Colourstar 3000 - Cyan LEDs

Nick and Ed and anyone else using the Colourstar, I am still "learning by doing" with my colourstar 3000. I have set up a channel for fully intergrated analysis and with a bit of "cut and try" most of my negs are turning out reasonably well but certainly no better than when I used the much less sophisticated Paterson analyser- at least not as yet but I am still a colourstar apprentice and hope that the colourstar will eventually prove to be a better tool than the simpler Paterson.

Not every time but more often than not when I press analyse I get some LEDs light up in all 3 colours. Manipulation of the Y and M dials will usually reduce the C LEDs but not eliminate them. I then increase the C dial from zero to that point at which the C LEDs go out. I then subtract the same values from the Y and M dials as I have dialled in C and then reduce C to zero again. This leaves me with a number of Y and M LEDs lit as expected which I then extinguish by dialling in both Y and M. Sometimes this eliminates all three LED columns but sometimes the C LEDs will re-light but usually only 1 LED and I go through the process again.

In the end I can always eliminate all three sets of LEDs but have noticed that I usually end up with a lower Y and M setting than would have been the case if I had ignored the C LEDs and simply dialled in Y and M to eliminate these Y and M LEDs.

So as an example purely for illustration let's say I have set the neg, opened the lens to max, focused and swung the diffuser across and put filters back in the light path. I press Analyse and get say two Y LEDs and two M LEDs and two C LEDs at a dial setting of 85Y 80M. I eliminate both Y and M LEDs at a setting of 92Y and 84M with two C LEDs still showing. I could print at this but don't. Instead I attempt to eliminate the C LEDs by the process described above and find that the final settings are 88Y and 82M. Invariably the final settings of Y and M and always less than they would have been had I ignored the elimination of C.

The resulting prints seem OK colour balance wise which indicates that had I not bothered to eliminate the C LEDs then the colour balance would have been too C as the Y and M settings would have been too high.

I hope all of the above makes sense.

Why is this happening? The manual doesn't seem to cover this issue at all. I couldn't see any mention of all three sets of LEDs being on together. So no mention of what to do when this happens.

As I mentioned in a reply to Roger Hicks, the manual was written by someone who perhaps was too familiar with it, such as the inventor of the colourstar who is usually the worse person to write a user manual. Computer geeks write computer manuals which are perfectly comprehensible only to other computer geeks which is why we now have astute publishers making money out of the series of the ABC/XYZ manuals for dummies. OK enough of my rant.

So:1. Why is this happening?
2. Am I correcting as I should.
3. Any observations which will improve my skills will be much appreciated

This is really a follow-on from a previous thread in which I went through the production of the neutral grey strip as suggested.

Thanks

pentaxuser

Yes, this is normal. You are doing everything correctly--- but I'll be damned if I can think of a way to describe what is happening "in three words or less".

Light is composed of three colors - magenta, cyan and yellow ... and your example starts with an imbalance - are you SURE it is equal in all three channels, or is one or two - or all three - "maxed out" - with all LEDS illuminated in that particular array? Correcting the magenta and yellow leaves you with excessive cyan -- but I'm puzzled that the amount of cyan imbalance does NOT change with magenta and cyan correction...???

Anyway, you are doing it right.

I'm going to try to duplicate what you have seen with my (one of my) ColorStars. Possibly, there is a "calibration" issue here???

If I could only stabilize this $%@$$% @#$@ PC! I've got to post my "ColorStar" article.

For me when I get all three lights it usually means not enough light. Too dense negative. Or I've stopped down.

If you're using the intergrator that's going to cost you some light.

Nick Zentena said:

For me when I get all three lights it usually means not enough light. Too dense negative. Or I've stopped down.

If you're using the intergrator that's going to cost you some light.

Click to expand...

If there is not enough light of any particular color, that LED array will flash, as an indicator.

A good test for proper operation is nulling the arrays at wide open aperture and then stopping down until the array flashes. The balance should NOT change more than one LED worth through the entire range.

Thanks Nick and Ed, especially Ed, knowing your computer problems. It's those geeks again.

Just to add. Yes if I have say 3 LEDs in both Y and M as well as in C then correcting both Y and M does eliminate C but often I am left with one C LED and occasionally say 2/3 C LEDs(never 4).

Nick You may have hit on something about the diffuser and light. If you re-call I had said in the original thread that I was still using the Paterson diffuser which compared to the Colourstar diffuser looks more diffuse, although not appreciably so. Maybe I'll try to use the Colourstar diffuser instead to see what difference it makes.

Do either of you know of a book which comprehensibly covers how to use semi-integrated and spot measuring and when to use either instead of fully integrated. I know the manual on page 32/33 mentions spot readings from a model's face( presumably for portraiture shots and for semi-integrated readings what it calls "market shots" using what appears to be a shot of a group of schoolkids and adults.

Well, neither seems to fit my average shot. It's not clear to me what constitutes areas of average density when looking at a colour neg. I am not at all sure that in my more difficult to print negs I have areas of average density. Neither is it clear to me what the colourstar is doing when you take several readings in white light at full aperture from different areas and then the final reading with the filters in place.

Is it averaging for exposure only with white light then in the final reading getting the colour right or is it trying to average for both exposure and colour? I am aware that if one colour predominates with full integrated analysis then the balance my be wrong BUT not how to determine when one colour is sufficiently predominant to affect the balance.

I need to know how to look at a neg on the easel and be able to say to myself. "Ah, a fully integrated reading could be inaccurate here and what I need to do is ........

I'll bet one evening watching either of your work could solve most of my issues but that's not to be.

One of the best books I have ever read on statistics was a book of worked examples by Murray R Spiegel. I could do with one on colour analysing! Getting back to my rant mode again. Sorry.

Thanks

pentaxuser

Ed Sukach said:

A good test for proper operation is nulling the arrays at wide open aperture and then stopping down until the array flashes. The balance should NOT change more than one LED worth through the entire range.

Click to expand...

Ed I'm not sure what you mean. Are you saying after getting the negative analyzed wide open then stopping down it should stay okay? My Durst is pretty wimpy when doing 35mm negatives. A dense negative can really cause problems at anything but wide open. But even a normal negative will cause problems more then say 2 stops closed down. Prints fine but not much use trying to analyze stopped down too far.

pentaxuser said:

Just to add. Yes if I have say 3 LEDs in both Y and M as well as in C then correcting both Y and M does eliminate C but often I am left with one C LED and occasionally say 2/3 C LEDs(never 4).

Well, neither seems to fit my average shot. It's not clear to me what constitutes areas of average density when looking at a colour neg. I am not at all sure that in my more difficult to print negs I have areas of average density. Neither is it clear to me what the colourstar is doing when you take several readings in white light at full aperture from different areas and then the final reading with the filters in place.

Is it averaging for exposure only with white light then in the final reading getting the colour right or is it trying to average for both exposure and colour? I am aware that if one colour predominates with full integrated analysis then the balance my be wrong BUT not how to determine when one colour is sufficiently predominant to affect the balance.

Click to expand...

If all you're seeing is one flashing C LED or other colour I'd say you're over worrying.

When you're in spot mode think of it like taking a bunch of intergrated readings. The differenace is it's only looking at the spot. The computer stores each spot. Colour and exposure. Once you've got all the spots it combines all those numbers into an intergrated number. Basically at that point it's like intergrated except you've managed to created your own idea of what the right combination is.

Think of a small black cat on a big white background. If you just let the computer figure it out it would go "The scene is 95% white and 5% black. So it should be greyish but mostly white. " But you know that the cat and the background have an equal worth. You may actually think getting detail out of the cat is more important. So you choose your spots to bias towards the cat. Even with the cat being only 5% of the whole scene.

It's the same idea with something more complicated. I think I said this before but try this simple idea. Stick a negative into the carrier. Focus etc. Then start taking spot selections. Watch the exposure number. You'll see how it reacts to different selections. It's not that hard to learn if something is considered dark or light. It's harder to see how it reacts to colour. But you can after awhile.

Pick the important parts of the scene.

Basically you're replacing the computer with yourself when it comes to selecting the important bits.

A good test for proper operation is nulling the arrays at wide open aperture and then stopping down until the array flashes. The balance should NOT change more than one LED worth through the entire range.
__________________
Carpe erratum!!

Ed Sukach, FFP.

Edand Nick. Ed. I have just tried your test above. Using full integration and then only at f16( min aperture on my 50mm El Nikkor lens) are there three LEDs lit. These happened to be M LEDs and also 1 C LED. No Y LEDs. This may be because of the particular neg. I can't say for sure as I didn't try other negs but trying to recall what I have seen with other negs when actually printing, I think it's always the M and C LEDs that light up. Usually its only two M LEDs and 1 C LED.

I have just ignored this in the past. This fails your test of no more than 1 LED until the row starts to blink due to lack of light. I have never had this happen. Mind you even at f16 the exposure is quite short on some negs and seldom more than 10-14 secs. I usually use an aperture that gives me between 5 and 10 secs. At these times no more than two LEDs light. It may be that due to the size of prints I do(normally 5 x 8 inch) the lens is quite close to the easel and so the light levels may be too high to trigger the signal for inadequate light even at f16.

If the colour balance on most negs is OK with full integration do I need to be concerned if two or even three LEDs in any one row light up at small apertures.

Just going back to the grey test if I may. I eventually arrived at log readings of 0.56 across the board and left things at that as it was pretty close to 0.55.

I had another look at my test strip today. With a blue bulb which is meant to replicate daylight in the darkroom the strip looks a green/grey. In a North facing room in overcast daylight by a North facing window it looks a little more grey especially if I shade it a little with my hand but it isn't a clear cut grey as I understand grey. I was expecting it to look the same shade as a pair of grey flannels I have but it doesn't.

Despite this, need I be concerned about the colour of the strip if the readings are 0.56?

Thanks for your patience and devoting time to my Colourstar apprenticeship

pentaxuser

pentaxuser said:

Despite this, need I be concerned about the colour of the strip if the readings are 0.56?

Click to expand...

Do you mean you've got three 0.56 instead of 0.55? In my expierence if you're at 0.56 you are golden. Trying to get better will just get you further away.

Nick Zentena said:

Do you mean you've got three 0.56 instead of 0.55? In my expierence if you're at 0.56 you are golden. Trying to get better will just get you further away.

Click to expand...

Nick. Yes that was my experience too. When I tried to get closer than 0.56 I only got further away.

However I was left with a strip that looks grey/green whereas at about 0.59 it was clearly a dark grey. In many ways my experience replicated that of Frances Schultz who also mentioned a dark blue grey at the start,becoming a dark grey then neutral grey. Unfortunately when she arrived at, I think readings of 0.57 and was satisfied with her results she didn't mention that the neutral grey has some green tinge because it presumably didn't, so I was left wondering if it's my eyes or if something had gone wrong with my calibration.

As neither you nor Ed have mentioned a good colour printing book which covers when to use integrated, semi integrated or spot measuring, I take it that no such book exists.

A pity. We seemed to much more poorly catered for than B&W printers where a host of good "how to" books exist. It does seem that analogue colour printing is poorly catered for compared to Photoshop.

It's almost as if a student seeks to learn how to make hand produced furniture like Chippendale and the instructor says "Look, son, it could take you 40 years this way. On the other hand here is a Chippendale machine which is numerically controlled and requires no real skills once you have pushed the buttons in sequence. So which would you prefer, a Chippendale machine that produces a perfect chair after a couple of weeks or nearly 40 years of disappointment at the end of which you produce, at best, the same chair."

pentaxuser

I don't know about books. I think Kodak used to have a book on colour but I doubt it would include something like which probe to use or how.

Your test strip should be grey. What sort of light are you using to look at it?

You might want to describe your steps for calibrating. I think you're using a diffuser over the lens. You take that off to do your reading with the probe?

Nick Zentena said:

I don't know about books. I think Kodak used to have a book on colour but I doubt it would include something like which probe to use or how.

Your test strip should be grey. What sort of light are you using to look at it?

You might want to describe your steps for calibrating. I think you're using a diffuser over the lens. You take that off to do your reading with the probe?

Click to expand...

Nick Yes. The calibrating steps were done minus the diffuser exactly as described in the manual and the Frances Schultz article. Ilooked at it in a North facing room in daylight with bright but overcast daylight outside. As I said it was slightly less green/grey looking in the room and even less so if i shaded the strip with my hand. When the readings had been about 0.59/60 it had looked dark grey, even in the darkroom which is equipped with a so called daylight bulb which looks blue if you look at it.For what it is worth I noticed that the grey spot probe calibration which gave me the 0.56 figures resulted in a channel density reading of 493 which I have had to reduce to 490 and even less for some negs for fully intergrated reading because 493 gave exposures which resulted in rather dark prints.

Maybe using the diffuser means that exposures have to be reduced anyway so 493 is OK for spot probe readings but not relevant for fully integrated readings.

If I am honest I cannot see how the time spent doing the grey test helped me get where I am. Maybe if I knew how to use the spot probe properly to get a correct colour balance and exposure I'd see the point of going through the process.

As things are I would have been better to use one neg and get the exposure and colour balance right and then calibrate on that which is what I have done effectively. The so called perfect print process which is what is required with the Paterson analyser.

As I said every so often I come across a neg which doesn't lend itself to a fully integrated reading. It gains a slight colour cast and can be wrongly exposed. The trouble is I only discover this after printing and then have to guess how to correct it.

So the analyser doesn't seem to help. OK neither did the Paterson but it was much simpler and I was hoping that the Colourstar would lead to less difficult to print negs.

To give two print examples:1. I took a shot of a statue of a dog which was a stone grey on a pillar surrouded by sky. On full integration the sky was a nice if very dark blue, helped by a polariser but really the sky was too dark and the dog was slightly brown and much too dark.

If I had used the spot probe where should I have taken readings?

2. I took a shot of a group of people standing in front of a red brick kiln. It was a little dark suggesting a slight over exposure by the enlarger. When I reduced the exposure by guessing the correct exposure, I had not only got it right, by chance, but noticed that on comparison the darker print had a slight blue cast in the concrete slabs which should have been a light grey. This cast disappeared in the lower exposure without any change in the filter settings.

So maybe in the second example the colour filters were correct but density channel reading was too high. There doesn't seem to be a happy medium which covers all the negs on the film. Maybe I shouldn't expect to have a set of channel settings which can cover all the negs but the problem is I cannot tell by looking which of the negs need a change of exposure. Hence my quest for a "How To" book.

Thanks

pentaxuser

IIRC every 10 points is 1/3 of a stop. So when you went from 493 to 490 that was a fairly small move.

On the dog picture I would have taken spot selections mostly of the dog. With limited readings of the sky. Maybe one or two readings of the sky with all the rest being the dog, the pillar and anything else.

On the second one. I would have used most of the selections on the faces. Then the kiln. Then maybe the clothes they were wearing.

It seems in both of these you've got a big more or less single coloured object. The sky and the kiln. When you intergrated those things get a heavy weight. Which then leads to problems.

Thanks to Ed and Nick from me as well

Wading through APUG and rec.photo.darkroom, and anywhere else I could find information on the Colorstar3000 led to a number of great discussions that usually involved Ed or Nick in some manner. It is a shame that they could have not been involved in writing the manual for the thing from the start, for it is a rather tortured thing to read the first time through.

The 3000 analyser is a joy though. Pre analyser I would work all night to get a good print out. Frequently now the finished print is one print past the test print. And the ability to master out aging chems and paper against all channels automatically after one grey print is a god send.

I bought mine new off ebay from a place in portugal, loaded the manual off of Nicks web site, re-configured the wiring to convert it to 120v (quite easy to do - one or two re-soldered jumpers) and have been pleased with it ever since. The money paid for it was repaid with the pleasure of printing colour again.