Banding with the Panasonic 20/1.7 MFT Lens

27 July 2015 (updated 6 January 2016)

The combination of the Panasonic 20/1.7 lens and certain Olympus and Panasonic MFT bodies can give unusually severe horizontal banding in images.

Discussion of the problem is spread over numerous posts in several threads in the DPReview Micro Four Thirds Talk forum. This page is my attempt to summarize the problem in one place.

If you have additional information, please get in touch.

If you are interested in using the Panasonic 20/1.7 lens on an Olympus body, you might also be interested in my summary of the problem with purple fringes.

Panasonic 20/1.7 Lens

The Panasonic 20/1.7 pancake lens was one of the first native prime lenses for the MFT system. Its focal length makes it a normal lens, with a field of view similar to a 40 mm lens on a 35 mm camera.

It was lauded as sharp and fast, especially compared to its somewhat disappointing contemporary, the Olympus 17/2.8. This high optical quality, combined with compactness and lightness, has made it one of the most popular prime lenses for MFT cameras.

Panasonic followed up the initial 2009 version with a very similar second version in 2013.

Problem

Unfortunately, the combination of either version of the 20/1.7 lens and bodies equipped with a particular Sony sensor has been shown to give unusually severe horizontal banding.

Now, all sensors have banding at some level. However, this particular combination of sensor and lens has banding that is often much worse than normal. That is, this particular combination of sensor and lens can give noticeable banding in circumstances in which other combinations do not. It is often obvious at ISO 1600 or higher, but can appear at lower ISOs if shadows are pushed. The severity seems to vary between different examples of apparently identical lenses and bodies and can vary with time and conditions.

There are occasional reports of similar banding with other lenses, but the accumulated evidence suggests that the 20/1.7 is significantly more prone to this problem than any other lens.

The root of the problem appears to be an anomalous electrical interaction between this particular lens and this particular sensor; such severe banding is not seen when this lens is used on bodies with other sensors or when bodies with this sensor are used with other lenses.

This problem is different to the banding caused by using a slow rolling shutter under fluorescent lights. Egregius V. has published an album showing the differences between the two kinds of banding.

Bodies

Olympus and Panasonic have officially acknowledged that this problem occurs when the E-M5 and the GH3 are combined with this lens.

More generally, it is believed that the bodies which suffer from this problem are those with the Sony IMX109 16 Mpx sensor. This sensor was introduced with the E-M5 and subsequently used in various Olympus PEN and OM-D bodies and in the Panasonic GH3.

Specifically, it is believed that the following bodies suffer from this problem when combined with this lens:

The susceptibility of the E-M10II is currently unknown, but given that the E-M10 and E-M5II suffer, it would not be surprising to discover that the E-M10II did too.

On the other hand, the combination of this lens with other bodies appears to be fine. Specifically, it is believed that the following bodies do not suffer from this problem when combined with this lens, either because of empirical evidence or because they do not use the 16 Mpx Sony sensor:

All Olympus MFT bodies with 12 Mpx sensors
All Panasonic MFT bodies with 12 Mpx sensors
All Panasonic MFT bodies with 16 Mpx sensors except the GH3
Olympus E-M1
Panasonic GX8

The E-M1 and the GH3 appear to be anomalous compared to their siblings. However, this anomaly can be understood by noting that the E-M1 is believed to be the only Olympus MFT body to use a 16 Mpx Panasonic sensor and the GH3 is believed to be the only Panasonic MFT body to use the 16 Mpx Sony sensor.

Solution

At the moment, there appears to be no complete solution.

Forum member Vlad S. reported that the banding was reduced, at least in the E-M5, by firmware version 1.5, but it was not completely eliminated.

Forum member Anders W. suggested the it can be mitigated by “warming up” the lens, but again this is not a complete solution.

Forum member secretworld reports that the Dfine plugin (in the Nik Collection by Google) can be used in post-processing to reduce banding. The debanding control is under the “»More” tab. Forum member Jean-Pierre Martel has shown an example of debanding with Dfine.

Forum member cainn24 also suggests using the free G’MIC software for debanding.

Alternatives

If you prefer to avoid the Panasonic 20/1.7, you might consider the Olympus 17/1.8, Sigma 19/2.8, Olympus 25/1.8, Panasonic 25/1.4, or Panasonic 25/1.7. None of these are so compact, though.

Acknowledgements

The understanding of this problem was largely developed by members of the DPReview Micro Four Thirds Talk forum.

I thank forum members Anders W., Jean-Pierre Martel, Egregius V., and secretworld for suggestions which have improved this summary.

Updates

6 January 2016: I have added the E-M5II to the list of cameras that suffer from this problem; miemo has posted convincing examples. I also added cainn24’s suggestion that G’MIC can be used for debanding. I noted that the susceptibility of the E-M10II is unknown. I added the Panasonic 25/1.7 to the list of alternative lenses.
30 July 2015: I removed reference to alleged banding with the E-M5II flicker-reduction mode, since this has now been withdrawn. I added more discusion on the E-M5II, following a suggestion by Anders W. I added more information on Dfine from secretworld.
28 July 2015. I added information from Egregius V. on the E-M5II. At his suggestion, I also added a note that this banding is not due to fluorescent lights.
27 July 2015. At the suggestion of Anders W., I removed reference to the Olympus 25/2.8, as it is a FT lens not a MFT lens. I added a suggestion from Jean-Pierre Martel on reducing the banding using Dfine.