Flatbed Scanning vs DSLR re-photographing?

[2F/2F]

Glad this thread is continuing. I'll admit, I don't use PS much for image editing, but I got curious when reading about Camera Raw. So I booted PS and loaded a raw image. Geez luiz, I've been used to Canon's DPP, which I've always thought was plenty good enough, but with Camera RAW, it appears that almost all necessary corrections can be done at the raw level before saving the file to a given file format. Wow. DPP is handy for translating quantities of files over to other formats, though. I guess CR can do this too?

Absolutely, it can.

1) Find files to work on using Bridge, and highlight them simultaneously.
2) Hit Command-R, and they will all come up in Camera Raw.
3) Do what you need to do to the pix. You can do pretty much everything, including burning, dodging, and sharpening (including sharpening masking). Be sure to mess with the tool...and forgive me, because I forgot it's name and I don't have the program in front of me...that looks like a brush. It basically allows you to make several separate burning, dodging, saturation, etc. adjustment layer masks using a brush tool, and then edit them after the fact with sliders, while looking at the image. It is a very powerful tool.
4) At the bottom of the Camera Raw screen, there is a line of text that will list the specifications of the image, such as 8 bit, so many by so many pixels. Click that. A window will pop up. Choose the bit depth of the files you want to convert to. Don't use any of the + or - settings unless you want to roughly resize the pix. For doing prints, I choose 16 bit conversion. For pix that I know are only going to be digital, I use 8 bit.
5) In the strip of thumbnails over to the left, highlight all the pix you want to convert.
6) Below the strip of images, click SAVE PICTURES, or whatever the button says.
7) Choose your file type, location, and batch rename the files if you want.
8) Hit OK in the popup window. You will be taken back to Camera Raw and you will see a tiny countdown in the bottom left that says "X remaining." When that finishes, all your files have been converted, renamed, and placed in the folder you chose.
9) Make sure you click DONE to save the changes you have made to your sidecar (.xmp) files. If you don't, all the adjustments you have worked so hard on in Camera Raw are thrown away.

It is a very powerful program. I only have to use Photoshop for resizing, major manipulation work (rare with what I shoot), minor adjustments when test stripping, and printing, of course.

P.S. When you first open your pix in Camera Raw, hit the F key. It is much easier to work this way with this particular program, IMO.

Also, be sure to tinker with the b/w conversion in Camera Raw. It is a very good way to convert to b/w. Use manual, not auto.

And make good use of snapshots. They can be lifesavers.

 

[cooltouch]

Wow, I think I should print out your comments. It'll take me a while to get through all of your suggestions. Except for a small time span a few months after I bought my DSLR, I've kept all my .CR2 files. Which means thousands. It makes me want to go back through all my photos, select the best ones, find their equivalent .CR2s, and redo the conversions. Geez, that'll keep me busy for a few months, I reckon. :cool2:

 

[pellicle]

Hi

To each his/her own, of course. I just find it incredibly convenient to work on my copies as if they were any other images from my digital camera. No extra equipment or software, and no need to change "modes" in my head when going from one to the other.

a very good point, when working on images taken of negative I would guess that you'll be working on a "lighting" which is unlike what you get in the field. Still may not make much difference to knowing the tool you know and like to work with.

I wonder, would you be interested in doing a comparison of results? You could copy a neg and I could scan it and we could compare notes; work flow; results ...

 

[pellicle]

Hi

No dispute here. If only the LS-4000 wasn't so dang expensive, I'd consider getting one.

I suppose, but the LS-4000 and the LS-V {aka LS-50} are quite similar in spec (mainly automation factors) and they aren't much over $1000 ... occasionally less

That's the thing, though. I'm still very much an active film user, both 35mm and medium format. Right now it's looking like I'm gonna have to sell a motorcycle to afford a 5DII

well as a motorcyclist I can't abide selling a classic for a consumer item either.
(a portion of my historical list) not that I've owned many classics (except the Ducati). Stuff like cameras is really consumable in many ways. Its starting to become an impossible task to determine "value" on that sort of thing.

Personally I struggle with buying a 5D for full frame (and my 21mm Olympus lens) but in the mean time I keep using a OM-10 an scanning the negs. For me at least there is only a subset of situations where that works better than my G1 (for what I do at least).

sell a bike for a scanner or camera .... me? ... nope ...

perhaps this one would have been a candidate ... a bit "consumable"

PS: should you be interested, I just found this LS-V for sale on slezebay

 

[cooltouch]

Thanks for the reinforcement, pellicle. In my case, the bike in question is an '88 BMW R100RS. It will only increase in value.

That's a very, um, eclectic selection of bikes you've had. Mine, I wouldn't call as interesting. My daily rider right now is an old '76 BMW R90/6. The R100 isn't being ridden at the moment because it needs a few minor repairs I haven't gotten to yet. I've also got few Yamahas -- the old XS650 -- I have three of 'em. None of them are running right now. One needs a bit of work before I can get it on the road, another is just a parts bike, and the other is a cafe bike project I've been working on for way too long. Then I have a '74 Norton Commando that ain't going nowhere. And an '87 Harley Sportster that has major engine problems. When I finally can afford to get the engine work done, I will be selling that bike, though.

I got a question for 2F/2F: do you ever shoot dupes of color negatives, and if so, what is the process you use to convert them so they look reasonably "normal?" I've recently discovered that with PS (CS5) after inverting the negative, if I go to Image/Auto Tone and then Image/Auto Color, I get a reasonably close image. A bit of posterization seems to happen, though, and I think it's happening because the negatives I've tried this on are only 8-bit. I've tried the demo version of the ColorNeg plugin (actually ColorPerfect, which includes ColorNeg) and it doesn't seem to do all that great of a job, really. Maybe once I get better at using it, I can get better results.

 

[pellicle]

Hi

love the old beemer ... always had a desire to get the R100RS, but after the time on the K100RS I found them a wee bit ... rubbery (and I have ducks disease so the saddle is a wee bit tall too)

My daily rider right now is an old '76 BMW R90/6. The R100 isn't being ridden at the moment because it needs a few minor repairs I haven't gotten to yet. I've also got few Yamahas -- the old XS650 -- I have three of 'em. ... And an '87 Harley Sportster that has major engine problems.

lovely set ... love the XS's too BTW ... like a Triumph that doesn't leak oil ;-)

F2, I'll also be interested in your C-41 copy technique if you will be so kind as to post that.

Thanks

 

[cooltouch]

Guys, have y'all by any chance read through any of the articles from this site?

Dead Link Removed

The author is used to working with very high-end scanning products, but the couple of articles I've read so far are quite dated. One comment he made, though -- actually he made the same basic comment several times -- was that the 2700 dpi Nikon LS-2000 did a better job than the Microtek or Polaroid 4000 dpi scanners, apparently mostly because the Nikon handled shadow detail better. D-Max, right? So anyway, I found that rather enlightening because my mindset has become "the more ppi (or dpi) the better" which it appears may not be the best way to approach this subject.

But getting back to the original topic of this thread, I have not found this to be a problem at all when duplicating slides with my DSLR. The DSLR seems to do a decent job of capturing shadow detail.

 

[cooltouch]

Hi

love the old beemer ... always had a desire to get the R100RS, but after the time on the K100RS I found them a wee bit ... rubbery (and I have ducks disease so the saddle is a wee bit tall too)

lovely set ... love the XS's too BTW ... like a Triumph that doesn't leak oil ;-)

Yeah, the XS is just plain fun to ride. Fortunately, there are very active ownership groups for the bike scattered around the world, and there continues to be an excellent supply of parts, including go-fast parts for it.

Not sure what you mean by rubbery regarding your K bike. Suspension issues? You know, it's like comparing apples and oranges when comparing the R bikes and the K bikes. The K bikes have it all over the old Airhead R bikes in terms of horsepower. That R100 only has about 70 or so, as I recall. But it's plenty enough to get me into a lot of trouble if I ever got caught. I've heard it said many times since I bought the R100RS that the fairing that was designed for the R100RS was probably the finest fairing ever designed -- by anybody, for any bike. Well, I dunno about that. But I can tell you this: the faster I go on that bike, the more stable it becomes. I've frequently caught myself doing 95 to 100 mph on the freeway without even realizing it. Very little impression of speed at high speed, and it's just rock steady. That's what I mean by getting myself in trouble.

 

[pellicle]

we're dipping quite a bit off topic here .... but

I've heard it said many times since I bought the R100RS that the fairing that was designed for the R100RS was probably the finest fairing ever designed -- by anybody, for any bike. Well, I dunno about that. But I can tell you this: the faster I go on that bike, the more stable it becomes. I've frequently caught myself doing 95 to 100 mph on the freeway without even realizing it. Very little impression of speed at high speed, and it's just rock steady. That's what I mean by getting myself in trouble.

heard the same thing said too. Never had the pleasure of an open highway trip on the R series, but the K was great (put around 200,000Km on mine). On bumpy backroads with crazy mates on GSX-R 1100's I found the K aquitted itself well. A mate on his R80 could do the dance well too and kept it on track wobbling through corners which just had me shaking my head watching his tail light

love the paralell twin sound heaps more than the inline 4 though

 

[John Poirier]

Guys, have y'all by any chance read through any of the articles from this site?

Dead Link Removed

The author is used to working with very high-end scanning products, but the couple of articles I've read so far are quite dated. One comment he made, though -- actually he made the same basic comment several times -- was that the 2700 dpi Nikon LS-2000 did a better job than the Microtek or Polaroid 4000 dpi scanners, apparently mostly because the Nikon handled shadow detail better. D-Max, right? So anyway, I found that rather enlightening because my mindset has become "the more ppi (or dpi) the better" which it appears may not be the best way to approach this subject.[/QUOTE}

The comments about shadow detail are also dated. The D-Max of scanners of that generation was indeed quite limited which resulted in shadow noise problems when scanning slides. Most newer scanners, even decent flatbeds, offer much better shadow performance and when used properly have advantages over DSLRs in terms of controlling rendition of both shadows and highlights.

John

 

[glhs116]

If I can briefly ride my hobby horse...

The thing that annoys me about the cessation of film scanner development is that it denies us the recent advances in sensor and electronics technology. Take my Coolscan 9000 and put a D3-grade sensor in it and what would you have? It would be fast with no other change, because its speed is not limited by its stepper motor but by the amount of exposure required by the current sensor. It would have better density range and more subtle colour rendition. It would be better in so many ways. And you wouldn't need to change a single other thing. People still buy these because the mechanics and optics are in another league from the mechanics and optics in any modern flatbed. This is why people still get better results out of the older gear than anything current.

But, gentlemen, just dream with me if we could have both!

Sam

 

[2F/2F]

You're right, Pellicle. It would be more challenging with color negs. I only copied black and white negatives and prints, and color transparencies and prints this way. My first try would probably start with white balancing to a frame of the orange mask. But since the orange mask is orange for reasons of attaining proper color, I am not sure if the color would be accurate after inverting the negs in Photoshop. Counteracting the mask this way might affect the color negatively.

Another possible approach comes to mind. If, instead of negating the orange mask, there were a way to make Photoshop interpret the color layers and the orange mask just the way that a piece of RA paper does, so it would use the orange mask to arrive at the proper color relationships, instead of just having it tint the whole picture. Perhaps some sort of plugin.

 

[glhs116]

If using a DSLR to scan colour negs I recommend making a filter pack with real colour correction filters to get the white balance close (doesn't have to be perfect). Remember, what you want is to get a good exposure on all three primaries. You can't do that with anything other than exposure. No processing after the fact will bring back information that was not there at capture time. The scanner does it by varying the exposure time for each primary as it scans the primary colours sequentially. Since the DSLR captures all colours at the same instant at the same aperture and shutter speed you will have to adjust the colour balance with filters if you want to achieve an even exposure.

My $0.02.

 

[cooltouch]

I was forwarded this link a few days ago:

Home > Resources > Film conversion

The most useful information -- for me at least -- I found at the bottom of the page.

Prior to being forwarded this link, I worked out a couple of different methods for coming up with reversed images where the colors were reasonably close. While I was struggling with this, I realized something: the only negatives I had been working with were Ektar. And I remembered that Ektar negs don't scan accurately with my 4990 and EpsonScan software. Grayish browns look sort of purplish. A single click using color correction takes care of it, but as it turns out, getting Ektar right when reversing the neg takes a LOT more than a single click. And in some cases, I'm just not successful at all.

I use two different software packages for my post processing: Photoshop CS5 and Paint Shop Pro X2. I've been using PSP since v4.0, whereas I'm still a relative newbie to PS. So I tend to use PSP X2 more than PS, and for some things it's quicker and works better than PS. But I wanted to see if I could come up with an easily reproducible way to reverse negatives with each. And here's what I came up with:

With PSP X2, I bring the negative in, reverse it, then go to Adjust/Color/Fade Correction, and then dial in the amount of fade correction that seems appropriate.

With PS CS5, I bring in the negative, reverse it, then go to Image/Auto Tone, then Image/Auto Color. Sometimes I'll do Auto Contrast as well.

Okay here is a scan of a photo of a caboose. Nikon F2, 24mm f/2.8 AI-s, exposure unrecorded. It was scanned on an Epson 4990 and I'm sure I didn't do any special tweaking within the EpsonScan software. I don't recall, but I may have adjusted the colors after the scan because of Ektar's color shift issues.

I have tried dozens of times to get an accurately colored duplicate of the above image. Even using the routines I suggest above doesn't do a very good job. I wind up with something that looks like this:

BTW, I know the two images are different size-wise. For some reason I cropped the dupe of this negative.

I have found that, in PhotoShop, after I do the Auto Color and Auto Tone stuff, if I bring up the Image/Adjustments/Color Balance and then Image/Adjustments/Hue-Saturation, and tweak the image with these settings, I can finally get pretty close. There's still some residual cyan that I can't entirely get rid of.

And in Paint Shop Pro X2, I can achieve even better results with less work. After doing the steps I outlined above there's still some residual cyan. But next I import the image into PSP's "Express Lab" where the default state when you enter it is with the eyedropper being active. So I select an area that is closest to 18% gray and click on it, and voila. Done. With PSP almost all the cyan is gone, but the sky is not the right color now. Oh well.

But wait, there's more! Just thought I'd also show you some results from non-Ektar negatives. These are considerably easier and in their cases all I have to do is the original steps I outlined above.

Fuji Superia 400

My daughter's cat, whom she's named "Squirrel" but whom I call "The Brat Cat" because that's what she is.

First Photoshop just doing the Auto Tone and Auto Color steps only:

And then PSP X2, doing the Fade Correction. I had to make an additional step and bring it into the Express Lab to get rid of some green tint that shouldn't have been there. Just one click.

Kodak Gold 200

A train trestle along the coast in Ft. Bragg, CA. First is a scan done with my Epson 4990.

Then PS:

And then PSP X2:

There are still slight color variations, but at least I'm getting close. I have not yet tried the steps outlined in that link at the top of this post, but I intend to do them as well.

 

[pellicle]

Michael

just a quick reply to your above post (well done too btw). Looking at the train carriage above I suggest you then play with the hue shifting in photoshop. Looks like the blue relationship is shifted (could be mismatch of sensor frequency responce vs film peaks) try shifting the cyan and blue a wee bit (note the numbers) and that should be usable for that film in future. Also look at red and yellow cross over points too

 

[indigo]

I found it's hard to get good quality images using a digital camera. It's much easier with the scanner and often with much higher resolution. But if I scan something like a book that I don't much care for quality as long as I can read it then the camera is much faster.

 

[cooltouch]

What sort of setup are you using, Indigo? It can make all the difference when you're shooting duplicates of slides or negatives.

 

[Peter De Smidt]

Here's a video of a semi-automated dslr film scanner: DIY Dslr Film Scanner - YouTube

The quality achievable with this method is better, both in resolution and dynamic range, than scanning with a consumer flatbed.

 

[RalphLambrecht]

some people need a hug more urgently than new equipment.

 

[haring]

There is a reason than some of the flatbed scanners cost close to $1000. Nobody would buy them if conventional DSLRs would outperform them in quality. This is that sad truth. I am just about buying an Epson printer. They go between $600-$900USD. I wish my DSLR would produce the same quality and I wouldn't have to spend money on it.

 

[ericdan]

getting better result from DSLR with macro lens than I ever got from my Opticfilm Plustek scanner.

 

[Jaf-Photo]

DSLR-scanning works, but it's slow. It works better for B&W than colour negatives, as it's very difficult to get the colour balance right.

It's also tricky to get the negative precisely parallel to the camera sensor, so there will often be some uniformity issues.

All in all, you will get much more consistent results in less time if you use a decent scanner and a good workflow.

I gave up on DSLR-scanning and got a 35mm negative scanner and a flatbed for medium format. If you divide the cost by the hours I have saved, it is very good economy, I think.

 

[colorfulbliss]

Hi all, first post here. Thought I would share my DSLR scanning results after a month and a half of practice. I'm still in the proof of concept stage and everything is just cobbled together at this point, but I'm convinced that this is a usable method. A little background...I currently have a PrimeFilm 7200 for 35mm and an HP G4050 for 6x7 and 4x5. I'm happy with the results from the PF7200, but not with the G4050. I was at that point of making a scanner purchasing decision and the choices are either very expensive or a compromise of some sort. A dedicated medium format scanner would have been nice, but would have orphaned the 4x5. I guess the goal was to get something that would equal the quality of the PF7200 in the larger formats. From what I have gathered, the v700 wouldn't quite do that. So I bumped into the DSLR scanning threads on LFF (Thanks Peter, btw!), and decided to give it a shot.

Without going into the gory details of my trials and tribulations, here is where I'm at now: Samsung NX100, Nikkor 55/3.5, Porta-Trace box with original bulbs removed and replaced with RGB LED ribbon strips, camera and light source fixed and hand panning the film in my other scanners holders. Workflow is - RAW > MakeTiff > Microsoft ICE > ColorPerfect > Lightroom. sRGB maintained throughout processing.

Samples are from a frame of Ektar 35mm, shot with a Maxxum 35-105 3.5-4.5 and self developed.

This is a 3-shot stitch, resulting file is 7236x4249 exported to 1000px in Lightroom.



And here is a crop:



Gotta get to work now, will post the comparisons to the PF7200 and G4050 later!