I was checking out Profotolife’s list of cheapo photo gifts and one of the suggestions was a set of grey cards, and that reminded me I always wanted to do a print my own grey card. Now a grey card, like the “real” Kodak Gray Card 18% R-27 or the cheapie Mennon ones I use, are actually painted pieces of card that are of exactly neutral grey (18% aka Zone 5) and are used for two purposes:
- To set Exposure — because meters are expecting 18% grey
- To set White Balance — because they are designed to be of completely neutral colour
Now before you tell me let me say I already know that it is impossible to print your own grey card that can serve either of these purposes accurately. The paper and inks that we use at home in ordinary inkjet printers are not going to give accurate results, there will be a colour cast, and the density is not going to be right.
If you have a top of the line printer, proper photo paper and inks, and are printing with a properly calibrated printer profile then it is going to be better — but it still won’t be accurate when seen under different types of lights. e.g. it might be right in sunlight, but in shade it’s wrong.
So, this is impossible – so what am I writing this about? Well I know it’s going to be inaccurate but the question is how inaccurate and is it better than nothing?
Also given that I’ve mention I have a real grey card why should I want to DIY print my own on an inkjet printer? Well I do have a valid answer to that — the cards themselves are a bit bulky and rather fragile (gets marked so easily) so I tend not to carry them around. Hence they don’t get used much which sort of defeats the purpose. If I could print my own I could carry them all the time because I would make small ones and be able to make new ones if they became damaged.
Therefore, on to the test!
First step is to make a Photoshop file and fill a rectangle on it with 18% grey. So what colour is that? If you think in 8 bits like me then I’d guess that would be RGB values of 127 127 127 — however I’ve seen a good argument made that it should be 118 118 118. So I tried both and at least for me the 118 is better — that may well be an accident of my printer/paper/ink combo.
This was on 3rd party noname coated matt “inkjet paper” which actually prints ok on my HP OfficeJet 5500 which is a 4 ink printer. All the printing was done with all colour management off (both photoshop and the printer driver) as I don’t have believable profiles for this printer.
Now the tests
Exposure
Testing at ISO 100 (well, “L1” on my D300 which is ISO100 equivalent) at f/16 in bright sunlight at midday gave me:
- Mennon grey card – 1/125s
- DIY 127 – 1/250s to 1/200s
- DIY 118 – 1/125s to 1/160s
That’s using spot metering in the middle of the card, getting up close and out of focus to average things out. And yes the DIY ones seemed more variable and sensitive to slight changes in position or angle than the real card.
So generally the DIY grey card at RBG of 118 118 118 was more accurate — close enough to be usable, certainly if shooting RAW with that stop or stop and a half lattitude you get with RAW.
But how about for colour correction? To me that’s more interesting really as exposure meters in matrix mode are really perfectly good enough, and you can always chimp the histo or look at the blinkies if you have some extreme cases.
Colour White balance setting
So using the DIY Grey card for WB correct is the next thing to test — and again results will vary wildly depending on printer and paper but for myself the results were mediocre at best. I printed out a sheet and then shot the real grey card beside the DIY one under different light sources — then in ACR I measured the WB using the dropper tool from the different greys and also from the “white” paper.
Results:
Sunlight
- Real grey card – 5150K +8 tint
- DIY grey card – 6000K -5 tint
- White paper – 5450K -10 tint
- Real grey card – 8700K +16 tint
- DIY grey card – 10750K +8 tint
- White paper – 12000K -1 tint
CFL – Compact fluorescent lamp
- Real grey card – 2750K +26 tint
- DIY grey card – 2700K +15 tint
- White paper – 2750K +18 tint
Tungsten
- Real grey card – 3150K +1 tint
- DIY grey card – 3250K +18 tint
- White paper – 3150K -9 tint
Fluorescent
- Real grey card – 5900K +40 tint
- DIY grey card – 6750K +28 tint
- White paper – 6050K +30 tint
I graphed a bit of this and it doesn’t show a lot of pattern I think — neither they DIY card nor the white paper gave accurate readings (which I define as being the same as the grey card). So is the DIY grey card better for this than simply white paper? Yes and no, it depends on the light.
For your amusement here are the test files I used:
So that’s the end of my print-your-own DIY Grey card adventure. I think I’ll start carrying a piece of folded up A4 white paper in my camera bag — yes printed fake 18% grey on one side — because it is better than nothing, or taking a WB reference point from someone’s shirt collar or a piece of concrete (which is often pretty “mid grey”).
But now at least I’ll have a good concept of how inaccurate it is — namely very!
Update:
Hm…. Maybe I’ll go back to “white” paper…