Tri-X @800: 3000 dpi and 6500 dpi

[Noll]

Hi all. I used to not think much of the resolving power of a film like Tri-X, but an experiment has me re-thinking this notion.

After getting fed up with my Epson V500 for 35mm, I started using a Panasonic G3 and OM Zuiko 50mm f3.5 macro lens for scanning B&W and some slide film. I use a mirror to parallel the camera to the film plane (thanks to someone's tip in another thread!) and shoot the whole roll in minutes. It works great.

I typically shoot the whole frame at about 1:2.5 magnification (and +1 exposure comp) and the resulting image is the equivalent of a 3000 dpi scan or about 13 mp digital image. These look great, but when viewing at 100% sometimes my inner pixel-peeping monkey demands more....

So for comparison today is an image shot with an OM Zuiko 135mm f3.5 lens either at f5.6 or f8, handheld. I developed it in a slight variant of Caffenol CM which used half of the normal Vitamin C and added some table salt. (if someone is interested I'll share my whole process)

Full image:

Here is a 100% crop of the above image:

Most people would look and think "Ah! I see speckles, that must be the grain. Clearly Tri-X at 800 is no match for 13mp. No need to look closer."

But, that's not actually the grain. It's more of a downsampled artifact of grain + digital noise that occurs for reasons I don't fully understand but no doubt someone here does (grain aliasing?). Anyway, I'm not concerned why right now...

If I throw on the extension tube and rack the lens out to 1:1, this is what we see (pardon my slight differences in processing):

3000 dpi

6500 dpi

And a crop of Lady Wisconsin:
3000 dpi

And again at 6500 dpi:

While the improvement isn't earth shattering, it's noticeably finer grained and a bit more detailed. If I scanned the whole frame like this and stitched it together, we'd be in 30 mp territory. Could a tad more detail be eeked out of the frame at 8000 dpi? Maybe, but probably not much. For now that question will have to remain academic as I'm at the limits of my technology. Hopefully some of you out there found this interesting - if not really all that practical!

I'll be happy to hear your thoughts and post some comparisons from another image if people are interested in this.

 

[mdarnton]

It's a totally viable way of scanning, that I've been using for a couple of years. I've been wanting to move up from my current 12Mp to 24 or 36, because yes, there's still something there, but in the meantime, I've been watching what the guys on the Large Format Forum have been doing--they've been setting up their rigs on rails, and using computers and stepper motors to shoot large format negs in bits, then stitching together extremely high resolution composites. I may still try copying a 35mm neg in, say, three sections, to double my current resolution without getting a new camera!

That mirror thing might have come from me, and I got the idea from someone on another forum. It solves a lot of alignment problems.

 

[Fernando2]

what the guys on the Large Format Forum have been doing--they've been setting up their rigs on rails, and using computers and stepper motors to shoot large format negs in bits, then stitching together extremely high resolution composites

Please, could you share a link? I would much appreciate it! 🙂

I'm working on a similar system, but I'm having some difficulties.

Thanks!

Fernando

 

[Gregoyle]

Here is a link to one of the consolidated pages on LFF:

http://www.largeformatphotography.info/forum/showthread.php?87536-DSLR-Scanner-Light-Sources

And here is a link to the 59 page thread that started the project:

http://www.largeformatphotography.info/forum/showthread.php?84769-Making-a-scanner-with-a-DSLR

Someone in one of the threads linked to a page showing a design using a wooden peg mount that looked like a promising way to do the stitching without expensive automation and rails.

Edit: here it is: http://www.explorablemicroscopy.org/