Anybody done scanning of LF Negs with a dslr?

[Adrian Pybus]

I was thinking of something along the line of taking 6 x 6 pictures of a 4x5 neg and then stitching.
(Thought of getting the soon to be released Nikon D800E as I don't have a proper Digital camera at the moment. That would give a ~0.5 GPix picture).

Has anybody tried something like that?

Adrian

[Jim Michael]

According to posts in this thread (http://www.largeformatphotography.info/forum/showthread.php?t=84769) it might be impossible. But some people are doing it anyway.

[Adrian Pybus]

According to posts in this thread (http://www.largeformatphotography.info/forum/showthread.php?t=84769) it might be impossible. But some people are doing it anyway.

Yes, that seems a lively thread on the subject!
Adrian

[Peter Langham]

In this article, Walter Blackwell talks about it and suggests it is more than viable.

http://theagnosticprint.org/future-of-scanning/

[Lachlan 717]

You can get a bloody good scanner for the price of a D800E. And get better results.

[Leigh]

Why not just shoot the scene with the dslr in the first place?

- Leigh

[Corran]

Obviously that's not the same thing Leigh. If one owns a nice DSLR and makes use of it for scanning, more power to them. I don't have the patience but it does pique my interest.

[alexn]

I have done this using my D3X and Nikkor 105mm F/2.8 Macro at minimum focus distance (1:1 magnification) and 6x17 slide film. (3x6 35mm frames to cover the slide giving fairly generous overlap) My tripod allowed me to hang my camera underneith it so I set it up over my light table and moved the slide around..

The results were neither great nor bad... Not as good as my Epson 4990 and total price of the setup is about 20x that of the 4990 too... The biggest issue was compiling the stitched image. My computer is very current, its a quad core Intel i7 with 16gb of RAM and it still struggled aligning, stitching and blending 18, 24.4mp images together.

All up it took about an hour to setup and shoot the slide, and about 3~4 hours of processing to get it aligned, stitched, blended and looking like one image rather than a stitch-up...

My advise, Buy a V750. Or, sell your LF gear and buy a D800E, 16-35 F/4 VR, 24-70 F/2.8 VR, 70-200 F/2.8 VR and pray that your $8500+ purchase produces better images.. (which it probably wont.)

[rdenney]

I have done this using my D3X and Nikkor 105mm F/2.8 Macro at minimum focus distance (1:1 magnification) and 6x17 slide film. (3x6 35mm frames to cover the slide giving fairly generous overlap) My tripod allowed me to hang my camera underneith it so I set it up over my light table and moved the slide around..

The results were neither great nor bad... Not as good as my Epson 4990 and total price of the setup is about 20x that of the 4990 too... The biggest issue was compiling the stitched image. My computer is very current, its a quad core Intel i7 with 16gb of RAM and it still struggled aligning, stitching and blending 18, 24.4mp images together.

All up it took about an hour to setup and shoot the slide, and about 3~4 hours of processing to get it aligned, stitched, blended and looking like one image rather than a stitch-up...

My advise, Buy a V750. Or, sell your LF gear and buy a D800E, 16-35 F/4 VR, 24-70 F/2.8 VR, 70-200 F/2.8 VR and pray that your $8500+ purchase produces better images.. (which it probably wont.)

You may be correct that stitching becomes the barrier. But as to achieving better resolution than a V750, we've already proved that using less sophisticated equipment than what you have. There was a thread linked earlier in this thread that you should read and participate in since you have experience with an attempt. Several of us have been exploring this issue in some depth, and your experience would be helpful in that thread. It's a long thread, but it does address some of your issues and hasn't yet reached others of your issues.

Rick "what if a V750 isn't good enough, and ten years from now nothing else is available at any price?" Denney

[Adrian Pybus]

You can get a bloody good scanner for the price of a D800E. And get better results.

That is the problem. You can't get a bloody good scanner for the price of a D800E. There isn't one, for LF at least.

[Adrian Pybus]

I have done this using my D3X and Nikkor 105mm F/2.8 Macro at minimum focus distance (1:1 magnification) and 6x17 slide film. (3x6 35mm frames to cover the slide giving fairly generous overlap) My tripod allowed me to hang my camera underneith it so I set it up over my light table and moved the slide around..

The results were neither great nor bad... Not as good as my Epson 4990 and total price of the setup is about 20x that of the 4990 too... The biggest issue was compiling the stitched image. My computer is very current, its a quad core Intel i7 with 16gb of RAM and it still struggled aligning, stitching and blending 18, 24.4mp images together.

All up it took about an hour to setup and shoot the slide, and about 3~4 hours of processing to get it aligned, stitched, blended and looking like one image rather than a stitch-up...

My advise, Buy a V750. Or, sell your LF gear and buy a D800E, 16-35 F/4 VR, 24-70 F/2.8 VR, 70-200 F/2.8 VR and pray that your $8500+ purchase produces better images.. (which it probably wont.)

I have a V750. The results are rather murky and have less DR compared to a drum scan but drum scanning costs so much it isn't feasible (about 100 euro per image).

Currently, my plan is to get my really good pictures drum scanned and use the V750 for the not so good pictures. This'll work out fine as long as I don't shoot too many good pictures! :confused:

Anyway, I'll probably find a good use for a D800E too! :)
I shoot both digital and film. I don't see it as an either or proposition.

Adrian

Adrian

[false_Aesthetic]

Trust Walker.
He knows his shit.
Really knows his shit.

[Peter De Smidt]

Trust Walker.
He knows his shit.
Really knows his shit.

I've corresponded with him through email. He's very excited about our dlsr scanner project.

[Nathan Potter]

I think Walker is on the right track and any DSLR with nominally 5 um pixel resolution can capture 5 um image detail at 1:1 given a high quality macro lens. The larger question from the hardware end is how does the contrast and purity of color from the DSLR capture compare with that at equal resolution from a drum scan. The answer rests in large part on the nature of the light source from the film and the angular dependence of capture. The DSLR will suffer from serious light scattering problems if a diffusion source is used so image contrast will be greatly reduced and cannot be recovered in PS without contrast expansion techniques. Scans will appear muddy with diminished tonal values.

I think one needs to use a highly collimated light source to preserve tonal values. This will lead to maybe too much increase in contrast but that is more easily adjusted in post processing than too little contrast. Proper light baffling is essential in early tests to maximize the captured color purity and density values really represented in the film emulsion. A high degree of collimation also reduces scattering from silver and dye in the emulsion, an exceedingly important aspect in the replication process.

I've been thinking a bit about this from the hardware POV and hence a sketch of a concept is inserted below. An obvious change would be to place the condenser pair right under the stage - sort of an upside/down D2 arrangement. For collimation a point source could be employed (shown).

Nate Potter, Austin TX.

http://farm8.staticflickr.com/7045/6851712881_d94fbc3ebd_z.jpg (http://www.flickr.com/photos/argiolus/6851712881/)
SCANNER-3jpg (http://www.flickr.com/photos/argiolus/6851712881/) by hypolimnas (http://www.flickr.com/people/argiolus/), on Flickr

[Peter De Smidt]

Nathan, you should post this sketch in the build thread, and we should probably discuss it there to avoid to much fragmentation of discussion.

I agree about the importance of the type of light source. I'm using a diffusion source with my prototype simply because I have all of the parts on hand. I don't have condensers or a halogen point source for a collimated source. (I do have a very nice 45" front surface mirror arrangement from Kodak.) If the stitching issues are resolved, then this would certainly be a good area of investigation.

Regarding the support structure, a multi-column approach might be preferable to a single column.

[Nathan Potter]

Yes Peter, I should be better organized. I'll drop this in the build thread.

Nate Potter, Austin TX.

[buggz]

VERY good read!
Thanks for the link.


In this article, Walter Blackwell talks about it and suggests it is more than viable.

http://theagnosticprint.org/future-of-scanning/

[buggz]

Does anyone have a setup to do this with 4x5 negatives/positives?
Is it better than scanning with an Epson V750-M Pro?
I am thinking of getting this Epson scanner, just don't know how effective this would work with 4x5.
I have a 5DMkII, I would LOVE to do this "alternative" method, if it is better than the Epson V750-M Pro output.

[rdenney]

Does anyone have a setup to do this with 4x5 negatives/positives?
Is it better than scanning with an Epson V750-M Pro?
I am thinking of getting this Epson scanner, just don't know how effective this would work with 4x5.
I have a 5DMkII, I would LOVE to do this "alternative" method, if it is better than the Epson V750-M Pro output.

Please, go to this thread. Much discussion and all your questions are at least addressed. It's a long thread but it's really digging into the possibilities here.

Making a scanner with a DSLR (http://www.largeformatphotography.info/forum/showthread.php?t=84769)

Rick "trying to avoid crossing threads" Denney

[buggz]

Thanks!