HU: B&W -- film vs. digital

[sircarl]

On Mike Johnston's excellent photoblog yesterday, there was an interesting discussion on the relative merits of film vs. digital for B&W photography, posted by him and contributor Carl Weese. (http://theonlinephotographer.blogspot.com/) It's nice to get reinforcement occasionally that I'm not crazy to be sticking with film!

 

[jrong]

🙂

Hear, hear!

 

[GeneW]

Good link, Carl. And it underscores why I like B&W film!

Gene

 

[Andy K]

There is also an interesting article in this week's 'Amateur Photographer' pitting two digital cameras (Fuji S3 Pro and Nikon D200) against colour negative film in a Nikon FM2. Guess what came out on top?

http://www.amateurphotographer.com/magazine/current.php






🙂

 

[jonasv]

There is also an interesting article in this week's 'Amateur Photographer' pitting two digital cameras (Fuji S3 Pro and Nikon D200) against colour negative film in a Nikon FM2. Guess what came out on top?

http://www.amateurphotographer.com/magazine/current.php

🙂

Only for dynamic range - no surprise.

 

[Michiel]

I mostly shoot B&W, so I'm glad to hear I shouldn't go digital 🙂 (like I can even afford it 😉 )

What I would like though is a digital rangefinder system with change-able (not sure if that's a word) sensors, so you can switch between color or dedicated b&w sensor.

 

[markinlondon]

There is also an interesting article in this week's 'Amateur Photographer' pitting two digital cameras (Fuji S3 Pro and Nikon D200) against colour negative film in a Nikon FM2. Guess what came out on top?

http://www.amateurphotographer.com/magazine/current.php

So that's a Nikon FM2. Add a 50 f1.8 to that and you've got change from 400 quid. How much is a D200? An S3? Good job they didn't use medium format, eh? 🙂

Interestingly (and I suspect because AP likes to hammer its point home) the first letter on the "AP Answers" page addresses the same issue. Quoting freely, Geoffrey Crawley's reply is that with a digital sensor you have one characteristic curve only. With film it changes with the medium.

I don't like film vs digital usually but this is an interesting aspect of the debate that hadn't occured to me before.

Mark

 

[fgianni]

There is also an interesting article in this week's 'Amateur Photographer' pitting two digital cameras (Fuji S3 Pro and Nikon D200) against colour negative film in a Nikon FM2. Guess what came out on top?

http://www.amateurphotographer.com/magazine/current.php

🙂

I think the comparison of the D200 vs S3 pro was quite useful and proved that Fuji's claims about their super CCD's dynamic range are well founded.

Testing the dynamic range of the two against Negative film is a bit silly since everyone knows that both CCD and CMOS sensor thend to behave more like slide film than negative thus having a much narrower dynamic range.

About resolution the reviewer did not do any test, but for what I gathered looking at the samples the D200 is very close to negative film quality, while probably an 1DS MK2 is needed if you want to get close to the capability of good slide film.

For B&W my guess is that we will need 20Mpixel+ to be able to compete with the finest grained monochrome film.

That said, I am quite happy with the resolution provided by my 6MPixel RD-1 since it is more than adequate at 10x8 or A4, and I never print anything bigger.

 

[jaapv]

You should read Fritz Pölking's newest book (sorry-it is only in German). He does not only reason that the printing medium is the real determining factor for needed resolution -i.e. 6-8 MP for 35 mm like prints up to A3 gallery-quality and 10-12 Mp for fine printwork for the press, and anything over 12 Mp is competing with medium format- not with 35 mm format, but the book has high-res illustrations that prove his point.

 

[fgianni]

I mostly shoot B&W, so I'm glad to hear I shouldn't go digital 🙂 (like I can even afford it 😉 )

What I would like though is a digital rangefinder system with change-able (not sure if that's a word) sensors, so you can switch between color or dedicated b&w sensor.

I always thought that myself, simply being able do remove the color filter array from a traditional bayer sensor will transform it in a B&W sensor, but I don't know the details, so it may be technologically impossible.

I remember my old Sony 717 had the possibility of taking pictures with the IR filter removed, but of course the CFA needs to be much closer to the sensor, and has a much more critical alignment.

 

[jaapv]

I don't really think so, the output of the sensor runs through the interpolation software in the camera, you'd have to change that as well. Having said that, there are plenty of examples of removed filter arrays in Canon DSLR's to turn them into IR camera's. You'd have to search FM-forums for that.

 

[JeffGreene]

One of the links for the conversions is here.

 

[fgianni]

I don't really think so, the output of the sensor runs through the interpolation software in the camera.

Well that is probably the easiest to do for a camera manufacturer, simply change the firmware.
I still think the most difficult part is the mechanical assembly for a removable CFA (since of course the user may want to have it back on later, not just remove and throw in the bin).

 

[jaapv]

If you remove the interpolation step from the firmware I wonder if the physical removal of the CFA makes any difference-apart from some optical effects-to the resolution, as all pixels will be treated equally instead of being interpolated to create colour.

 

[Kevin]

You should read Fritz Pölking's newest book (sorry-it is only in German). He does not only reason that the printing medium is the real determining factor for needed resolution -i.e. 6-8 MP for 35 mm like prints up to A3 gallery-quality and 10-12 Mp for fine printwork for the press, and anything over 12 Mp is competing with medium format- not with 35 mm format, but the book has high-res illustrations that prove his point.

A 135mm frame is 0.96 inch by 1.44 inch (24x36mm).
Scan in a frame of Fuji Velvia at 4000dpi (dots/pixels per inch).
The result is an image with pixel dimensions 3840x5760.
That is a 21.6 megapixel image.

So how can anything over 12MP not be competing with 135?

You make me curious.

 

[Bertram2]

On Mike Johnston's excellent photoblog yesterday, there was an interesting discussion on the relative merits of film vs. digital for B&W photography, posted by him and contributor Carl Weese. (http://theonlinephotographer.blogspot.com/) It's nice to get reinforcement occasionally that I'm not crazy to be sticking with film!

Quote:
MIKE JOHNSTON responds: The short answer is, highlights. Digital just doesn't have enough dynamic range in the highlights. This is not so serious with color, where tonal distinction is not as important as color distinction (and because we're used to color slides, which also generally have poor highlight gradation). But in black-and-white it becomes very obvious in many cases—especially to people who are familiar with good craftsmanship in traditional optical-chemical B&W—because the richness and subtlety of the gradations in the highlights just isn't there.
Quote end

Still about a year ago this criticism ( tho based on a simple fact everybody could prove by watching the result) was either neglected, not heard or even denied each time it came up in those silly discussions titled usually "Not the medium counts, it's the pic only" .
It has been something like a Taboo in the times of digital hype to say yes, exactly because the pic counts, therefore digital B&W is NOT acceptable for me, this has been a no-no issue. Watching all this converted stuff people offered as a proof that digital B&W works fine one could make yo think that buying a digital SLR made them blind !? Fact was that this was also the time where we discovered, how many photogs had no clue at all of what a good B&W neg is ! Stunning sometimes and sometimes very very disappointing because people in which you had always trusted began to talk plain BS, as if they had not learned their craft ever.

One photog once tried to explain me that his digital camera has even MORE dynamic range than film because it offers ISO values from 50 to 3200 and I thought o.k, if this is level the discussion runs here I must give up !!
It would be plain stupidity to discuss blown out highlights with somebody who obviously has decided not to see them.

I am really glad that the digital hype now has calmed down enuff that one can publicly say that A is still A and NOT a ! Who thinks he can live with digital B&W shall do so, that is o.k. for me. But thank god it seems the times are over now when you got attacked for saying you are not contended with it and it is reason enuff for you to stay away from digital.

Thanks for pointing out this link, makes me optimistic for the future film niche in a more sober digital world !

bertram

 

[fgianni]

The purpose of removing the colour array is to have real monochrome pixels, so the advantage in resolution is probably similar to the advantage of a foveon sensor compared to a bayer sensor with the same number of pixels, clear as mud as usual 😉

 

[jaapv]

A 135mm frame is 0.96 inch by 1.44 inch (24x36mm).
Scan in a frame of Fuji Velvia at 4000dpi (dots/pixels per inch).
The result is an image with pixel dimensions 3840x5760.
That is a 21.6 megapixel image.

So how can anything over 12MP not be competing with 135?

You make me curious.

The maths are correct, but there is a vast difference between film and sensor, the main in this context being that film, due to its nature, has all kinds of diffraction and diffusion effects inside the film layer itself and inside the anti-halo layer, reducing its actual resolution to 35 % of the theoretical maximum, if that. That means that the most one can hope for is about the equivalent of 8 MP. That is not even thinking about irregular grain size and differences in MTF reproduction because of the different edge behaviour of pixels vs. grain.(and grain aliasing in scanning at 4000 DPI, but that does not go for prints -obviously.)

 

[Socke]

A 135mm frame is 0.96 inch by 1.44 inch (24x36mm).
Scan in a frame of Fuji Velvia at 4000dpi (dots/pixels per inch).
The result is an image with pixel dimensions 3840x5760.
That is a 21.6 megapixel image.

So how can anything over 12MP not be competing with 135?

You make me curious.

We're talking B/W here, Tri-X is not as fine grained as Velvia and so the scanner resoluiton does not reveal any more detail.

And another thing, you'll get some 13 MPixel out of a Canon 5D at ISO 1600, try pushing Velvia 100 four stops 🙂

There is no film with the latitude of traditional B/W combined with the fine grain of Velvia and it grows worth with increasing sensitivity.

The only difference I see in digital vs. film is you make a compromise when you buy the camera. One has to make an informed decision if it is worth the money and does what one needs.

 

[jlw]

I always thought that myself, simply being able do remove the color filter array from a traditional bayer sensor will transform it in a B&W sensor, but I don't know the details, so it may be technologically impossible.

It isn't really necessary to physically remove the filter -- all that's needed is to change the software to read only the luminance value for each photosite on the sensor.

There used to be some digital cameras with a b&w mode that worked this way, but I think it's a feature that has fallen by the wayside. However, I think you can get the same effect by making a grayscale conversion from a raw-format image, which retains the sensor's full original bit depth.

There's been some discussion of the effect of interpolation here, but note that interpolation is used only to derive the COLOR value of each pixel -- if you're making a grayscale conversion, it isn't a factor.