3000 scan res for 35mm the max?

Was looking at the Photo-I website, and he made a comment in one of

the reviews that 3000 dpi resolution was the max that 35mm was able

to retain. Anything above that just enhanced the grain? Any comments

on that?

I can get a measurable 80lp/mm to film. No way you can resolve that with 3000 dpi. 4000 dpi is about 60 lp/mm best case and you always have to have a lot more capability in your copy system that what you are copying to minimize your losses.

 

Look here for more test information:

 

http://www.fbi.gov/hq/lab/fsc/backissu/april2002/swgitfield1.htm

 

No doubt for hand held, point and shoot type photography, 3000 dpi is probably overkill.

 

In some of the scanning I do, I get what looks like grain with 1200 dpi scans of 4x5 negatives. Obviously it is a artifact of the scan as it can't possibly be grain at that magnification. That may be what they are talking about.

The information recorded on film is in the size and spacing of grain and grain clumps. In the case of transparencies, the grain information gets transfered to dye clouds. It's more or less the same thing. Whichever form, the information is stochastic. That is, from the scanner's perspective, it is more or less randomly distributed over the film.

 

When you scan, the scanner applies a virtual grid to the film. The size and locations of the holes in the grid is rigidly deterministic, and is determined by the scanner resolution setting.

 

The question is, where do these two things meet? Clearly, once the size of the grid becomes smaller than the size of the grain, you have the ability to capture all the information recorded on the film. Higher scanner resolutions will only tell you what the film substrate looks like between film grains - this is more data, but it is not more information. There is a difference between the two, yes?

 

Turns out, with a really fine grained film (say, Velvia), you can get everything there is to get at about 5000dpi. With a high speed film (say, Tri-X) you can get everything there is to get about about 4000dpi. This is, of course, independent of film format, and applies from Minox to 20x24 and beyond.

 

Now, this presumes that your optics and your technique, and your processing can keep up with the film. If you are doing the full large format, tripod mounted, zone system, modern lenses, cable release, wait-until-the-wind-dies-down thing, the above applies. If you are doing the 35mm hand held commercial processing thing, then scanning to these high levels is probably a waste of effort because the image recorded on the film will not be taking full advantage of the sharpness capability of the film.

 

But the sharpness of the image is very subjective. What it really does is determine how much enlargement you can give an image before the lack of sharpness begins to annoy you. Utterly subjective, and utterly image dependent.

 

If you are doing your own scans on your own scanner, you might as well scan at full optical resolution of the scanner. If your image is really sharp, you avoided a rescan at higher resolution. If it's not really sharp, you can down rez it in an image editor. What you might loose is some disk space and a few minutes of scan time.

 

Clearly, YMMV.

It depends on the film you scan, and which lens you used. Scan Velvia or TMX (as an example of fine-grained film) (or even finer go for TP) and you'll need way more than 3000dpi to capture all that the film can record.

If anybody claims that 35mm films is worth scanning at more than 4000dpi, then they must also claim that Kodak 110 film looks just fine scanned at 2000dpi. This is a relative example of just how bad the exagerated claims of film sharpness are.

 

My limit is 2000dpi if I can avoid it, and that's with 100 speed slide film.

You need to size the scan to the final output. Definitive statements like, "I never scan 35mm at over 2000 dpi" are useless and don't address the issue of the scan size needed for the final print.

 

I've actually made poster sized prints from 110 film - and without resolving the dye clouds, you couldn't get the effect of abstract pointilism that changes into a photo at the correct viewing distance. For that image, I scanned at 6300 ppi - my scanner's maximum resolution. You can easily see the dye cloud formations long before you get to the image's individual pixel level. Without that level of scanner resolution, the entire point of making a very small film image into a big print would be lost.

 

Lately, I've been scanning a lot of Kodachrome 200 - a film with huge dye clouds compared with something like Provia 100F. Yet, regardless of the film's construction and grain, what I'm interested in is the number of pixels I need for the final print.

 

In this case, I'm making prints that have an image area about 19-inches in length. They are printed on an Epson 9600 printer. For that size, if I'm expecting a 300 dpi output on the Epson, I need a scan of approximately 6100 ppi. Since my scanner maxes out at 6300 ppi - that's what I scan a 35mm at for a 19-inch long print, as that is the scanner size available to me nearest the "perfect" 6100 ppi.

 

If I limit myself to the "expert recommended maximum" of 2000 ppi - then I have to enlarge the file size through something like Genuine Fractals or use a stepped (<10% per step) up-size in Photoshop using "bicubic smoother" interpolation. I don't get the point or see the advantage of that approach. I've tried it, (stepped / bicubic smoother in Photoshop), and to me, the final print is not what I'm looking for when compared to the image scanned at a higher resolution.

 

This really has nothing to do with "resolving details" on the film, but generating the correct file size required for a certain print size. Scaling your scan to the final print size is the way to correctly determine the number of pixels needed, and not some abitrary "limit" based on someone's idea of acceptable resolution.

 

Yes, you will see film "grain," actually the dye clouds being reproduced, but - so what? That's no different than making the same size print in a wet darkroom. When the print is viewed at the correct viewing distance, you are not aware of the grain. When viewed at a closer distance, you become slightly aware of the dye cloud pattern, but, that is the nature of film-based photography isn't it?

 

If you want "grainless" images at a certain size, then you must change formats and work in a film size that lends itself to grainless reproduction at the print size you desire. I recently made a 36-inch long photo from a 6x12 transparency. At close range viewing (less than 1-foot), you still could not see grain or dye clouds. But, if you are working in 35mm, and you want a fair size print, then you live with the results of choosing that format to be reproduced at a large size - and, there's nothing wrong with that.

 

However, taking a 35mm image to a larger size does take a certain level of craftsmanship and work. It is not as easy as limiting yourself to 8x10 or smaller where you can slop along and get fairly good results.

 

Because scanners generate some inherent electronic noise, and sharpening to compensate for the scanning process also generates some noise (accentuates grain) - one must be careful when to apply sharpening, how much to apply, and how to counteract it through noise reduction. People write books on this stuff. You can learn it through trial, error, and observation, and it really doesn't take long if you make accurate notes and are good at discriminating changes in the final print.

 

My advice is to ignore the self-styled "experts" and most of the reviews and information you find on the web. A lot of it is just plain bias coupled with "facts" manipulated to support the person's contention.

 

I think what you'll really find, is that if you approach the imaging problem with an open mind and the realization that certain goals (grainless 30x40 print from a 35mm, for example) are not possible because of the materials involved. Then you either scale back your expectations or learn to exploit the materials to their maximum with the best craftsmanship possible, and accept the final results.

 

 

Steve, you have just laid out the best approach to enlarging 35mm format I have ever read on photo.net. I am tired of this renewed wave of 35mm-bashing. It's still a darn good format, considering its versatility and yes, quality. No one argues that MF or LF are better, I just don't understand this belligerence when it comes to 35mm. It has turned into some kind of anathema here, especially among the 6MP DSLR shooters. Give me a break!

 

As for Chip's question, it looks to me that most of my 3000 ppi scans are good enough to see all the detail, however, when it comes to enlarging your 35mm to a certain print size, I'll second what Steve has to say on this (read above).

 

Best regards.