For what camera???
(The D800 for example apparently has a 4.88µm pixel size)
Then it can be said that film has a higher resolution than the D800.
Then what's your point?
The "molecular radius" of silver bromide does not matter at all (nor the lattice constant, which is 0.57745 nm for pure AgBr if that was meant), it is the size of the crystal grain in the emulsion. The size of the original silver halide crystal is related to the size of the "grain" that we later see, but it is not the same. If a silver halide grain gets enough exposure (usually at least 4 photons) it will be completely changed to silver in the development process. That silver is usually a bunch of silver filaments whereas the original grain was either something close to a cube shape or triangular to six-sided plate in the case of T-grains. What we see in the end as "grain" in the print, are the holes between the silver filaments in the negative.
Silver halide grains have a pretty wide range of sizes depending on the film. It ranges from about 0.03-0.04µm (30-40nm, http://www.ilfordphoto.com/holofx/holofx.asp) for holographic film with a claimed resolution of 7000lp/mm (so-called Lippmann or micrate emulsions) to T-MAX 3200 with a grain width and length of 5-6µm (thickness of 0.2µm), which is about the pixel pitch of a D800 sensor.
Here's where I think conversations like this have a problem. The attached image (one of many that I have with this phenemenon) is a 100% from Velvia RVP scanned at 4000 PPI with a Nikon CoolScan scanner (not as good as a drum scan but better than an Epson flat bed). This is a reflection of a sunset on a calm lake The tonal gradations should be smooth but they are blotchy. With digital they would be smooth. This is very subjective. Some people will prefer the look from film, some from digital. In this case I would prefer the digital because to my eye it is more true to reality. All the other science and theory becomes secondary because the proof is in the pudding and the pudding is the final product plus interpretation of the final product involves personal tastse.
No confusion Tim. I and others have simply chosen to put the airy disk diameter equal to a linewidth in order to simplify the results from a scanning densitometer reading. In doing that we determine at what spacial frequency the contrast drops to 50%. And that spacial frequency can be tied directly to a diffraction limit rather than an Abbe condition. It's just a different choice of measurement technique.
The idea is to eliminate individual judgement by instrument scanning through a series of lines. The resolution figures obtained thusly may not agree at all with figures obtained by other means - for instance using an Abbe limit or even edge deconvolution using Imatest software. I think this is all OK as long as we understand the different data collection techniques. We are all, in effect, setting our own standards for our own pedestrian purposes.
Generally I would not favor working sideways. For instance if a set of data showed higher resolution numbers (yours) than a set taken by another method (mine) I'd be nervous fitting the result from the former to the latter. But discussing the difference is a good intellectual exercise nonetheless.
That very high resolution number for the Zeiss bothers me. It is way above any diffraction limit at f/4 which for green light is about 5 µm. The only reasonable way to achieve that is to reduce the wavelength but perhaps Arne knows some details. Of course the method of measurement could depart greatly from what we have been discussing.
Nate Potter, Austin TX.
Yeah that's fair enough Working with a fixed MTF50 does give a universal reference point but I do honestly think that MTF10 or even MTF5 would be more realistic when comparing the visual resolution of 'system'. However in reality it's probably productive to show both.
What would be ideal would be a curve showing contrast drop off with resolution..
p.s. I don't think it's possible to compare our data at all in this case... unless there is an agreed fall off in contrast as resolution increases (which I think is possible but unlikely)
Still Developing at http://www.timparkin.co.uk and scanning at http://cheapdrumscanning.com
Bookmarks