PDA

View Full Version : Colors of architectural surfaces



Leonard Evens
13-May-2006, 12:13
I've been photographing buildings using Portra VC 160. I scan using Vuescan on an Epson 3200. After some adjustment in a photoeditor, I don't have too much trouble getting vaguely plausible colors of building surfaces using neutral elements of the scene. I also photograph a gray card with a digital camera, and that helps me get a bit closer. But I still am not close enough to what I feel the actual colors should be. This is particularly a problem when I use wide angle lenses and the building surfaces show significant variations anyway. Strong colors usually don't present as much of a problem as various shades of "off white".

There are a variety of problems, which I feel I don't completely understand. I use a center filter, which reduces intensity drop off from center to edges, but it appears there may also be a slight color shift which remains. In addition, I think the quality of the reflected light may vary depending on the angle to the lens. On top of this, the natural variations in color already present complicate matters, as do deficiencies in the film, the lens, and the scanning process.

I would appreciate any pointers to help me understand these issues and websites showing examples I can copy and measure.

tim atherton
13-May-2006, 12:40
leonard,

have you tried using NegPos with vuescan for colour negs? I find it excellent

(bit of a learning curve - lots of info on colour negs on the site - quirky vuescan like interface...)

tim

Gordon Moat
14-May-2006, 14:05
Hello Leonard,

Recall that colours on a computer monitor will only be a simulation of what a printed output might show. If you are only processing the images for internet, broadcast, or multimedia display, then don't worry too much about slight variances; most viewers will not have their monitors or televisions even remotely close to proper specifications.

With print output, you should be judging the actual print, or an approved proofing output. Anything near light green, to near Cyan, just will never properly display on any computer monitor. Certain inkjet printers and paper combinations will show differences in subtle tones depending upon several combinations, or even environmental factors.

If you can be more specific, or post a link to an example, I might have more specific suggestions. The best way to test a centre filter is to photograph the same scene with the filter in place, and then with it off. If you are doing adjustments in post, you might find that shooting without the centre filter, then creating a mask to mimic the filter might work better.

Feel free to contact me off group, or to reply here. I don't check here that often, so if this is time sensitive, just e-mail me.

Ciao!

Gordon Moat

Tom Westbrook
14-May-2006, 14:34
I second the NegPos recommendation. I checked it out after Tim recommended it a while ago and am very happy with it's perforance. They have an evaluation verion on their site (tho you get a newer version if you buy) at http://www.c-f-systems.com/Plug-ins.html.

Kirk Gittings
14-May-2006, 21:09
I don't find Neg/Pos useful over a good scanner software combination like Silverfast and one of the higher end Epsons or Nikons.

I need to know a little more about your shooting methods. Are you using strobe fill in daylight or just natural light or quartz fill or what?

tim atherton
14-May-2006, 22:03
kirk and I are at opposite ends of the spectrum on this - I've never really liked silverfast, nor have I found their film profiles to be that great (for colour or B&W - and I've used it with a number of different higher end scanners) - he's not so keen on vuescan and negpos....

adrian tyler
14-May-2006, 22:34
tell you what thiough leonard, whatever architectural surfices you shoot it can't he worse than frank gehery's "marques de riscal" wineary in the north of spain. i have been working on the project for 3 years and he has uses 2 different tones of titanium, stainless steel and sandstone. the colours in the titanium change subtly evey minute of the day, 365 days a year.

http://adriantyler.net/ficha.cfm?idcategoria=35

(very crude edit)

back to your question, use a reference in the picture that you are familiar with when printing, grass, flowers, granite, rust, whatever. i find that if i print to get that right the rest more or less falls into place.

Kirk Gittings
14-May-2006, 22:42
True Tim, I don't find anything in Neg/Pos that I can't do in SF, including developing and saving very detailed personal profiles for particular films. The Sf profiles are a good starting point but I can tweak them to my hearts content.

Plus scanning the negative basically as a neg transparency in a raw state and converting it to a postive in a PS plug-in (Neg/Pos) generates more noise in the transition tonal areas than being able to apply curve adjustments at the scanning stage in SF. Any workflow that includes significant contrast enhancement after the scan in PS will greatly enhance noise. The more noise in the original scan the more it will be enhanced by a steep curve in PS, so the Nikons are better for instance than the Epsons because they have less noise in the original scan. No matter what I did the N/P workflow produced significantly inhanced noise over the traditional SF workflow because the major curve adjustments are applied in the scanning with the SF method I suggest vs. a flat raw scan for the N/P workflow with contrast enhancements latter in PS. Based on yours and others recomendations, I really wanted N/P to work but after a couple of weeks of testing it I came to the conclusion that it was an inferior workflow and that was true whether I used Sf, ES or Vuescan to generate the original scan.

Some have suggested in the past discussions on this forum that only in N/P can you tweak the individual color channels while viewing all channels simultaneously and that only in N/P can you develope and save your own profiles. These are false assumptions by people who do not fully understand the capabilities of SF, because both are easily accomplished in SF.

Kirk Gittings
14-May-2006, 22:52
A couple of the tricks we apply to control color shifts (beyond problems with scan profiles).

Put a small greycard in the scene in an place it is easy to remove later in PS. Use the middle grey eyedropper in Curves to nuetralise the mid tones and key the colors overall or do the same in SF in the scan.

COLOR MECHANIC PRO, COLOR MECHANIC PRO, COLOR MECHANIC PRO, COLOR MECHANIC PRO. Let me be clear on that-use the PS plugin COLOR MECHANIC PRO. It is the production architectural photographers best friend.

julian
14-May-2006, 23:57
Some have suggested in the past discussions on this forum that only in N/P can you tweak the individual color channels while viewing all channels simultaneously and that only in N/P can you develope and save your own profiles. These are false assumptions by people who do not fully understand the capabilities of SF, because both are easily accomplished in SF.

Didn't ay that Kirk - or if i did i didn't explain myself well. I said that only in negpos can you calibrate and manufacture your own convert profiles. I've seen this discussion a lot and I think it boils down to whether you like to treat a film scanner like a drum scanner and do all the colour in the scanner software, or if you want the scanner software to not meddle with the data and manage the data in photoshop. IMO there are damn good reasons for the drum scanner workflow IF you use a drum scanner and less so for a film scanner. I know Kirk uses smaller print sizes than I do, but a 1mx1m print from a neg I see much richer tonal values, smoother transitions in negpos than in sf. But hey, whatever works for you

julian
15-May-2006, 05:07
In fact re reading my last post I realise I still haven't articulated myself well, so...
SF IMO has a fundamental confusion between neg conversion and colour processing. You have canned convert settings and then you adjust or meddle with the histos to your hearts content. A RAW scan gives you everything on the neg you can get, negpos allows you to calibrate the convert process, photoshop then does colour. On top of that negpos allows you to compress hilight tails to give fatter histos where it counts. If you like the SF workflow is a bit like having to compensate for bad exposure or development in the print in the analogue workflow...

Kirk Gittings
15-May-2006, 07:13
Maybe I didn't understand your full workflow, Julian can you describe it in general again please.

Kirk Gittings
15-May-2006, 08:24
Julian,
Maybe I misunderstood your comments. Someone did argue specifically what I said above saying he could not do those adjustments in SF.

But explain for me again your workflow. Maybe we can learn something.

julian
15-May-2006, 11:12
Julian,
Maybe I misunderstood your comments. Someone did argue specifically what I said above saying he could not do those adjustments in SF.

But explain for me again your workflow. Maybe we can learn something.

Might be interesting if other negpos users chime in here .-)

I've found lots of differences between scanners, so what i 've found only really applies to a dimage scan multi pro. I do a 16bit linear raw scan using the minolta twain. I've found the vuescan 'raw' to be not as good as the minolta, and negpos needs a linear raw 16bit scan to work best (not all 'raw' scans are equal IMO) - note a straight conversion from this gives a very washed out image. My aim is to get as much untouched data as I can into photoshop - I'm in the 'art' area so having to do batch scans, or high volume isn't an issue. I've followed all the calibration routines on the negpos site and and set up my convert routine, I've also got various setups in negpos to compress the hilight detail and give me more mids. So, I do a linear raw scan, convert using a fully calibrated negpos which gives me a fat 16bit histo in PS and then I colour correct in ps and any artistic stuff I need. I basically go scan/convert/interpret with the scan/interpret stage being fully calibrated. The convert stage isn't a matter of fiddling but of measuring - this comes later. The convert also isn't an ICC profile, its a way of controlling the mapping of the bits. Its the same as sorting out ISO/dev times for BW film processing. However all this stuff is akin to Mac v PC and it doesn't really matter as long as you like what you get
!

Gordon Moat
15-May-2006, 14:45
Concerning the Neg/Pos workflow. I mostly found that very useful when scanning B/W films, though I should state that I rarely shoot colour negative. It seemed to be more scanner dependant, on when to use this. My adjustment in the scanner software for B/W would be to get an even Histogram across range, the idea being capturing a wide range of information. Then invert in PhotoShop, and further adjust as needed (if needed).

On a Creo (or older Scitex), I found little need to do this in oXYgen scanning software. Seems that the higher end software (and scanners) do a better job at converting negatives. It is more common for me to have everything correct in the scanning software, often to the point of only needing PhotoShop to add in a copyright notice or do resizing.

Regarding the Neg/Pos I do for B/W, it seemed that the more limited the scanner in actual Dmax, the better this method worked. However, the benefit was largely visible in shadow details on the low to mid range scanners on which I used this method.

Comments on SilverFast: I find this to often be better than the scanner software included with some low to mid range scanners. However, it is not easy to learn, often not intuitive, and depending upon the scanner brand might not be worth the extra expense. I think this is why there is a free download trial available, and people should try before buying. I have also used Vuescan, though I have yet to find a scanner on which I liked using it, though that is mostly my personal preference.

Ciao!

Gordon Moat

Kirk Gittings
15-May-2006, 15:07
I currently work with four scanners for a variety of reasons, an Imacon 949, an Epson 4990, a Microtel 1800f and a Nikon 8000. I have found that any workflow that starts with a fairly flat scan and enhances contrast in PS causes noise ehancemaent. Every scanner I have ever used has this problem except for when I pay for a drum scan. The Neg/Pos workflow caused this in spades with the above scanners. Michael Mutmansky explains my experience with this best:

"I believe it is important to get is reasonably close to right in the scan, and not leave it to PS manipulation later on. Every manipulation to the file after it's creation destroys data in some manner, with the exception of some basic rotaional transpositions and other minor things like that.

Curves, Levels, color adjustments, etc. all ultimately are destroying data that was in the original file, which is why it is so important that the manipulations are made as Levels adjustments, and are not applied directly to the background.

In the case of Curves, as you have suggested, the adjustment of the midtones causes a seperation in the steps somewhere in the curve, and a compression of the steps in another part of the curve. When the seperation is drastic enough, it will cause an increase in the apparent 'grain' that is visible. In this case, I mean noise more than grain, but it does ultimately come off looking like digital grain.

This is partially exacerbated by the fact that most people consider a digital file to be a continuum, rather than a stepped data source. Further, PS is apparently not truely 16 bit, it actualy operates at 15 bits. Even more, most B&W files don't have anywhere near the theoretical number of possible levels of gray in them, most have much fewer than 1/2 and high bit files (16 bit) may often have 1/3 or 1/4.

Additionally, the problems most easily crop up in areas where there is a discontunity in the tonal structure of the file, because this is where a large step is already being made in the data.

This can be made worse if the source file is a general scan to cover the entire density range (the raw scan output approach), because then the useful data does not cover the entire range of the image file, and if there were possibility for some interpolated data within the useful data set, it was thrown out in favor of capturing useless data beyond the image density range.

Suddenly, 16 bit doesn't sound like too much, does it? Image how careful you have to be when scaning at 8 bit? My drum scanner has a 12 bit output, and writes a 16 bit file so it is useable in PS, but it only has 12 bits of true data in it. Because of the fine control over the input data, I am able to easily produce good scans that require only minor adjustments to the fie, and these types of scanning artifact problems do not show their head too much. I still have to be careful in open sky areas, though."

Because Neg/Pos rellies on a flat scan which gets boosted in PS, I found the Neg/Pos workflow to be a massive noise generator.

Patrick Quinn
15-May-2006, 17:46
Because Neg/Pos relies on a flat scan which gets boosted in PS

it relies on a linear scan, which is basically what your scanner software is relying on.

The problem with all this is that what is one mans meat is another mans poison. There are so many variables (not least of which is input - are the negatives being produced for wet printing or scanning?).

It is the same with so many of these kinds of issues - take Imageprint. Some people think it is the best thing since sliced bread. Others (myself included) think it is an overpriced piece of rubbish that doesn't really do any of the jobs it claims to do well (I can't think of another high end RIP that gets it's customers to do half the work of producing profiles - sometimes at their own cost - with all the inaccuracies and variations that can introduce).

There are such differing variations in input, workflow, requirements and so on that in most cases we will never come to a definitive conclusion. I have never found Neg/Pos to be a massive noise generator, quite the opposite. It has given me some of the smoothest most beautifully balanced and toned photographs I have ever got. What works for one doesn't necessarily work for another

Kirk Gittings
15-May-2006, 20:02
Patrick, Just for the record, I never said Neg/Pos was a "massive noise generator". I said "I found the Neg/Pos WORKFLOW to be a massive noise generator" (as is any workflow that starts with a raw scan unless that raw scan is close to the tones you want in a final print or have a scanner that generates little or no noise).

Patrick Quinn
15-May-2006, 20:22
Patrick, Just for the record, I never said Neg/Pos was a "massive noise generator". I said "I found the Neg/Pos WORKFLOW to be a massive noise generator" (as is any workflow that starts with a raw scan unless that raw scan is close to the tones you want in a final print or have a scanner that generates little or no noise).

agreed I slightly misquoted you, but you make the same point. Your scanner software starts off with the same raw linear scan as the Neg/Pos workflow. It isn't causing or forcing the scanner to "make a scan close to the tones you want in the final print"

I just have not seen the Neg/Pos "workflow" generate particularly noticable noise at all, never mind massive noise. It gives less noise than any other workflow for color or black and white negatives I have used.

Kirk Gittings
15-May-2006, 20:41
Patrick, the plain fact is that with all the scanners I mentioned above using both Vuescan and Silverfast for a raw scan (but using the Imacon's software of course), then converting it with neg/pos I end up with significantly more noise in the transition areas than if I apply my major tone adjustments in Silverfast and allowing Silverfast to do the conversion. I had great hopes for Neg/Pos. I tested it extensively. That workflow created more problems than it solved. It is not particular to Neg/Pos it is the workflow that starts with a raw scan. I have never seen a raw scan that was not flat. Show me some prints. Do you print low contrast or use very low noise drum scans? That would explain the different experiences.

Patrick Quinn
15-May-2006, 21:01
That workflow created more problems than it solved. Show me some prints. Do you print low contrast or use very low noise drum scans? That would explain the different experiences.

which is why I started off with the one mans meat another mans poison thing - your workflow works for you, but there are so many variables (especially, but not only, in terms of input) that you can't make a definitive statement about either approach.

It is not particular to Neg/Pos it is the workflow that starts with a raw scan. I have never seen a raw scan that was not flat.

though this is the self same raw linear scan (with the caveat that not all raw scans are necessarily created equal) that silverfast starts off with - it is just different software making adjustments in different ways (note that raw scans really aren't "flat" as you call them, they are linear and what your software is doing is basically applying a curve)

Kirk Gittings
15-May-2006, 23:21
Do the tests as I have.

Patrick Quinn
16-May-2006, 06:11
Do the tests as I have.

I have tested this approach extensively, with different scanners and software. A Raw scan and Neg/Pos generally gave me the better result. Silverfast did well too, but it offered no great advantage in comparison and in some examples wasn't as good. In most cases both methods were generally superior to the various proprietary scanner software.

Paul Coppin
17-Jun-2006, 06:14
I wonder if you would be better off going into the work with NC rather than VC? I can't comment on your workflow, but its been my qualitative opinion that the hue range of NC is more "natural" than VC, and might make it easier to control in the computer...???