Banding with the E-M5 and the 20/1.7: A potential work-around

Started Oct 19, 2012 | Discussions thread
Shop cameras & lenses ▾
tsi Forum Member • Posts: 54
Re: Banding with the E-M5 and the 20/1.7: A potential work-around

Anders W wrote:

Hi tsi and welcome to the forum. See comments below.

Thanks for the welcome.

I also thought, at one point, that the aperture actuator might be responsible. What I say about the matter in that post early in this thread is:

"My own best guess currently is otherwise interference from the aperture actuator, which, as far as I can tell, is operative during the read-out phase. I have in some tests seen indications that aperture movement might be responsible (no banding if I shoot wide open so that it won't have to open up again after exposure but banding over a vertically increasing area of the frame with increasingly smaller apertures)."

It is very time-consuming to read everything that has been written about this topic.

However, it is good to know that your system also shows the variation of the affected area with the aperture.

"However, in other cases, I have seen banding with the lens wide open too so perhaps this is a blind alley, a partial truth, or a valid theory but with some important bits still missing. We'll see."
[...]
Did you try shooting wide open when the lens was completely "cold"?

You are right. I have found one ISO-series of images shot with E-M5 and 20/f1.7 at 1/10 and f1.7. All images including the ISO200 image show a regular pattern of bands over the entire frame. I have repeated this series after many shots for "warm up". The bands were still there. The intensity of the banding is similar to the case, where the area depends on the aperture. It shows up only in this series, but not in any other image that was shot wide open.

Note these posts describing banding problems with lenses other than the 20/1.7:

http://forums.dpreview.com/forums/thread/3217040

http://forums.dpreview.com/forums/post/41473019

http://forums.dpreview.com/forums/post/41731894

http://forums.dpreview.com/forums/post/50163367

I agree with Timur Born: "one difficulty in identifying the banding might be that the E-M5 produces banding on its own, too."

This is true in the case of the 10-50mm lens, but the banding of the 20/1.7 is very specific. Except for the banding at f1.7, it has a start and an end-line. It is color neutral. It usually produces a very regular pattern which creates this awful artificial look. Sofar, I have not seen any direct comparison showing that the banding with another lens was on par with that of the 20/1.7.

The 20/1.7 has shot 90% of all images that I have shot during the last years. I love the small size and the sharp images produced at f2.0. With the E-PL3 the images are acceptable up to ISO1600. With the E-PM2 the quality at ISO1600 is better, but ISO3200 is not acceptable due to the artificial look created by the banding. Without the banding, ISO6400 is usable. Since I cannot afford a new camera every year, I have decided to wait for a solution to this problem and returned the camera.

Is your high-frequency banding with the 20 any worse than what you can see with other lenses?

As I wrote, "Banding is not visible with 45/1.8, 75/1.8, and Walimex 7.5/3.5 lenses."

Does the frequency increase when the lens is "warmed up" (by exercising the AF mechanism) as it does in my case?

In one case, I have observed that the frequency decreases, i.e. the width of the bands becomes larger. After turning off the camera for 1 min, the frequency was about the same as in the first shot. I have not saved a second series, where the frequency decreased continuously, but in the next frame the frequency was high again. For most shots taken under controlled conditions with manual focus, the frequency did not change so much.

Did you try shooting wide open when the lens was completely "cold"?

Another problem with the aperture-actuator theory is that I didn't see much of an effect on banding if I tried to "warm up" the lens [...].

I am not convinced about the "warm up" theory. It did not work in my case. Maybe it is more important close to the Arctic Circle.

I hope that Olympus will solve this problem with a firmware update, which prevents the movement of the aperture actuator during read out.

This update would be worth at least one more E-PM2 customer.

On the other hand, this problem is known for such a long time and nothing has happened.
They could at least provide a software tool for correction, which is relatively simple to do.

-- hide signature --

Here is another banding effect:

I was able to produce colorful horizontal lines with the E-PM2 and the 20/1.7 when used in ISO bracketing mode. The lines extend across the entire frame. There was also a TV in the image. I think that this effect is not unique for the 20/1.7, but a general problem of the ISO bracketing mode.

Do you know if this has been discussed elsewhere?

Post (hide subjects) Posted by
tsi
tsi
tsi
tsi
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow