Okay, all you armchair photographers...

[cliff_rames]

Lakhinder's last set of ColorChecker examples shows a saturation anomaly. This would happen if the lightnesses were adjusted in Lab mode, due to the uncoupling of lightness from chroma. To avoid this I would suggest adjusting the middle gray patches in RGB mode.

 

Interesting thread.

[l_a_k_h_i_n_d_e_r]

<i> I didn't read your posts. </i><br><br>

I like your style, Eric! However, my style is to do some ground work and then present it. I do not mind being proven wrong. But not by empty assertions.

[l_a_k_h_i_n_d_e_r]

Cliff you are right. That is what I did. I do not use LAB mode, but

decided to use it to normalize the values, since my reference values were in LAB mode. I shall present again with RGB mode correction.

Many thanks for catching this anomaly.

[cliff_rames]

I take it back. There are other things going on here.

 

How are these negative scans being inverted? Invert in Photoshop produces a tone curve that is far from anything you would get by making a photographic print.

[rowland_mowrey]

Lakhinder;

 

That might explain what was confusing me about those last samples. I think that might clarify things more. I'm looking forward to the reposts. Now, I'm for bed.

 

Its late here.

 

Ron Mowrey

[EricM]

"I like your style, Eric! However, my style is to do some ground work and then present it. I do not mind being proven wrong. But not by empty assertions."

 

Cheers Lakhinder, but "empty assertions", i don't get it?

[l_a_k_h_i_n_d_e_r]

Some general comments:

 

1. Going back to the very first 4 images I corrected and posted above, let me add that when I did the curves based correction in PS, it is not individual color correction. But only the overall curve. Which should mean brightness and contrast.

 

 

2. Scans are handled by Imacon software-- the variable I can not

control. I have corrected the individual curves there of R,G,B to

produce consistent gray scale value. (That change is a new variable

by itself! . No contrast changed while scanning, but that would perhaps be unavoidable if I correct individual R,G,B curves for gray scale.)

 

3. I am not sure I can correct just by handling brightness-control of Photoshop even a properly exposed negative

to get its image to tally with the gray scale reference values --

I quoted above page 221 of Real world Color Mangagement. Let us forget the other colors in the chart for the moment. I need to adjust the contrast as well. And that is what I was doing originally! The first four images.

 

To summarize the net result of #1 and #2 and #3 is, the whole process

of scanning is non-linear I would imagine. (Plus the fact that the film curve is non-linear in the two extremeties.). And just controlling the brightness is not helping here. I am not presenting these images to say that this is THE way to find the film dynamic range or latitude! Just like printing off Frontier is hardly a scientific method to evaluate film response. I have already pointed out the anomaly where 3-stops underexposed' has lot of shadow detail-- the images posted by James.

 

BTW, Cliff, these scans are from Imacon scanner, and are not inverted in Photoshop.

[rowland_mowrey]

Weather permiiting, I'm going to shoot an exposure series Sunday or Monday. I will be using EK Portra 800 film for the test. If all goes well, I hope to post some results. I may start another thread. It will take me a while to expose, process, print, and scan these results, so don't expect results this Sunday or Monday.

 

I may be able to include other films in this test. IDK yet. I'm working on it.

 

Ron Mowrey

[elliot_n]

You might want to test Portra 800 under tungsten lighting (without filtration), as this

is how it is often used (by low light photographers).

[rowland_mowrey]

Elliot;

 

Thats what was in my camera. I was shooting a moving 2 year old in heavy shadow and sunlight Saturday, and I didn't want to waste the rest of film. As it was, my 220 back had Portra VC 160 in it, so I shot that as well. I have the film ready to process tonight. I used my RZ instead, but had some trouble with the 800, as the shutter is limited to 1/400th, so I got 1600, 800, 400, 200, 100, 50, and 25 speeds. I got the same series with the 160, but the 160 exposure is in there as well. I used even stop increments otherwise. I just hope they turn out.

 

I metered through the lens, and varied the f stop for some of the exposures. I used the speed setting on the backs to vary the speed.

 

Wish me luck. The hardest part will be to decipher my notes and match them to the negatives. I suspect I'll have to repeat this again to make doubly sure of my results.

 

Ron Mowrey

[rowland_mowrey]

Today, I shot 3 rolls of film. Portra 800, Portra 400, and Portra 160 VC. I used my Mamiya RZ. The scene was composed of the MacBeth checker and a Kodak color step scale and grey card.

 

The 800 and 400 did not allow a good enough exposure range due to the shutter / f stops available to me, but I got some data.

 

The 160 gave a range from ISO 800 to 25.

 

I have 7 exposures coming up soon. They are drying. I will scan them in first hoping for the best. I have 800, 400, 200, 160, 100, 50, and 25. That is approximately 3 stops over, and 2 stops under using ISO 100 as the center point, and 160 as the 'correct' exposure.

 

I also photographed the setup so you can see it.

 

Ron Mowrey

[rowland_mowrey]

Here are three scanned frames from the Portra 160 VC series.

 

The left picture was exposed at ISO 25, the middle is 160, and the right one is ISO 800.

 

Epson scans with exposure correction enabled. No photoshopping.

 

Pasted into Paint, and saved as one JPEG image.

 

Ron Mowrey<div></div>

[l_a_k_h_i_n_d_e_r]

<i> Epson scans with exposure correction enabled. No photoshopping.</i>

<br><br>

Rowland, thanks for the experiment. I am impressed by the response at ASA 25. That is nearly 3 stop under exp.

<br><br>

To clarify my results above, I did no exposure correction in scanning stage, except to match the R,G,B guns to handle gray chart accurately (Since Imacon has no profiling built in).

<br><br>

I could be wrong, but looks like Epson software takes the extreme black and extreme white into account in order to correct the scan. Something which I tried to do in Photoshop, just so that I could control better. I really don't know anything about Epson software.

[bernhard]

Ron, again a scan suggesting that underexposure by 2 stops looks even better than over exposing, but we all (think we) know, that overexposure benefits C-41 films.

What usually happens with underexposed neg film is that they start to get grainy, so how does your Portra look in full rez scans?

 

Another note: I wonder whether scanning (hardware, software, operator) is a big factor here in these results as both you and Lakhinder get better looking scans from underexposed negs and you both seem to rely on what your Epson or Imacon software is doing.

[rowland_mowrey]

Well, first, Epson scanning software has the 'adjust for exposure' button turned on as default. It cannot be turned off. If you click 'set to default' it turns the button off, but as soon as you click on 'scan' the button turns itself on again. A 'depressing' state of afairs, to be sure. I wanted no image modification. I guess the only way to get that is in the conventional proofs.

 

I intend to print a proof sheet of all 7 negatives on one 8x10 sheet of paper, and use 3 - one stop brackets around the center point to show the degree of variation.

 

I expect that there will be severe sharpness and grain changes with that extreme of exposures. I have seen it before. Flare alone, at 2 stops over should be a big factor in sharpness. Grain at 2 stops under should be bad. But when I make 8x10s of these, that should resolve that part of the issue as well.

 

Remember, I did this with 800 and 400 as well, and they showed me similar ability, but I could not get the range of underexposure due to camera limitations. Therefore, I did not post them.

 

The best bottom line I can come up with right now is that virtually every color negative film can be over or under exposed by up to one stop with no significant problem for most casual users. The discriminating professional may want to push or pull process to trim the results, and they may be unhappy with the results wrt grain and sharpness.

 

In a pinch, negative films that I tested seem to have a 2 stop full range latitude on either side of the true ISO in which you can get a usable (not good necessarily) result. In other words, the film won't be blank. You will have an image. Grain and sharpness are waiting in the wings for proof. I'll be getting around to that in a few days.

 

Regards.

 

Ron Mowrey

[EricM]

Ron, can you get the Silverfast software for your epson?

[rowland_mowrey]

Eric;

 

I just bought the Epson a few months ago. I'm still getting used to it and have not looked into alternative software.

 

Ron Mowrey

[rowland_mowrey]

To answer an e-mail request, here are the originals, unresized in Paint, as they came from the scanner.

 

Don't be too disappointed if there are any problems. Don't be overelated if they look better.

 

Ron Mowrey<div></div>

[rowland_mowrey]

And the next.

[rowland_mowrey]

And the last.<div></div>

[bill_tuthill]

Portra 160VC as rated: dark skin is too dark, top-row green is too

brown, red is too dark, magenta too purple, but grays are excellent.

Does this film have too much midtone contrast, or is it the scanner?

Similar color problems at EI 25.

[rowland_mowrey]

Bill;

 

IDK. Wait until I get proofs on Endura.

 

Then we can see what traditional prints look like. They could be awful. I hope not.

 

Ron Mowrey

[rowland_mowrey]

Bill;

 

I have the chart on my lap, and compared to what I see on the monitor for all of the scans, the pictures at 25, 50, 100, and 160 are all too light. All of them are bluer than the original. They need more yellow.

 

Ron Mowrey

[elliot_n]

'Wait until I get proofs on Endura...'

 

Hmm, I can tell you the results:

 

Prints from the frames over-exposed by one, two or three stops will all look the same

as the print from the properly exposed frame.

 

Prints from the underexposed frames will look flat and lifeless.

[bill_tuthill]

If anyone wants to see an sRGB-colorspace Macbeth chart to compare,

see dpreview.com's review of the Nikon D70, page 14, or their review

of the Olympus E-1, page 15. On my monitor these look closer to a

real Macbeth chart than those produced by the Canon D60 (way too dark)

or 10D (still too dark).