PDAF Striping: can you help me list cameras?

[Andrea B.]

After talking with another member about the PDAF striping problem in mirrorless conversions, I have written up the following to add to the Sticky :: UV Photography: Cams, Mods, Lights, Links .

 

I need to know of any mirrorless conversion you have used which shows serious PDAF striping and which I do not have listed below. Also I need to know which mirrorless conversions do *not* exhibit PDAF striping and which I do not have listed Thanks for any input!!

---

 

This is the updated version to be placed into the Sticky.

 

Warning about mirrorless cameras using PDAF autofocus!

Any mirrorless camera using phase detect autofocus is prone to striping artifacts. In ordinary Visible photography, PDAF striping is rare. Wide dynamic range scenes and/or brightly illuminated, low ISO scenes can often induce the banding in a non-converted camera. And if the banding is there, then shadow lifts make it look worse.

 

In a full spectrum converted mirrorless camera, PDAF striping seems to be exacerbated. UV photos typically have wide dynamic range, and shadow lifts are often sometimes used to bring out dark details in UV. The UV false color white balance step induces a radical color shift. If striping is present, it will look worse in the processed UV file. Any striping artifacts in an IR file can also be further worsened in the converter with the typical IR processing of stronger contrast, channel swapping, enhancement of edge acutance and/or a white balance step. There are some software fixes available which can clean up some of the PDAF striping with varying degrees of success.

 

Unfortunately, we have not seen warnings from the major retail conversion shops about exacerbated PDAF striping in some mirrorless conversions. So, again unfortunately, you are on your own about deciding whether to convert a particular mirrorless camera. Please first check to see whether your mirrorless camera uses phase detect autofocus (always, sometimes?) and whether striping has been reported for that camera.

 

Here are a few full spectrum mirrorless conversions shown by our members to have PDAF striping. This is by no means a complete list. For some cameras the striping problem is reported by some and not by other users of that camera.

 

Also listed are some mirrorless cameras for which the PDAF striping problem has not been reported (at least so far).

• DO NOT CONVERT Nikon Z6, Z7. See Link https://www.ultravio...n-the-bad-news/
  The artifacts from a converted Nikon Z-type camera are called "PDAF banding".
  

• MAY SHOW Striping Artifacts
  
  • Sony A7R-type
    Some have seen minor PDAF striping with A7R-type bodies, others have not.
    
  • Olympus EM-type
    Some have seen minor PDAF striping, others have not.
    

• OK TO CONVERT
  
  • Panasonic Lumix G-type not having internal IR shutter monitor.
    
  • Sony NEX. No reports of striping so far.
    

[Andy Perrin]

Sony A7S so far seems fine. I have not observed anything like nfoto's issue.

[colinbm]

Sigma fp, 35mm full-frame (35.9mm x 23.10mm) Back-illuminated Bayer CMOS sensor uses, Contrast detection system.

 

https://www.sigma-global.com/en/cameras/fp-series/specifications/

[dabateman]

I only see PDAF striping on the EM1mk1 when taking extreme 300nm images using a weak light or filter. Using a 302nm light or a filter that lets through more than 20%, then I don't see it in the 300nm images.

Interestingly, the PDAF striping I did see on the 300nm Em1mk1 images are only in one of the 2 green channels. You can rip that green channel out and merge back an RGB image at 4Mpixels using 4Channels. Haven't figured out how to trick software to interpolate it back up to 16Mpixels yet. But now just use my better 303bp10 filter, which has 60% transmission at 302nm mercury line.

[enricosavazzi]

Sony A7 II seems fine. I have seen no striping in over one year.

[ulf]

My Sony A7 III seams fine too. At least so far I have not seen any trace of striping.

I have been using it for almost three months now.

Alan Burch also made an excellent job shielding the internal monitor LED. No leakage at all.

 

Please remind me of how to provoke the phenomenon.

My images might be taken with settings that not show the striping.

[GaryR]

I haven't seen any of those banding issues on my a7R (2013 model). I purchased it to avoid the internal IR focus diode, that was added to later models.

[Andrea B.]

Thanks everyone. I'll work on this little warning message. I don't want to scare anyone off of conversions, but it seems only fair to let readers know that PDAF striping might be a problem in some conversions.

 

For now, I removed the Sonys and Olys from the list until I can rewrite it.

[Andrea B.]

I put the updated text in the first post.

 

I do have one more camera to add to the list.

 

I'm trying to be very fair about reporting this. So please tell me if anything looks wrong or extreme.

[eye4invisible]

For the Sony A7 series, there's a note on KV's website that lists those models that have an internal LED:

 

Full Frame: RX1, RX1R, Rx1R II, A99, A99 II, A7*, A7R*, A7S*, A7 II†, A7R II†, A7S II†, A7R III†, A9†, A7 III†, A7R IV, A9 II

 

† These cameras contains an internal diagnostic IR Led. This can cause light leaks when using very long exposures and high ISOs in the H-alpha and IR enabled conversions, in the range of 30 second exposures at 6,400 ISO. The A7RII and A7RIII additionally can have fine horizontal artifacts across the frame in highlighted areas in the IR enabled conversions due to the different sensitivity of some rows of pixels.

 

https://kolarivision.com/product/sony-dslr-slt-mirrorless-full-spectrum-conversion-service/

[bvf]

Hi, Andrea.

 

I passed you some info. on the Canon EOS M3 which seems to suffer from this problem.

 

My first full spectrum conversion was a Canon EOS M. This had a related problem in that there 2 rows of tiny tiny black dots which I believe were associated with the focussing mechanism. These were not visible in visible light photography, and generally didn't cause any problem in UV - you could only see them if you went looking for them. But they could occasonally cause a problem when focus stacking, when you would get lines as in the image below. Fortunately only a few images were spoilt in this way, and I had lots of successful focus stacked images on that camera.

 

My current conversion is a Sony A6000, and I've detected no problems with that.

 

[Andy Perrin]

For the Sony A7 series, there's a note on KV's website that lists those models that have an internal LED:

 

Full Frame: RX1, RX1R, Rx1R II, A99, A99 II, A7*, A7R*, A7S*, A7 II†, A7R II†, A7S II†, A7R III†, A9†, A7 III†, A7R IV, A9 II

 

† These cameras contains an internal diagnostic IR Led. This can cause light leaks when using very long exposures and high ISOs in the H-alpha and IR enabled conversions, in the range of 30 second exposures at 6,400 ISO. The A7RII and A7RIII additionally can have fine horizontal artifacts across the frame in highlighted areas in the IR enabled conversions due to the different sensitivity of some rows of pixels.

 

https://kolarivision...ersion-service/

 

Andrea, if my A7S has an internal LED, I've never seen it. Admittedly I rarely go past a 30 second exposure, just because the A7S has the "star-killer" problem where the auto-noise suppression kicks in after 30 seconds. Also, LifePixel did NOT note it to be a problem on this camera, and believe me I checked!

[Akira]

No Panasonic cameras, G (m4/3) or S (full size) series, have sensors with PDAF photosites. As Colin mentions, Sigma fp uses a BSI sensor without PDAF photosites. All cameras mentioned here (and Leica SL series) employ CD-only AF system.

 

I haven't seen any banding in the images shot with Sigma fp even at high ISO settings like 12800, or 6400 with the shadow area raised, or in IR with a quite dense IR90 filter without modification, The stock fp is practically unusable for UV.

[ulf]

For the Sony A7 series, there's a note on KV's website that lists those models that have an internal LED:

 

Full Frame: RX1, RX1R, Rx1R II, A99, A99 II, A7*, A7R*, A7S*, A7 II†, A7R II†, A7S II†, A7R III†, A9†, A7 III†, A7R IV, A9 II

 

† These cameras contains an internal diagnostic IR Led. This can cause light leaks when using very long exposures and high ISOs in the H-alpha and IR enabled conversions, in the range of 30 second exposures at 6,400 ISO. The A7RII and A7RIII additionally can have fine horizontal artifacts across the frame in highlighted areas in the IR enabled conversions due to the different sensitivity of some rows of pixels.

 

https://kolarivision...ersion-service/

This means that you have to use a conversion service that is able to solve the internal LED-problem.

[Akira]

Won't the electronic shutter mode solve the IR-LED issue?

[dabateman]

Won't the electronic shutter mode solve the IR-LED issue?

 

That is the solution recommend for Panasonic cameras with the IR shutter monitor problem. However, Panasonic cameras limit the electronic shutter to maximum duration of 1 second. I typically need 2 to 16 seconds for some of my exposures. So a Panasonic camera will not work.

But check the manual. The new Panasonic S1/S1H/S1R may not be limited to 1 second electronic shutter. Just for super resolution shot mode. However, these cameras also don't have a IR shutter monitor.

I hope Panasonic lifts this limit on all cameras now with a firmware. Panasonic please provide a firmware update to allow up to 60 seconds electronic shutter in any mode on every camera you have.

Adding it to my GM5 would be great, currently stuck a 1 second maximum.