A2 size prints and aliasing some experiment

[Erik Kaffehr]

Hi,

Back in 2013 when I started shooting medium format, I found that there was little difference in IQ in my A2-size (16"x23") prints from my 24 MP 24x36 mm gear and my 39 MP 37x49 mm back.

At larger size the larger sensor with the higher resolution was at an advantage.

For some reasons, mostly related to curiosity, I wanted to repeat the experiment. I tried it yesterday:

Here is the protocol:

Cameras compared:

• Sony A7II (24 MP, 24x36, single layer OLP filter). Anno 2015, bought 2015 (1k$US)
• Sony A7rII (42 MP, 24x36, no OLP filter). Anno 2015, bought 2015, (3k$US)
• Sony ArIV (61 MP, 24x36, no OLP filter). Anno 2019, bought 2020, (4k$US)
• P45+ (39 MP, 3749 mm, no OLP filter, 0.55 fill factor). Anno 2007, bought 2013 (10k $US)

Lenses:

• Sony cameras used with the Voigtlander 65 mm, intended aaerture f/5.6.
• P45+ used with Zeiss Planar 100/3.5 at f/5.6 or f/8.

Shooting:

• Overcast sky
• RRS Versa 33 tripod
• Arca Swiss 4D head

Focusing:

• Sony A7 cameras, peaking at magnified live view
• P45+, split image on vertical edge with 90 degree prism and extra 3X magnification using a Zeiss 3X monocular.

Processing:

Color profiles created with Lumariver Profile Designer

• White balance on white road marking
• Exposure adjusted to be the same on the same road marking
• All adjustments except color profile, white balance and exposure are intended to be the same (copied from the first sample to all others)
• A7II (24 MP + OLP filter) has more sharpening, others have Adobe Lightroom 'Light sharpening' preset.
• Central crop exprted at 21x30 cm image size at 180 PPI

Shows the crop used and road marking used to match white balance and exposure.



Resolution in an A2 size print:

In my experience, looking at a 180 PPI representation of print on a 24" 2k monitor is a good 'proxy' for viewing real prints. That would also apply to 48" 4K monitors.

This is comparison for detail. What I would think I would see in an A2 size print. View image at actual pixels below.

Aliasing and color rendition

The 'naked' tree tops in the background have color aliasing, best seen at actual pixels, linked below.

I see significant color aliasing in all four images. These images were cropped at 360 PPI and may be representative of a large print, like A0 size, or the A2 size print viewed with a decent loupe.

Just to say, this is not a Jim Kasson quality experiment. But it reflects the way I work.

Best regards

Erik



--
Erik Kaffehr
Website: http://echophoto.dnsalias.net
Magic uses to disappear in controlled experiments…
Gallery: http://echophoto.smugmug.com
Articles: http://echophoto.dnsalias.net/ekr/index.php/photoarticles

 

[nameless0ne]

I really do not see much a difference between this cameras here... even 24 Mp A7II is still totally fine

 

[Erik Kaffehr]

I really do not see much a difference between this cameras here... even 24 Mp A7II is still totally fine

Yes, one way to see it...

Pretty much what I have seen in my prints back in 2013-2015.

Best regards

Erik

 

[DavidMillier]

Is the colour aliasing in the trees visible in an actual A2 print - I know sometimes printing masks things that can be seen when pixel peeping (noise for instance). How about the aliasing?

 

[Erik Kaffehr]

Is the colour aliasing in the trees visible in an actual A2 print - I know sometimes printing masks things that can be seen when pixel peeping (noise for instance). How about the aliasing?

Well, there is a way to find out, print it.

I may do that tomorrow.

Best regards

Erik

 

[Barry Twycross]

I don't notice much difference between micro-4/3rds and medium format (GFX 50R) at print sizes up to 30", up to about A1.

You're also over sharpening, the halos are distracting on your tower.

 

[Erik Kaffehr]

I don't notice much difference between micro-4/3rds and medium format (GFX 50R) at print sizes up to 30", up to about A1.

Thanks for that comment!

You're also over sharpening, the halos are distracting on your tower.

Thanks for that comment, too.

This setting was used on the A7rII, A7rIV and the P45+

This was used on the A7II.

Printing larger than A2, I would typically set sharpening to zero in LR and use FocusMagic for sharpening.

Best regards

Erik

--
Erik Kaffehr
Website: http://echophoto.dnsalias.net
Magic uses to disappear in controlled experiments…
Gallery: http://echophoto.smugmug.com
Articles: http://echophoto.dnsalias.net/ekr/index.php/photoarticles

 

[Laci55]

Very nice work Erik!° I think the best images is coming from the 61 MO Sony with the less color aliasing,,, I never use 1,0 radius at the sharpening... Mine is 0,4-0,6 always... Have a nice day!

 

[Erik Kaffehr]

Very nice work Erik!° I think the best images is coming from the 61 MO Sony with the less color aliasing,,, I never use 1,0 radius at the sharpening... Mine is 0,4-0,6 always... Have a nice day!

Hi Laci,

Sharpening is tricky. A way to see it is that the radius should be a match for the blur that needs to be restored. So, shooting medium apertures small values may be a good match, stopping do to f/11 - f/16 larger radius would be needed.

With a smaller radius we can use a higher amount.

There is an issue with sharpening at actual pixels and that visual impression of sharpness is dominated by what is known as low to medium frequencies. So sharpening at actual pixels we may end up with non optimal sharpening where it really matters.

I tried to start a thread on optimal sharpening strategies:

A sharpening excercise

Hi, Would be interesting what kind of sharpening different photographers prefer. Here are four sample images, two from an older MFD camera, Hasselblad 555/ELD with a Phase One P45+ back, 37x49 mm and 39 MP. The other two were shot on a Sony A7rII with a Voightländer 65/2.0 APO lens at f/5.6...

www.dpreview.com

But it was immediately 'derailed' as the discussion focused on aliasing instead of sharpening.

On the other hand, aliases are also a part of sharpening.

Back in 2014, I had some input from Bart van der Wolff on sharpening P45+ images processed in Capture One. One of Bart's suggestions was to sharpen quite strongly with FocusMagic but dampen out the artifacts it causes by some advanced blending techniques.

I have learned at lot and I am still learning.

Personally, I feel that looking at proper tests and MTF-plots gives good understanding of the imaging process.

This is some good research from Jim Kasson:

GFX 100 sharpening in Lightroom - the last word

This is one in a series of posts on the Fujifilm GFX 100. You should be able to find all the posts about that camera in the Category List on the right sidebar, below the Articles widget. There’s a drop-down menu there that you can use to get to all the posts in this series;

blog.kasson.com

Looking at MTF curves, it is essentially the left side that matters, according to almost any image research. Sharpening at actual pixels, we focus on the part of the image that causes artifacts.

This example was shot with a very sharp lens (on top) and a no so sharp normal zoom (at bottom). The sharpness in the bottom image was 'boosted' by a bit to aggressive sharpening that causes artifacts of it's own



Best regards

Erik

--
Erik Kaffehr
Website: http://echophoto.dnsalias.net
Magic uses to disappear in controlled experiments…
Gallery: http://echophoto.smugmug.com
Articles: http://echophoto.dnsalias.net/ekr/index.php/photoarticles

 

[marco1974]

Why not at least try to reduce the ugly colour Moire artifacts by using the raw converter's dedicated function?

I use DxO Photolab with my GFX50r and it does a very good job of minimizing Moire when needed, to the point where it is essentially invisible.

M.

 

[Stan Disbrow]

Hi,

Derailed? Here? No! Say it isn't so!

The only rail around here would be a dragster. With slick tires. On ice. Sliding all over the map.

Like right now.

Stan

 

[Erik Kaffehr]

Why not at least try to reduce the ugly colour Moire artifacts by using the raw converter's dedicated function?

I use DxO Photolab with my GFX50r and it does a very good job of minimizing Moire when needed, to the point where it is essentially invisible.

M.

What they do is to blur the color information.

From this article .

Look at the text 'Raymarine' on the radar dome of this boat, it is some kind of red. The center image that has 'moiré reduction' it turned black. The image on the right was shot with a longer lens, that solves the issue in this case.

On the same thread, there was a comment by a DxO-user, his raw processing also turned the red 'Raymarine' text into black.

DxO is not available for Phase One, which is the worst offender regarding color aliasing among my systems.

Best regards

Erik

--
Erik Kaffehr
Website: http://echophoto.dnsalias.net
Magic uses to disappear in controlled experiments…
Gallery: http://echophoto.smugmug.com
Articles: http://echophoto.dnsalias.net/ekr/index.php/photoarticles

 

[marco1974]

Sure, but I much prefer the occasional casualty of an inoffensive black Raymarine (which doesn't detract from the image at all), than having to put up with the horrid colour moire...

Marco

 

[Erik Kaffehr]

Hi,

Derailed? Here? No! Say it isn't so!

The only rail around here would be a dragster. With slick tires. On ice. Sliding all over the map.

Like right now.

Stan

OK!

I didn't want to use the term 'derailed' and just to say, the discussion on aliasing was interesting.

With Lightroom/ACR I often see some artifacts, even with sharpening set to zero, it is a weak 'halo' along high contrast edges that occurs with systems without OLP filter when used with sharp lenses near optimal aperture.

All the images I have chosen for that thread have some issues and I hoped to have some discussion on how to handle that in sharpening.

But, I have noticed that threads often start of in a different direction than intended by the OP, at least as long as the OP happens to be myself.

But, most of the time it is a learning experience.

Best regards

Erik

 

[Erik Kaffehr]

Sure, but I much prefer the occasional casualty of an inoffensive black Raymarine (which doesn't detract from the image at all), than having to put up with the horrid colour moire...

Marco

Personally, I would prefer properly OLP filtered sensor with pixel apertures matching the lenses I want to buy.

Best regards

Erik

 

[Erik Kaffehr]

Hi,

Derailed? Here? No! Say it isn't so!

The only rail around here would be a dragster. With slick tires. On ice. Sliding all over the map.

Like right now.

Stan

BTW,

I was just at the post office and picked up a Voigtlander 35/2 APO Lanthar that in all probability also will generate aliasing artifacts on my 3.8 micron pitch sensor with no OLP filter.

Best regards

Erik

 

[marco1974]

Sure, but I much prefer the occasional casualty of an inoffensive black Raymarine (which doesn't detract from the image at all), than having to put up with the horrid colour moire...

Marco

Personally, I would prefer properly OLP filtered sensor with pixel apertures matching the lenses I want to buy.

Best regards

Erik

Yes, sure. But failing that... I still see the software correction of moire as the lesser evil.

Marco.

 

[Erik Kaffehr]

Sure, but I much prefer the occasional casualty of an inoffensive black Raymarine (which doesn't detract from the image at all), than having to put up with the horrid colour moire...

Marco

Personally, I would prefer properly OLP filtered sensor with pixel apertures matching the lenses I want to buy.

Best regards

Erik

Yes, sure. But failing that... I still see the software correction of moire as the lesser evil.

Marco.

Yes, I would agree.

In the end, I sort of of found that I don't use images with a lot of aliasing. Would I print large, I would probably put effort into reducing aliasing.

Most of the time, it may not be a huge problem.

Best regards

Erik

 

[Laci55]

Hi Erik,

Thank You so much for taking the time to answer to me! As usual You teaching me about this "science of sharpening" a lot more than I thought it was possible! Thank You so much!

It is interesting to know a little more about the "radius" too. My only feedback at the sharpening process is the actual pixel view what is not the best way for any evaluation as I got to know now... I have a very lot to learn indeed but I am not the real pixel peeping person even if I love that aspect of the photography too. I do want to use all these "technicalities" together with the esthetical part of the overall picture quality too...Have a wonderful day!

 

[Stan Disbrow]

Hi,

Raymarine is a maker of marine electronics, in this case a radar unit. Their logo is a very dark red. Almost black and hard to distinguish from black unless in bright light. And even then it helps to have the model name, where they use black, visible which then helps one see that the logo is that dark red and not black.

And this with the Mark-I Eyeball standing on the deck looking at the Radome (the white cover over the actual radar unit inside - think Ray Dome).

I'm not surprised any sensor has trouble with that color.

Stan

Holder of FCC General Radiotelephone Operator's Licence with Radar Endorsement.

Yeah. Something else I work on from time to time.

--
Amateur Photographer
Professional Electronics Development Engineer
Once you start down the DSLR path, forever will it dominate your destiny! Consume
your bank account, it will! Like mine, it did!