Spyro P. wrote:
denoir, to be honest I'm not sure I've understood why you'd bother. From the little I've seen from your photography it looks like you enjoy lots of detail with consistent, realistic colours. Film is going to set you back in all these areas, you can still achieve it but with more pain than digital because film is generally much less massageable after the fact. Some film stocks are great for resqueing highlights but thats about it. Film is not realistic, its an interpretation of reality which you have to accept and love, otherwise you'll be frustrated.
Well, if anything, that sounds like an excellent reason to learn it! I like to explore new fields of photography. A few months ago I got an M9 even though I at the time 1) disliked rangefinders 2) didn't like Leica lens drawing style and 3) I used live view for most of my photos. It took me some time to learn, but now I'm starting to get images that I'm happy with. I made a similar transition before from wild life photography (long lenses, fast AF etc) to landscape photography with manual focus lenses. I've always had some nominal excuse (size of the M9, and the rendering style of the Zeiss SLR manual focus lenses etc). Medium format could be my excuse for film.
In fact, I'd probably like to try large format but I don't know how realistic that is though..
I haven't decided yet though as I'm really not all that excited about the whole development and scanning thing.
Simon Kennedy wrote:
Why?
They are both simply methods of transforming information into different form, film has a 100 plus year head start on developing technology
Well, because a diode is tuned to an exact range of frequencies while film relies on a much more imprecise chemical process. With film you get even variations between the same stock of film. Anyway, precision isn't everything and I'm sure it's all been discussed to death before.
Post processing negative scans is a pain. I believe the scanning software you use is an important factor in getting the color right. I spent 5 days nailing down my work flow and finally decided to use VueScan version 9. It can save the scan as a dng raw file to be used in Photoshop RAW plugin. During preview, the software allows me to correct many colors. I say close enough. I like the fact that the software can scan up to 16 passes on each negative. Time consuming but using multiple passes you can better dynamic range. This is great if you have one shot you really want it to look good and to print it out. Maybe I am only using a lower mid range scanner, Epson V600, or 400 ISO film, off focus area can be grainy. Far distance subjects lack details. Maybe using a lower speed film may create sharper images.
Shooting digital is easy and too clean too. There are scenes being shot digital may not have certain look or feel to them. When shot with film, they become something different. They begin to look and feel interesting or moody.
TooManyShots wrote:
Post processing negative scans is a pain. I believe the scanning software you use is an important factor in getting the color right. I spent 5 days nailing down my work flow and finally decided to use VueScan version 9. It can save the scan as a dng raw file to be used in Photoshop RAW plugin.
You know you can just use tiffs or even jpgs in Camera Raw? No need to save it as a dng...
denoir wrote:
Well, because a diode is tuned to an exact range of frequencies while film relies on a much more imprecise chemical process.
Tuned? CCDs and CMOS sensors use filters in front of each pixel. I'm sure they are not exactly the same as filters you put in front of your lenses, but the concept is the same; they just filter light.
As far as I know, digital sensors have fancier and better filters primarily because they are reusable. With film, you need to be able to make 36 times the area of filters for a roll of 35mm film as you do for a digital sensor, be able to coat that at high speeds, and still keep the costs down to a couple bucks a roll.
Also, large format probably isn't that much more expensive. You can get a decent setup for less than $1k. You seem to have enough money to try lots of different gear out, so 4x5 is totally within your reach. You tend to shoot less too which might help keep film costs down. Also, something like an Epson V700 does comparatively better with 4x5 than it does with MF.
You are much more limited with film choices though.
My advice to you though would be to try a couple different rolls of slide film. Shoot the same scenes with them. Get some decent scans (not the ones you posted). Most importantly, adjust to the fact that its a) not digital and b) not immediate. You might find that you develop some photographic skills that you didn't have before.
Yup. I think there's a setting in Camera Raw too that lets you tell it which tiffs you want to open - all, ones with camera raw settings saved in them, or none. To initially open a tiff, go to the File>Open menu, select your tiff, and change the file type to Camera Raw. I think you can open it directly from Bridge too.
kidtexas wrote:
Tuned? CCDs and CMOS sensors use filters in front of each pixel. I'm sure they are not exactly the same as filters you put in front of your lenses, but the concept is the same; they just filter light.
Yes, you have a CFA (color filter array) in front of the photodiodes that have a bandgap adjusted for the output from the filter. The CFA is in a GRGB arrangement and each filter element lets through three different ranges of wavelengths.
As far as I know, digital sensors have fancier and better filters primarily because they are reusable. With film, you need to be able to make 36 times the area of filters for a roll of 35mm film as you do for a digital sensor, be able to coat that at high speeds, and still keep the costs down to a couple bucks a roll.
Um. No. Film doesn't have pixels (descrete photodiodes) so having a CFA on a film camera would not work. Color film consits of three levels of dyes stacked upon each other (Blue-Green-Red). Each layer contains crystals that are sensitive to a range of wavelengths. The light goes through all three layers each reacting for a part of the spectrum. As chemicals are used, the cutoff between the three spectral ranges are partially overlapping.
Also, large format probably isn't that much more expensive. You can get a decent setup for less than $1k. You seem to have enough money to try lots of different gear out, so 4x5 is totally within your reach. You tend to shoot less too which might help keep film costs down. Also, something like an Epson V700 does comparatively better with 4x5 than it does with MF.
You are much more limited with film choices though.
My primary concern is not so much with the cost, but simply that I don't know where I can get 4x5 film and where I can get it developed.
My advice to you though would be to try a couple different rolls of slide film. Shoot the same scenes with them. Get some decent scans (not the ones you posted). Most importantly, adjust to the fact that its a) not digital and b) not immediate. You might find that you develop some photographic skills that you didn't have before.
I think that's good advice but I would have to figure out where I can get good scans locally. I really don't want to buy a big and bulky scanner that won't be used.
--
Anyway, I shall not tax this thread further with my pondering on the subject. Thanks all for the tips and suggestions, they have been very valuable.
denoir wrote:
Um. No. Film doesn't have pixels (descrete photodiodes) so having a CFA on a film camera would not work. Color film consits of three levels of dyes stacked upon each other (Blue-Green-Red). Each layer contains crystals that are sensitive to a range of wavelengths. The light goes through all three layers each reacting for a part of the spectrum. As chemicals are used, the cutoff between the three spectral ranges are partially overlapping.
I'm quite aware of the technical aspects. Color film uses a combination of color filter layers and sensitization to different wavelengths. You missed my point I think. I didn't mean that color film uses a CFA or has pixels. The main point was that color accuracy is partially compromised in film because you had to do it VERY cheap, because each 'filter' pack was only ever used once. Compare that to a sensor chip that costs $500 to make and might be used for 200k exposures. Joseph Wisniewski explains this aspect in detail.
Getting past the Bayer matrix aspect of most digital cameras, one could conceive of a digital camera that takes 3 separate exposures through different filters, etc. Obviously, systems like this exist and are used. You can also shoot B&W film with 3 different color filters (R, G, B) and reconstruct a color image. I would imaging the color accuracy would be similar if shot through similar filters, compensating for the native spectral sensitivity of the film and sensor.
And if you think digital camera CFAs don't have overlapping spectral ranges, I've got news for you. Here is the spec sheet for the Kodak KAF-18000. It's not the Leica M9 sensor, but it's probably closely related. Spectral sensitivity curves are located on page 14.
denoir wrote:
6x7 minimum and with live view (60 megapixels is pointless if you can't get the focus exactly right). Name one.
I won't argue with you about the live view, but with, for example, the Leica S2, once it is adjusted correctly, the focus works (if you take the time to learn it). Tethering is another possibility.
I cannot imagine what reason you could possibly have for wanting to specify the sensor size? The quality of the output of the current MFD cameras is not good enough for you? Do you want less depth of field than almost none?
There are some Phase One backs that have it, but they have other shortcomings.
No digital medium format cameras have live view, so far. They have tethering, which they call live view...
Nope, I don't think so. I'm waiting for a 6x7 CMOS sensor based medium format camera to come on the market. I'd prefer a reasonable price, but I'd probably be willing to pay as much as the current one costs.
You are willing pay as much for a 6x7" sensor as you are for one half that size? How generous of you
carstenw wrote:
I won't argue with you about the live view, but with, for example, the Leica S2, once it is adjusted correctly, the focus works (if you take the time to learn it). Tethering is another possibility.
I've heard of AF inaccuracies with the S2. Lloyd Chambers did a big test and his conclusion was that with the difficulty of nailing focus the output resolution of the S2 was lower than of the D3x.
I cannot imagine what reason you could possibly have for wanting to specify the sensor size? The quality of the output of the current MFD cameras is not good enough for you? Do you want less depth of field than almost none?
The latter. I couldn't care less about resolution. I'd be perfectly happy with standard DSLR resolution. What I want it for, and why I'm interested in large format photography, is wide aperture landscape photography. I want to be able to have a tree 50 meters away in focus and with a wide background out of focus. I have been doing that so far by using a Gigapan to stitch images together but it's a lot of work. I'd like to be able to achieve it with a single shot.
You are willing pay as much for a 6x7" sensor as you are for one half that size? How generous of you
denoir wrote:
I've heard of AF inaccuracies with the S2. Lloyd Chambers did a big test and his conclusion was that with the difficulty of nailing focus the output resolution of the S2 was lower than of the D3x.
The ones he tested were off. For some reason, the incidence of S2s shipped with the AF off appears to be high. Everyone over on the S2 forums are worried about it and are testing like he does. When someone gets a good one, they are generally quite happy and can live with the way the AF works. It isn't AF like a Nikon, but it is useable and accurate. Marc Williams (fotografz) just got his third S2, and this one is finally accurate as well, it appears.
The latter. I couldn't care less about resolution. I'd be perfectly happy with standard DSLR resolution. What I want it for, and why I'm interested in large format photography, is wide aperture landscape photography. I want to be able to have a tree 50 meters away in focus and with a wide background out of focus. I have been doing that so far by using a Gigapan to stitch images together but it's a lot of work. I'd like to be able to achieve it with a single shot.
Have you read the Joseph Holmes article on MF accuracy? I wouldn't go that direction for this kind of shots. If you can live with a scanning back, look at the Betterlight devices. Otherwise tethering appears to be the only game in town. I wouldn't hold your breath for a CMOS MF sensor with live view. Few people seem to want to go in that direction. The only two large sensor fabs are Kodak and Dalsa, to my knowledge, and neither do anything but CCDs for the kinds of cameras we buy.
Even with good AF, you can't point the camera accurately enough to get the focus accurate on a pixel level. It's problematic enough with a 20 MP DSLR, but with a 60 MP camera such as the H4D-60 it really becomes problematic.
Anyway, the development I'm hoping for is that the prices will drop (Pentax 645D is a good start) and that more amateur photographers will buy them. Then there will be more pressure (and more development resources) to develop MFD cameras that are as good from a usability point of view as a modern DSLR. That would include CMOS & Live View. I know it won't happen any time soon.
As for a scanning back, well, then I might as well continue using the gigapan. It's about as big and weighs as much as an LF camera. And to be honest, I'm starting to feel that I have too much camera gear already.
I'm not too into the measurbating, but apart from very demanding commercial photographers, I can't think of anyone needing focus accuracy on a pixel level from a 60MP file. I see things go to print in international magazines all the time that are ~20MP and not sharp on a pixel level, and they look impeccable.
Old shot on Provia, M645 and 80/1.9. Might have had a grad orange filter looking at it.
