• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

Handheld, long focal, comparison between 5D Mk2 and Pentax 645D

Yes, I encounter it everytime it often as it is just above the high-pass filter that I use a lot, but admit I have never used the custom one before. I now need to understand how it actually works instead of just guessing. I'm assuming that when you do your sharpening thread that you might explain more; I hope so anyways. And Nicolas, Sorry to have hijacked your thread. :p
 
Yes, I encounter it everytime it often as it is just above the high-pass filter that I use a lot, but admit I have never used the custom one before. I now need to understand how it actually works instead of just guessing. I'm assuming that when you do your sharpening thread that you might explain more; I hope so anyways. And Nicolas, Sorry to have hijacked your thread. :p

Hi Maggie,

No hijacking occurred, just a discussion about the practical differences between two camera platforms, and some differences in how the resulting images look after proper sharpening.

I have to add a remark about the Custom filter though. I have found that while it functions correctly with the settings that I suggested for LR downsampled images, it does require that the image mode is in 16-bit/channel (which is a given for me, but others may underestimate its importance). When the image mode is 8-b/ch, the image also changes to an overall darker image, which the given filter settings shouldn't do. In 16-b/ch mode there is no overall brightness reduction, things function as intended.

So, as with all image manipulation, keep your images in 16-b/ch mode (or convert to it before editing) until the final output file is created. At that point it is usually no big issue any more to go down to 8-b/ch mode.

Cheers,
Bart

P.S. I routinely add the 'convert to 16-b/ch mode' to the start of all my Photoshop actions, like here:
LR4_PS_SharpenAction.png
 

nicolas claris

OPF Co-founder/Administrator
Hi Maggie,

No hijacking occurred, just a discussion about the practical differences between two camera platforms, and some differences in how the resulting images look after proper sharpening.

Of course!

@ Bart
Thanks again.
And Yes, 16 bit is the way I use for long time, with the exception of the images posted here, as I wanted to save time and disk space… A 16 bit file from the 645D is 242 Mb…
 

nicolas claris

OPF Co-founder/Administrator
…We shouldn't underestimate that, and I don't. Nicolas comes from a camera system designed for action, with a sensor that uses AA-filters and Microlenses, to a camera system without AA-lenses wich behaves a bit differently. On top of that, he is forced by the lack of 645D support in his favorite C1 Pro Raw converter to start using Lightroom/ACR as Raw converter, and that Raw converter just switched to a whole new Process version that requires unlearning old habits and learning new ones for everybody.

Let's call it a challenge.
Agreed!

Take your time, take a shot of my resolution target at f/8, and send me the file. Ideally, shots at all apertures that can be used should be taken and analysed but that would take me some time. However, we can start with one (a target shot at f/8, ISO 100 or 200)
Sorry Bart, I must have missed something, but where can I get the resolution target file?
 

Joachim Bolte

New member
Thank you, that sure gives a better comparisson between the two. In these pictures you can clearly see that the Pentax's sensor is less prone to noice, and that the bokeh effect is nice.
One negative point I see, is that (maybe due to lack of an AA filter, maybe because of the higher resolution) the fine contrasty lines in the Pentax's picture are prone to anti-aliassing and Moiré-effect. You can see it in the cables hanging over the railing on the second picture.
 

Asher Kelman

OPF Owner/Editor-in-Chief
...........
One negative point I see, is that (maybe due to lack of an AA filter, maybe because of the higher resolution) the fine contrasty lines in the Pentax's picture are prone to anti-aliassing and Moiré-effect. You can see it in the cables hanging over the railing on the second picture.


Nicholas,

What of that Moiré? Can you fix it? Presumably you'd do that on a layer and not have the filter alter anything else, but what software would you personally choose? Capture One?

I must admit, I was not as careful as Joachim as I was emotionally drawn to the greater presence in the Pentax images, which of course can be explained in part just by the use of longer focal lengths at any given camera to subject distance than the 35mm 5DII format.

It would haver concern me to have Moiré on a model's dress, (at those times when present) and having a fine repeating texture. Thank goodness it is not ever likely to occur on a thigh or face! :)

Asher
 
One negative point I see, is that (maybe due to lack of an AA filter, maybe because of the higher resolution) the fine contrasty lines in the Pentax's picture are prone to anti-aliassing and Moiré-effect. You can see it in the cables hanging over the railing on the second picture.

Hi Joachim,

Camera sensors without an AA-filter will be more sensitive to producing aliasing artifacts than sensors with such a filter. But even the latter can show some aliasing artifacts, although they're less likely to occur, or less obvious.

However, it is important to separate a few effects because some can be remedied easier than others.

  • Moiré is caused when 2 regular patterns with a different spatial frequency are mixed. This can happen when the subject has a highly periodic and regular structure, and it is mixed with the highly regular pattern of the sensor array. Slight changes in orientation/position can make a difference in how the interference patterns look. A slight defocus, or narrow aperture diffraction, will blur (reduce the amplitude) particularly in the finer detail pattern and can reduce the modulation of the mixed pattern enough to make the pattern unnoticeable.
  • Aliasing is caused by subjects with spatial frequencies that exceed (are smaller than) the highest sampling frequency, they are imaged as larger features (aliases) than the real thing. They need not be in a repetitive pattern, although they may be easier to spot when they are (because they will deviate from the expected pattern). A typical example is a brick road or wall, shot at an angle. As the detail gets smaller with distance, the smaller detail will grow in apparent size due to aliasing where we expect it to just become smaller and ultimately unresolved. There is no real remedy after the shot is taken, it should be prevented (AA-filter, defocus, diffraction) by eliminating the higher frequencies that the sensor cannot resolve.
    The aliasing can also manifest itself as jagged edges or lines (stairstepping or jaggies), because a very sharp edge or line is either registering on one pixel, or the next one, without smooth transition.
  • False color artifacts are caused by the different sampling densities of Red and Blue, versus Green, in our Bayer CFA filtered sensors. This means that all of the above phenomenae occur at different detail levels for Red/Blue versus Green. Upon demosaicing this will confuse the Raw converter, which has no knowledge of the real scene. This can lead to the brightly colored waves that can be seen on the mostly vertical cable strands. Fortunately, the better Raw converters are relatively good at suppressing those bright colors (most noticeable against neutral colored surfaces) by local desaturation.

All of those side effects are less prominently visible when a proper AA-filter (Optical Low-Pass Filter, or OLPF) is used, but such filters are very expensive at large sizes, and for non-retrofocus lens designs the effect varies visibly with the angle of the light, IOW the image corners will be affected differently compared to the center. This is why AA-filters are usually omitted from Medium Format (say 48x36mm, or 44x33 mm as in the 645D) sensors, or relatively large sensors in range finder cameras where the lenses have very short exit-pupil to sensor distances (and thus more oblique edge and corner rays).

What of that Moiré? Can you fix it? Presumably you'd do that on a layer and not have the filter alter anything else, but what software would you personally choose? Capture One?

The Pentax 645D's PEF Raw format is not handled by Capture One, so Nicolas had to abandon his favorite Raw converter. Lightroom 4 or Photoshop's ACR 7 offer a reasonable alternative, although each converter has its particular strong and weak points. Fortunately, false color artifacts can be relatively successfully removed. The other optical phenomenae are inherent to a non-OLP fitered sensor array, but due to the large number of sensels, the output magnification is smaller, hence some of the artifacts less noticeable. We're pixel peeping here at an image that would measure 73x54 inches (185x138 cm) at an average display resolution of 100 PPI ...

It would haver concern me to have Moiré on a model's dress, (at those times when present) and having a fine repeating texture. Thank goodness it is not ever likely to occur on a thigh or face! :)

Yes, fashion/fabrics are a difficult subject to shoot with an non-AA-filtered sensor. Often these subjects are shot tethered, so the issues could be spotted during the shooting session on a larger display, and by repositioning or refocusing or narrower DOF or diffraction blur one can try to subdue the detrimental effects. The rest is handled by the post-production retouching crew.

Cheers,
Bart
 
Top