Micro Four-Thirds, film comparison

[alanrockwood]

The advantage to shooting film and scanning is simply flexibility. With film you can optically print the important negatives if you want/need maximum quality, and scan the others for less demanding use.

Or one could go one step further in the maximum quality direction and drum-scan the most important negatives. This is not to say that a drum scan should replace the original negative, but only that a drum scan can probably recover more image information from a negative than a normal optical print can recover. From the drum scan one could then recover a hard copy using various technologies.

 

[faberryman]

Or one could go one step further in the maximum quality direction and drum-scan the most important negatives. This is not to say that a drum scan should replace the original negative, but only that a drum scan can probably recover more image information from a negative than a normal optical print can recover. From the drum scan one could then recover a hard copy using various technologies.

Except how many film users actually drum scan their negatives?

 

[alanrockwood]

Except how many film users actually drum scan their negatives?

Very few, though drum scanning is available for those who want to pay for it. For example, I have never had a drum scan made of any of my film shots. (None of them are worth a drum scan.) Of course, drum scans also vary in quality, depending on the make/model of the scanner, the skill of the operator, and how much quality the customer is willing to pay for.

 

[Luckless]

I found the paper and started reading it. I have not made my way through the entire paper yet. The premise of the paper is very interesting and thought provoking. It is also wrong in at least one important way.

The premise of the paper is that a pixel is a point, not a square. The problem is that the concept that a pixel is a point does not map onto any physically realizable image acquisition system because there is no such thing as a point detector in the physical world. All detectors have size.

You have it backwards: The camera sensor does not create or establish the points, and does not need to, it records values for many points by mass sampling.

Light heading to your camera comes from an infinite number of points, but since we don't have a way to record each individual photon on its own, we have to bin them together to establish a new value for a mathematically modeled point instead. The values from the camera sensor are not the exact values given to an individual pixel of the final output - Neighbouring sensor sites will be polled to define the final pixel colour.

Pixels are mathematical models of data, not 'little squares with a colour'.

 

[alanrockwood]

You have it backwards: The camera sensor does not create or establish the points, and does not need to, it records values for many points by mass sampling.

Light heading to your camera comes from an infinite number of points, but since we don't have a way to record each individual photon on its own, we have to bin them together to establish a new value for a mathematically modeled point instead. The values from the camera sensor are not the exact values given to an individual pixel of the final output - Neighbouring sensor sites will be polled to define the final pixel colour.

Pixels are mathematical models of data, not 'little squares with a colour'.

I don't think we are in disagreement, but only using different ways of expressing how reality maps onto a conceptual framework.

A couple of other things may be relevant. One is that even if there was such thing as a perfect point detector, individual points from an image cannot map onto individual points on a detector if one uses a lens as part of the imaging system because all physical lenses are subject to diffraction, which spreads the image onto the detector somewhat. (Again, I don't think we are in disagreement on the science. It's more of a case of how to describe the problem, and also what is meant by the abstraction of the concept of a pixel as a point.) This is separate from whether a lens is imperfect due to aberrations.

The issue of detecting individual photons is not actually a fundamental issue. Single photon counting is an old technique. I have done it myself. However, photon counting is not something that is very relevant to the silicon image detection systems found in current consumer cameras. In my case, I used a single photomultiplier tube in a non-imaging application. There has actually been some imaging using photon counting with photomultiplier tubes, at least I think they used photon counting. I am pretty sure they used arrays of photomultiplier tubes and that the project was called "fly's eye" and was used to detect tracks of light in the atmosphere caused by cosmic rays. I actually once owned a photomultiplier tube that supposedly came from that project.

 

[Ste_S]

Except how many film users actually drum scan their negatives?

Very few - but then that's because a 6MP Noritsu/Frontier scan from a lab is perfectly fine for home ink jet prints up to A4.

 

[blockend]

Film suffers from practical limitations rather than theoretical ones. In the 1970s Eggleston used dye transfer printing, arguably the finest colour print medium that has yet been invented. I read somewhere that a large professionally made dye transfer print cost $500 in the mid-70s. Today you can have a drum scan made of a 35mm negative that will exhaust every grain of information, but very few people do. In the video Jem Southam said the demise of his favourite Kodak paper was one of the reasons he gave up the 10 x 8 camera. The background to film photography is one of gradual contraction and withdrawal. Even getting a quick turnaround C41 isn't easy unless in most areas, you process the film yourself. The Magnum agency printer (who produced Cartier-Bresson, Koudelka and other's work and whose name escapes me for the moment), said black and white printing was so much more difficult since cadmium rich paper finished.

People persevere with film despite the issues. By contrast a straight out the camera jpeg can be printed on fine archival rag paper at the touch of a button. The theoretical boundaries of any medium of format are much larger than their usual application. The difference between theory and practice can be a costly business.

 

[trendland]

Or one could go one step further in the maximum quality direction and drum-scan the most important negatives. This is not to say that a drum scan should replace the original negative, but only that a drum scan can probably recover more image information from a negative than a normal optical print can recover. From the drum scan one could then recover a hard copy using various technologies.

Hi Alan - concerning drum scans (Heidelberg) I got the information from an experience operator
that the best drumscan with highest resolution isn't able to record all information a negative have!
These highly expensive machines are discontiniued from "Heidelberger Druckmaschinen AG"
since decades! A good scanner in probably good used condition (with much operating hours)
will cost between 25000,- up to 35000,- bucks (relative to the service wich is done : replacing parts a.s.o)
A rate of 60% the information within a film negative can be recorded (possible more from best workflow) but there is a real great remaining lost from information!

So I would compare : To print from negative film can't be done without a lost!
But I would expect that a lost (relative to best printing workflow, optical condition = best lens a.s.o.) from printing is the workflow wich serve the best resulting quality also in 2019!

Would you agree?

with regards

 

[trendland]

Film suffers from practical limitations rather than theoretical ones. In the 1970s Eggleston used dye transfer printing, arguably the finest colour print medium that has yet been invented. I read somewhere that a large professionally made dye transfer print cost $500 in the mid-70s. Today you can have a drum scan made of a 35mm negative that will exhaust every grain of information, but very few people do. In the video Jem Southam said the demise of his favourite Kodak paper was one of the reasons he gave up the 10 x 8 camera. The background to film photography is one of gradual contraction and withdrawal. Even getting a quick turnaround C41 isn't easy unless in most areas, you process the film yourself. The Magnum agency printer (who produced Cartier-Bresson, Koudelka and other's work and whose name escapes me for the moment), said black and white printing was so much more difficult since cadmium rich paper finished.

People persevere with film despite the issues. By contrast a straight out the camera jpeg can be printed on fine archival rag paper at the touch of a button. The theoretical boundaries of any medium of format are much larger than their usual application. The difference between theory and practice can be a costly business.

I agree : this discussion is indeed extreme academical! For normal practice (prints of smaller formats in 5x7 inch for example and with higher format) it has no impact for usual practice!

But (for me) there is a strange background from history of digital!
Around 2000 I was told from a head lighting technician (he owned a digital cam) that the quality
of a 800.000 pixel digicam has best resolution up to 30 x 40 cm prints!
And he was impressed from anountsments that there is next a digicam from 1 Megapixel!
He stated to me : 1 000 000 pixel thats "film quality"!
Aha - I didn't care about!
But one a half year later first professional camera operators shot digital with 2 Megapixel cams
and stated : The quality of their 2MP Nikon is much better than with film (and they had experience from own darkroom) !
I realy wonder about their darkroom experience since 20can't!

So there is a serious need to clearify that nonsense in 2019!
(And this nonsence proceed )

To me the 20MP class full frame digital cams reached indeed the resolution of best 35mm film!
But 40MP can't double resolution and 80MP digital isn't able to win in comparison with 4times
value!

So my conclusion is format in comparison of the media!

That should mean : More format = more quality (with film and with digital sensors)

Because the restriction isn't from total value of megapixel - it is from total size of sensor space!

And that issue brings me into trouble with discussions about quality of sensors from smaller size than full frame !

with regards

It is like a discussion about the immense quality and resolution of 110 film format!

 

[trendland]

Hi Alan - concerning drum scans (Heidelberg) I got the information from an experience operator
that the best drumscan with highest resolution isn't able to record all information a negative have!
These highly expensive machines are discontiniued from "Heidelberger Druckmaschinen AG"
since decades! A good scanner in probably good used condition (with much operating hours)
will cost between 25000,- up to 35000,- bucks (relative to the service wich is done : replacing parts a.s.o)
A rate of 60% the information within a film negative can be recorded (possible more from best workflow) but there is a real great remaining lost from information!

So I would compare : To print from negative film can't be done without a lost!
But I would expect that a lost (relative to best printing workflow, optical condition = best lens a.s.o.) from printing is the workflow wich serve the best resulting quality also in 2019!

Would you agree?

with regards

Sorry I must add : from printing with optical darkroom workflow in comparison to print from
drumscan media! = best method should be optical printing and that is also in concern of
a drumscan (80,- bucks per single scan)!

 

[Ces1um]

I'd be interested to hear your thoughts. It's worth mentioning that a printed surface is more forgiving of sharpening than a computer screen. Images that look pin sharp on a monitor can appear to lack sharpness in a print. It's a complicated business!

I haven't forgotten about this- I just haven't had time to get to the computer at night (it's May, time for soooo many outdoor chores).

 

[blockend]

I haven't forgotten about this- I just haven't had time to get to the computer at night (it's May, time for soooo many outdoor chores).

After recent experiments I've uprated my maximum M4/3 print to 24 x 18"!

 

[Ariston]

I am not familiar with 4/3 cameras, but am curious... isn't it more difficult to get a shallow depth of field with the smaller sensors? I know with my RB I get shallow depth with smaller apertures than I can with full frame. It seems that with 4/3 cameras it would have to be a lens with an awfully wide aperture. I'm not sure phones can do it at all, which is why they have to fake it with software.

 

[MattKing]

I am not familiar with 4/3 cameras, but am curious... isn't it more difficult to get a shallow depth of field with the smaller sensors? I know with my RB I get shallow depth with smaller apertures than I can with full frame. It seems that with 4/3 cameras it would have to be a lens with an awfully wide aperture. I'm not sure phones can do it at all, which is why they have to fake it with software.

Micro 4/3 sensor, kit zoom at maximum length of 42mm, f/8, ISO 8000.
Handheld too!

 

[Ariston]

Micro 4/3 sensor, kit zoom at maximum length of 42mm, f/8, ISO 8000.
Handheld too!
View attachment 223571

Nice! Does it work that well at the focal length that is the 50mm equivalent for full frame? ... is that 25mm? I guess that is my real question. At equivalent focal lengths, do they have to have wider apertures for the same bokeh?

Also, is micro 4/3 the same size sensor as a point and shoot? I have an old Panasonic LX7 - is that the same? I've been seeing a lot about micro four thirds, and I am not sure I understand where they fit in the camera ecosystem.

I guess I could just stop being lazy and ask Google.

 

[alanrockwood]

Also, is micro 4/3 the same size sensor as a point and shoot? I have an old Panasonic LX7 - is that the same? I've been seeing a lot about micro four thirds, and I am not sure I understand where they fit in the camera ecosystem.

I guess I could just stop being lazy and ask Google.

The LX7 sensor size is 7.44 x 5.58 mm.
The micro 4/3 sensor size is 17.4 x 13 mm, which is quite a bit bigger.

Just for comparison, 35mm half frame is about 24 x 18 mm, and 35mm full frame is about 36 x 24 mm.

 

[MattKing]

Nice! Does it work that well at the focal length that is the 50mm equivalent for full frame? ... is that 25mm? I guess that is my real question. At equivalent focal lengths, do they have to have wider apertures for the same bokeh?

Also, is micro 4/3 the same size sensor as a point and shoot? I have an old Panasonic LX7 - is that the same? I've been seeing a lot about micro four thirds, and I am not sure I understand where they fit in the camera ecosystem.

I guess I could just stop being lazy and ask Google.

For depth of field questions, the 25mm setting on the zoom will give the same field of view as a 4/3 crop of a 50mm lens on 135 film, but the same depth of field as a 4/3 crop of a 25mm lens on 135 film.
So to get shallower depth of field, use a bigger aperture.
I relatively infrequently use shallow depth of field for effect, so I wasn't able to quickly find an example in my stuff of a digital shot that uses it with a more typical subject to camera distance. If I was doing it for effect, I probably would pop a 50mm f/1.4 OM lens on the camera, via an OM to M4/3 adapter. That would give me the same field of view as a 4/3 crop of a 100mm lens on 135 film, along with the same depth of field as a 4/3 crop of a 100mm lens set to f/2.8 on 135 film
A micro 4/3 sensor (or for that matter, a 4/3 sensor) is almost exactly the same size as a frame of 110 film.

 

[blockend]

I am not familiar with 4/3 cameras, but am curious... isn't it more difficult to get a shallow depth of field with the smaller sensors? I know with my RB I get shallow depth with smaller apertures than I can with full frame. It seems that with 4/3 cameras it would have to be a lens with an awfully wide aperture. I'm not sure phones can do it at all, which is why they have to fake it with software.

M4/3 manufacturers have addressed the portrait lens issue by using wider apertures - f1.2 and below. While such lenses are not cheap, compared to their full frame equivalent they are complete bargains! Panasonic's 50mm f1.4 for their new full frame system costs around £2.3k.

Comparing M4/3 wide aperture lenses with full frame equivalents. Prices listed by UK dealer Wex Photographic:

Standard 25mm lenses M4/3 (50mm full frame equivalent)
Voigtlander 25mm 0.95 Nokton £719.00
Olympus 25mm 1.2 Pro £1199.00
Panasonic 25mm 1.4 Leica Summilux £449.00

Portrait lenses (85-90mm full frame equivalent)
Voigtlander 42.5 0.95 £749.00
Panasonic 42.5 1.2 Leica Nokticon £1149.00

Full Frame mirrorless lenses
Canon RF 50 1.2 £2349.00
Panasonic S 50 1.4 £2299.00
Canon RF 85 1.2 £2799.00

 

[jtk]

As of last week, hand-holdable 100mp is available. Mere 50mp will surely be plenty for those of us willing to compromise