Will more pixels ironically result in less sharp images?

[JerryReed]

Perhaps, I am late coming to this impression, if so, please know that I am a giclée maker whose practice is quite limited to working in a very controlled environment where I am consistently utilizing Live View (LV).

I am wondering if professional photographers whose practices are much broader than my own may helpfully extend my understanding of the implications of a recent project that I undertook for one of my very excellent painter clients.  On this recent occasion the painter had completed a oil painting (12 by 15 inches) that was to be used as an illustration (one-page) for a child's book.  For this reason, and to keep the cost of the reproduction down the artist, whose typical original oils are 3 by 5 feet,  asked that I not use my Sinar 54H with the 16-shot standard practice, but instead do the job in one shot.

I mounted my Hy6 with the Schneider 90 macro on a tripod in the camera room, pushing the the camera stand mounted Sinarcam out of the way.  Lighting was quickly established and I MANUALLY focused the lens using the camera's indicator light that show when the highest point of contrast has been achieved at the single focus spot.  Later on the monitor the image was judged okay and the painter client left.  As I was tidying up the files, I reopened the 200 meg TIF file and saw that at the 100% level the image as not critically sharp.

This experience belatedly brought home to me the situation that others, whose practices are broader than my own, have no doubt considered earlier.  With the expected arrival of the Leica S2 and the likely responses from Canon and Nikon, resulting in these cameras use in a broad rage of situations, where it is unlikely that LV will be practical, how can a means of dependably achieving critical focus be achieved?

Will more pixels ironically result in less sharp images?

Jerry Reed

[ThierryH]

hi Jerry,

A few questions:

- What do you mean with "not critically sharp"?
- Which f-stop did you use?
- I suppose you did take a "Preview" before the actual full-res one-shot and check the focus?: was it sharp?

I presume that your subject/painting was flat, or considered as flat and without depth. Thus, by stopping down the lens at e.g. f 8, the whole should be sharp without any doubt.

- If not, and if you stopped the lens down for the shoot, then it might be vibrations un-sharpness
- If you did use the lens full open, was the full image entirely of the same "not critically sharp"?: if yes, then I suppose the reason has to be vibrations or a misalignement of the sensor/CCD, if not evenly unsharp, then you might not have been a 100% parallel to the subject with the image/lens plane.

Just a few ideas in which direction to look.

Best regards,
Thierry

Perhaps, I am late coming to this impression, if so, please know that I am a giclée maker whose practice is quite limited to working in a very controlled environment where I am consistently utilizing Live View (LV).

I am wondering if professional photographers whose practices are much broader than my own may helpfully extend my understanding of the implications of a recent project that I undertook for one of my very excellent painter clients.  On this recent occasion the painter had completed a oil painting (12 by 15 inches) that was to be used as an illustration (one-page) for a child's book.  For this reason, and to keep the cost of the reproduction down the artist, whose typical original oils are 3 by 5 feet,  asked that I not use my Sinar 54H with the 16-shot standard practice, but instead do the job in one shot.

I mounted my Hy6 with the Schneider 90 macro on a tripod in the camera room, pushing the the camera stand mounted Sinarcam out of the way.  Lighting was quickly established and I MANUALLY focused the lens using the camera's indicator light that show when the highest point of contrast has been achieved at the single focus spot.  Later on the monitor the image was judged okay and the painter client left.  As I was tidying up the files, I reopened the 200 meg TIF file and saw that at the 100% level the image as not critically sharp.

This experience belatedly brought home to me the situation that others, whose practices are broader than my own, have no doubt considered earlier.  With the expected arrival of the Leica S2 and the likely responses from Canon and Nikon, resulting in these cameras use in a broad rage of situations, where it is unlikely that LV will be practical, how can a means of dependably achieving critical focus be achieved?

Will more pixels ironically result in less sharp images?

Jerry Reed

[JerryReed]

So, I am understanding you to are saying that there must be something wrong with my work, and that the question that I posed is groundless, save for the possibility of the mis-alignment of the sensor and the 45 degree finder's representation of the image,  Which raises for me the question, ...if the contrast method of achieving focus is adequate, and the optical path of the reflex viewing is accurate, why LV?

hi Jerry,

A few questions:

- What do you mean with "not critically sharp"?
- Which f-stop did you use?
- I suppose you did take a "Preview" before the actual full-res one-shot and check the focus?: was it sharp?

I presume that your subject/painting was flat, or considered as flat and without depth. Thus, by stopping down the lens at e.g. f 8, the whole should be sharp without any doubt.

- If not, and if you stopped the lens down for the shoot, then it might be vibrations un-sharpness
- If you did use the lens full open, was the full image entirely of the same "not critically sharp"?: if yes, then I suppose the reason has to be vibrations or a misalignement of the sensor/CCD, if not evenly unsharp, then you might not have been a 100% parallel to the subject with the image/lens plane.

Just a few ideas in which direction to look.

Best regards,
Thierry

[ThierryH]

Dear Jerry,

No, I was not meaning to say that you did something wrong, nor that your question is groundless, far from me this idea: I was just wondering or asking what you understand by "not critically sharp".

Let me clarify myself: if it is focused on the spot, then it should be sharp, provided that all other factors influencing the focus/sharpness are not an issue (and there are many, other factors). That counts as much for the "contrast method" as for the LV focusing. Although I must say that the LV method is my prefered, since no optical path interfers: one is focusing on the CCD level, with LV.
This being said, focusing manually for this type of subject, flat, certainly shows immediately any mis-alignment of the sensor, at full open aperture, thus my question if it was stopped down or at full open aperture.

In other words, I do suspect that the sensor is not at the right distance.

Sorry if I have offended you with my answer, that was not my intention.

Best regards,
Thierry

Quote:

"So, I am understanding you to are saying that there must be something wrong with my work, and that the question that I posed is groundless, save for the possibility of the mis-alignment of the sensor and the 45 degree finder's representation of the image, Which raises for me the question, ...if the contrast method of achieving focus is adequate, and the optical path of the reflex viewing is accurate, why LV?"

[JerryReed]

Good, then with the question that I might not know that I need to stop down out of the way, and the recognition that an optical path to the sensor may not accurately represent a point of focus that when viewed at that same point at the pixel level may not be critically sharp, may we return to my question?

Which was,... if the sensor's ability to record is such that with LV I can consistently see the fibers of the canvas when sharply focused, but cannot when focused using an optical reflex system, since the progress of digital photography is toward ever larger sensors containing greater numbers of pixels, which have the unfavorable result to (when fully enlarged) show that the image is not critically focused, am I the only one who views this paradoxical result of progress with irony?

Jerry

[JDG]

The answer is yes and no.  At very wide and very small apertures where diffraction begins to come into play, then smaller pixel backs (higher MP) will begin to appear a bit softer.  At the same time shooting at the sweet spot of the lens, say f/8-f/11 will be much sharper.

[ThierryH]

No, unless you enter the diffraction limits: shoot at the best/adequate aperture, and it will be sharp + in theory showing more details (lens resolution limit).

Best regards,
Thierry

... may we return to my question?

Which was,... if the sensor's ability to record is such that with LV I can consistently see the fibers of the canvas when sharply focused, but cannot when focused using an optical reflex system, since the progress of digital photography is toward ever larger sensors containing greater numbers of pixels, which have the unfavorable result to (when fully enlarged) show that the image is not critically focused, am I the only one who views this paradoxical result of progress with irony?

Jerry

[tho_mas]

I MANUALLY focused the lens using the camera's indicator light that show when the highest point of contrast has been achieved at the single focus spot.

my advice is to use a split image screen - especially when shooting flat subjects. Once adjusted accurate you will always have perfect focus.
Just was shipped 2 new split image screens of Bill Maxwell and these screens are really extremely usefull (no, not better than the Brightscreen split image screens but very, very well made). But you have to adjust the screen exactly to the film plane of your back and once you've done AF might not work perfectly anymore (it doesn't either way IMHO).

[paulmoorestudio]

Good, then with the question that I might not know that I need to stop down out of the way, and the recognition that an optical path to the sensor may not accurately represent a point of focus that when viewed at that same point at the pixel level may not be critically sharp, may we return to my question?

Which was,... if the sensor's ability to record is such that with LV I can consistently see the fibers of the canvas when sharply focused, but cannot when focused using an optical reflex system, since the progress of digital photography is toward ever larger sensors containing greater numbers of pixels, which have the unfavorable result to (when fully enlarged) show that the image is not critically focused, am I the only one who views this paradoxical result of progress with irony?

Jerry

I am not sure this gets to the heart of what you speak jerry, but it is my impression that yes, the focused and out of focus areas of your file at 100% look softer than when viewed either through an optical viewer or on the lcd lv.. and this is assuming the sensor is properly aligned and no lens shift.
 I remember my film days I was always surprised by the films sharpness as compared to the polaroid.. and likewise was always amazed at the increase of apparent sharpness in the final print.. somehow the converting my film to a dot pattern had the effect of sharpening.. but this was most clearly seen when I provided an image with some lens diffraction due to shooting at f45 or smaller with no area of perfect sharpness and voila, when in print, crispy looking.. maybe something similar is going on here.  I share your pain as it is my experience that besides aging eyes the digital sensor is brutal on focus and lack of...and the ability to nail it not a perfect science in my world..does the thickness or physical aspect of the sensor, as compared to film have something to do with this?  whatever the cause this new digital world we are exploring has its cons as well as the pros.

[Dick Roadnight]

Quote:

"So, I am understanding you to are saying that there must be something wrong with my work, and that the question that I posed is groundless, save for the possibility of the mis-alignment of the sensor and the 45 degree finder's representation of the image, Which raises for me the question, ...if the contrast method of achieving focus is adequate, and the optical path of the reflex viewing is accurate, why LV?"

The way I see it is that you told us you have a problem and asked us for our input: I have some very soft Sinar images, and I have the same problem.

If you normally use Live Video, you know what "critically sharp" looks like. With Live video there is no possibility of sensor position focus errors, as you focus on the sensor, nothing else could be as good... and you use LV to test lenses as it illuminates any possibility of focus errors.

Could you not have used the 54h in single shot or 4 shot mode? Don't the movements of a Sinar P2/3 make it easier to get the plane of sharpest focus where you want it... and the non-DSLR Sinars are completely free of mirror shake.

If you view a high-res picture at 100% (on the same monitor, and the same res), you are looking at a smaller proportion of the picture, but with a good AA filter free CCD, it should still look good... (but I think that, with 16 shot, the effective pixels are so small that you are beginning to loose it as the effective pixel size is getting close to the wavelength of light). Did you have more pixels/(square inch of subject) than with the larger pictures - if so, you should be able to see the canvas texture.

If you do not normally use that set-up for flat copy work, then if the focus distance was not bang on, you might not be aware of it.

Is the macro lens you used optimized for 1:1 (like the Schneider Apo Digitar Macro) or is it designed to perform from 1:1 to infinity (like the Zeiss Macro Planar 120)?

As per my thread about mounting Apo-digitars, is is easy to check focus - for short ranges put a meter rule (or yard stick) on a table and focus on the mid point, and see what is sharp.

Getting back to the basics (without accusing you of making basic mistakes)... what can cause unsharp images?
sensor res/pixel count
lens res/mtf (assuming lens up to spec)
faulty/dropped lens - out of position elements.
Focus error
alignment of plane of sharpest focus
camera shake: mirror, cable induced, hand, wind, earthquake, people walking round the studio etc. or (relative) subject movement.
misted /dirty lens or sensor
lens out of it's optimal reproduction ratio
Diffreaction

You can, of course immediately illuminate most of these, but if you got an unsharp image, there must be a reason for it.

[evgeny]

Like you, I compose the image in viewfinder and manually focus the Contax 120mm lens using the camera's indicator light.

I found that the quality of live view in 54H (and similarly LV in 54M, which I already sold) is very owful, completely useless for true judment, especially for critical sharpness.

How do you get that acceptable Live View image from 54H to C-L-E-A-R-L-Y see the area of focus?

Edit to add: the Green indicator in LV is not always right.

[Dick Roadnight]

Like you, I compose the image in viewfinder and manually focus the Contax 120mm lens using the camera's indicator light.

I found that the quality of live view in 54H (and similarly LV in 54M, which I already sold) is very owful, completely useless for true judment, especially for critical sharpness.

How do you get that acceptable Live View image from 54H to C-L-E-A-R-L-Y see the area of focus?

I thought that live view on Sinars was good (Live view has not been implemented on my Hasselblad)

[BernardLanguillier]

Will more pixels ironically result in less sharp images?

My personal experience is that it is impossible to focus reliably my D3x with manual focus, whatever the lens and time I spend composing, accurate AF or live view are the only 2 ways.

Cheers,
Bernard

[evgeny]

I thought that live view on Sinars was good (Live view has not been implemented on my Hasselblad)

Live Video in a $200 camera is Superb comparing to LV in Sinar 54H/54M.

OR I do it wrong with Sinar. Here is how I do LV in Sinar 54H:

1. Light a subject using one Modeling Light in a studio strobe. My Modeling Lights are 300W with output proportional to flash power.
2. Open Shutter in Contax 645 ("B" mode, for unlimited time)
3. Enable Live View in Capture Shop.
4. Enable Focus mode in Live View in Capture Shop.

What I see is a chaos of pixels where I hardly see the Subject, and ever more harder see the magnified part to perform focus.

[JeffKohn]

My personal experience is that it is impossible to focus reliably my D3x with manual focus, whatever the lens and time I spend composing, accurate AF or live view are the only 2 ways.

Cheers,
Bernard

That's my experience as well, both with a D300 and now the D3x. LiveView with a high-quality LCD is the only way to go if you want consistent critical sharpness at these pixel densities. Even for my AF lenses, I prefer the sensor-based contrast AF in LiveView over traditional phase-detect AF, because it tends to be more accurate (and I can place the focus point anywhere in the frame).

[JerryReed]

My post was more philosophical (I saw it as ironic) than was probably useful in this setting.  To try and tidy up the mess that I have made, here goes.  

Usually, my client and all other clients get the FULL BOAT reproduction so that I can later make giclée prints at full size.  That means that I use LV and my Sinarcam2 using Captureshop, not eXposure that I use on the Hy6.  Since the client will be giving the publisher a file to be printed 4 by 5 inches, giving him a 200 meg file at 30% of what I usually charge to meet this needs seemed to be the best way to meet his needs.  I was able to give him the file in 30 minutes including time to burn the disc.  He was happy.

The FULL BOAT process involves using ZigAlign carefully lighting Preview shot, Live View to focus, calibrated focus targets, a light trap to get black point, well you get the idea.  This would have meant that I would be looking for three time the money to do the job.  So, I grabbed my Hy6, knowing that the image would be plenty adequate for his purpose.  As I mentioned it look great on screen at four times his intended resolution.

Now, only later for a reason that maybe I now wish I had left alone, I looked critically at the image as if it were going to be printed 36 by 40 inches instead of 4 by 5 inches.  Which brings me to this point.  

1.Progress in the industry has led to larger sensors with greater numbers of pixels which reveal imperfection in focus abundantly especially were the desired subject is flat not three-dimensional.
2. Better sensors are harder to focus without LV.
3. LV is not something easily implemented in the field (I do not work in a field situation)
4. Hy6 is one very nice camera
5. Schneider lenses are terrific
6. I am a pretty good photographer in my narrow end of the profession

Therefore be it resolved that:

I wish there might be a means like some DSLRs (Canon) to adjust in some manner, for each lens the relationship of the lens to sensor light path so that it could be equal to the viewfinder to sensor path, thus allowing me to reliably focus using the eye piece.  I promise not to ever shoot anyone's art with a single shot ever again.

Jerry

[mattlap2]

Live Video in a $200 camera is Superb comparing to LV in Sinar 54H/54M.

OR I do it wrong with Sinar. Here is how I do LV in Sinar 54H:

1. Light a subject using one Modeling Light in a studio strobe. My Modeling Lights are 300W with output proportional to flash power.
2. Open Shutter in Contax 645 ("B" mode, for unlimited time)
3. Enable Live View in Capture Shop.
4. Enable Focus mode in Live View in Capture Shop.

What I see is a chaos of pixels where I hardly see the Subject, and ever more harder see the magnified part to perform focus.

Evgeny,

Do you have an LC shutter on the front of the lens?   An LC shutter is definitely needed for quality live video on the 54H.

[jimgolden]

i've used LV on a Sinar P3/54 setup a bunch and never had a problem. not sure what the shutter config is, but we have to manually open up to focus and stop down to shoot. and my clients are fairly critical folks, the objects are enlarged quite a bit and look great...

[ThierryH]

Without LC shutter with the 54H or 54M/MC, the quality of the live image is certainly not what it is with (an LC shutter): forget about focusing accurately without LC shutter.

Best regards,
Thierry

I found that the quality of live view in 54H (and similarly LV in 54M, which I already sold) is very owful, completely useless for true judment, especially for critical sharpness.

How do you get that acceptable Live View image from 54H to C-L-E-A-R-L-Y see the area of focus?

[ThierryH]

It is, indeed, see my previous post: you need an LC shutter, with 54 series = Kodak 22 MPx sensor.

Best regards,
Thierry

I thought that live view on Sinars was good (Live view has not been implemented on my Hasselblad)