera sensor works on pixels. Except …. well we photographers in particular, it being such a large element in the marketing of the quality of capture our cameras can make (at least according to the industry’s marketing departments), actually confuse the physical elements on the sensors which are, more correctly, photo sites, the receptors that sit behind tiny , tiny lenses and convert light into binary electronic signals with something that is fixed in size.
Logically that is not the case (necessarily). The actual, physical, size of these picture elements is determined by the resolution the designer is engineering for. So if you are looking at this on a 800 x 600 screen the answer is, half the size of looking at this on a 1600 x 1200 screen. Or if you have 16 mega pixels of an APSC sensor Nikon D3100 the sensor on the 24 mega pixel Nikon D500 has 50% more pixels, but they have to be smaller to fit in the same physical space. You have more points, so, like having more dots in your final print, you have more chance at rendering gradations etc more accurately and subtly. So there is no definitive answer to how big is a pixel, it depends upon the medium you are using for the display, and the display size itself is dependent upon how far it is expected to be viewed from. The rough dimensions are one and a half to two times the diagonal (on the diagonal is how TV screen sizes are measured).
The basic rule – and I mean rule – is that the initial quality of the file you are displaying dictates the maximum overall quality you can achieve. So, being both Pro cake and pro eating it both the Get-It-Right-In-The-Camera-istas and Ye-Accolytes-Of-Photoshop are right. Firstly you need to get your picture as right as you can at source and if you have more information to work with you can do more at the extremes of editing. This is important, because a low resolution produced-for-the-web image is not going to make much resizing before it becomes useless. No amount of expensive dots per inch squirting is going to change that. Interpolation, where your editing programme takes each pixel in relation to the pixel bordering it makes a decision on shade and contrast of it and fills in the space with a guess-of-a-pixel (we will look at that in a little more depth another time) has its limits. It also takes up space – the very reason that your average web image is of low resolution, it’s good enough for the purpose its designed for given data limits and user load times.
Oh, I could go on, but I won’t.
N E X T M E E T I N G
In the Footsteps of Shackleton.