E-M1X Ghost images?

Mack

Mu-43 All-Pro
Joined
Jan 14, 2018
Messages
1,936
I was having issues trying different editors with an image from of the 29mm f/0.8 Voigtlander Noct lens taken with the E-M1X. It seemed to be making a double-image or an offset ghost image which seemed odd.

I was thinking "This thing needs to go into service to check out if it has some funky optics issue," but I cannot find out who services Voigtlander in the USA. Nothing in their minimal literature either. Website is void of any Service locations too. :shakehead:

So I put the thing on the E-M1 Mark II and got no ghost image! Tried with a 12-100mm f/4 too on the same ruler on the ColorChecker Digital SG card and no ghost there either. It will only "ghost" on the E-M1X body. :confused-53:

I also went through all the Stabilization settings (5 I think, counting Off) on the E-M1X thinking one might be better, but same issue happens with all. A faint offset ghost appears on all, and of the three editors I tried none would address that (i.e. ON1 Photo RAW 2021, DxO PhotoLab 4, and Exposure X7). Aside, Exposure X7 has the better control of CA red/blue fringing than the other two by far, imho.

Below are some over-sharpened images out of Olympus Workspace that shows the E-M1 Mark II with the 29mm f/0.8 Voigtlander and the 12-100mm f/4 that show no ghost. But the E-M1X appears soft (Same settings in Workspace) and has the ghost look to it. All shot on tripod, fwiw.

1.jpg
Subscribe to see EXIF info for this image (if available)

2.jpg
Subscribe to see EXIF info for this image (if available)



The E-M1X. Look at letters, lines, and numerals which appear soft and with an offset ghost (Ghost is lower and on the left side.):
3.jpg
Subscribe to see EXIF info for this image (if available)


What the ghost looks like to me:
4.jpg
Subscribe to see EXIF info for this image (if available)


The only thing that I can think of that might cause this is the Stabilization maybe shifting the image during exposure, but it was also tried Off. Another might be a secondary reflection caused by the AA filter in front of the sensor, but nothing shows with the E-M1 Mark II. I've seen secondary ghosts made off filters mounted in front of lenses and why I dislike them (also for introducing side flare), but none was used here.

Puzzling.
 

Mack

Mu-43 All-Pro
Joined
Jan 14, 2018
Messages
1,936
Are these crops? If so, how much?

Looks more like slight CA to me.

Your 12-100 shot appears to be a little out of focus.

What ISO were you using?
Thanks John.

Yes, they are all crops out of Olympus Workspace at 300% to show the offset ghost of the E-M1X. Its a small section of the card, but it works well to show things like CA and sharpness, but the secondary image overlay and the softness is what bothers me.

Yes, there is a blue cast in the E-M1X image, but even when corrected it shows the secondary ghost slightly down on the left as a blur. No amount of CA adjustment in any editor got rid of the offset ghost, but they can get rid of the blue CA cast. I fought this all AM while coming up with a DCP color correction preset (DCP) and wondered why I was having sharpness issues, but that ghost is what I found. The E-M1 Mark II is a snap to fix and sharpen/CA fix so that ruled out the lens which I was suspect of.

I suspect the 29mm Voigtlander might be sharper than the 12-100mm too. The 12-100mm was in AF mode and the Voitglander was MF, but neither shows the ghost and offset that I can find easy on the E-M1X. Almost like it pixel-shifts down and to the left during the exposure somehow. All the small tertiary vertical marks show a step in the top of the E-M1X image from the ghost, but no step is in either of the E-M1 Mark II images.

ISO was same for all, and ~1200 or so.

Don't know if there is some other deeper menu item other than Stabilization that might cause this.

I did use a 12 second timer on the tripod too for all to rule out me bumping the thing.
 

John King

Member of SOFA
Joined
Apr 20, 2020
Messages
3,695
Location
Beaumaris, Melbourne, Australia
Real Name
John ...
Sounds as if you have covered the bases, Mack.

A couple of things:

1) Stop looking at images at greater than 100%!! I use 25% and then 100%.

2) At 300%, the image is being re-sampled at uneven pixel multiples. When I print, I always ensure that I am using 2x2 multiples. This ensures minimal interpolation errors.

3) Carefully test your new lens for sharpness at centre focus, looking at corners and short edges in particular. I have the USAF test images and instructions uploaded somewhere on my web site, I think. Will look tomorrow. REMIND me if I forget, as I probably will !!!

Just a few thoughts, mate. HTH.
 

Mack

Mu-43 All-Pro
Joined
Jan 14, 2018
Messages
1,936
Sounds as if you have covered the bases, Mack.

A couple of things:

1) Stop looking at images at greater than 100%!! I use 25% and then 100%.

2) At 300%, the image is being re-sampled at uneven pixel multiples. When I print, I always ensure that I am using 2x2 multiples. This ensures minimal interpolation errors.

3) Carefully test your new lens for sharpness at centre focus, looking at corners and short edges in particular. I have the USAF test images and instructions uploaded somewhere on my web site, I think. Will look tomorrow. REMIND me if I forget, as I probably will !!!

Just a few thoughts, mate. HTH.
Good idea on the chart.

I have one of those USAF charts from Edmund Scientific framed in a 2x3 foot wooden frame for flatness someplace, but damned if I can find the bloody thing! I must'a hidden it well - maybe in the garage. I'll tear the place up today and see if I can find it. Too big to misplace, but those little lined targets in its corners might tell a story.

Still, comparing the two bodies side-by-side, and both in the corner, I wouldn't expect a ghost in one and not the other. I don't know if there are multiple filters stacked in front of the sensor, but if there are then that might be the reason. I guess I could also try the Pen-F, and IR converted E-M1 Mark I too.

I may call Lifepixel and ask the setup since they tear into the sensor filters daily to convert cameras to IR. I know from playing with the IR camera that a filter on the front of it can cause a ghost image which is often to one side and inverted when I shot a table lamp in a dark room. When I took the (cheap) lens filter off, I didn't get the ghost. But this E-M1X ghost isn't inverted, it just shifts down slightly and to the left. If there is a stack of filters on the E-M1X sensor I could see how that might cause the issue if they all are not plano to the sensor.

Might be one of those "Just live with the E-M1X being softer -- and haunted by ghosts!" :eek:
 

Mack

Mu-43 All-Pro
Joined
Jan 14, 2018
Messages
1,936
Suspect is bad lens.

I forgot to try the 12-100mm f/4 on the E-M1X and it is flawless and no ghost as below. You can see the small green AF square made from Workspace near the #4 showing it nailed focus in the corner, but I did calibrate all 25 of the AF fine-tune points too.

5.jpg
Subscribe to see EXIF info for this image (if available)


Bad part is there appears to be no service mentioned for Voigtlander in their literature or the Japanese website either. I emailed B&H Photo to see what to do since it has been 3 weeks since I received it.

I mentioned earlier in the thread for this lens I thought it odd the box had no silica gel pack in it as usually comes with a new lens. One edge of the box was upturned a bit so it may have been a return and resold as new. Bah!
 

John King

Member of SOFA
Joined
Apr 20, 2020
Messages
3,695
Location
Beaumaris, Melbourne, Australia
Real Name
John ...
Good idea on the chart.

I have one of those USAF charts from Edmund Scientific framed in a 2x3 foot wooden frame for flatness someplace, but damned if I can find the bloody thing! I must'a hidden it well - maybe in the garage. I'll tear the place up today and see if I can find it. Too big to misplace, but those little lined targets in its corners might tell a story.

Still, comparing the two bodies side-by-side, and both in the corner, I wouldn't expect a ghost in one and not the other. I don't know if there are multiple filters stacked in front of the sensor, but if there are then that might be the reason. I guess I could also try the Pen-F, and IR converted E-M1 Mark I too.

I may call Lifepixel and ask the setup since they tear into the sensor filters daily to convert cameras to IR. I know from playing with the IR camera that a filter on the front of it can cause a ghost image which is often to one side and inverted when I shot a table lamp in a dark room. When I took the (cheap) lens filter off, I didn't get the ghost. But this E-M1X ghost isn't inverted, it just shifts down slightly and to the left. If there is a stack of filters on the E-M1X sensor I could see how that might cause the issue if they all are not plano to the sensor.

Might be one of those "Just live with the E-M1X being softer -- and haunted by ghosts!" :eek:
Mack, USAF.PDF file is here:

https://canopuscomputing.com.au/Downloadable_Files/General_Photography/USAF.pdf

And instructions here:

https://canopuscomputing.com.au/Downloadable_Files/General_Photography/USAF_pdml-procedur.pdf

Both will download immediately from clicking on the link. They are fairly small - less than 150KB in total.
 

Mack

Mu-43 All-Pro
Joined
Jan 14, 2018
Messages
1,936
Mack, USAF.PDF file is here:

https://canopuscomputing.com.au/Downloadable_Files/General_Photography/USAF.pdf

And instructions here:

https://canopuscomputing.com.au/Downloadable_Files/General_Photography/USAF_pdml-procedur.pdf

Both will download immediately from clicking on the link. They are fairly small - less than 150KB in total.
Thanks John for the PDF. I managed to find my old Edmund Scientific USAF Resolution Chart stuffed under a bed in its frame.

A080378-small.jpg
Subscribe to see EXIF info for this image (if available)


It does seem to be a CA issue. I shot the chart with the Voigtlander 29mm f/0.8 against my Olympus 25mm f/1.8 and found it sharper than the Olympus as far as the three sagittal lines (i.e. Ones pointing towards center of chart.) which lost definition to the meridian ones on the Olympus. The Voigtlander seemed to hang onto the three pairs on both S/M axis at the corners of the chart.

As to the CA, the Voigtlander has a lot of it. The red seems to point towards the center of the lens where its controlled. Also, the barrel distortion pulls the CA towards the outside near the middle and maybe the ghost is the red CA being pulled outwards and towards the center which causes the editing programs to come out weak to correct it. PS was really bad for correcting this lens and DxO Photolab 4 didn't have enough CA correction to address it.

The trial version of Exposure X7 has more controls that seem suited to this lens, imho. Just turning the Defringe to ON fixes much of the matter, and they have more CA adjustments that can move the CA colors around a lot more than other editors can - and further too.

Below is a unedited CA of the lower right corner targets on the chart. Second one was edited in Exposure X7 which shows the extra CA controls to the side. Even their Distortion has some refinements for the Barrel distortion that also has Midpoint and Asymmetry (Corner) refinement.

Corner-CA-29mm-f0.8.jpg
Subscribe to see EXIF info for this image (if available)


Exposure X7 Trial version and Distortion, CA Controls, and Defringe controls set to correct the CA above:



Exposure-X7-Trial-CA.jpg
Subscribe to see EXIF info for this image (if available)


I looked at the ORF images first in FastRawViewer at 1,000% and can see this lens seems to like f/5.6 at the corners and f/2.8-f/4 in the center. I could scroll through the series from f/0.8 to f/16 and the diffraction softness starts to show at f/8 and is pretty bad at f/11-f/16.

Aside, B&H Photo did respond today - other than first email saying they are slow a few days due to Covid - but I am unsure if I should return this lens and maybe get a worse one? Too bad there isn't a USA Service center that might take a look at it.

Knowing I might be able to correct it in Exposure X7, and maybe set up a Preset in the software too, and that the Olympus 25mm f/1.8 was worse than this, maybe it's a keeper after all. Dunno. Lots of cash for a manual lens though. :hmmm:
 

BPCS

Mu-43 Regular
Joined
Oct 15, 2017
Messages
191
I have noticed that in extreme flare situations, there are what appear to be artefacts showing up from the sensor... a pattern of sorts... a reveal of the sensor structure? Its rare and the photo is usually a throw away anyways... ugly extreme high contrast and huge highlight to shadow differences is when it can show up that pattern... like when the sun is in the shot and there are also deep shadow areas.

However my EM1X and Voigtlander 29mm f0.8 Super Nokton show this artefact more than any other lens and more than if using my EM1.2.

Some things to consider:

The microlenses on each sensor pixel have a speed of about f1.2... that's why you don't see faster exposures on lenses wider than that. Maybe there is an interaction at play here with f0.8/ 0.95 lenses?

The EM1X gained a better sensor coating over the EM1.2 which improves IQ in general, but maybe it makes those artefacts more obvious as well?

As far as ghosting goes, I do sometimes see a shadow effect with a very sharp lens focused perfectly, and in high contrast situations... more so on the EM1X... like the image does a tiny double exposure shift. The m43 sensor filter glass is unusually thick... 4mm vs 1mm or 2mm with other sensors. Maybe there can be a tiny ghost visible... a shift in image reflection caused by this?

All of my above observations are if looking for them and viewing at 200/ 300% or more. They are interesting but don't worry me as at any normal type of viewing, are not noticable.

The Voight 29 0.8 is an interesting lens and needs to be shot wide open, or else together with any other f1.2/ 1.7/ 1.8 lens shot at f2.8 - 11 etc is not too different... although I think the Voight at f4.0 could be the sharpest possible option across the frame in that focal length range... but only by pixel peeping. It could also be, at f1.4, better than any other similar FL at f1.4. Wide open, it is wrought with aberrations, most problematic is colour fringing, but RAW processing can eliminate 85% of this and make it acceptable. That leaves one free to enjoy the unique look of the lens image at f0.8... vintage yet modern... imperfect by modern standards but noticable better than the best fast vintage lenses wide open.

I won't use it often, but its nice to know it is there for when you want it. A luxury maybe, but thats always the hard decision to make.
 

Latest threads

Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Mu-43 is a fan site and not associated with Olympus, Panasonic, or other manufacturers mentioned on this site.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Forum GIFs powered by GIPHY: https://giphy.com/
Copyright © Amin Forums, LLC
Top Bottom