Epson V700 Scans blown out or blocked up

[kevs-2323668]

I had an intern scan 90 images over seven hours only to discover most of highlights were really blown out.

I just scanned some of the same images with same setting and they look ok, except then the shadows are blocked up.

What to do? I'm using Epson scan. thanks!

[Jon Shiu]

What are all the settings you are using?

If you click on the icon that looks like levels (a histogram) you can make adjustments to the auto-exposure levels done in the prescan.

Jon

[Lachlan 717]

Adding to Jon's comments, you can use the eye-dropper tools in the histogram to set the dark, mid and high points.

While I set these fairly accurately to the range displayed, some people like to leave the points just outside the range to make sure that they do capture all possible dynamic range.

Also, you sometimes can get a "false positive" on the black end of the range if you include some of the film border in the scan. This, to me, nearly always gives a darker low end than I have in the exposed area.

[photobymike]

Sometimes a cheap or not-tuned monitor will do that. If your monitor is not up to the task it will look like a bad image... then again sounds like some pratice is needed on scanning adjustments....

[kevs-2323668]

Thanks Jon, have not been paying attention. I thought the scanner know what to do? So you laboriously do this for each and every scan? I see levels, so like Photoshop, you just play around.

Lachlan, thanks.

I see image adjustment too. I was hoping not to mess with all this stuff and save that for Photoshop, but the scanner is so off sometimes, that I guess I should now start tweaking before scanning.

So in the future, it's just good to have an intern do all just all of them in auto just to see them, no tediou dusting, and then do the tedious work on final picks.

L, I was not even cropping like I do on documents. Just viewing preview in thumbnail view, so best to use normal preview and do a crop to block out black areas?

I'm comfortable with the histogram and image adjustment. The color palette I cannot make sense of, and always have hated curves.

Also, how do you get Epson to stop checking the unsharp mask! And anyway to get them scanned in Adobe 1998? any way to save scan presets?

Finally, a friend (Rick D) uses Vuescan, have you tried it, is is any better to deal with? I took a peek and could not make sense of it at all. Same question Silverfast.

[Frank Petronio]

With Epson Scan or any of the other scanning apps... if you do Auto, it depends on the area selected. It needs to be within the image area, not including the border or scanner bed, so that it can gather only good "film" information to be "automatic". If the selection area for the scan has some film border or empty non-image area, it skews the automatic results. Click and drag the selection area manually, hit the Auto button, watch the preview scan... even an intern can do it. If the preview still looks bad or you don't like the results, redo the preview and go into the manual adjustments and either set black and white points, plus grey balance, with the eyedroppers - or move the Histogram around until you like it.

[polyglot]

a) don't use Epson Scan
b) adjust the scanner exposure appropriately so that you don't clip the image
c) check the raw (scanner data, not image data) histogram for each image before saving it to ensure that the exposure was correct

[Lachlan 717]

I just use the Epson software (no idea why some consider it the wrong programme to use), and do the rest in PS.

I don't find the histogram workflow too tedious, nor too time-consuming. If you have a decent holder that keeps the film square (which I find the 4x5 proprietary one does), then I do a prescan, put the area guide over the required part of the film, then adjust the histogram, then scan. Scanning 6x17cm is a bit more of a PITA as the Betterscan holder has quite a bit of room around the 120 film that can leave the image off line. Just like when shooting the image, a quick tap on the holder's base will help to "seat" the film against the base of the holder's film slot.

Note that I consider the histogram a rough tool here; the fine tuning needs to be done in PS. I use it to make sure that I have the best range of tones in the initial scan, and also to make sure that the file doesn't contain useless info (i.e. those areas outside the histogram's tone range).

[rdenney]

Finally, a friend (Rick D) uses Vuescan, have you tried it, is is any better to deal with? I took a peek and could not make sense of it at all. Same question Silverfast.

Ken, I don't know the Epson software--I never even installed it on my machine when I bought my V750. I have used Vuescan since about 1999, so I'm accustomed to its weirdness. I once installed Silverfast AI--highly recommended by many photographers--and could not make heads or tails of it. Vuescan is the result of 10,000 modifications, so the user interface tends to be rather non-intuitive and tedious. But you can get everything set up and then save your parameters to be called up next time you are scanning that particular film.

If you set Vuescan to clip the white at about 0.01% and the black at 0%, and if you output to a wide-gamut color space like ProPhoto, then it will capture the full density range of the film without blowing out highlights or jamming the entire range into one end of the histogram. Leave the other settings on their defaults. The very slight clip on the white will find the few specular highlights in your picture, and keep the image from being too compressed in the histogram.

Frank is right that you MUST select the scan are to include only image area if you use ANY of the automatic level-setting tools (even if Epson Scan doesn't tell you its applying such automation).

From this distance, I can't help you learn Vuescan. You just need to put in the time to play with it until it makes sense. The Vuescan Bible will help (google it and buy it). But lots of controls just don't seem to make any sense. Without experimentation. It is not immediately intuitive, but it does provide access to lots of capabilities.

Vuescan does not scan color negatives perfectly without a lot of fiddling, and I just get it close and do the rest in Photoshop.

Rick "who could not live without curve adjustments" Denney

[Frank Petronio]

Frank is always right. Unless he's scared of Kirk.

[RichardSperry]

Scanner software(EpsonScan) adjustments are edits to the scanned image after its scanned, it doesn't actually change the scanning itself.

Letting it make those changes is akin to making undo'able unrepairable edits to a raw digital file and saving them permanently.

It makes more sense, to me at least, to scan raw with no edits, then do repairable edits in PS or LR. Besides ICE, which I don't use, I can't think of a scanner edit which can't be better done in post with more control.

[Frank Petronio]

Scanner software(EpsonScan) adjustments are edits to the scanned image after its scanned, it doesn't actually change the scanning itself.

Letting it make those changes is akin to making undo'able unrepairable edits to a raw digital file and saving them permanently.

It makes more sense, to me at least, to scan raw with no edits, then do repairable edits in PS or LR. Besides ICE, which I don't use, I can't think of a scanner edit which can't be better done in post with more control.

Where/How did you learn this? I have never heard of this being the case and it doesn't jive with my experience.

[RichardSperry]

Frank,

1) One can't make most Adjustments changes to the scanner image until after one has done the preview scan. If one were actually making input changes, they would be available irrespective of the need for a preview scan. Original and Destination changes can be made prior to an image being scanned, and cover changes to actual scanning itself, scan size, resolution, reflective or negative, etc.

2) One then makes the Adjustments edits to the preview scan and they are instantly viewable on that image(with no need to actually rescan the preview image) just like an image editing program.

3) Does it make sense that EpsonScan is actually making optical scanner(hardware) changes on a $160(V500) scanner for Adjustments? EpsonScan for the v500 is essentially the same program as it is for the v700.

4) There are no Adjustments 'settings' available that are not part of any image editing program. Contrast, exposure, curves, histogram, sharpening tools, etc., are all the same as in any basic image editing program.

5) Comparative adjustments need all of the pixels in the image scanned before taking place. Take unsharp mask for example, it uses an algorithm of comparison of adjacent pixels for it to make changes to a pixel. If this adjustment were actually being done at scantime how would the scanner 'know' what the adjacent but yet unscanned pixels look like to make changes to the pixel being scanned at that moment? Auto tone is another comparative tool which needs the whole image file before changes to pixels can take place.

No, I did not learn that Adjustments is an image editing tool. There just is no evidence to back up the notion that EpsonScan makes any changes to the scanner, on the input side prior or during the scan for Adjustments.

[polyglot]

The one critical hardware adjustment is exposure and Vuescan allows you to manually set this consistently. The scanner will vary the duration for which the CCD is illuminated in order to deal with more- or less-dense negatives, which is required to fit the image into the dynamic range of the scanner's CCD.

If you get the exposure wrong, you get (for negs) clipped shadows or noisy highlights. It's not fixable in post.

[Ken Lee]

You might find this brief article (http://www.kenleegallery.com/html/tech/scanning.php) helpful.

[RichardSperry]

Ken,

You write


We want to get an image directly from the scanner, which comes as close as possible to the final print. We want to avoid performing lots of digital adjustments in our image editing tool, like Photoshop.

Why do you want to do that?

In PhotoShop you can undo those edits you are making to the file


Minimal manipulation in Photoshop was required.

Manipulate the file in the scanner program or manipulate in PhotoShop, it's still manipulation. What advantage does one get by doing it your suggested way, rather than scanning flat and making those same edits in PhotoShop?

[rdenney]

Manipulate the file in the scanner program or manipulate in PhotoShop, it's still manipulation. What advantage does one get by doing it your suggested way, rather than scanning flat and making those same edits in PhotoShop?

Perhaps not relevant in this case, but it is true that the data is passed through a data bottleneck on its way from the scanner to Photoshop. In the days of 24-bit color, it was really important to get the histogram approximately right before squeezing the data through that narrow place, to avoid having to expand the histogram after doing so.

It's less important with 48-bit color, but I still find that if I try to use Photoshop to expand contrast in, say, a sky area where banding is easy to see, I run a greater risk of posterization than if I make that same move before it was committed to the TIFF file. I do not have this problem with the Nikon film scanner, but I occasionally do with the Epson.

I have not had any luck trying to bring color negatives across as a DNG file (pseudo-raw in the case of Vuescan) and invert them in Photoshop. Getting good color balance and filtering out the orange mask seems to defeat me. Here, the algorithm in Vuescan has already been tweaked to cover much of this terrain.

It would be nice if Photoshop had tools for inverting raw color negative scans, with an associated database of film types. But it does not. It's no problem inverting black and white negatives, though I still find that the scanner compresses the density values into a narrow histogram (and my negatives from that era are rather thick) that requires lots of expansion in Photoshop. I'd rather do that expansion before it's written to an intermediate data file, even if it is less of an issue with 16-bit depth than with 8-bit depth. With color slides, I'm just trying not to clip the histogram.

Rick "who tries to make the big moves before writing a disk file, just out of habit" Denney

[Frank Petronio]

Richard, your point that the scanning software applies sharpening after the scan sounds correct, and judging sharpening from a prescan is folly ~ which is why many of us recommend not doing any of the editing "add-ons". But I am confident that the software does tell the scanner to adjust the exposure based on the where the white and black points are set, even on a cheap flatbed. Perhaps Gamma and color adjustments are adjusted after the actual scanning takes place, I am not sure how that works.

Whether in the scanning, Lightroom, Bridge, or Photoshop ~ when you edit the file, each step is slightly destructive to the file because you are removing information. That's why if you can get to where you want to be in one step rather than three, it's better. And that is why you see most experienced users use the Curves adjustment to adjust exposure, Gamma, tone, and color (and maybe turning to Hue/Saturation to fine tune). That's pretty much it for global adjustments. When I've taught, I will see students fidget back and forth with multiple tiny adjustments and tweaks but a few ruler wacks on their knuckles gets their attention.

If you watch the Histogram as you work, you can actually see it degrade and lose blocks of information if you do too much, over too many steps. In the extreme you can see actual gaps.

I do try to get the look of the image roughly in place during the pre scan so that when I open the file it looks somewhat like I intended it. It just saves me time. I know I can't get it perfect and the scanning software is probably never going to get past a 1998-era level of refinement, but as long as I can set the black and white points - or simply pull the triangles out to the edge of the preview Histogram - then I'll have something to work with.

[Ken Lee]

Just as Frank says.

From the article:

1. Make as few adjustments as possible, since every adjustment is "destructive".

2. Make important adjustments as early as possible in the process.

3. Make adjustments in the analog state where possible: adjust exposure and development first. Then scan. Photoshop last.

[bob carnie]

If you are using a mac system to operate as I do so I can only speak for mac driven scanner but I know you can somehow get this meter on a windows based system.

Go into Macintosh HD > open applications> open utilities> open digital color meter......DCM

Keep this on your desktop for many purposes. Reading scans , reading files sent to you , pre flighting colour and density settings and so on.

I use the Cie Lab setting but you can use RGB if you understand that better.
Set the aperture to a smaller setting so you are taking samples of a smaller range of pixels.

The L channel or luminosity channel is a range of numbers from 0 - 100 where 0 is absolute black pixels and 100 is absolute white pixels.
The RGB channels go from 0 - 255 *** I never use these numbers for anything so I cannot give you the drop off points***
The K channels go from 0 - 100 ** you may want to use this if you are only scanning black and white work and do not do any colour editing.


When you have done your preview on any scanner , put the DCM curser over the image, you will see the L number go up and down depending on the density of the area
you are working on... highlights will be high numbers and shadows will be low, numbers.

You can take my word on this that if your highlight area is above 94 you will start to blow out highlights.
if your shadow area is below 6 you will start filling in your shadows.
this is using the L channel.... so for you Zonies out there think about this, place your tones much like you meter and you will get great results.

I prefer not to make the image look good in the scanner software as PS offers incredible tools to do this , and I have not seen scanner software as good for this purpose.
I do not sharpen in scanning for the same reason, also if you are in colour mode and your colours are not perfect you will get fringing effects due to the sharpening function.

I use Epson flat bed and Imocan and a friends Drum for our work and each system though different is greatly helped by using the digital color meter.

I use this meter {DMC} for everything as it is just like using the eyedropper and L readings in PS and is a powerful tool to use with scanning.

[kevs-2323668]

Frank, thanks do you use auto? I’ve always just used auto with my Nikon film scanner, and old Epson 1660 for document and scanning magazine images. But this 700 really gets out of whack. I do wish auto could just see the chrome or whatever and scan the exposure so all the info is there, and the color matches the film – so I don’t have to mess with it -- (I also don’t understand what Richard is saying)

Rick, no problem. I think now that I have the levels and color tools with Epson, I should be able to get a bland scan that has both highlights and shadow area. Vuescan does seem like it will take dozens of hours to learn, so I’ll probably pass for now. Silverfast is ridiculous. It takes an hour even with a CD to open the software and after that I’m lost like you. I’m scanning 24 bit, I don’t see a problem. Tell me about 48 bit. I’m just worried those scans will double in wait time and maybe not be necessary.

Lachlan, good comments.

Ken L, will read article tonight, had trouble printing it out.

Richard, agree, it’s best to get the thing flat so you’ve got the goods. But Ken L has a point in that if you can get the color and exposure optimal, it’s less tweaking later, but you have to get all the info to start out before PS.

PS again, is there a way in Epson scan to assign a non Epson profile? And how to turn off the dang unsharp mask. Thanks!

Bob, will PM, did not understand most of that.

[Frank Petronio]

Post some screen shots of using Epson Scan because setting the auto scan and turning off the sharpening should be obvious

[rdenney]

48-bit color uses 16 bit numbers to describe each of red, blue, and green levels--64K levels. 24-bit color uses 8-bit descriptions for each color--255 levels. If the image is just where you want it, 24-bit is fine. But if you take a portion of the histogram and really stretch it out, you'll leave visible gaps in the histogram that will show up as posterization in the image, where what should be a smooth gradation shows edges.

Most everybody works in 48-bit color to make big moves more possible. But if you use 24-bit color files, then it's essential that you get the image close to where you want it before you write that file, which means doing it in the scanner software.

Rick "warning: excellence always has a steep learning curve" Denney

[sanking]

"Rick "warning: excellence always has a steep learning curve" Denney"

Great comment.

And even mediocrity requires a bit of work.

Sandy

[kevs-2323668]

Frank the unsharp mask is just always checked. I close software open it up. It's checked. How can I default it to not be checked? and can I insert a color profile other than Epson's? thanks.

thanks Rick, I'll test it, I imagine it's double the scan time....

[Frank Petronio]

You can save the settings in Epson Scan, I make a set for Portra, Tri-X, etc. and it retains keeping the USM box unchecked.

Looking around under "advanced" under color it only offers three profiles, not any custom ones... So I guess you can't.

It's not perfect ;-/ I wouldn't wait for a fix either.

[kevs-2323668]

Frank, thanks, how do you save settings? Yes I shudder at having to call them...

[Frank Petronio]

81488

Fly me out, $800/day plus expenses, jam packed one-on-one hardcore live nude training. Just for you.

[kevs-2323668]

Frank, does the monitor compensation help/ matter? I'll test these thanks! I'll screenshot later that pesky unsharp mask if it returns... Have you tried Vuescan or Silverfast. I don't believe I need them, but curious if you tried them.

[Frank Petronio]

What monitor compensation are you talking about? If it is check box somewhere that I missed then try it with and without, one test and you'll know ;-p Nothing beats just experimenting. The basic concept is to scan a wide bit depth, full range file and from there it doesn't matter that much how you muck up the settings on the various apps, as long as you are saving good data (and can revert back to the original) then just experiment until it looks good. It really shouldn't be all that different from working with your DSLR files once you get it scanned and think about how much work you need to do to a RAW DSLR file to make it good... same thing.

I wouldn't worry about profiling the scanner or anything, just get the scans done and saved with Adobe RGB, 48-bit and then convert them on import into Bridge or LightRoom to whatever wide-gambit profile preference you have (ProPhotoRGB, Adobe RGB, not sRGB) you use in your other editing (like digital camera files).

As long as you are consistent with your workflow then you'll be OK.

I use VueScan instead of the OEM software on my 35mm film scanner and it works fine, but I do not think I get a better scan with a different app. Epson Scan is the leanest/simplest of the scanning apps and I wouldn't worry too much about it once you figure out how to get a decent scan.

It's only a scan, just like the other files you've been working on except larger. If you aren't confident that you're making decent scans then go get a drum scan from Eiger to compare, or work with someone doing good work and confirm you're on the right path.

[bob carnie]

Frank I assume you are meaning you will pay the fees if its live nude training.


81488

Fly me out, $800/day plus expenses, jam packed one-on-one hardcore live nude training. Just for you.

[rdenney]

Frank I assume you are meaning you will pay the fees if its live nude training.

No, Frank's threat of showing up nude is an attempt to scare Ken straight.

Rick "be afraid, be very afraid" Denney

[kevs-2323668]

Frank, I'll think I'll be ok, because of this thread. thanks.
The monitor compensation was a checkbox I saw checked in your screenshot. I'll test it.

So you think 48 bit is critical? Also, profile question to Adobe 1998, is just one less conversion I have to do later. Although Epson does not offer the Prophoto, do you convert later to Prophoto? I know Rick believes in that.
Finally, I was going to just make 720 res scans that come out to 23 mb or so for the website. . But maybe I should do massive 1GB scans (on 25 favorites) also just so I have them in case the house burns down. I'm definitively going to do the more expensive drum scans on a few favorites (3-5), but it would not be a bad idea to do super large ones to with the Epson no? It's free. Eventually I'll print a large 40" print comparing the Epson vs, the drum vs the Canon 5D2

[rdenney]

Frank, I'll think I'll be ok, because of this thread. thanks.
The monitor compensation was a checkbox I saw checked in your screenshot. I'll test it.

So you think 48 bit is critical? Also, profile question to Adobe 1998, is just one less conversion I have to do later. Although Epson does not offer the Prophoto, do you convert later to Prophoto? I know Rick believes in that.
Finally, I was going to just make 720 res scans that come out to 23 mb or so for the website. . But maybe I should do massive 1GB scans (on 25 favorites) also just so I have them in case the house burns down. I'm definitively going to do the more expensive drum scans on a few favorites (3-5), but it would not be a bad idea to do super large ones to with the Epson no? It's free. Eventually I'll print a large 40" print comparing the Epson vs, the drum vs the Canon 5D2

ProPhoto is a wider gamut than Adobe, so some software will use that to advantage and spread out the potential histogram. Adobe is wider than sRGB, which was intended to be a common denominator. I have found that Vuescan will clip the histogram of scan less when using a wide-gamut color space.

You're probably fine with Adobe. That's what I used to use.

Whether you scan for storage or scan each time you need a new target is up to you. I scan for storage and make my basic corrections and overall interpretations on that scan. Then, when I need to display it, I'll make a new file and target it to that display medium. It keeps me from having to go through that basic correction and interpretation all over again, though there are times when it's good to start over and look for a fresh approach.

Rick "thinking you'll find that 4x5 scanned on an Epson is no more a match for a 40" print than a 5D2, but the drum scan will be" Denney

[Frank Petronio]

Yes use 48-bit (16-bit means the same thing, 3 RGB channels x 16-bit =48) for everything that you can.

The difference between Adobe RGB and ProPhotoRGB isn't going to hurt you, I convert and edit in ProPhotoRGB but I doubt it matters. I know converting between profiles is degrading but I figure if I do it once early on then I am affecting the image the least possible, and going to a wider gambit ProPhotoRGB isn't clipping anything.

[kevs-2323668]

Whether you scan for storage or scan each time you need a new target is up to you.

Thanks guys, great info. Rick, not sure I understand those terms, storage and target!

But interesting. I shot a lot of backup on the 5d. You don't think 4x5 Epson scans even at 1GB will blow up to large prints much better than the 5D. Hmmmm maybe I do a test before belaboring all that -- just scan for the web for now.

[bob carnie]

Rick
we are a few years away from the inkjet manufacturers getting us to the gamut of pro photo, even though it is larger than adobe 1998 , I prefer to stay in a gamut that the inks and papers are capable of and that is Adobe 1998.
I swear Kodak and Fuji reversed engineered their RA4 papers as Adobe 1998 as it is a perfect colour space for the Ra4 papers.

The slow bleed of finances (yours and mine ) until is the inks that start reaching the larger colorspaces, I do not thing we are there yet to be thinking pro photo is the right space to be in... yes its bigger, but so is Lab.. now that will be some incredible array of ink sets.

Bob .. who hates the manufacturers taking advantage of our dreams...

ProPhoto is a wider gamut than Adobe, so some software will use that to advantage and spread out the potential histogram. Adobe is wider than sRGB, which was intended to be a common denominator. I have found that Vuescan will clip the histogram of scan less when using a wide-gamut color space.

You're probably fine with Adobe. That's what I used to use.

Whether you scan for storage or scan each time you need a new target is up to you. I scan for storage and make my basic corrections and overall interpretations on that scan. Then, when I need to display it, I'll make a new file and target it to that display medium. It keeps me from having to go through that basic correction and interpretation all over again, though there are times when it's good to start over and look for a fresh approach.

Rick "thinking you'll find that 4x5 scanned on an Epson is no more a match for a 40" print than a 5D2, but the drum scan will be" Denney

[rdenney]

Thanks guys, great info. Rick, not sure I understand those terms, storage and target!

But interesting. I shot a lot of backup on the 5d. You don't think 4x5 Epson scans even at 1GB will blow up to large prints much better than the 5D. Hmmmm maybe I do a test before belaboring all that -- just scan for the web for now.

Epson performance on a 40" print:

My experience with the Epson is that 4x enlargement gives up nothing in resolution, while 6x still looks outstanding to my eye. I've seen 8x enlargements that were quite good.

But a 40" print from 4x5 is a 10x enlargement, and that goes beyond what the Epson scan can deliver. It does not exceed what a drum scan can deliver, however.

Thinking about it from a math point of view, if we set a standard of 5 line-pair/mm in the print as being sufficiently resolved, then a 4x enlargement needs 20lpm on the film. A 10x enlargement needs 50lpm--quite attainable on the film but beyond the scanner, which is probably limited to about 30. And some set a higher standard more like 8 lpm on the print.

The 5DII has to provide over 200 line-pairs/mm on the sensor to achieve that level of quality on a 40" print. There may be a lens or two in the line that can attain that if used absolutely perfectly, but probably not, especially for color. That 40" print would need over 5000 line pairs, which is easily more than double what the Canon sensor can provide.

Maybe the Epson scan would still outperform the 5DII in terms of resolution, though it might not outperform it in tonality, on a 40" print. But neither will provide that large-format sense of endless detail. A drum scan of the 4x5, however, will.

Storage vs. targeting

I scan first at high resolution to get the best record I can of what's on the negative. I store that scan in my archive the same way I store my negatives. The manipulations I make to that file include basic corrective sharpening (to minimize the fuzziness caused by the scanner at the pixel level) and whatever overall manipulations I want to make.

Then, when I want to print it, I target it for the print. So, my purpose shifts from scanning for storage to altering the stored file to target it for a particular display medium.

For example, I'm doing a couple right now. I've scanned a 4x5 color negative at 2400 spi, which makes a file about half a gig. In Photoshop, I performed basic corrective sharpening with the unsharp mask radius down around a pixel. The effects of this sharpening process are only visible at high enlargement. Then, I add all my basic adjustment layers to make the print look right on my calibrated screen. Finally, I save the file as a PSD file in my archive.

Then, let's say I want to put it on the web. I pull up that archived PSD file, downsample it to maybe 800 pixels wide, convert it to sRGB so that it will look right on a Windows display without color management, adjust curves, color and saturation if necessary to make it look right at that size and color space, sharpen it (maybe--often don't need this). This sharpening is done on a low-res file, and therefore the effects of the sharpening will be visible at display resolution. Then, I overlay my copyright mark on it and post it on the web. That's a targeting process.

Or, let's say I want to make a 16x20 print. I pull up the original file, view it with my Epson printer preview (in Photoshop), make any adjustments to curves, color and contrast necessary to make it look right in the printer color space. I probably won't sharpen it here--usually the 4x5 scans are already sharp enough at print sizes I can make anyway. Then, I'll snip a bit out of the middle and make a test print, iterating until I'm happy with the color and tonality on the print. Then, I'll print a full work print. I may make several more adjustments. I may crop it to fit the print size I want exactly. I then save that file under a new name that tells me that it's been targeted for 16x20. Next time when I need a 16x20, I pull up that targeted file and just hit the print button.

If my interpretation changes, I go back to the archive copy, make the changes, and then retarget it for whatever display media I've used. Thus, the corrections I make because my Epson always makes things a little more green won't get propagated to the file I send to Costco to make 15 prints to give family members, and so on.

Rick "who has little time to do thing over again" Denney

[rdenney]

Rick
we are a few years away from the inkjet manufacturers getting us to the gamut of pro photo, even though it is larger than adobe 1998 , I prefer to stay in a gamut that the inks and papers are capable of and that is Adobe 1998.

Agreed. I'm going to start converting the ProPhoto file to Adobe before printing on the Epson. I've noticed some gamut problems when working in ProPhoto and your post has turned on the light as to what I've been seeing.

But I'm still using ProPhoto as the output space for my scanner software, because it seems to open a wider envelope for the histogram and avoid clipping, especially for transparencies. I'll keep my storage copy in ProPhoto in hopes of better printers in the future, maybe, and convert to Adobe as a targeting process for printing on the Epson (just as I convert to sRGB as a targeting process for web display).

Rick "who is dealing with a photo that uses more gamut on the blue end than ever before" Denney

[kevs-2323668]

Great Rick, got the gist of that. Ok, I may now do 2400 instead of just for the web at 720, though those will take a lot longer, I know that I can make great 11x14 and 16 x20's without having to buy a drum scan file. Of course I'll buy drum scan files for the real biggies.

[bob carnie]

The out of gamut issues are indeed the problem, if I was a betting man I would say Fuji and Kodak selected srgb to use as a guide for their papers as you will never ,, well almost never go out of gamut
or have any screen to print issues . this is why those minilabs can crank out the files so fast.

I have noticed the Cannon printer I have has a wider gamut than the Epson printer, only a year or two between them. I imagine within a couple of years we all will need to be editing in Pro Photo to take advantage of the inksets.
I do not see the day when LAB the required space for inksets, but who knows, if so Peter Lik watch out.


Agreed. I'm going to start converting the ProPhoto file to Adobe before printing on the Epson. I've noticed some gamut problems when working in ProPhoto and your post has turned on the light as to what I've been seeing.

But I'm still using ProPhoto as the output space for my scanner software, because it seems to open a wider envelope for the histogram and avoid clipping, especially for transparencies. I'll keep my storage copy in ProPhoto in hopes of better printers in the future, maybe, and convert to Adobe as a targeting process for printing on the Epson (just as I convert to sRGB as a targeting process for web display).

Rick "who is dealing with a photo that uses more gamut on the blue end than ever before" Denney