PDA

View Full Version : detail expectation



LotusEsp
21-Sep-2016, 06:38
Hi

I am beginning with LF and trying to work out whether I am expecting too much from my LF camera/films, or if there is a failure in my process, with regards to focus and detail.
I am thinking when shooting test shots that I am not getting the level of detail and sharpness with my negs and scans.

in a scenario like shooting in an average sized back yard, with trees & grass with a 90mm rodenstock lens. Let say the the trees are between 15 and 20 feet away.

In a correctly exposed image on 4x5 Ilford delta 100 b&w negative, how much details in trees & grass should I be seeing in the neg? When I scan in at 2400, and I look at 100% crop - what is that likely to be equivalent to when printed out (i.e. is that my nose up against a 10 foot high print?)

I am wondering if I have too high expectations from my equipment or if I need to revise my methods

Leigh
21-Sep-2016, 07:17
Achievable resolution depends on a whole bunch of factors, including but not limited to:
- subject contrast
- quality of the lens
- shooting aperture
- camera movement (shutter shock, wind, etc)
- film being used
- how the film is developed

The best you can hope to achieve with common materials is about 100 lp/mm for high-contrast subjects.
The resolution drops dramatically as subject contrast goes down.

100 line pairs/mm = 200 pixels/mm. 4" = 100mm, so 20,000 pixels along the short dimension of 4x5.
The 5" dimension would be 25,000 pixels, so the full image is 20,000 * 25,000 = 500 megapixels

Remember, this is absolute max theoretical resolution. I doubt you could achieve it in the general case.

However, scanning at 2400dpi reduces the resolution by a factor of 4.
2400 * 4 = 9600 dots; 2400 * 5 = 12000 dots. 9600 * 12000 = 115 megadots.

To accurately reproduce any analog subject (photograph, drawing, song, etc) you need at least two samples of each original data point*. So for the 200 pixel/mm negative you need 400 samples/mm. The resulting data file would be 500 * 4 = 2,000 Megapixels.

So your 2400dpi scan is actually reducing the resolution by a factor of 8 rather than 4.
Sorry if this is a bit technical, but the subject is technical.

- Leigh

*NB: This is from classical sampling theory and the Nyquist limit.

Randy Moe
21-Sep-2016, 07:28
Go to a museum and look at LF prints.

Download a known good LF scan.

Maybe your eyes are bad, mine are, with lots of floaters.

Then examine everything Leigh has posted.

faberryman
21-Sep-2016, 07:48
Achieving proper focus is essential. Are you using a loupe to verify focus on the groundglass?

Dan Fromm
21-Sep-2016, 08:05
At 15-20 feet @ f/16 your 90 mm lens should cleanly separate blades of grass on film. Forget about what you can see on the GG, resolution will improve somewhat when stopping down a little from wide open and anyway the GG has less resolution than your film can deliver.

Leigh is right. Poor technique kills image quality. Shoot from tripod. Use a cable release. If y'r focus is off a little, no matter because somewhere in the frame the grass will be in focus. And you can't judge what's in a negative from a scan. If you want to know what's in a negative, look at it. You'll quickly find that at 12x everything is mushy. And this is why we use LF. A larger negative allows a larger print before the mush shows.

Corran
21-Sep-2016, 08:21
When I scan in at 2400, and I look at 100% crop - what is that likely to be equivalent to when printed out (i.e. is that my nose up against a 10 foot high print?)

Typical monitor @96 DPI will show at 100% from a 2400 DPI scan a rough approximation of a print 10 feet wide from a 4x5 inch negative (check your settings, if your monitor is at 144 DPI it would be closer to 7 feet). That's a 24x enlargement, which may be a bit much to expect. Plus assuming your scanner is a typical Epson you are pushing the limits of its resolution. Back out to 50%, add a reasonable amount of unsharp masking, and it should look pretty sharp on screen. If not you may have issues at capture, but a lot of people think they are going to get super-sharp scans at 100% from an Epson but that is a tall order. If you want try scanning at 3200 or 4800, adding the unsharp masking to taste at 50% or even 25% view and then resize the file down to 2400, and see if you get a bit more from the scanner.

100lp/mm is high quality small format lens stuff. The best lenses tested HERE (http://www.hevanet.com/cperez/testing.html) show at most ~70 stopped to typical apertures (diffraction limited @ f/22). 75mm Rodenstock (they didn't test the 90) shows 54 in the center/middle, even less on the edge. 50lp/mm x 4x5 dimensions (120mm x 96mm) = 11700x9600 or about 2400 DPI so your scan has to be perfect to get every last drop of resolution, which is going to be hard. For the highest rez scan I do 6000 DPI on a high-end flatbed and resize down to 3000 DPI or appropriate resolution for printing after some slight sharpening.

Drew Wiley
21-Sep-2016, 08:31
With the right film and lens, you should be able to clearly see an aphid on a blade of grass fifteen feet away. But just by the nature of subject matter in the real world, it's pretty rare that everything in the scene can be in focus at the same time, even with the use of judicious view camera movements, which is one of the first things you need to learn. And how detail is itself used in an aspect of composition, involving aesthetic choices and priorities. Perception of detail is also related to contrast. Then you hit all kinds of secondary logistical problems. How flat is your film plane to begin with, including how flat the film lies in the holder? How rigid is your complete camera/tripod system? How much visual information are you potentially losing in the scan and printing protocol afterwards? All this stuff adds up, and you need to isolate each variable in succession to know what is really going on. This starts with a good groundglass loupe, not with a bunch of hypothetical math. Get to first base first.

Ken Lee
21-Sep-2016, 09:40
Mathematical analysis aside, if you already have the equipment, test it out and see if it meets your actual subjective requirements.

You don't have to print an entire 8x10 foot print to see how parts of it will appear at full size: print sample sections instead.

Sharpening can't add visual information, but it can change the impression of visual information. With appropriate sharpening I've been able to make sharp prints at sizes surprisingly larger than I would have expected based solely on theory.

LotusEsp
21-Sep-2016, 09:54
Download a known good LF scan..

thanks
do you have any suggestions?

LotusEsp
21-Sep-2016, 09:55
Thanks all
I'm going to take another shot and print that out at the max I can at home (a3+) and see what it looks like and go from there

Randy Moe
21-Sep-2016, 10:03
thanks
do you have any suggestions?

http://www.loc.gov/pictures/

tgtaylor
21-Sep-2016, 10:07
Go to the last image posted on this thread: http://www.largeformatphotography.info/forum/showthread.php?132537-The-Uranotype-Thread/page3
This is a Uranotype (circa 1850's) the emulsion of which was handcoated with a hake brush on a matte surface 100% cotton rag paper that was sized but untreated with any other chemistry. The finished print was scanned at low resolution on an Epson 3200 scanner going on 15 years of age now. The tripod-mounted 8x10 camera was fitted with 300mm Nikkor-W and it is positioned about 10 or so feet in front of the first row of grave stoned (i.e., Adam Neder). Individual blades of grass are visible through the 3d row and the engraving on the tombstones are clearly legible on the 2d row.

Thomas

Corran
21-Sep-2016, 10:23
any suggestions?

I went through some scans from the last year and found what I think is a pretty sharp example for you. Right click and save >this< (http://www.garrisaudiovisual.com/photosharing/3000dpi-scan.jpg) jpg and you can look at it. That's from a bargain $100 90mm f/6.8 Angulon (so the corners suck) and stopped down to f/32 so fairly handicapped really. It's scanned @ 3000 DPI though on a high-end scanner, so perhaps it evens out.

Perhaps you can post a scan or crop from said scan? Might help.

Ken Lee
21-Sep-2016, 11:04
Here (http://www.kennethleegallery.com/html/tech/8x10Detail.html) is an example of an 8x10 negative scanned on a modest Epson 4990 scanner and sized to 8x10 feet.

Detail alone may not be an adequate criterion: depth of field with a 450mm lens at f/18 (sharpest aperture) is fairly shallow. In large prints we can really notice it, unless we are shooting at infinity. In such cases, detail is diminished.

Here (http://www.kennethleegallery.com/html/tech/RangerDetail.html) is an example of a 5x7 portrait scanned on an Epson V800 and sized similarly: life-size or larger. Depth of field from a 300mm lens at sharpest aperture is also fairly shallow: better than a 450mm lens, but nothing to boast about.

Therefore there's a sweet spot we all must find for ourselves, where portabilty, affordability and image quality converge to our liking.

Kevin Crisp
21-Sep-2016, 11:13
The advice about using a loupe is really sound. Even when my eyes were better up close I could not nail focus on a wide angle lens like that. Use a 4 or 6 power loupe.

Bruce Watson
21-Sep-2016, 12:40
I am beginning with LF and trying to work out whether I am expecting too much from my LF camera/films, or if there is a failure in my process, with regards to focus and detail.
I am thinking when shooting test shots that I am not getting the level of detail and sharpness with my negs and scans.

It's all the things that Leigh said, and more. Primarily though it's your inexperience. We've all been there and gone through it. LF isn't point-'n-shoot, it takes skills, and learning those skills takes time. Few of us have been able to just "step off the bus" and shoot winners. Doesn't work like that. For the record, it took me more than a year and many hundreds of sheets of film to get to the place that I could routinely make exposures that I could print at 10x enlargement if I wanted.

Learning the craft side of LF takes practice. A lot of practice. Shoot and process a lot of film. Put that film on the light table and examine it with a 10+x loupe. Learn from it and try again. Make some prints (real physical prints), put them up on your proofing wall, light them well, and live with them for a while. Learn what works for you and what doesn't. If what you're doing isn't working, change.


In a scenario like shooting in an average sized back yard, with trees & grass with a 90mm rodenstock lens. Let say the the trees are between 15 and 20 feet away.

Oh, ouch. That 90mm lens is half your problem. It's really difficult to learn to use a view camera with a really short lens. Why? A short lens doesn't give you enough image magnification for you to be able to see detail well on the ground glass. Even with your focusing loupe (and you are using a focusing loupe of some kind, yes? If not, start). This leads to focus problems, and in particular difficulty in judging camera movements.

If you're a newbie, you should be learning with a "normal" lens. That is, something close to a 150mm lens in 5x4. Yeah, yeah, I know you never used a normal lens with 35mm/digital. But once again, this isn't 35mm or digital. Learning with a normal lens will shorten your learning curve. Been there, done that, am the (hated) voice of experience.


In a correctly exposed image on 4x5 Ilford delta 100 b&w negative, how much details in trees & grass should I be seeing in the neg? When I scan in at 2400, and I look at 100% crop - what is that likely to be equivalent to when printed out (i.e. is that my nose up against a 10 foot high print?)

First thing -- you can't tell jack from a computer monitor. You can't accurately judge graininess, acutance, tonality, much of anything from a computer monitor. Why? A computer monitor is a light source, where a print is a reflective source. A computer monitor has the wrong dot pitch, usually a lot wrong, and this effects everything from acutance to color, saturation, tonality, etc.

The only way to judge what a print is going to look like is to make one. You don't have to make the entire print (but in the end you'll have to make a full size proof print to judge overall tonality with), you can learn a lot from printing a small section of the full print. But you need the reflection of light instead of the light source, and you need the correct dot pitch.

To answer your question about how much detail you should be seeing, much depends on the quality of your scan. You can get a lot more from a decent drum scan than you can from a consumer level flat bed scan. The consumer flat bed scan is really only good for prints less than 4x enlargement in my book. A professional flat bed will get you to 8x, and a good drum scan can take you to the limits of what the film has to offer.

But if your scene has good visual contrast, with a first class capture (spot on focus, spot on movements, steady tripod, cable release, spot on exposure, spot on film processing, etc.) you should be able to see every blade of grass, every pine needle, the texture of the bark on the trees (if your really short lens leaves you enough image area on the film -- IOW, enough film grain clumps to allow the film to capture that texture).


I am wondering if I have too high expectations from my equipment or if I need to revise my methods

Both. Film (regardless of format) is typically good for 10-12x enlargement. As you exceed this, graininess becomes unavoidably intrusive, acutance decreases, details get mushy, etc. So making a 10' high print from a 4" high negative is... unlikely going to satisfy you.

Also, if you're a newbie, you most likely don't have sufficient control of your craft yet to be able to maximize what your equipment can do. All of us end up revising our methods as we learn workflows that give us the results we are looking for. You will too.

---------------------------------------------------------------------------

All that said, don't get discouraged. It's certainly possible to make razor sharp prints, big prints, from 5x4 film. I've got a few hanging from my walls. If I can do it, anyone can. Learning to use a view camera just isn't that hard, but it does require effort, practice, and a willingness to learn.

LotusEsp
21-Sep-2016, 13:02
Thank you for that Bruce, I appreciate such a comprehensive reply.

Leigh
21-Sep-2016, 13:16
All that said, don't get discouraged.
Very good post, Bruce.

Thank you.

- Leigh

Jim Noel
21-Sep-2016, 13:29
A scan will not show the fine detail that a contact print of the negative will show.

Bruce Watson
21-Sep-2016, 13:41
Thank you for that Bruce, I appreciate such a comprehensive reply.

I got into LF for some of the same reasons as you, including my desire to make a good big print. If I think I can help someone, I try, even though I'm worthless as a rhetorician. If I was good with language, I wouldn't need photography so much!

As you go don't be afraid to ask questions here. There's a number of very generous souls here; a lot of knowledge freely shared on LFP.info.

Leigh
21-Sep-2016, 13:52
I got into LF for some of the same reasons as you, including my desire to make a good big print.
This is funny.

I got into LF because I needed to ensmall the negatives, down to the size of a newspaper column or two.

Shooting basketball available light was challenging.
I needed film speed, and Kodak Royal-X Pan was only available in 4x5.

- Leigh

Jac@stafford.net
21-Sep-2016, 14:07
I got into LF because I needed to ensmall the negatives, down to the size of a newspaper column or two.

Shooting basketball available light was challenging.
I needed film speed, and Kodak Royal-X Pan was only available in 4x5.

Wow! That is an eye-opener! I sucked as a sports photographer, and the worst in the world was basketball in the indoor light of the Sixties. You have my respect.
.

Bruce Watson
21-Sep-2016, 14:29
This is funny.

I got into LF because I needed to ensmall the negatives, down to the size of a newspaper column or two.

Shooting basketball available light was challenging.
I needed film speed, and Kodak Royal-X Pan was only available in 4x5.

- Leigh

Funny indeed -- my first job was sports photographer for the local newspaper. I'm familiar with the trials and tribulations of shooting basketball in a dark high school gym. I've pushed many a roll of Tri-X. Finally gave that up and started lugging around one of the paper's old Singer strobes. Remember those?

Then there was the joy of M3 flashbulbs. I used to use those for night HS football games. Catch plays run to the sidelines.

Ari
21-Sep-2016, 17:53
This is funny.

I got into LF because I needed to ensmall the negatives, down to the size of a newspaper column or two.

Shooting basketball available light was challenging.
I needed film speed, and Kodak Royal-X Pan was only available in 4x5.

- Leigh

I suspect you got into LF so that you could one day use the word "ensmall" in a sentence.

Leigh
21-Sep-2016, 18:41
I suspect you got into LF so that you could one day use the word "ensmall" in a sentence.
Hi Ari,

We're all in LF because we like challenges. The challenges differ. :cool:

- Leigh

Randy Moe
21-Sep-2016, 19:44
I shot 1969 HS B-ball with a $120 Super 8 kit. I have no idea what film speed but it still looks good now even digitized. Outside high jump looked about the same. My brother was a star for 2 years...


Funny indeed -- my first job was sports photographer for the local newspaper. I'm familiar with the trials and tribulations of shooting basketball in a dark high school gym. I've pushed many a roll of Tri-X. Finally gave that up and started lugging around one of the paper's old Singer strobes. Remember those?

Then there was the joy of M3 flashbulbs. I used to use those for night HS football games. Catch plays run to the sidelines.

bloodhoundbob
21-Sep-2016, 20:16
I shot stills for the yearbook with a 3x4 SG and film packs. When I wasn't playing football, I filmed the games from a rickety scaffold with a Bolex 16mm. We would put the film on a bus late Friday night to Chicago for processing and had it back for viewing by Monday afternoon. Also shot basketball games with the Bolex and SG, not at the same time, of course. When I wasn't filming or playing football, I used the SG with those big ol' #22 bulbs, which turned night into day!

Randy
26-Sep-2016, 13:33
When I scan in at 2400, and I look at 100% crop - what is that likely to be equivalent to when printed out (i.e. is that my nose up against a 10 foot high print?)
I am wondering if I have too high expectations from my equipment or if I need to revise my methodsI haven't read every reply in this thread so forgive me if this has already been addressed - I have at times experienced trouble with my flat-bed Epson scanners (4990 and V750) and their crappy plastic film holders - namely holding the film flat and at the correct distance for sharp scans. So I would advise you to look at the negative through a magnifying loupe to see if it is sharp - if it is, then see if you are getting sharp scans. It is because of these scanning problems that I pretty much gave up on digital printing - I contact print only until I can eventually set my enlarger up again.

Jac@stafford.net
27-Sep-2016, 07:57
[...] I've pushed many a roll of Tri-X. Finally gave that up and started lugging around one of the paper's old Singer strobes. Remember those?

Sure! I had two, one slaved. Had both the bare tube and reflectors. The battery was what, 510V? It was a mass of 9V batteries soldered in series.

Desperate for speed, we would use 2475 Recording Film processed in HC-110. My boss would joke that he couldn't tell the basketball from the grain.

One of our staff (who went on to National Geographic) shot down-court with a Nikkor 180mm F/2.4 (at that time a rangefinder lens) modified for the Nikon F. It was an eye-opener!

Bruce Watson
27-Sep-2016, 08:40
Sure! I had two, one slaved. Had both the bare tube and reflectors. The battery was what, 510V? It was a mass of 9V batteries soldered in series.

All I remember is how much the battery pack weighed. And shooting HS football at night I didn't get many chances to set it down. One of the advantages of being young and stupid; it didn't bother me that much, because it let me really throw light.


One of our staff (who went on to National Geographic) shot down-court with a Nikkor 180mm F/2.4 (at that time a rangefinder lens) modified for the Nikon F. It was an eye-opener!

Oh yes. I just found my old 180mm f/2.8 that I used to use on my old F2 for the same reason. It's a beast, but when you need to reach out to the middle of the field it's nice to have. Like a 600mm on 5x4 I guess. I put a number of pix in the paper from that lens. But I never went higher than 360mm with 5x4. Different tools for different jobs.

Ted R
1-Oct-2016, 09:58
Film imaging has always been subject to limitations in the detail that can be recorded. When we start to look at detail very close up everything turns to mush, this is normal. Some of the limitations arise from the following factors. Film has a tendency to buckle slightly and may not be completely flat in the film holder when this happens this causes the image to be slightly out of focus in some places. The lens has a resolution limit due to the physics of optics and the finite wavelength of light. A very good lens focuses the corners and the center at the same film plane, an inferior lens focuses the corners and the center at different distances so on the flat film some parts of the image are blurred due to loss of focus. Cameras vibrate, especially when mounted on tall tripods, the vibration causes motion blur of the focused image, eliminating this requires the careful use of a shutter release cable or air operated release or delayed release. With very fine grain film developed in fine grain developer exposed correctly and with the problems described previously under control then the best negative sharpness is achieved, however even this will turn to mush when examined using powerful magnification.

There is an approximate rule of thumb for enlargement factors with film under favorable conditions a print size about five times the negative size will show good sharpness and contrast. The same negative enlarged ten times may start to show problems with lack of sharpness and film grain. More than ten times demands the ultimate in care and the use of the very best taking lens and the enlarging lens.

Note that viewing distance is a factor. It is very interesting to visit a museum where high quality original paintings are on display. The majority turn to mush when viewed from close up however when viewed from a distance of five or ten feet the effect intended by the painter is visible and the lack of detail is not apparent.

PS
Viewing scans of negatives is a good way to find disappointment, viewing the image on a monitor with the scale of the image size absent (the scan may be easily enlarged on-screen to be many feet across) the mush will be apparent. On the other hand making a print five times enlargement and viewing it from five feet may show an entirely satisfactory situation.

Jac@stafford.net
1-Oct-2016, 10:42
All I remember is how much the battery pack weighed.

Oh! You were referring to the big blue pack! It was like strapping a 6-pack of beer on your shoulder, but had no such utility after the event. :)

Simos Xenakis
17-Oct-2016, 17:07
To answer your question about how much detail you should be seeing, much depends on the quality of your scan. You can get a lot more from a decent drum scan than you can from a consumer level flat bed scan. The consumer flat bed scan is really only good for prints less than 4x enlargement in my book. A professional flat bed will get you to 8x, and a good drum scan can take you to the limits of what the film has to offer.

Bruce, would you mind expanding on what class of scanners you would consider "consumer level flatbed" vs. "professional level flat bed?"

And to clarify, when you say 4x enlargement, you're referring to getting 16 inches of print size out of 4 inches of film length, is that correct? Therefore, a pro-level scanner (capable of 8x enlargement) should be able to produce a satisfactory 32"x40" printed image from a 4x5 scanned neg, right?