Why Do My Film Scans Suck?

Maybe the camera shop deliberately put a cheaper lense on the camera (used) to save money and re-sell the more expensive lense. But still, I would think a "cheap" lense would be capable of an 8 megapixel film scan. Looking through the viewfinder, everything looks sharp if it's in focus. Are you sure the lense model (still a Canon SLR) is a limiting factor more than the film? I would expect blurriness instead of pixelation.

The film in that shot was extra-color Kodak Elite Chrome, which is a consumer version of Kodachrome I believe. The vivid Kodachrome (another roll) also had some grain, but I did let it sit around at room temperature for a couple weeks before processing. I'm going to experiment with ISO 50 Fuji Velvia (recently discontinued) that I've been keeping in the fridge and it was kept in the fridge at the camera shop. I figure with a heavy tripod, timer, mirror-lock, and auto-exposure bracketing (slide film), I should be able to get a high-res shot of at least 8 megapixels, maybe 24 megapixels non-interpolated w/ 12 bits/channel color depth...

While skill does play a factor, the optical dpi is the limitation of the scanner's potential performance. I have asked various photo labs various questions about their scanners, and half of them try to weasel around my technical questions and avoid giving me a straight answer. I would like to find somebody in Palm Springs or Los Angeles who can say, "Yes, we can do a 4,400 optical dpi drum scan." Any ideas?

The lens is pretty soft. If you were handholding I would not expect to see anything great.

Also - if you're shooting something like kodak gold, you might be limited by the grain. I think I can see it in the scan but I might be wrong.

The lens is pretty bad also.

That cheap lens if used wide open won't deliver sharp pictures at 8 megapixels.

Thank you so much, Doc! Is there any online reference for that? Manufacturers don't like to publish limitations of their equipment.

manufactors don't publish on purpose... most lenses are sharpest between f/8 and f/11 or so, some @ f/16

as for film - elite chrome is a consumer version, but it is not a consumer version of kodachrome, which is the most neutral balanced positive, period. at its truest nature, it is a b&w film and cannot be processed in e-6 chemicals.

as for performance... stop worring about dpi. as i stated before, scanning is as much an art as photography. having someone who is skilled at it makes all the difference.

that said, a drum scan doesn't mean its the best. we tested the imacon $20k model) and it didn't hold up to our Noritsu production scanners.

id look into a better lens, and try shooting some better films. pick a film based on what you want.. example, velvia doesn't do the best things for skin tones, so don't use it to shoot people

On the Canon web site, they show a different lense, but my camera is the previous model:
http://www.usa.canon.c​om …tegoryid=138&mo​delid=9831

So it's sounding like the camera shop sold me a high-end camera with a cheap lense. It would have been ethical to inform me that it was not the original lense for the camera.

It's a night photo of traffic, so f/11 would work well and possibly give enough clarity, you think? I want to re-take this with a slightly higher perspective and more traffic:
http://www.palm-springs-photography.com/palmsp​ringsnight.html
That photo is good enough for a postcard, but not much larger. I used a Canon Powershot A70 (3 megapixels).

The thing about a drum scan is that it's hard to screw up, right? While your lab might do really good scans, some other lab might not be as skilled.

funnypicmaker wrote in post #2137136
So it's sounding like the camera shop sold me a high-end camera with a cheap lense. It would have been ethical to inform me that it was not the original lense for the camera.

The camera shop didn't switch lenses from "good" to "bad"

The lens on the site is same level (just as crappy). But it satisfied most consumers and people shooting film don't really know the difference in terms of sharpness, contrast and quality... especially since you have to count the variables like processing, films... etc.

With digital it's different. You get to see the results of your lens directly on the monitor and suddently investing a couple grand in lenses suddently makes sense.

If you want good quality, get a 50/1.8 lens. Stopped down, it's stellar from 2.8 to around f/11 giving best quality at around f/5.6.

Thanks, Doc. I didn't realize that when you buy a high-end camera you also have to buy a high-end lense separately. For this shot, I'll give it a shot and see what it can do with f/11 and ISO 50, heavy tripod, timer, mirror lock, AEB...

Here's a great article somebody emailed me:
http://www.luminous-landscape.com/reviews/​shootout.shtml

That's a comparison of 6cm x 7cm film camera to 11 MP digital SLR. I think it's reasonable to say that 35mm can't compare to 8 MP digital SLR, but it is encouraging that the Fuji Velvia and Provia will have finer resolution and grain than consumer films and vivid saturation films. It seems that the only reason to prefer film these days is if you have a large format camera and can't afford to spend several thousand dollars on a digital equivalent. The only other reason I can think of is to use a negative film to reduce blown highlights and increase contrast of dark objects when there are huge differences in lighting. But there may be a digital equivalent to that.

in my film days I had nikon coolscan ls2000 - this was about as good as you could get without getting a drum scanner. The scans out of it were not great until I purchased software from Silversoft - they offer a free trial - if you try it out, it may make a difference. It went from okay to oh my god and 1 scan was all I needed to shell out the $400 it was then.........

funnypicmaker wrote in post #2137849
Here's a great article somebody emailed me:
http://www.luminous-landscape.com/reviews/​shootout.shtml

That's a comparison of 6cm x 7cm film camera to 11 MP digital SLR. I think it's reasonable to say that 35mm can't compare to 8 MP digital SLR, but it is encouraging that the Fuji Velvia and Provia will have finer resolution and grain than consumer films and vivid saturation films. It seems that the only reason to prefer film these days is if you have a large format camera and can't afford to spend several thousand dollars on a digital equivalent. The only other reason I can think of is to use a negative film to reduce blown highlights and increase contrast of dark objects when there are huge differences in lighting. But there may be a digital equivalent to that.

thats a fantastic read, thanks for posting the link !

Generally, a prime lens will be better than a zoom. Maybe you should think about a macro lens if you plan to do a lot of that work.

Something was done drastically wrong.

IF, this is a big IF, a drum scanner was used and the film was properly mounted some settings on the scanning software were set wrong.
I've seen the type of thing you are describing when the USM settings are cranked or the scan spot size is not applied correctly on our Tango.
Definitely ask them what scanner they are using. If they say Imacon run away FAST!

I'd personally take the Luminous Landscape article with a good pinch of salt....
Here's why.

Following the 6x7 vs. 1Ds logic.....
Shoot two "identical" images one with the 1Ds and one with the 67.
Scan the film twice; once maxed out at say 5000 spi which gives you a 11,600x13,750 pixel file, and one at 1811± spi to equal the native resolution of a 1Ds(4992x3328 pixels) which gives you 4992x4202 pixels.
Process the 1Ds RAW file to produce two files, one at native resolution(4992x3328 pixels) and up-rez the other to where the long edge equals 13,750 pixels( 2.75x factor), 13,750x9167 pixels(comparable to the maxed out 67).

Prints will tell the true tale here if they are all printed on the same printer at the same resolution, 300ppi.
1. Take the maxed out 67 and downsample to the native resolution of the 1Ds to make like sized prints of course you will have film grain. You are scanning at 5000, capturing grain and downsampling - of course the grain will be evident. This is how the Luminous Landscape article test was performed.
2. Compare the 1811spi 67 scan to the native 1Ds shot. This is the apples-to-apples comparison. My bet is on the scan. 1811 spi is well below the threshold of film grain or grain clumps so grain will not be a factor what so ever.
3. Compare the upsampled 1Ds shot to the maxed out 67. This is the rebuttal to the original test in the Luminous Landscape article. Once again my bet is on the scan.

On the surface the test seems legitimate, but when you actually pencil things out it is not all that it seems to be.....

Wouldn't it depend on the downsampling method? Anybody in their right mind would use bicubic (possibly after filtering out the grain) or similar method. Do you think the author used bilinear or nearest neighbor, thus introducing grainyness into the lower res photo on purpose? I do think zooming in on a white object is misleading because now we can't see the pixels in the digital camera.