Originally posted by TME
I am aware that this is the case for audio data capture where the speech or singing of a person can be only fully captured by an analog device like a cassette tape since all nuances and discernible and out-of-our-range audio frequencies are captured on the tape. Given this analogy, I believe that this may be the case for digital video and still photography devices too.
Tapes are far less linear and have a more limited bandwidth than what can be achieved digitally. The main problems with digital lies in the A-D D-A process where distortions are introduced in the form of jitter and and in the analog reconstruction process.
Its very easy to get a very wide bandwidth in digital, just that other problems such as jitter are then magnified due to the high sampling rate.
Conversion of analogue audio data to digital depends on 2 variables, sampling rate(Hz) and precision(number of discrete levels or "bits"). The sampling rate determines wat frequencies(nyquist theorem- must sample at >2x of capture frequency) can be captured and the precision determines how closely that event in time mirrors the waveform amplitude at that instance.
To relate this to film/imaging, think of the number of pixels as the sampling rate and for each pixel, there is the number of bits which determines the colour precision. The equivalent of frequencies in imaging would LPH( lines per picture height).
Though film may seem superior due to it's "analogue" nature, do not forget that the amount of useful detail it carries is limited by the film grain(analoguous to the noisefloor in audio). The size of the film grain determines the useful limit of the info, likewise for digital images where digital noise may occlude detail.
The best 35mm slides may still hold the resolution edge against 6MP DSLRs, however this is assuming that it is kept in its native form, or scanned at a resolution far higher than the useful detail it may contain.
If one's workflow involves digital editing whatsoever, then he is better off working digital direct to avoid conversion losses.
For the majority of film users who may not have access to a megabuck 4000dpi film scanner, the results may usually be worse off than a 6MP DSLR, especially in the case of high ISO noise where DSLRs do seem to hold an edge.
There have been reports on the web of people who compared 20-30MP drum scans of 35mm film to the 1Ds with merely 11MP and the conclusion was clearly in favour of the 1Ds. Only medium format film with over 4x the area can still compete in this regard.
What Nightshade is trying to point out is that most users are not able to get the best out of their film unless at a great cost, which makes all the supposed effective MP rating of film moot.
Where I differ is that I'm not too optimistic about the real resolution of film in equivalent MP. Methinks its only at the 10MP level at best.(Reference to 1Ds sensor)
