Adox CMS 20 II: An Ultra-High Resolution 35mm Film
Test Results and Sample Close-Ups From a 400-Megapixel Scan7 min read by
No other film is sharper, no other film is more finegrained, no other film resolves more lines per mm (up to 800 l /mm)… The film achieves grain free enlargements of up to 2,5 meters diagonally. This equals mathematical about 500 Megapixels.
— product description from the Adox’ website.
In this article, I will be putting Adox’s claims of ultra-high resolution on 35mm film to the test using a couple of modern digital imaging techniques.
Adox CMS 20 II is a fine-grained monochrome emulsion that was originally used for archiving documents on microfilm. It has since been made available for 35mm cameras to create high-resolution captures without having to go medium- or large-format. Adox claims that each frame from their 35mm film can yield you a whopping 500 megapixels: incredible, considering that most modern digital cameras are in the 20-40-megapixel range.
✪ Note: You can get Adox CMS 20 II in 4x5 large format to capture in what I would imagine be a mind-blowing amount of detail.
Rating and developing Adox CMS 20 II.
The challenge of working with this film is managing its extremely high contrast and miserable dynamic range when developed in standard black and white chemicals at box speed. Usable contrast levels can be achieved when the film is shot and developed at 3 or 6 ASA/ISO, according to Adox.
This effectively means that you may need to shoot at 𝒇2.0 in the full sun to get the 1/125s shutter speed for motion blur-free images. Maybe 𝒇2.8 or 𝒇4.0 if you have a really steady hand. For greater depth of field or to shoot in the shade, you may need to use a tripod and longer exposure times.
Adox’ recommended Adotech developer helps with the contrast issue, allowing you to shoot your CMS 20 at ISO 20-80 — with the higher contrast effect appearing at higher speed ratings. I’ve also seen someone try Rollei RLC Low contrast developer with good results while shooting at the box ISO 20.
For this little experiment, where I try to squeeze every bit of digital resolution from the film, I had The Lab develop my Adox CMS 20 in Ilford DD-X after I’ve exposed it at ISO 12. It was certainly too high of a rating as most of the images on my roll didn’t turn out well. Especially the ones in high-contrast scenes. But there’s one that stood out. It was a photo of a large heron bird in the distance along with a few ducks; taken in strong overcast with a bit of fog there was very little contrast, which translated nicely onto the emulsion. I managed not to shake my Voigtländer Vitessa too much and captured plenty of detail in the middle of the frame.
I’m sure that with better chemistry, maybe a sharper lens, a tripod, and a scene full of minute detail you will get even better results.
Scanning Adox CMS 20 II into a 400MP file with PrimeFilm XAs and Adobe Super Resolution.
To get the 500-megapixel resolution Adox says CMS 20 can produce, I’d need to extract 0.5 billion pixels from a 24 ✕ 36mm film strip (1.3348in²). That means I’d have to have a scanner that’s capable of 23,000DPI — which I do not have.
But I do have my PrimeFilm XAs scanner with its 5,000 ✕ 10,000DPI resolution which can get me 137-megapixel files after some interpolation. And I’ve got Adobe’s new Super Resolution tool that can artificially boost this number to 409 megapixels.
✪ Note: I use Adobe Photoshop to make all my image manipulations in this article. The link above has a few special offers on Adobe plans and products. Last time I checked, you can get the “photography” package for $10/mo.
It’s true that a lot of these pixels are “fake.” Starting with the interpolation that stretches one of the image’s dimensions using a bit of clever math. Adobe’s Super Resolution tool draws on its data from machine learning sessions to “guess” what the neighbouring pixels would look like.
The “purest” resolution I could get with my PrimeFilm XA — i.e., no interpolation, no machine learning enhancement — is 5,000DPI or 34 megapixels on 35mm film. Which is already a lot — but why not push it a little further, even if it’s with a bit of help from the machine brains? That should still give a nearly complete idea of what it’s like to work with the “highest resolving image recording system in the world.”
How I got this 400-megapixel scan: I had my PrimeFilm XAs scanner output a positive 24-bit uncompressed TIFF via VueScan driver at full interpolated 10,000DPI. This generated a huge, unprocessed image that I opened using Adobe Bridge — the only way to get to the “Super Resolution” tool if you are working with TIFFs at the time of this writing. From Bridge, I’ve right-clicked on the image and chose “Open in Camera Raw…” Then, right-clicked on the image again and “Enhance...” with “Super Resolution” checked. At this point, I’ve got an enormous negative that I then negated and equalized in Photoshop.
At 100% crop on a 13-inch high-pixel-density monitor, there’s no discernable grain, as expected. The file is 24,800 ✕ 16,500 pixels, which translates into 10 image pixels for each physical display point on my Macbook Pro’s 2560 ✕ 1600 Retina display.
If you’re reading this article on a laptop or desktop, you may be looking at images that are 1,800 pixels wide. This is not the best way to appreciate the resolution of this film. Let’s zoom in.
At 25% crop, the file is a more manageable 6,200 ✕ 4,125. Here, I’m starting to see a bit of a texture, but I certainly can’t pick out individual grains yet. The resolution of this scan is still more than double that of my laptop’s display’s ability to resolve points.
After enlarging the image by the factor of eight, I’m not sure if there’s any more detail than when it was 4x. It’s hard to tell whether it’s because the interpolation and machine learning software failed to “fill in the blanks” or if it’s the film’s grain or, maybe, it’s my camera’s lens that limits this film’s true resolving power. But there’s still plenty of pixels as this is 3,100 ✕ 2,063.
Finally, at 7% crop, I’m beginning to recognize Adobe’s “fake” pixels. They look like tiny swatches of patterns stitched together.
A drum scanner or a darkroom print may reveal a lot more detail than this. Nevertheless, pretty impressive results. This is certainly more detail that I was able to capture on my PrimeFilm XA from any other emulsion.
If I were serious about printing Adox CMS 20 large, I would probably limit my scan to 5,000DPI and take “Super Resolution” out of the equation. Even at that level, I should be able to print two-feet-wide (0.6m) at 300DPI, which is good enough to be viewed as close as 15 inches (38cm).
Or, I’d look for someone with a really nice enlarger to just blow up it HUGE!
❤ By the way: Please consider making your film and developer purchase using this link so that this website may get a small percentage of that sale — at no extra charge for you — thanks!