New TMax P3200

[crypt47]

Good CD quality audio is not digitized at CD sample rates....[skipped]

This forum is about image in first place, let's do not stray to another territory. (Equally ignorent in this regard too.

In the case of ADOX CMS 20, the frequency of that films 100% contrast response is significantly higher than what pretty much every scanner can sample at, which means that you're only digitizing it at the limit of your scanner and not actually digitizing it at a resolution that does the film justice, so how exactly is that a good scan?

Haha.) But with your definition of sharpness I've got maximum edge contrast. You suggest doing it hard way (first scan with multiple time more resolution then edit then downscale anyway) but either you definition of sharpness is flaved or I've got 100% sharp image and that confirms what I say about grain size and pixel from the start.
I don't do scan for the sake of [frequency] theory or for a film justice (using your words, but phrase came like in a hollywood movie not a native speaker). The higher an output resolution the better, of course. But...

They are used to seeing images that have a contrast response of 100% all the way up to the Nyquist limit and therefore confuse sharpness with resolution. Resolution and sharpness are related, but they are not the same thing. Sharpness is the contrast difference along edges.

As far as I know english word 'acutance' is used to define edge sharpness. Generaly when people speak about sharpness they mean acutance _and_ presence of small details.

Back to the films with larger grain. I tried to encrease resolution of scan with Delta 3200 by 2. After small details are already captured increasing resolution doesn't provide more sharpness (as I understand this term), it provides only another representation of grain. I also compared downscaled images: no difference to my eye. Grainy noise is just a noise. So what does 'working resolution would be 4800x6000, and the source scan should ideally be 9600x12000 pixels' means for films like Delta and P3200? More details? No. More acutance? No. Better represented grain projection? Maybe, it depends.

In your experience, how do you deal with a grain that is the same size as your pixels but happens to be right on the boundary of two different pixels? That is the definition of grain aliasing ...

I believe I have the very same case with some EI100 films. And given I have the same scan resolution for it and for EI3200 films 100% magnification looks better with aliasing of EI100 films. Do you know why? Because grain is not a dot. It's projection has peculiar form, sharpness and location. And it may be the case where your theory of frequinces doesn't apply that well. This is also the reason I think P3200 is not that fan to scan as Delta: more irregular grain (no prof picture though).

Image with 1000x magnification of grain (do you see it? i doubt i've uploaded it right):
Dead Link Removed

And here is another image i pixelated with a pixel size of Acros film:

Dead Link Removed
and another has a bigger pixel:
Dead Link Removed

Looking from a normal distance can you say which is more messy? Acros now looks more even, fp4+ too. Does this impact accutance? Not that obvious because of how grain pattern impacts our perseption of sharpness. With finer grain it's aliasing less impacts details. With faster films grain comes closer to the size of a detail (the same output resolution).




p.s.

Digitize at the maximum resolution of your sensor, low pass filter it either optically, or after the fact digitally to reduce visual artifacts and aliasing, then scale down to an intermediate working resolution to do the bulk of your work, then scale down to your destination resolution. It looks better once at the destination resolution.

Well, let's say for briefness I've heard other opinions (about final sharpening etc) on this particular matter, but it's out of my scope at the moment.

 

[Adrian Bacon]

This forum is about image in first place, let's do not stray to another territory. (Equally ignorent in this regard too.



Haha.) But with your definition of sharpness I've got maximum edge contrast. You suggest doing it hard way (first scan with multiple time more resolution then edit then downscale anyway) but either you definition of sharpness is flaved or I've got 100% sharp image and that confirms what I say about grain size and pixel from the start.
I don't do scan for the sake of [frequency] theory or for a film justice (using your words, but phrase came like in a hollywood movie not a native speaker). The higher an output resolution the better, of course. But...



As far as I know english word 'acutance' is used to define edge sharpness. Generaly when people speak about sharpness they mean acutance _and_ presence of small details.

Back to the films with larger grain. I tried to encrease resolution of scan with Delta 3200 by 2. After small details are already captured increasing resolution doesn't provide more sharpness (as I understand this term), it provides only another representation of grain. I also compared downscaled images: no difference to my eye. Grainy noise is just a noise. So what does 'working resolution would be 4800x6000, and the source scan should ideally be 9600x12000 pixels' means for films like Delta and P3200? More details? No. More acutance? No. Better represented grain projection? Maybe, it depends.



I believe I have the very same case with some EI100 films. And given I have the same scan resolution for it and for EI3200 films 100% magnification looks better with aliasing of EI100 films. Do you know why? Because grain is not a dot. It's projection has peculiar form, sharpness and location. And it may be the case where your theory of frequinces doesn't apply that well. This is also the reason I think P3200 is not that fan to scan as Delta: more irregular grain (no prof picture though).

Image with 1000x magnification of grain (do you see it? i doubt i've uploaded it right):
Dead Link Removed

And here is another image i pixelated with a pixel size of Acros film:

Dead Link Removed
and another has a bigger pixel:
Dead Link Removed

Looking from a normal distance can you say which is more messy? Acros now looks more even, fp4+ too. Does this impact accutance? Not that obvious because of how grain pattern impacts our perseption of sharpness. With finer grain it's aliasing less impacts details. With faster films grain comes closer to the size of a detail (the same output resolution).




p.s.



Well, let's say for briefness I've heard other opinions (about final sharpening etc) on this particular matter, but it's out of my scope at the moment.

I can't see the images until I reply with quote. Not sure what you're trying to accomplish with the pixelating stuff.

Food for thought: In simplified terms, when you look at a frame of film illuminated from behind, what you are seeing is high frequency luminance noise combined with the amplitude of said noise that makes up an image that you can recognize. There's that word again: frequency. The source of the noise is the film's silver particles. When you scan film, you are illuminating it from behind and sampling the high frequency luminance noise to create pixels. The high frequency noise amplitude is largely in response to how much light hit the silver particles in the emulsion which is what we perceive and measure as density. As the frequency increases, the contrast or amplitude decreases. This is what you are looking at when you look at a film's MTF chart from a film manufacturer. This is why on some films you don't see an improvement in fine detail or sharpness when scanning at ever higher resolutions. You're already sampling at a higher frequency than what the film's 0% contrast response is. This doesn't mean that the higher scan resolution is pointless, it just means that you have more precision at the higher resolution, which will generally make a better looking version when downscaled, though how much difference will vary depending on the film.

In other films, like Adox CMS 20, the film's frequency response is significantly higher than what pretty much every scanner can do, and therefore, you are getting a scan that is the maximum resolution of your scanner, and not the maximum resolution of the film. Again, it might be a good scan for you because it's all the resolution you want, but by my definition, a scan that doesn't extract all the resolution the film offers (which the whole point of shooting high resolution films is largely to get more resolution) is not a good scan.

Therefore a good scan samples this high frequency noise at a significantly higher sample rate (or DPI) than the film's maximum frequency where the contrast drops to 0. If the film's zone of 100% contrast response exceeds the scanners sample rate, then to get a good scan you need a higher resolution scanner.