[architorture]

I saw the thread by Zplus and it spurred me to get this posted.

About a month ago I spent an afternoon/evening making a comparison for myself out of curiosity, and also b/c i thought others might be interested as well. I will post small-size pictures here, with links to full-size jpegs (saved @11 in PS).

Conslusions
I am disappointed/surprised. BASED ON MY RESULTS IN THIS SPECIFIC TEST WITH THESE LENSES, MY SCANNER, MY LAB, ETC. in terms of resolution 35mm sucks, k10d and 645 are very close - at 10mp size k10d looks sharper, but 645 shows a hint of more beyond, while k10d looks digital - you see all the detail but that's all there is. the 645 at full 3200dpi scan has more detail depth than k10d, but is not as crisp. Beforehand I was confident the 645 would be much better than the APS-C.

Certainly, this is alot of magnification when the files are viewed @100% on screen, but the digital handles it much better than the 35mm.

k10d is very clean! 35mm slide film is very clean! 645slide film would be outstanding, while print film is rather grainy.

colors - not gonna go there in this test.

contrast range - k10d wins. i did not try at all to get more out of the film - and i could, but with RAW the k10d makes it child's play to get so much detail from dark to light.

pleasure of using - 645 wins hands down for the great view through the viewfinder and just as easy as 35mm slr. digital wins for quick satisfaction.

best lenses: choose your own
on bronica i can see the 75 is sharper than the 150 by a little.

Please scroll down past these screeenshots for all of the info and the real pictures

k10d w/ sigma 28mm @100% vs. ETRSi w/ 75mm lens @ 50%



k10d w/ sigma 28mm @200% vs. ETRSi w/ 75mm lens @ 100%


k10d w/ p-fa 35mm @ 100% vs k1000 w/ p-k 55mm @100%




Test Details
Cameras:
Pentax k10d
Pentax k1000
Bronica ETRSi

Lenses:
Pentax DA12-24 F4 ED AL (IF) @18mm & 24mm
Sigma Mini-Wide II 28mm f2.8
Pentax FA-35mm f2.0 AL
Pentax K-55mm f1.8
Pentax SMC Takumar 135mm f2.5 (6/6 version)
Bronica Zenzanon PE 75mm f2.8
Bronica Zenzanon PE 150mm f3.5

Image Capture:
Pentax .PEF --> Adobe ACR 5.x (latest) with no sharpening/NR --> jpeg & 11 in PS
35mm Fuji Astia ISO100 --> Epson v700 & Silverfast se 8bit 3200dpi with no sharpening/NR --> jpeg & 11 in PS
645 Fuji NPL 160 (expired tungsten film, always refrigerated) --> same as above (resized to 2592px image height by PS)
Of course, for any real usage these would all need sharpening, especially the scans (flatbed scans always need a lot of sharpening.)

Technique:
Light via 1 Speedlight through umbrella at camera left ~ 45deg up/left. power adjusted to attempt to get even exposure as aperture was changed.
Tripod, cable release, MLU (except k1000), Self-timer (k10d only).
Test at every full-stop from f11 and open, + max. ap. if not a whole stop.
Focusing: microprism collar with k1000, split-image screen with ETRSi, focus bracketing with k10d (about 30 frames through 2ft+ range).

I tried to match image framing, but I realize I did it wrong - I matched with the front of the table, which is ~1ft in front of the subject, oops. This only matters for comparisons btwn focal lengths, btwn formats it should not matter imo.

Images
for comparison I have:
APS-C---35mm------645
18---------28----------x
24---------35----------x
28----------x----------75
35---------55---------75
55----------x---------150
x----------135----------x
135---------x-----------x
*multiply 645 lenses by .6, APS-C by 1.5 for equiv. 35mm fov

Shutter speed was at x-sync for Pentax cameras, 1/500 for ETRSi (but the light is all flash anyway so shutter speed is really 1/1000-1/10000?). EXIF should be intact for the digital, and I have placed a note identifying what the shot is in each picture anyway (though a couple times i forgot to change it - I know 100% the captions i put here are correct).

Pictures Below:

[architorture]

without further ado:


k10d, 18mm, f5.6

original

k1000, 28mm f8

original


k10d, 24mm, f8

original

k1000, 35mm, f11

original


k10d, 28mm, f8

original


ETRSi, 75mm, f11

original


k10d, 35mm, f8

original

k1000, 55mm, f8

original


k10d, 55mm, f8

original


ETRSI, 150mm, f11

original


k1000, 135mm, f11

original


k10d, 135mm, f8

original

[architorture]

additional conclusion:
how amazing would it be to have med.fmt. digital!
for now i will stick with stitiching images together, i guess.

[irchan]

...in terms of resolution 35mm sucks

Your scanner sucks
in any terms.

have some nikon scan will ya..

Z1+FA 20-35mm f/4



crop


crop again

[architorture]

i would prefer to leave attacks out of this you could have said there are better scanners... which i do know. i may go use an imacon tomorrow to settle this once and for all...

yes, my $400 scanner is obviously not as good as a 1-2 thousand dollar scanner.

but that should make no difference in this case, since both negatives were scanned on the same scanner with the same settings?

I would expect the 35mm to have more detail at the same scan resolution, since 35mm lenses are reputed to resolve better than those for larger formats.

should not scanning both at 3200dpi and viewing at 100% eliminate the advantage the larger format has when printing or viewing at the same output dimensions?

[architorture]

maybe this is more useful / less unnerving?

I have applied my usual sharpening routine for scanned images (plus a little extra) on the 35mm scan. the digital still has no sharpening. it still shows the same thing though - the 35mm does not resolve as much as the 10mp digital APS-c, imo (with my equipment, of course).

ps sorry, but i swithced their positions for this one - the 35mm is on the left here.

[irchan]

with all do respect sir, your scanner sucks.

given this kinda test, even ixus would be way better than flatbed scanned 35mm film

[irchan]

but that should make no difference in this case, since both negatives were scanned on the same scanner with the same settings?

yup, no difference, both are sucks

if using decent scanner, 35mm at least can make 20mp and 120mm maybe 90mp, plus the grain offcourse.

so 35mm should be better than aps-c, and 120mm would be waaay better than full frame digital 35mm but digital cameras will always 'cleaner' though.

[pingflood]

yup, no difference, both are sucks

if using decent scanner, 35mm at least can make 20mp and 120mm maybe 90mp, plus the grain offcourse.

so 35mm should be better than aps-c, and 120mm would be waaay better than full frame digital 35mm but digital cameras will always 'cleaner' though.

Not even going to go into the debate of how many megapixels are in a film frame (it's been debated to death many a time), but what matters is the PRINT. Take prints from a digital file and blow up to say 20x30 -- then do the same with a film scan (at these sizes doing optical enlargements are a disadvantage, you'd get better results from a good scan). Now go compare the two, give it the Pepsi challenge or what have you, and see what comes out. You'd be surprised how few MP it takes to equal or better a 35mm print. (I'd put my 1Ds2 files up against medium format film for prints any day, and its 16MP are a far cry from the 90 "needed" to compete even for large, large prints...)

Then again, how silly to compare real world results to pixel peeping...

(edit: though if you want to pixel/grain peep, check out http://www.luminous-landscape.com/reviews/shootout.shtml for some drum scans of 6x7 vs 1Ds 11 mp files.)

Also, if you REALLY want to get into detail, Norman Koren has a good series of articles on digital vs film. WELL worth a read and a MUST read before you start debating pixels vs film "pixels".
http://www.normankoren.com/Tutorials/MTF7.html

[irchan]

^^ya ya ya.. digital is good, great, superior, whatever man.. i'm just saying that op really need at least a proper scanning before making any "comparison" thread

[pingflood]

^^ya ya ya.. digital is good, great, superior, whatever man.. i'm just saying that op really need at least a proper scanning before making any "comparison" thread

I shoot film as well including 4x5 so kindly kiss my butt. I was commenting on your statement as to how many MP you'd get out of a scan. Read the links I posted and see if you still cling to that.

[irchan]

samples images from your link





i like the 2nd better, but that just me.. like i said before, digital is cleaner and even better. Look at my image posted before, it's grainy.. but 100% crop from it still.. big

so kindly please stick your 1ds mkII up your ass, sir.. don't make this thread go any longer, and please.. please.. tell the op to get more proper scanning

[irchan]

shouldn't mention "mp" in the first place..

[pingflood]

samples images from your link





i like the 2nd better, but that just me.. like i said before, digital is cleaner and even better. Look at my image posted before, it's grainy.. but 100% crop from it still.. big

so kindly please stick your 1ds mkII up your ass, sir.. don't make this thread go any longer, and please.. please.. tell the op to get more proper scanning

Keep in mind that the DSLR sample there is from a camera released in _2002_. I just thought it was a good example of how well even a moderate resolution DSLR performs compared to MF. Now pit a more recent one (1DsII or III or 5D II) against MF scans and you can imagine what happens... and mind you, that was a 6x7 camera, 6x4.5 is not going to do as well.

And yes, the sample you posted shows that there's plenty of detail in a film scan, but it doesn't translate directly to usable detail (the last crop sample is pretty 'mushy' so the detail that's present is fairly useless due to the grain / low contrast). That's why I pointed out that one needs to look at the print to get a true comparison. Sure, you might be able to eek out X lp/mm resolution with Tech Pan in a medium format, but in real world shooting with real world film and scenes the story's fairly different.