Question concerning blues and purples with D70

[ozimax]

These are a couple of shots from some recent weddings I shot. On the two screens at our place, one LCD and CRT, both blues were identical, but I got some prints back yesterday and the second one is unmistakably purple.

Is it true that dlsr cameras struggle with reproducing purples and blues correctly? Anyone have experience with same?

Max

Blue dress


Purple dress

[Alpha_7]

On my LCD the second shot looks more purple then the first.
I'll let someone wise inform you about the blues and purples, as it's beyond my knowledge, but atleast on my LCD there does seem to be a marked different in the colour of the dress.

[gstark]

Max,

Are your monitors calibrated?

I suspect that we're seeing minor WB issues.

The first of these images looks a tad blue-ish. Look closely at the lady's shoulders amd necklace and you may see what I'm seeing.

The second of these looks slightly purple-ish (still blue, but there's a touch of red in it) but the whites and the skins, which is where I make my primary judgement, look ok. Looking very closely at the faces, I'd be willing to suggest that there's a bit too much magenta here, which would explain the "redness" I describe above and the apparent purple that you're seeing.


I note that the first of these was shot in shade, whereas the second one was shot in bright sunlight, and that can certainly have an effect on the WB, and especially if you're not adjusting or compensating.

I'd like to see these WB issues fixed before we go further.

[Matt. K]

Max
Your printer can't print all of the colours captured by the D70. Your blue is probably out of gamut. This can be a bit of a problem for folk who shoot in ADOBE RGB colour space...they are capturing more colours than the computer can display or the printer can print. I have had the same problem and if the printer can't print it then you just have to wear it or change it to a different hue. You can preview out of gamut pixels by clicking VIEW/GAMUT WARNING in PShop.
Cheers!

[timbo]

Now I could be putting my socks on over my boots here Max...

... but isn't it the same chick?

Identical jewellery, hairdo, lipstick... the dress even looks the same but for a slight shift in hue. Were you shooting the weddings of identical twins, or are you pulling a swifty?

Tim

[timbo]

These are a couple of shots from some recent weddings I shot.

I get it now, it HAS to be the same girl. I'm guessing it was the same wedding

[gstark]

Tim,

Same wedding, same lady, same day, same dress.

Max's concern is that in the actual prints of these two images (and probably others), the colour of the dress looks different, whereas it shouldn't.

[Zeeke]

Dont forget the fabric has a metallic sheen to her dress.. it would be probably having different tones as it is hit by light from different angles and that is possibly affecting the way you are seeing things... might be blue to your eyes on the day.. but.. if you held the fabric on different angles it would change tones..

Tim
p.s. sorry.. prolly didnt help.. but just an idea on why you might be seeing things

[timbo]

Thanks Gary: it took me a while but I got it at last. D'oh...

I agree it's probably a WB issue. Max, were you shooting in RAW? If so you can tweak the first one and select the "Previous Image" option on opening the second file in PS. It's a great way to get colour fidelity in a sequence of shots.

[sirhc55]

One must also consider that certain colours, and in particular certain materials, will reflect light differently under certain conditions.

The dress appears to be made of satin or even sateen and the reflective qualities of these materials DO differ greatly under different lighting conditions even to the point of changing colour perception.

This would not be a fault of colour balance or even WB, but of the material itself.

[Aussie Dave]

Hi Max
I assume that you had your prints done at a lab & not at home on your own printer ?

In either case, if your monitor(s) are not calibrated to be the same as the output from the lab (or your home printer), then discrepancies will begin to creep in. If they are wildly out of calibration, then you will notice a huge difference.

Also, WB issues and using colours "out of range" for printers & monitors will also hinder your ability to print what you see onscreen (or vice versa).

Occasionally, when I have prints done they are a touch too red or too green etc... I don't have my monitor calibrated to what the lab prints, so I can't complain. Sometimes when you sit in front of the PC for too long, your eyes play tricks on you and images look fine - until you walk away and come back and then you see that something is not quite right.

My advice would be to either calibrate your monitor(s) or re-PP your images and allow for how the lab is printing (obviously, you're off on the red channel a bit).

Quickly, another concern can sometimes be that smaller labs don't keep their units calibrated as well as some others do, so their printers can also fall out of specs - causing another problem....

Isn't all this technology fun

[Raskill]

I can't offer much in the way of technical advice but I recall a thread similar to this on Dpreview when someone took a photo of a guy at a function who was wearing a horrid purple suit. The D70 flipped out big time and produced almost a negative type image (the image was posted for all to see).

http://forums.dpreview.com/forums/read.asp?forum=1034&message=12881748

The images no longer appear but you might get some advice in the posts that follow.

Hope this helps.

[ozimax]

There's a lot to take in here. Some further comments.

WB was auto. I always forget to change it if I use manual WB settings so I keep it on auto. I don't shoot in RAW as my 800mhz iBook, while working perfectly well, struggles to process JPG files. To process RAW files is not an option as I don't have all year to wait for files to open and process. Maybe I would use RAW if I had the processing power.

Having said all that, if I come across a great surfing situation or the prospect of a wonderful sunset/landscape, then I usually switch to RAW.

My wife made the aforementioned dress, I think from satin or something similar. She says that it definitely changed colour depending on the light source, be it sunlight, flash in shade, incandescant or flouro.

The prints were from a lab, and the second print was the "purpleist" purple you've ever seen.

No, I haven't bothered about calibrating my iBook screen. It's not the greatest screen, just adequate, and I use many different printing establishments who all use different profiles, paper etc. If I'm printing a large 30"x20" or similar, then my printer will usually colour correct for me and match his printing setup.

Thanks for the input all.

Max

[ozimax]

The first of these images looks a tad blue-ish. Look closely at the lady's shoulders amd necklace and you may see what I'm seeing.

The second of these looks slightly purple-ish (still blue, but there's a touch of red in it) but the whites and the skins, which is where I make my primary judgement, look ok. Looking very closely at the faces, I'd be willing to suggest that there's a bit too much magenta here, which would explain the "redness" I describe above and the apparent purple that you're seeing.

It's funny you mention this Gary, but on my monitor, the whites and skin of the second shot look red, and I can't see any blueishness in the first. As I said before, I don't think my iBook screen is very accurate but it's all I have at the moment.

Max

[gooseberry]

The D70 suffers a little from IR contamination - the colours of clothing especially will appear slightly off depending on the amount of IR present. If you are really concerned, you can try using a hot mirror filter which will improve the colour response dramatically. Remember seeing some examples on the net sometime ago.

[gstark]

The first of these images looks a tad blue-ish. Look closely at the lady's shoulders amd necklace and you may see what I'm seeing.

The second of these looks slightly purple-ish (still blue, but there's a touch of red in it) but the whites and the skins, which is where I make my primary judgement, look ok. Looking very closely at the faces, I'd be willing to suggest that there's a bit too much magenta here, which would explain the "redness" I describe above and the apparent purple that you're seeing.

It's funny you mention this Gary, but on my monitor, the whites and skin of the second shot look red, and I can't see any blueishness in the first. As I said before, I don't think my iBook screen is very accurate but it's all I have at the moment.

Max, it's a shame you're not closer and could see this the way I'm seeing it. What happens if you pull some the reds that you're seeing back a little? Is doing that and getting the print redone from that a viable option?

[darb]

Dont forget the fabric has a metallic sheen to her dress.. it would be probably having different tones as it is hit by light from different angles and that is possibly affecting the way you are seeing things... might be blue to your eyes on the day.. but.. if you held the fabric on different angles it would change tones..

Tim
p.s. sorry.. prolly didnt help.. but just an idea on why you might be seeing things

I find some dark dark reds and browns do this ... i have some leather shoes that are dark dark brown, but in most light you would swear theyre a deep burgundy. (and im not meaning when a glass of red goes over... )

for the record i can almost barely notice the difference ... and wouldnt have unless it was poinetd out and closely scrutinised ... plus a warmer White Balance to flatter the girl, isnt a bad thing.

[paulvdb1]

Max - on my LCD the first shot looks too cool. The second shot looks quite a bit warmer. That said though - in both cases the dress looks blue, even if somewhat different. I'm running my monitor a little warmish so I'd have expected the second shot to be truer-to-life.

[Dug]

It has been my experience that most cameras films and prints have problems with some shades of blue. in the 70's we used to hate jacaranda season because people would always complain we printed the colour wrong.

it was just that no paper made could reproduce that shade of blue at that stage.

It has got a lot better but the change in light source is probably the main cause in the colour change. Do you have any scraps of the material left to do some tests on?

Below are some Iris shots I am pretty happy with the way the colour turned out on them.

[ozimax]

Do you have any scraps of the material left to do some tests on?

Good idea, might try some experiments.

[the foto fanatic]

I suspect that lighting and white balance are the issues.

The lighting is obviously different in the two pix (looks loke shade for #1 and direct sun for #2), and AutoWB may not have quite captured the difference.

[gstark]

Max,

Regarding reshooting with just a sample of the fabric, consider also including a reference chart (see our front page for one) within the image.

Download and print it out, and include this within the image, and then match your colours to the reference print, rather than to the fabric.

Also consider using a preset WB rather than auto.

Mind you, this won't exactly replicate the shooting conditions encountered at the wedding, but it can give you some guidance as to what is going on.

[ozimax]

Will give it a try Gary, I don't think I will get time today but we'll see how things go.

I might also shoot in raw for a change. Will definitely use manual WB settings.

Max