Minolta Dimage Scan Elite 5400 mounting direction
Moderators: rjlittlefield, ChrisR, Chris S., Pau
- Enoplometopus
- Posts: 64
- Joined: Wed Mar 31, 2010 1:48 pm
- Location: Germany
- Contact:
Minolta Dimage Scan Elite 5400 mounting direction
Hi all,
I wanted to find out what orientation is best for sharpness. Since i have this lens inside the housing of an old Leitz RMS objective, I can easily reverse it. I put it on the bellows of my stacking setup, put an old, dry and dusty bumble bee in front and run two stacks (100 frames each, zerene) with everything set identical, just the Minolta objective in two positions. The magnification should be somewhere between 1:1 and 2:1.
The resulting picture are always arranged the same way: stack 1 on the left and stack 2 on the right. The diagram shows the position of the Minolta.
The first twin picture shows that the magnification is more or less the same (though the bumble bee is more on the right in picture two). The start and end of the stacking was always done with certain hair of the bumble bee sharp in focus, so it should really be similar.
The second picture pair shows both in a 100 percent crop
The third picture pair shows a 200 percent crop
Picture pair number four shows a 400 percent crop
My impression is that the objective position seen on the right is always a bit sharper. It is only a little bit of difference, but it is this little bit of sharpness that I don’t want to miss – the salt in the soup.
Best regards,
Daniel/Enoplometopus
I wanted to find out what orientation is best for sharpness. Since i have this lens inside the housing of an old Leitz RMS objective, I can easily reverse it. I put it on the bellows of my stacking setup, put an old, dry and dusty bumble bee in front and run two stacks (100 frames each, zerene) with everything set identical, just the Minolta objective in two positions. The magnification should be somewhere between 1:1 and 2:1.
The resulting picture are always arranged the same way: stack 1 on the left and stack 2 on the right. The diagram shows the position of the Minolta.
The first twin picture shows that the magnification is more or less the same (though the bumble bee is more on the right in picture two). The start and end of the stacking was always done with certain hair of the bumble bee sharp in focus, so it should really be similar.
The second picture pair shows both in a 100 percent crop
The third picture pair shows a 200 percent crop
Picture pair number four shows a 400 percent crop
My impression is that the objective position seen on the right is always a bit sharper. It is only a little bit of difference, but it is this little bit of sharpness that I don’t want to miss – the salt in the soup.
Best regards,
Daniel/Enoplometopus
- enricosavazzi
- Posts: 1475
- Joined: Sat Nov 21, 2009 2:41 pm
- Location: Västerås, Sweden
- Contact:
The orientation that you identify as "left" corresponds to the actual orientation of the lens in the scanner (in which it works at a magnification of 1.78x). It would be useful to know with more precision the magnification of your test subject, since 2x is rather close to 1.78x (which we can assume is the design optimum for the lens), while 1x is quite far from the assumed optimum.
The crops you show are from near the center of the image, which is usually the part of the image circle that displays the best image quality and also tends to be more tolerant of changes in magnification. It would perhaps be more informative to compare 1:1 pixel crops near the corners of the frame, since these regions usually degrade faster when a lens is pushed out of its optimal magnification. My tests of this lens at 2x on 42 Mpixel full-frame do show a significant field curvature, for example. Even cropping my test images to 18 Mpixel APS-C format does not eliminate the field curvature.
I think I see a significantly lower contrast and possibly a little veiling flare in the "left" images, compared to the "right" ones. I don't really see much difference in outright resolution between "left" and "right".
The crops you show are from near the center of the image, which is usually the part of the image circle that displays the best image quality and also tends to be more tolerant of changes in magnification. It would perhaps be more informative to compare 1:1 pixel crops near the corners of the frame, since these regions usually degrade faster when a lens is pushed out of its optimal magnification. My tests of this lens at 2x on 42 Mpixel full-frame do show a significant field curvature, for example. Even cropping my test images to 18 Mpixel APS-C format does not eliminate the field curvature.
I think I see a significantly lower contrast and possibly a little veiling flare in the "left" images, compared to the "right" ones. I don't really see much difference in outright resolution between "left" and "right".
--ES
There is another factor to consider when comparing orientations: the thicknesses of the glass that the lens is designed to shoot through, versus the thickness of the filter pack on the camera sensor. Some scanners image the slide directly, and project it through the glass on top of the scanner's sensor. The lens would be designed to shoot through that thickness of glass. This orientation will then have an advantage over the other orientation when used with a camera with similar-thickness glass above its sensor. For a given subject, this may or may not coincide with the correct orientation for the desired magnification.
- Enoplometopus
- Posts: 64
- Joined: Wed Mar 31, 2010 1:48 pm
- Location: Germany
- Contact:
- Enoplometopus
- Posts: 64
- Joined: Wed Mar 31, 2010 1:48 pm
- Location: Germany
- Contact:
- enricosavazzi
- Posts: 1475
- Joined: Sat Nov 21, 2009 2:41 pm
- Location: Västerås, Sweden
- Contact:
It does since we know that the f (focal length) of this lens is 36 mm. To record an image at 2x, one needs to place the rear pupil of the lens at a distance of 3f from the sensor, i.e. 108 mm (an f distance focuses the lens at infinity, while 2f focuses it at 1x). The rear pupil is inside the lens, in this case not far from the rear element. This confirms that you are pretty close to 2x.Enoplometopus wrote:@Enrico: I know next to nothing about this objective. The distance between the last lens of the objective and the sensor is exactly 100 mm. Does this help in calculating the magnification? The camera is Canon EOS 5D Mk IV
--ES
- Enoplometopus
- Posts: 64
- Joined: Wed Mar 31, 2010 1:48 pm
- Location: Germany
- Contact:
I tried again to compare the two positions of the Minolta 5400 objective. Since I have read that the original position in the scanner is what I have named „position 1“ (in my comparison pictures always on the left), but my stacking results yesterday have been better in „position 2“ (in my comparisons always on the right), I got curious.
I checked the scanner again, and my impression is the following: The golden paint symbol on one of the edges seems to mark the light intake of the objective. Consequently the other end is supposed to face the sensor. In the scanner there is a little glass mirror (1); I believe that this mirror reflects the light ray coming from the film, and it directs the light into the objective (2), and the sensor is about 7 cm far away on the right (3).
This is the position that gave me by far better results (see the golden mark on the edge). Since I have the objective in the housing of an old Leitz objective, I can easily chance the direction for comparison.
So what I did was three comparisons with both objective positions, at first around 1:1 (distance of the last lens to sensor 70 mm), then 1,5:1 (distance last lens to sensor 85 mm) and then 2:1 (distance last lens to sensor 100 mm). The magnification factors are not really exact, but you can always see the mm scale and calculate. All magnifications are 400 percent crops.
All comparative pictures show objective position 1 on the left and position 2 on the right.
The object is a beautiful butterfly wing of Urania ripheus.
In objective position 1 the magnification is always a little smaller than in position 2. At 1:1, the center shows a slightly better sharpness in objective position 2, but the corner crop shows a noticeable difference in sharpness.
At 1,5:1, the center also shows a slightly better sharpness in objective position 2, but still a small difference. But the corner crop shows a dramatic difference in sharpness.
At 2:1, the center also has better sharpness in objective position 2, but I would not call that difference dramatic. But the corner crop shows a really dramatic decline in sharpness in objective position 1 on the left.
This comparison shows that the objective can only be used in one direction, with the light entering on the side carrying the golden paint mark.
Daniel Knop/Enoplometopus
I checked the scanner again, and my impression is the following: The golden paint symbol on one of the edges seems to mark the light intake of the objective. Consequently the other end is supposed to face the sensor. In the scanner there is a little glass mirror (1); I believe that this mirror reflects the light ray coming from the film, and it directs the light into the objective (2), and the sensor is about 7 cm far away on the right (3).
This is the position that gave me by far better results (see the golden mark on the edge). Since I have the objective in the housing of an old Leitz objective, I can easily chance the direction for comparison.
So what I did was three comparisons with both objective positions, at first around 1:1 (distance of the last lens to sensor 70 mm), then 1,5:1 (distance last lens to sensor 85 mm) and then 2:1 (distance last lens to sensor 100 mm). The magnification factors are not really exact, but you can always see the mm scale and calculate. All magnifications are 400 percent crops.
All comparative pictures show objective position 1 on the left and position 2 on the right.
The object is a beautiful butterfly wing of Urania ripheus.
In objective position 1 the magnification is always a little smaller than in position 2. At 1:1, the center shows a slightly better sharpness in objective position 2, but the corner crop shows a noticeable difference in sharpness.
At 1,5:1, the center also shows a slightly better sharpness in objective position 2, but still a small difference. But the corner crop shows a dramatic difference in sharpness.
At 2:1, the center also has better sharpness in objective position 2, but I would not call that difference dramatic. But the corner crop shows a really dramatic decline in sharpness in objective position 1 on the left.
This comparison shows that the objective can only be used in one direction, with the light entering on the side carrying the golden paint mark.
Daniel Knop/Enoplometopus
- Enoplometopus
- Posts: 64
- Joined: Wed Mar 31, 2010 1:48 pm
- Location: Germany
- Contact:
Then did you resize the screenshot to fit the forum requirements? Or did you let the forum software do that automatically? Either way, these won't be 400% crops as seen by the viewer. You need to take 400% crops in your editing program and then crop small enough pieces of them that they don't get resized on the path between you and the viewers. Then we will see true 400% crops.
But for your comparative purposes (Condition 1 vs Condition 2, which is better?) this doesn't matter too much. Where it matters is in comparing the Dimage to other lenses tested by other people.
But for your comparative purposes (Condition 1 vs Condition 2, which is better?) this doesn't matter too much. Where it matters is in comparing the Dimage to other lenses tested by other people.
- rjlittlefield
- Site Admin
- Posts: 23605
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
About the scale issue, I do not know what "400%" means in Capture One, as it is being used by Daniel.
Working backward from published specifications and what is shown in the images...
The red boxes enclose an area that is about 45 pixels wide, where the whole frame is about 503 pixels wide. Canon 5D Mark IV is 6720 pixels wide at highest resolution.
So the crops seem to represent about 6720*(45/503) = 601 pixels of sensor width.
But as posted, the crops are only about 503 pixels wide also.
So, it seems that what is being displayed is actually scaled a little smaller than "actual pixels". Certainly this would explain the lack of pixelation noted by Lou.
Daniel, we have a FAQ: What is '100% crop' or 'actual pixels crop' ?, located at http://www.photomacrography.net/forum/v ... hp?t=36168 .
Can you please review the FAQ and see if you can make sense of it for the software that you're using?
--Rik
Working backward from published specifications and what is shown in the images...
The red boxes enclose an area that is about 45 pixels wide, where the whole frame is about 503 pixels wide. Canon 5D Mark IV is 6720 pixels wide at highest resolution.
So the crops seem to represent about 6720*(45/503) = 601 pixels of sensor width.
But as posted, the crops are only about 503 pixels wide also.
So, it seems that what is being displayed is actually scaled a little smaller than "actual pixels". Certainly this would explain the lack of pixelation noted by Lou.
Daniel, we have a FAQ: What is '100% crop' or 'actual pixels crop' ?, located at http://www.photomacrography.net/forum/v ... hp?t=36168 .
Can you please review the FAQ and see if you can make sense of it for the software that you're using?
--Rik
- enricosavazzi
- Posts: 1475
- Joined: Sat Nov 21, 2009 2:41 pm
- Location: Västerås, Sweden
- Contact:
I went back and checked all my notes and sources, and found that I was mistaken about the orientation of the lens in the Scan Elite 5400: the brass-colored marking is oriented toward the film (i.e. away from the sensor) as mounted in the scanner.
I will have to test again at 2x on full-frame with the lens in correct orientation, and this time I expect to get a better flatness of field and peripheral resolution.
This should explain the results of your tests, since you got the best results with the lens in the same orientation as in the scanner. In retrospect, this is exactly what one should expect.
I will have to test again at 2x on full-frame with the lens in correct orientation, and this time I expect to get a better flatness of field and peripheral resolution.
This should explain the results of your tests, since you got the best results with the lens in the same orientation as in the scanner. In retrospect, this is exactly what one should expect.
--ES