Pages: 1 ... 9 10 [11] 12 13 ... 15   Go Down

Author Topic: Can you see a difference in small prints between an MFDB and a DSLR?  (Read 79687 times)

tho_mas

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1799
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #200 on: March 08, 2014, 09:05:52 am »

What you miss is the Contax Apo Makro Planar is a different lens.  (...) Nice for you to have an excellent general purpose lens
I know it is a different lens.
But I also know that we use pretty similar equipment (Hassy V & Zeiss lenses & P45+ / Contax645 & Zeiss lenses & P45). You have issues with focussing (you tend to shoot with front focus), moire, aliasing, exposure, sharpening, noise and other artifacts. You stop down until your captures get soft. You use DNGs and Lightroom (which definitely shows more aliasing and stronger halos with your P45+ files even without sharpening applied).
I do not have those issues. For me the workflow and especially the results are really satisfying. Back then it took me some time to get familiar with MFD and I also tuned my kit (shims, screens...) to work as supposed to. But finally I made it work for my purposes.
I think our completely different expierence with a very, very similar kit can not be explained by MTF charts and/or the number of lens elements in a certain lens.
For whatever reason... there is something simply not working well in your workflow. But instead of trying to get the best out of your kit you are doing questionable comparisions again and again. You are wasting your time... and your money. Shoot your A99 and be happy.

BTW: the ZCX Apo Makro Planar is not an excellent general purpose lens. It's somewhat weak at very long distances (but that's expected with a Makro). But field of curvature is certainly a non-issue in practice... it's a non-issue with all my ZCX lenses (even the Distagon 3.5/35).
« Last Edit: March 08, 2014, 09:12:47 am by tho_mas »
Logged

jerome_m

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 670
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #201 on: March 08, 2014, 09:18:52 am »

I was actually answering Jeremy. He suggested that the problem was noise in shadow detail, due to noisy sensor. I said I didn't think so and suggested that aliasing may be the cause.

My name is Jérôme, not Jeremy. And I don't think that the effect is caused by colour aliasing or front focus. There is a bit of front focus, but not on the wooden house and it looks muted as well. The pines which are well lit also look sharper than the places which are in the shade, even when the first are further away than the second.

Why don't you use Capture One for your back?
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #202 on: March 08, 2014, 09:53:46 am »

Hi Jerome,

Sorry for the miss-spelling.

I don't really want to discuss about the "Fischleinboden" image. I wanted DoF from the flower in the front to the mountains in the back. it is shot at f/22, while I normally use f/11.

There are a couple of reasons I don't use C1 as those may mainly be ideological.

- I prefer DNG
- I like Lightroom
- I am very much opposed to the idea that if I buy a camera/back I would need to use the manufacturers software

On the practical side

- I have a mighty lock in with 70000 images in Lightroom and I want a single solution
- I am not really happy with some aspects of Capture One, but that may change
- I don't like the user interface

There are a couple of areas where I find C1 superior. Mainly aliasing related but also better handling of CA.

Best regards
Erik



My name is Jérôme, not Jeremy. And I don't think that the effect is caused by colour aliasing or front focus. There is a bit of front focus, but not on the wooden house and it looks muted as well. The pines which are well lit also look sharper than the places which are in the shade, even when the first are further away than the second.

Why don't you use Capture One for your back?
Logged
Erik Kaffehr
 

tho_mas

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1799
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #203 on: March 08, 2014, 10:10:22 am »

- I prefer DNG
- I like Lightroom
- I am very much opposed to the idea that if I buy a camera/back I would need to use the manufacturers software

On the practical side

- I have a mighty lock in with 70000 images in Lightroom and I want a single solution
- I am not really happy with some aspects of Capture One, but that may change
- I don't like the user interface

There are a couple of areas where I find C1 superior. Mainly aliasing related but also better handling of CA.
in other words: you don't care about image quality.
I wonder why you do all these comparisions when IQ is not what you are aiming at... I don't get it.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #204 on: March 08, 2014, 10:11:44 am »

OK, you don't get it.

Best regards
Erik

in other words: you don't care about image quality.
I wonder why you do all these comparisions when IQ is not what you are aiming at... I don't get it.

Logged
Erik Kaffehr
 

synn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1235
    • My fine art portfolio
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #205 on: March 08, 2014, 10:19:49 am »

None of us do.

You do every possible test under the sun for image quality, but you deliberately choose a workflow that delivers inferior quality.

Just sell that back, Erik.you and your bank account will be much happier. Trust me.
Logged
my portfolio: www.sandeepmurali.com

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #206 on: March 08, 2014, 10:25:19 am »

Hi,

Thanks for your suggestion.

Actually, I am testing Capture One and I am also testing RawTherapy.

Best regards
Erik

None of us do.

You do every possible test under the sun for image quality, but you deliberately choose a workflow that delivers inferior quality.

Just sell that back, Erik.you and your bank account will be much happier. Trust me.
Logged
Erik Kaffehr
 

jerome_m

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 670
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #207 on: March 08, 2014, 02:10:28 pm »

He suggested that the problem was noise in shadow detail, due to noisy sensor. I said I didn't think so and suggested that aliasing may be the cause.

I had a look at the "forest" CF044568.IIQ file with raw photo processor. That software may not be the best, but I trust it not to apply any noise reduction when I tell it not to. You may be right that most of the problem comes from aliasing. Noise, while present in the darkest part of the image may be a bit higher than the one from an H3D-31 of which I had some files laying around (and which is about of the same era), but does not seem to be a real problem. Aliasing, on the other hand, is visible on some objects, for example: the rocks.

I would still suggest to try Capture One. I know that it is inconvenient to be forced to use the manufacturer's software: I don't like to use Phocus myself, because I also have my images in something else. But I also know that Phocus gives better results on Hasselblad files than anything else I tried. Therefore I suppose that Capture One would make your Phase One back sing. The manufacturers invest enormous resources  into optimising their software for their backs and third parties may not even know all their little secrets. I believe that using the manufacturer's software is part of the secret of getting the quality one pays for when buying these cameras.
Logged

EricWHiss

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2639
    • Rolleiflex USA
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #208 on: March 08, 2014, 03:25:19 pm »

Hi

Your posting:could be interpreted like you are not well informed on Planars and Makro Planars. What would your point be? If you think that Makro Planars (the MF lenses) have planar field than you are obviously wrong. If you are familiar with the lens, your statement is absolutely pointless.

The papers by Hubert Nasse on Zeiss lenses are an interesting read.

Best regards
Erik

Erik,
Thanks for the laugh!  You are absolutely hilarious!
Logged
Rolleiflex USA

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #209 on: March 08, 2014, 03:57:38 pm »

Hi Jerome,

The latest samples I have posted were processed using Capture One 7.2, the great news is that Bart van der Wolf sent me a preview on the processing he figured for the image, it is astonishing. It is fascinating to see what is possible and than to learn doing myself. A great learning experience! Bart hopefully shares his processing, but I think that C1 v7 processing was an essential first stage of his work and it involves multiple tools. I guess I need to change attitude towards C1 and other tools.

Anyway, thanks for your observations. They are most helpful. Lots of food for thought!

Best regards
Erik

I had a look at the "forest" CF044568.IIQ file with raw photo processor. That software may not be the best, but I trust it not to apply any noise reduction when I tell it not to. You may be right that most of the problem comes from aliasing. Noise, while present in the darkest part of the image may be a bit higher than the one from an H3D-31 of which I had some files laying around (and which is about of the same era), but does not seem to be a real problem. Aliasing, on the other hand, is visible on some objects, for example: the rocks.

I would still suggest to try Capture One. I know that it is inconvenient to be forced to use the manufacturer's software: I don't like to use Phocus myself, because I also have my images in something else. But I also know that Phocus gives better results on Hasselblad files than anything else I tried. Therefore I suppose that Capture One would make your Phase One back sing. The manufacturers invest enormous resources  into optimising their software for their backs and third parties may not even know all their little secrets. I believe that using the manufacturer's software is part of the secret of getting the quality one pays for when buying these cameras.
Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #210 on: March 08, 2014, 04:00:08 pm »

You are most welcome Eric!

Best regards
Erik

Erik,
Thanks for the laugh!  You are absolutely hilarious!
Logged
Erik Kaffehr
 

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8914
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #211 on: March 08, 2014, 04:09:38 pm »

Hi Bart,

The raw images are here: http://echophoto.dnsalias.net/ekr/Articles/PrintSize2/RawImages/

The DNG files actually contain bitwise binary copies of the raw images.

I much appreciate your efforts, and I must say I am very much interested about what you find out.

Hi Erik,

Thanks for making the original files available. I'll do a number of tests on them and I'll share the steps, so anyone interested can follow along. It may take a few posts due to attachment # limitations. My preliminary conclusion is that there is indeed a limit to the amount of detail this lens has to offer, but there is still a lot that can be utilized. The issues do involve aliasing, but I do not think that stopping down further is the best remedy if we can find a way to work around it and keep more MTF response at all lower spatial frequencies.

Okay, here's the first approach I liked better than some others. I started off with Capture One Pro, version 7.15. Since I was not familiar with the original scene, I had to guess to get some parameters in the ball-park, others may prefer a different starting point.

As the attachments show, I only did very minimal adjustments in C1 on image CF044568.IIQ, pulled exposure a tiny bit, and added a small amount of HDR recovery to add a bit of a smoother roll off to the linear tonecurve response. The White balance was a bit of a guess, because as usual under a leaf canopy everything takes on a green cast, and part of the image was in the shadow while part was lit directly.

I added no Clarity yet, but did add a minuscule bit of 'Structure' in an attempt to give a slightly more defined micro contrast basis to work on with sharpening, just enough to give something to work on, but not enough to cause problems. Less is more ...

Because the image was well exposed, and at ISO 50, I reduced all noise reduction to zero. I left the Detail control at its default average setting.

I exported the image with all sharpening disabled. The reason for the minimal adjustments is not that C1 is that inferior in its toolset, but it's just that the tools I am going to use to extract as much detail out of the image data are so much better than anything else has to offer (and I've torture tested many of them).

That was the basis I would work different strategies on with Photoshop as a command center, also because I needed a small trick to help avoid some issues that cannot be done with C1, LR, ACR, or other tools as easily.  

To be continued...

Cheers,
Bart
« Last Edit: March 08, 2014, 04:16:50 pm by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8914
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #212 on: March 09, 2014, 08:28:00 am »

Continued from reply #217

Attached you will find screen grabs at 300% (and therefore pixelated) of the Original unsharpened Raw conversion with CaptureOne Pro, and FocusMagic restored versions. The first screen grab shows that there is very little contrast in the finest details (even in Green), which suggests that the lens is struggling. Diffraction does not help, so it might be possible to improve that micro-detail contrast by using a wider aperture, but then obviously there will be less DOF, unless one resorts to focus stacking.

The first screen grab shows the result after using FocusMagic with a radius setting of 2, on a duplicate layer in Photoshop, because a smaller radius has less effect, and 3 creates ringing artifacts. A good method to achieve the optimum radius setting is to start with an overdone Amount setting of 300%, and then slowly increase the Radius. There will be an increasing effectiveness, until suddenly there will apprear 'fat' edges instead of sharper edges. That's the signal that one needs to step back 1 Radius unit.

Very sharp lenses can require a Radius of 1, and not more, but 1 or 2 are common values for detail in the plane of optimal focus.

The FM sharpened result looks too contrasty, and a bit blocky (part of which is because of the 300% screen zoom), but there are no obvious edge halos. That means that the Radius setting was indeed optimal. However, the image does not look very organic anymore, so it seems to be pushed too far (amount to high). But I did that for a reason, with the following modifications in mind.

The FM deblurring was done on a separate layer, because that allows to use two useful adjustments. The first adjustment is to turn the layer Blend If mode into the following basic settings (which can be tweaked for further refinement):


This Blend-if setting is based on the following assumptions. When edge contrast is very high already, additional sharpening can cause clipping on images that span the full brightness range, and because the contrast is already high, more contrast is not really needed. Therefore we gradually reduce the increased edge contrast effect the closer we get to the extremes of edge contrast. We also use a Luminosity blending mode to reduce the risk of color shifts caused by the boosted contrast. This already takes away some of the initial harshness, and the start and end of the transition zones can be tweaked for even smoother or more sudden transition. The initial 127/128 switch is a good starting point, so I just used that for simplicity. It is also possible to use a Photoshop action to create a duplicate layer and set the layer blending mode and invoke FocusMagic with a single click to play the action.

The second screen grab attachment compares the straight FM deblur, with the Blend-if version, and I also added an 85% opacity to the layer to let some of the original smooth transitions to blend in. That opacity setting is also a matter of taste, and depends on the subject matter, so use you imagination.

The Blend-if layer also tends to reduce noise that may be accentuated by sharpening, although Focus Magic attempts to not boost noise as much as detail.

To be continued...

Cheers,
Bart
« Last Edit: March 09, 2014, 08:30:28 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8914
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #213 on: March 09, 2014, 11:08:01 am »

Continued from reply #218

The deblurred shot, while sharper at the detail level, can still be significantly improved visually. It currently lacks a lot of local contrast, for which I prefer to use Topaz Labs Clarity, because it produces no halos and avoids color shifts that are usually introduced with changes in brightness / contrast. Topaz Clarity produces much higher quality results than LR/ACR or C1 do, that's why I recommend to give it a serious look. It's like Clarity on steroids, and is IMHO an essential tool which seems to lift a veil of drabness from images and let them come to life, as if someone added another lightsource to accentuate surface structure.

I've added screen grabs that demonstrate the positive effect that Clarity also has on perceived sharpness. I also used its HSL controls to adjust saturation a bit, which produces more natural looking colors as I know them from where I live. This is of course also subject to personal taste, so feel free to make other choices. Many of these changes can also be applied only locally, with the built in masking functionality.

Additionally one can adjust how different sizes of detail are rendered. For that I prefer to use Topaz Labs Detail, another must-have tool to enhance our image quality. It is also very useful for Output sharpening, because it also allows to use deconvolution sharpening (might even be used as an alternative if one doesn't have Focus Magic), even after upsampling to native printer output size.

To be continued...

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8914
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #214 on: March 09, 2014, 02:53:22 pm »

Continued from reply #219

And to round off my series of posts, I attached a 'before and after' finishing touch with Topaz Labs Detail.
The 'before' is the Clarified image, and the 'after' is the same after adding some Detail.
 
It's an example of how one could prepare for print on Matte paper, opening up the shadows a bit,
and adding some more definition to pre-compensate for ink diffusion.

Hope you've found it instructive.

Cheers,
Bart
« Last Edit: March 10, 2014, 06:40:39 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

Geoff_C

  • Newbie
  • *
  • Offline Offline
  • Posts: 7
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #215 on: March 09, 2014, 03:04:18 pm »

Hi Everyone,

I don't use a MF system, but have always lusted to try it.  I may treat my self and rent a system for a weekend one day.  One of the things that about MF that intrigues me is that you can get a wide field of view without using "wide angle" lenses.  The look of having a wide field of view without the wide angle look (less compression, distortion etc) really appeals to me.

I didn't see that mentioned in the previous posts. Wouldn't that be evident in small prints?  If not, all my lusting for MF may be in vain.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #216 on: March 09, 2014, 04:25:35 pm »

Hi,

Can you explain a bit more?

Best regards
Eik


Hi Everyone,

I don't use a MF system, but have always lusted to try it.  I may treat my self and rent a system for a weekend one day.  One of the things that about MF that intrigues me is that you can get a wide field of view without using "wide angle" lenses.  The look of having a wide field of view without the wide angle look (less compression, distortion etc) really appeals to me.

I didn't see that mentioned in the previous posts. Wouldn't that be evident in small prints?  If not, all my lusting for MF may be in vain.
« Last Edit: March 09, 2014, 04:28:14 pm by ErikKaffehr »
Logged
Erik Kaffehr
 

Geoff_C

  • Newbie
  • *
  • Offline Offline
  • Posts: 7
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #217 on: March 09, 2014, 04:51:42 pm »

I can try...

I use an APS-C sensor (Fuji S5pro).  If I want to take a picture that encompasses some foreground elements and background elements, I'll be using a wide angle lens. For the example, let's say 14mm.  The foreground elements will be very large relative to the background elements. It will look unlike what our eyes see.

If using a MF system, in order to frame the photo similarly, a 35mm (I think?) lens would be used. With the 35mm lens, the foreground and background elements will be more similar in relative size and the background will appear "closer" than the image taken using the APS-C sensor with 14mm lens.

One of the recent images that I've seen posted in the forums here is SailRonin's picture of a person in a red jacket walking in a snow covered forest "out for a walk".  It was taken with an 80mm lens.  As a result, there is a subtle, slight "telephoto" effect, yet there is a wide field of view (relative to an APS-C or 35mm sensor).  To frame the image similarly with an APS-C camera, one would have to use a 35mm lens.  The image would have a totally different look/feel. 

Does the above help to explain the differences that I would expect to see, even in a small print?  Maybe I over estimate the difference in the look/effect between the different sensor sizes and lens focal lengths.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #218 on: March 09, 2014, 05:06:53 pm »

Hi,

No I don't think so.

The angle of view is proportional to the size of the sensor. So, if you shoot with say 14 mm on APS-C you would need 21 mm on full frame (24x36) and 30 mm on 49x37 mm (the format of my P45+). Proportions and perspective would stay exactly the same. With the larger sensor you would need to stop down more for having the same depth of field.

It is very simple. If you don't move, perspective will not change. Proportions will not change. If you use a different lens, the angle of view would change but proportions would be the same.

Simple test. Shoot a picture from where you stand with a wide angle. Shoot the same subject with a telephoto lens. Crop the wide angle image to telephoto crop. The proportions will be the same, but its is probable that DoF on the tele shot will be shorter.

Best regards
Erik



I can try...

I use an APS-C sensor (Fuji S5pro).  If I want to take a picture that encompasses some foreground elements and background elements, I'll be using a wide angle lens. For the example, let's say 14mm.  The foreground elements will be very large relative to the background elements. It will look unlike what our eyes see.

If using a MF system, in order to frame the photo similarly, a 35mm (I think?) lens would be used. With the 35mm lens, the foreground and background elements will be more similar in relative size and the background will appear "closer" than the image taken using the APS-C sensor with 14mm lens.

One of the recent images that I've seen posted in the forums here is SailRonin's picture of a person in a red jacket walking in a snow covered forest "out for a walk".  It was taken with an 80mm lens.  As a result, there is a subtle, slight "telephoto" effect, yet there is a wide field of view (relative to an APS-C or 35mm sensor).  To frame the image similarly with an APS-C camera, one would have to use a 35mm lens.  The image would have a totally different look/feel. 

Does the above help to explain the differences that I would expect to see, even in a small print?  Maybe I over estimate the difference in the look/effect between the different sensor sizes and lens focal lengths.
Logged
Erik Kaffehr
 

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8914
Re: Can you see a difference in small prints between an MFDB and a DSLR?
« Reply #219 on: March 09, 2014, 05:20:33 pm »

I use an APS-C sensor (Fuji S5pro).  If I want to take a picture that encompasses some foreground elements and background elements, I'll be using a wide angle lens. For the example, let's say 14mm.  The foreground elements will be very large relative to the background elements. It will look unlike what our eyes see.

If using a MF system, in order to frame the photo similarly, a 35mm (I think?) lens would be used. With the 35mm lens, the foreground and background elements will be more similar in relative size and the background will appear "closer" than the image taken using the APS-C sensor with 14mm lens.

Hi Geoff,

As Erik said, the relative sizes of foreground and background object/subjects remains unchanged when the entrance pupil position is unchanged, regardless of focal length differences. The entrance pupil position is the only thing that determines projection perspective.

There is a perceived perspective issue when viewing output from the 'wrong' distance. Wide angle shots will have a stretched wide angle 'look' when viewed from relatively far away, farther away than the ratio of focal length to sensor size would dictate. But that has more to do with viewing distance than focal length in the case of comparing smaller or larger prints. Smaller prints are usually viewed from shorter distances, so you'll get the wide angle 'look' faster than from the same image printed larger and viewed from the same distance.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==
Pages: 1 ... 9 10 [11] 12 13 ... 15   Go Up