I know that there are attachments that allow you to "scan" a neg or slide using a digicam. But, I've never seen any such attachment for anything larger than 35mm. Why is there no adaptor that allows the user to "scan" just a portion of a neg (either 35mm or 120) and then stitch in PS? I tried building a rig to do this using a copy stand, but moving the neg around to get 6 shots that could be stitched was way too imprecise and the results were far from what I wanted. Anyone out there got ideas or the skill to build something like this?
--Joe
A 24MP APS-C dslr to digitize a whole frame of 120 isn’t enough resolution? How much resolution do you need? 4000x4000 or 4000x5000 will make a pretty nice 16x16 or 16x20 inch print. Need to print bigger than that? You should probably go to a bigger negative as 16x20 is a pretty big enlargement from what is essentially a 2x2.5 inch negative. If you step up to a 4x5 neg, just use a flatbed scanner.
The whole stitching a bunch of photos of a frame of film is totally overrated, as it is extremely time consuming, prone to less than great results, and at the end of the day isn’t generally as good as just taking one picture of the whole film frame with a reasonably modern sensor digital camera. If you really want to hit it with resolution, buy a used Canon 5DSr and take 50MP photos of each film frame.
I think you’ll soon discover that 24MP in one shot is plenty of res for pretty much anything within reason. Trying to scan and stitch multiple frames isn’t really going to give you more usable resolution. For 35mm, 6000 pixels to scan 36mm of film is ~80 line pairs per mm of film.
While I agree that photographing frames single shot with a high resolution camera is the preferred option (when your budget allows this), in most cases even a 50Mp camera is not sufficient. I think your numbers are a bit optimistic. Here are some examples.
Example 1 - 35mm film, 16Mp camera, single shot.
Camera: Pentax K-5II
Sensor: 4928x3264 pixels (3:2)
Film format: 24x36 mm (3:2)
1" = 25.4 mm
Horizontal 36 mm, that is 36/25.4 = 1.417"
Resolution: 4928/1.417 = 3470 dpi
Example 2 - 35mm film, 24Mp camera, single shot.
Camera: Pentax K3-II
Sensor: 6016x4000 pixels (3:2)
Film format: 24x36 mm (3:2)
1" = 25.4 mm
Horizontal 36 mm, that is 36/25.4 = 1.417"
Resolution: 6016/1.417 = 4240 dpi
Example 3 - 35mm film, 16Mp camera, stitched in 4 parts.
Camera: Pentax K-5II
Sensor: 4928x3264 pixels
Film format: 24x36 mm
Stitch: 2x horizontal and 2x vertical + overlap 3 mm
1" = 25.4 mm
Sub frame horizontal 36/2 + 3 mm = 21mm, that is 21/25.4 = 0.826"
Resolution: 4928/0.826 = 5960 dpi
Example 4 - 35mm film, 50Mp camera, single shot.
Camera: Canon 5DsR
Sensor: 8688x5792 pixels (3:2)
Film format: 24x36 mm (3:2)
1" = 25.4 mm
Horizontal 36 mm, that is 36/25.4 = 1.417"
Resolution: 8688/1.417 = 6130 dpi
Example 5 - 120 film 6x4.5, 24Mp camera, single shot.
Camera: Pentax K3-II
Sensor: 6016x4000 pixels
Film format: 56x41.5 mm (Pentax 645N)
1" = 25.4 mm
Position frame horizontal, fill out vertical in sensor, 41.5/25.4 = 1.63"
Resolution: 4000/1.63 = 2450 dpi
Example 6 - 120 film 6x4.5, 50Mp camera, single shot.
Camera: Canon 5DsR
Sensor: 8688x5792 pixels
Film format: 56x41.5 mm (Pentax 645N)
1" = 25.4 mm
Position frame horizontal, fill out vertical in sensor: 41.5/25.4 = 1.63"
Resolution 5792/1.63 = 3550 dpi
Example 7 - 120 film 6x4.5, 50Mp camera, stitched in 2 parts.
Camera: Canon 5DsR
Sensor: 8688x5792 pixels
Film format: 56x41.5 mm (Pentax 645N)
Stitch: 2x vertical + overlap 5 mm
1" = 25.4 mm
Sub frame vertical 56/2 + 5 = 33mm, that is 33/25.4 = 1.30"
Resolution: 5792/1.30 = 4450 dpi
You can read ppi or lpi instead of dpi.
Some conclusions:
1. A 16Mp camera is not sufficient to scan a 35mm film (24x36) single shot (Example 1).
2. You can scan 35mm film with a 24Mp camera with more than 4000 dpi (example 2).
3. When you are stitching even a 16Mp camera for scanning 35 mm film can give plenty of resolution (example 3).
4. A 24Mp camera is not sufficient to scan 120 film (6x4.5) single shot (Example 5).
5. A 50Mp camera is not sufficient to scan 120 film (6x4.5) single shot (Example 6).
6. With a 50Mp camera you can scan + 2-part stitch a 120 film (6x4.5) with 4400 dpi (Example 7). This number is reduced for 6x6 and 6x9.
So even with a 50Mp camera you can not get enough resolution in most cases and the only solution is to stitch.
On average you need at least 4000 dpi for a good scan, but it depends what you are doing with it. For a large print you need more resolution. When scans are downsized and only shown on a monitor even scanning a 35mm film with a 16Mp camera single shot can be sufficient.
Edit 25/04/19: Example 5 corrected.
@Adrian Bacon. I know that prints are 240..300 dpi, but the principal point is that the function of a scanner is to reproduce all the details in the negative, preferably including the details of the grain (grain clumps), so you need 4000 dpi as a minimum to accomplish that. It depends what you are doing with the scan after that, with options to resize to the dimensions of a print or to show it on a screen.
It makes a difference when you are scanning at 240 dpi and then print at 240 dpi, or scan at 4000 dpi and then resize to 240 dpi for a print.
I agree that a scan of a 35mm negative with a 24Mp camera gives sufficient resolution without stitching (see my example 2), but not for 120 film. You can easily calculate in advance what you get with the gear you have or when you want to build a scanner yourself, the numbers are speaking for themselves.
Example: When you are scanning a 24x36 mm negative at 240 dpi and print the result (at 240 dpi) you get a postage stamp with a size of 0.84x1.42".
If you really want to make 4000 dpi the minimum scan resolution you’re willing to accept, I’m not going to disagree as only can decide what is acceptable for you, however, I own and operate a film lab and process and scan *a lot* of film. 24MP shows film grain in most emulsions in 120 format. It’s a simple fact. Once you’re seeing film grain, more resolution is diminishing returns.
The point of scanning film is that ideally you fully need to reproduce the grain because it is a very important feature that can't be missed. So, the grain is an easy benchmark to check if you are scanning correctly.
I have read somewhere (on Photrio?) that there are persons who actually try to remove grain from scans artificially.
It really comes down to: You can throw out extra data you don't need, but its a lot harder to create data you didn't have if you decide you need it.
Line pairs per mm is a resolution standard.
But there are a few other factors (e.g. noise and the affect of certain post-processing techniques) that do make it beneficial to have more data than less.
You may be only able to see a difference with certain presentation options - large, backlit prints come to mind - but if you expect to be able to make those presentation choices in the future, you need to either make your scans at those higher resolutions, or be prepared to make additional scans later.
I looked into the suggestion of @Luckless (#14) to modify a 3D printer for stitching, with an X-Y adjustment for a film-holder and height adjustment for the camera, and came across the Velleman K8200 3D printer which is a relatively cheap kit to build yourself.
View attachment 219997
Here are some thoughts for modifications and possible limitations.
The printer has an open frame which allows modifications for mounting a camera and a film-holder. I think the best place for a camera is on the position of the nozzle (which is left out) because it has a height adjustment, and the print bed is replaced by a film holder where it can be moved in X-Y directions by the gear below. In this configuration a flat LED light source could be placed below the film-holder between the feet of the frame (for example a Kaiser Slimline LED panel or equivalent).
The camera is relatively heavy compared to the nozzle so perhaps a contra-weight is needed to level that out. I think the camera should also have an angle finder to allow focusing looking from the front.
If the space between the X-Y guide rails for the film-holder is too narrow for the light source to illuminate the film completely, you could perhaps replace some parts to increase the distance.
All adjustments X, Y and Z can be made manual by leaving out the stepper motors and by mounting knobs, but you could theoretically make an automated system too (all software is open source).
The overall dimensions of the frame are width=50cm, depth=42cm, height=60cm, which seems sufficient for 35mm and medium format films. However, the height of the vertical guide rails is much shorter than 60cm so perhaps they must be mounted higher up to get sufficient distance between camera and film.
As the camera is the heaviest, I'd have it fixed to the bed. As if on a tripod with the lens horizontal. Have the neg carrier controlled by the x and z axes. The y axis would control focus by jogging forward and backward. afaik, some cameras have a video output so focus could be achieved by looking at a monitor.
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?
We use cookies and similar technologies for the following purposes:
Do you accept cookies and these technologies?