Why we have aliasing and does it really matter?

[SeeingSomeThingsMissingMany]

The GFX 50 had undersize micro lenses and that contributed to aliasing on the GFX 50. My guess is that it was not really a GFX 50 feature, more a feature of the sensor. The X1D images showed similar aliasing as the GFX 50S in DPReview's test images, or even worse.

The GFX 100 can also have aliasing, Jim's testing shows that. But it has obviously much less of that issue than the GFX 50 models. Objectively speaking, that is a good thing.

Hmmm this has considered my attention towards the 100S, whereas I was previously looking at the 50R.

With a perfect lens, shooting at f/11, the ideal pixel size would be 2.75 microns. The GFX 110 has 3.8 micron pixels, so it would take around f/11 * 3.8 / 2.75 -> f/15.2.

The corresponding aperture for the GFX 50 would be f/21.6

Useful info.

Yikes! Although I'm not sure that's aliasing.

--
Photography after all is interplay of light alongside perspective.

--
Photography after all is interplay of light alongside perspective.

 

[Erik Kaffehr]

Took these screen shots from the JLC website. I’m really surprised they would allow this atrocious level of aliasing to be published.

The face of the watch is incredibly important and the starburst pattern or other pattern can make the watch desirable or not, so to have aliasing creating a false pattern on this critical part of the watch is bad form for the photographer.

I think this image would have been much better if taken with higher MP or with slight diffraction to eliminate the aliasing. What do you think?

Hi,

What I think we see are not aliasing on the sensor, but downsizing artifacts. Sensor aliases would normally be colorful, essentially because the aliases of the RGB channels don't overlap.

Whenever we downsize the image, we need to apply some kind of low pass filtering to the original image to avoid aliasing in the downsampled image. The bicubic interpolation does add some blur.

The circular patterns on the face of the watch are very difficult to handle in downsampling.

Best regards

Erik

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

 

[JimKasson]

The GFX 50 had undersize micro lenses and that contributed to aliasing on the GFX 50. My guess is that it was not really a GFX 50 feature, more a feature of the sensor. The X1D images showed similar aliasing as the GFX 50S in DPReview's test images, or even worse.

The GFX 100 can also have aliasing, Jim's testing shows that. But it has obviously much less of that issue than the GFX 50 models. Objectively speaking, that is a good thing.

Hmmm this has considered my attention towards the 100S, whereas I was previously looking at the 50R.

With a perfect lens, shooting at f/11, the ideal pixel size would be 2.75 microns. The GFX 110 has 3.8 micron pixels, so it would take around f/11 * 3.8 / 2.75 -> f/15.2.

The corresponding aperture for the GFX 50 would be f/21.6

Useful info.

Yikes! Although I'm not sure that's aliasing.

What else could it be? But the aliasing doesn't look like Bayer aliasing. It probably happened later in the processing chain.

--
https://blog.kasson.com

 

[SeeingSomeThingsMissingMany]

That was my initial reaction it didn't look as Bayer aliasing.

--
Photography after all is interplay of light alongside perspective.

 

[Mads Bjerke]

I have shot a lot of watches with the GFX100 and these photos are definitely showing downsampling rather than aliasing.

 

[JimKasson]

I have shot a lot of watches with the GFX100 and these photos are definitely showing downsampling rather than aliasing.

To be more precise, the aliasing is created by the downsampling, after demosaicing.

 

[Erik Kaffehr]

I have shot a lot of watches with the GFX100 and these photos are definitely showing downsampling rather than aliasing.

To be more precise, the aliasing is created by the downsampling, after demosaicing.

We see aliasing in luminosity. The original image has a lot of high frequency detail, which the downsized image cannot hold. The downsampling does not handle that correctly.

So we get aliases, with the bayer pattern the aliases occur at different locations for different colors, which is quite obvious.

On the demosaiced image the aliases don't yield false color, so it is far less noticable.

I don't think there is a real good solution to the problem, the surface of the watch has a lot of detail, that would not be visible in a small image.

Best regards

Erik

 

[JimKasson]

I have shot a lot of watches with the GFX100 and these photos are definitely showing downsampling rather than aliasing.

To be more precise, the aliasing is created by the downsampling, after demosaicing.

We see aliasing in luminosity. The original image has a lot of high frequency detail, which the downsized image cannot hold. The downsampling does not handle that correctly.

So we get aliases, with the bayer pattern the aliases occur at different locations for different colors, which is quite obvious.

On the demosaiced image the aliases don't yield false color, so it is far less noticable.

I don't think there is a real good solution to the problem,

Pre-resampling filtering.

the surface of the watch has a lot of detail, that would not be visible in a small image.

 

[SeeingSomeThingsMissingMany]

Well I can say this thread your posts Erik has shifted my focus from 50R to 100S. Which means waiting another 7years for it's price to drop to become affordable for me. Which is fine it's already 7years from when Gfx was released, kinda flown by these 7years.

--
Photography after all is interplay of light alongside perspective.

 

[DJF55]

Well I can say this thread your posts Erik has shifted my focus from 50R to 100S. Which means waiting another 7years for it's price to drop to become affordable for me. Which is fine it's already 7years from when Gfx was released, kinda flown by these 7years.

I'd really like to see a GFX 100x image compared to a Fujifilm X 40 mp image resampled to 4k, to check for color or other artifacts.

 

[bobby350z]

Well I can say this thread your posts Erik has shifted my focus from 50R to 100S. Which means waiting another 7years for it's price to drop to become affordable for me. Which is fine it's already 7years from when Gfx was released, kinda flown by these 7years.

There is a diff between real life vs sample charts. Honestly I have both (50s and 100s) and except a few shots I see no difference in aliasing. Now it could be I don't know what to see. But then I saw some moire on my 100s and 110mm f2 shots. Maybe I don't know how to focus. Most of my portrait shots wide open. If I do landscapes, then f8/f11.

 

[JimKasson]

Well I can say this thread your posts Erik has shifted my focus from 50R to 100S. Which means waiting another 7years for it's price to drop to become affordable for me. Which is fine it's already 7years from when Gfx was released, kinda flown by these 7years.

There is a diff between real life vs sample charts. Honestly I have both (50s and 100s) and except a few shots I see no difference in aliasing. Now it could be I don't know what to see. But then I saw some moire on my 100s and 110mm f2 shots. Maybe I don't know how to focus.

The better you focus, the more likely aliasing becomes.

Most of my portrait shots wide open. If I do landscapes, then f8/f11.

 

[Mads Bjerke]

What are you all photographing where aliasing is such a big problem?

I shoot a lot of fabrics and fine detail and I just don't come across these issues in normal life. Moire almost never happens and colour aliasing isn't something I have seen in a way that it has ever been noticeable.

I shoot mostly around f8 with my X2D and XCD 80 and XCD 45P.
Beautiful, crystal clear and clean images.

 

[JimKasson]

What are you all photographing where aliasing is such a big problem?

I shoot a lot of fabrics and fine detail and I just don't come across these issues in normal life. Moire almost never happens and colour aliasing isn't something I have seen in a way that it has ever been noticeable.

I shoot mostly around f8 with my X2D and XCD 80 and XCD 45P.
Beautiful, crystal clear and clean images.

f/8 is enough to greatly reduce aliasing with the X2D. F/11 will virtually eliminate it.

 

[Adam Kingston]

What are you all photographing where aliasing is such a big problem?

I shoot a lot of fabrics and fine detail and I just don't come across these issues in normal life. Moire almost never happens and colour aliasing isn't something I have seen in a way that it has ever been noticeable.

I shoot mostly around f8 with my X2D and XCD 80 and XCD 45P.
Beautiful, crystal clear and clean images.

They're photographing test targets and scenes they've set up specifically to demonstrate an issue that has little to no practical impact on actual photographic practice.

I have thousands of files from MF sensors and had to dump a grand total of one file ever due to aliasing, because it wasn't a good enough photograph to warrant the effort required to save it.

 

[JimKasson]

What are you all photographing where aliasing is such a big problem?

I shoot a lot of fabrics and fine detail and I just don't come across these issues in normal life. Moire almost never happens and colour aliasing isn't something I have seen in a way that it has ever been noticeable.

I shoot mostly around f8 with my X2D and XCD 80 and XCD 45P.
Beautiful, crystal clear and clean images.

They're photographing test targets and scenes they've set up specifically to demonstrate an issue that has little to no practical impact on actual photographic practice.

That's not my experience. I should point out that some people don't seem to see aliasing, and others actually like the effect.

--
https://blog.kasson.com

 

[Greg7579]

What are you all photographing where aliasing is such a big problem?

I shoot a lot of fabrics and fine detail and I just don't come across these issues in normal life. Moire almost never happens and colour aliasing isn't something I have seen in a way that it has ever been noticeable.

I shoot mostly around f8 with my X2D and XCD 80 and XCD 45P.
Beautiful, crystal clear and clean images.

Mads,

I have always thought this aliasing issue was way overplayed here on this forum, but I will admit this....

The first week I had my original GFX 100, I did some test shots on my condo roof with the 50r vs 100 on a tripod (same scene, same lens and equiv F stops) with very detailed busy city scenes.

Jim and Erik pointed out aliasing on the 50r test shots in some distant windows that had cross hatch patterns and close parallel lines on some distant patterns on buildings.

It was definitely there on the 50r test shots and not the 100.

That said, I have been shooting tens of thousands of GFX 100 MP images since then and I just don't see it, and I view my images at full res on good 45K and 6K monitors.

So, we read about it constantly, and I know it is probably there, but I just don't see it.

It seems to always be listed a s negative with the GFX and Hassy MF cameras, but I don't see it if it is.

Yet as much as I argue on this forum about image fidelity and what we see vs don't see, I stay away from aliasing discussions because I believe the science about what a problem it is with GFX gear, but I shy away from discussing it because I know it is there, but I never see it. Well, I guess it is there. Erik and Jim say so a lot and I believe them.


Greg Johnson, San Antonio, Texas

Greg Johnson’s albums | Flickr

www.flickr.com

 

[JimKasson]

What are you all photographing where aliasing is such a big problem?

I shoot a lot of fabrics and fine detail and I just don't come across these issues in normal life. Moire almost never happens and colour aliasing isn't something I have seen in a way that it has ever been noticeable.

I shoot mostly around f8 with my X2D and XCD 80 and XCD 45P.
Beautiful, crystal clear and clean images.

Mads,

I have always thought this aliasing issue was way overplayed here on this forum, but I will admit this....

The first week I had my original GFX 100, I did some test shots on my condo roof with the 50r vs 100 on a tripod (same scene, same lens and equiv F stops) with very detailed busy city scenes.

Jim and Erik pointed out aliasing on the 50r test shots in some distant windows that had cross hatch patterns and close parallel lines on some distant patterns on buildings.

It was definitely there on the 50r test shots and not the 100.

That said, I have been shooting tens of thousands of GFX 100 MP images since then and I just don't see it, and I view my images at full res on good 45K and 6K monitors.

So, we read about it constantly, and I know it is probably there, but I just don't see it.

You could consider yourself lucky for not seeing it.

It seems to always be listed a s negative with the GFX and Hassy MF cameras,

Compared to what? The IQ4 150?

but I don't see it if it is.

Yet as much as I argue on this forum about image fidelity and what we see vs don't see, I stay away from aliasing discussions because I believe the science about what a problem it is with GFX gear, but I shy away from discussing it because I know it is there, but I never see it. Well, I guess it is there. Erik and Jim say so a lot and I believe them.

Greg Johnson, San Antonio, Texas
https://www.flickr.com/photos/139148982@N02/albums