jake_hossfeld

<p>William,</p> <blockquote> <p>Much more important is that the histogram be reasonably balanced. If your histogram is pushed way over to the right, it means you're losing information. Instead of 256 gradations from pure black to pure white, you may have only 150 or fewer gradations that you can actually use. When you correct the exposure in Photoshop (which, by the way, you're probably better off doing in Camera Raw, before you even get to Photoshop), if the histogram looks jagged with lots of gaps, it means you've lost information.</p> </blockquote> <p> <br> I was worried about the loss of highlight information since the histograms were so right biased but am feeling better about it now that I'm starting to understand the gamma issue. Assigning the gamma 1 profile to the linear files and then converting to ProPhoto has shifted the histograms to the left and expanded the highlight range, so I feel reasonably confident that I'm capturing everything my scanner is capable of. I don't know if this is the best way to work with linear files, but it's definitely an improvement.<br> <br> Why do you think starting in ACR is better? Since the files aren't truly raw I tend to skip ACR (if for no other reason than to avoid constantly hopping back and forth between it and Photoshop whenever I want to make a change), but if there's an advantage to using it I'm all ears.<br> </p> <blockquote> <p>The image may look OK, but you may have trouble recovering shadow or highlight detail, and your midtones may not be as smooth as you'd like. (In your excellent flickr images, for instance, I'd be unhappy with the tones in that very pretty girl's blouse.)</p> </blockquote> <p> <br> Thanks very much for taking the time to look at the Flickr images and for the compliment. What about the blouse tones bothers you? Do they look as if they've been stretched too far and lost a sense of smoothness? These are more or less the first pictures I've put out into the world, so advice on improvement is more than welcome.<br> <br> Thanks again for taking the time!</p>

<p>Andrew,</p> <blockquote> <p>Not if you're getting high bit data into PS. Use the tools that you prefer to get to your final results. Generally speaking, IF the scanner software has the tools and controls to get to that goal, it's preferable because it can be faster since the 'corrections' are applied at the scan stage. But if the scanning software doesn't provide the controls, doing the edits in high bit elsewhere is just fine.<br> </p> </blockquote> <p>The majority of the software controls seem to be changeable post-scan, which leads me to believe that I can't make a significant impact on the data during the scan itself given my current setup, but I admittedly don't know for sure. I'll investigate this further, as what you say about optimizing the image during the scan itself certainly makes sense.<br> </p> <blockquote> <p>You need to <em>assign</em> the profile that defines whatever the scanner fed you. It's not ProPhoto RGB so assigning that is pointless. IF the profile you created produces decent color appearance after assigning, it's fine; it is defining the RGB values as it should.<br> </p> </blockquote> <p>That makes perfect sense. Assuming that the profile I'm now assigning is accurately defining the RGB values, does changing the profile gamma from 1.72 (it's default value) to 1 mess anything up? My thinking is that assigning a gamma of 1 is necessary in order to communicate to Photoshop that the file is linear so that the image is not darkened to compensate for gamma correction that was never made in the first place. This seems to be working - when I convert the linear profile to ProPhoto the images become brighter (and thus darker when inverted) and feature a better initial tonal distribution than before - but I don't know if it creates any problems elsewhere.<br> <br> Sidenote: Are you the author of 'Color Management for Photographers?' If so, I've read good things about your book but have hesitated to buy it just because it's older. Do you think the color management landscape has changed significantly since it's publication or do you feel it's still a useful reference?</p> <p>Thanks very much for your previous response - makes me think I'm finally on the right track.</p>

<p>Hi everyone,</p> <p>I recently migrated from the traditional darkroom to the digital darkroom in order to produce a small digital portfolio and have a question regarding the color management of film scans...</p> <p>I'm currently using an Epson v700 with Vuescan. Previously, I outputted a 'raw' tiff file (I understand that Vuescan's 'raw' files are not equivalent to true, digital raw files) and then reopened them in Vuescan for further processing. It then occurred to me that this further processing was post-scan software manipulation, so I decided to skip Vuescan in this regard and import straight to Photoshop, where I prefer to process my photos (if anyone sees a hole in this logic, please let me know). The scans, when inverted, were very light with histograms scrunched to the far right and thus somewhat difficult to work with. I figured I might not be doing something right, but I was on a deadline so I rolled with it. For anyone interested, the resulting portfolio (my first real foray into digital output, so plenty of room for improvement) is posted here...</p> <p>(Photos #1, 5-8 & 10-12 were made with the aforementioned process)</p> <p>https://www.flickr.com/photos/112010584@N04/albums/72157661830539494</p> <p>So, now that that's done, I'm revisiting my workflow. I've since discovered that the 'raw' tiff files I was using were untagged, linear gamma files, which I think explains why they were so light once assigned to ProPhoto (my Photoshop working space) and then inverted. After some color management research, I tried this: I made a custom RGB profile based on the 'SFprofT (PerfectionV700)' profile I found (which I believe is the appropriate canned profile for transparency scans with the v700), and assigned a gamma of 1 to it to correspond with the gamma uncorrected tiff files. I then 'assigned' this profile to the scans, and subsequently 'converted' them to ProPhoto. This has significantly altered the coloring (I'm scanning black and white film as if it were color - I've sometimes found it useful to use different channels) and has also moved the histograms to the left, decompressing the highlights and making the initial global curve easier to make.</p> <p>Does this seem like a 'correct' workflow to everyone? Working with the linear files confuses me - was I essentially working with an unnecessarily light image (after inversion) because no gamma correction was done upfront? Does 'assigning' the linear profile and then 'converting' it to ProPhoto truly solve this problem or is there something else I should do before beginning to implement creative manipulations?</p> <p>My goal is to do as absolutely little in Vuescan as possible. If Vuescan's image processing is truly post-scan, I would much prefer to do it in Photoshop. I've only scanned black and white film thus far so I've had it easy, but I intend to dive into color soon, where I assume that figuring this out and having a solid color management workflow in place will become truly critical, especially as I plan to begin printing digitally.</p> <p>Thanks so much in advance to anyone who actually managed to read this entire post and can tell me whether or not I'm on the right track!</p> <p>-Jake</p>