Weird sharpness issues with Panasonic 100-300

CaspianCanuck wrote:

Hi,

I have a Panasonic Lumix G Vario 100-300 lens (the original one, not the newer mark II) that I use with a couple of Olympus bodies (E-M5 II and E-M1 III). My typical shooting scenario is:

• RAW only
• Limit the focal length to 275-280 mm to avoid the IQ degradation at 300 mm
• Holding the lens under by the hood to reduce the shaking
• Lens OIS turned on and both my cameras set to lens OIS preference if available (i.e. the body's own sensor stabilization is off)
• Using the low-speed sequential mode + electronic shutter (L-heart) or first-curtain antishock (L-diamond).

When I view the resulting sequential shots of the same scene shot at focal lengths above 200mm I always see strange artefacts that look like waves of visual distortion and floating areas of poor sharpness.

I can see how a wave of optical distortion can be caused by a microscopic movement of the OIS element inside the lens as it tries to compensate for the lens shake while the camera is in the process of reading out pixels from the sensor. But I struggle to come up with an explanation for the floating areas of poor sharpness.

I can best illustrate the latter problem with the following three crops from larger images, all taken at 275mm, 1/2000s and f/7.1. If you look at the grassy area in the center of the frames and compare it from one shot to the next you can see how different parts of it are blurry in one shot but sharper in another. (Best viewed in full screen mode at 100% resolution.)

I know this problem has nothing to do with the shimmying air currents as it appears in all shots regardless of the ambient temperature or the proximity of the subject to the ground.

Any idea what could possibly cause this strange effect?

I have seen that effect also.

I have that same lens and I get better IQ when I use the IBIS of my EM5II rather than the OIS of the lens.

Allan

drj3 wrote:

Check the RAW files in an Exif program (I use Phil Harvey's ExifTools) and check the focus distance or focus steps. The three files look like they are all focused at somewhat different distances.

I assume you used a single focus point on the same object for each image. If not, then this can cause significant focus variation.

Even if focused on a single distinct object, all cameras and all lenses will have some focus variability.

The E-M1.2 and E-M1.3 show less focus variability with CAF focusing than with SAF focusing with my 300mm f4 lens. This may be different for other lenses.

The size of the focus points are also larger in SAF mode (PDAF then CDAF focusing) than with CAF (PDAF only focusing) which can lead to more variability in focus.

Agreed. Try using CAF and a burst of about ten shots of a static test subject at about ten to fifteen metres (to test focusing but not introduce heat shimmer) using a single smallest AF point. Then pixel peep the subject you focused on. You will find some shots are critically focused, and others are not. It would be the same result with using SAF and ten shots in succession, but with Olympus I will get an extra shot or two with SAF that are slightly soft. Depending on the camera and lens, there might be more out of focus than critically in focus. If there are 7 or 8 out of 10 critically in focus, everything is working as well as can be expected with current gear. Only 4 out of 10 critically sharp would not be unusual. The hit rate does also vary depending on the chosen subject. The out of focus shots should also only be obvious at 100% viewing... so just missing focus and depth of field tolerances. Use of more expensive lenses does help the hit rate in my experience. So a PL 100-400 would give you one or two more critically sharp shots on average in every ten than the 100-300. Using Olympus, I have even noticed that with FW upgrades (EM1.2 and FW3 or later and EM1X FW2), the hit rate has improved about 10-20% (particularly the EM1.2). Of course lens firmwares are also kept up to date, so some credit to improvements might be there as well.

I have noticed the drifting focus too, and your reply does a good job explaining how it happens. But in this forum thread I wondered about the weird out of focus areas that seem to float around from one burst shot to the next.

Thanks for an interesting theory!

CaspianCanuck wrote:

I have noticed the drifting focus too, and your reply does a good job explaining how it happens. But in this forum thread I wondered about the weird out of focus areas that seem to float around from one burst shot to the next.

It could be that tiny changes in focus difference (due to the focus "drift") have an effect on the rendering of other areas, particularly out of focus areas due to depth of field, or that the Lens IS element being in a different position affecting the rendering (sharp areas "move" around). Even IBIS could do this as IBIS will move around relative to the lens image circle. It could also be stabilisation in general as whether electronic shutter or mechanical, not the whole image area is exposed at exactly the same time. Picking a static subjective, maybe a sturdy tripod and electronic shutter (to eliminate shutter shock) with a timer delay and no stabilisation at all, could be a way of isolating stabilisation as a cause? My gut feel is that Lens IS especially could cause random weird patches of very slight blurriness. But I also, from practical experience, have found that the higher end the gear is, the less this comes into play.

With my PL 100-400, at very fast shutter speeds (1/2500 - 1/5000) and a static subject, I get a slightly higher keeper rate with Lens IS Off (no IBIS either).

mak_kawa wrote:

Are you using a protect/UV filter for this lens?

Once I had been annoyed by this strange "wavy" artifacts in the tele-end (300 mm) shot from my P100-300II. I had attached a lens protect filter to the lens, and when I removed it, the artifacts are surprisingly gone.
I suspect that the artifacts are caused by micro-scale heterogeneity in the filter surface/coating, but just a speculation.

Yes... 150mm and over, any flaws in a filter will be apparent and can be in the form of "patches".

CaspianCanuck wrote:

Hi,

I have a Panasonic Lumix G Vario 100-300 lens (the original one, not the newer mark II) that I use with a couple of Olympus bodies (E-M5 II and E-M1 III). My typical shooting scenario is:

• RAW only
• Limit the focal length to 275-280 mm to avoid the IQ degradation at 300 mm
• Holding the lens under by the hood to reduce the shaking
• Lens OIS turned on and both my cameras set to lens OIS preference if available (i.e. the body's own sensor stabilization is off)
• Using the low-speed sequential mode + electronic shutter (L-heart) or first-curtain antishock (L-diamond)

With your bodies and this lens, I would reconsider the use of OIS. I tested it extensively a couple years ago and found that the EM5ii IBIS was substantially better than Mega OIS. Also, on my copy of the lens at least, there was no discernible falloff in sharpness at the long end - just gradual decline across the full range.

https://www.dpreview.com/forums/post/57524794

I know this problem has nothing to do with the shimmying air currents as it appears in all shots regardless of the ambient temperature or the proximity of the subject to the ground.

Atmosphere was definitely what came to my mind first, but looking at the shots, it seems like it might be as simple as slight variations in the plane of focus.

Leo3 wrote:

The photos you posted are midday with plenty of sun, perfect for atmospheric haze/ thermals to cause random pockets of mush between frames.

If you had examples from a calm cloudy day that would be different....

Here are examples from a sequence of frames shot in the low-silent mode on a cool day. These are approximately 15% crops of larger frames.  I don't think air convection currents could be responsible here because they move too slowly and couldn't possibly create this effect given the short time delta between the frames. Besides, convection currents can introduce local distortions in an image but can't make an entire frame or a large portion so badly fuzzy for a fraction of a second. You can clearly see how the focus drifts between the four frames, especially between 2 (ok), 3 (terrible) and 4 (ok again).  This in my mind could only mean the thrashing of the lens' OIS.

I have now added a new thread with my comparison of this lens with my brand-new Olympus 75-300 here: https://www.dpreview.com/forums/post/64999609

CaspianCanuck wrote:

Leo3 wrote:

The photos you posted are midday with plenty of sun, perfect for atmospheric haze/ thermals to cause random pockets of mush between frames.

If you had examples from a calm cloudy day that would be different....

Here are examples from a sequence of frames shot in the low-silent mode on a cool day. These are approximately 15% crops of larger frames. I don't think air convection currents could be responsible here because they move too slowly and couldn't possibly create this effect given the short time delta between the frames. Besides, convection currents can introduce local distortions in an image but can't make an entire frame or a large portion so badly fuzzy for a fraction of a second. You can clearly see how the focus drifts between the four frames, especially between 2 (ok), 3 (terrible) and 4 (ok again). This in my mind could only mean the thrashing of the lens' OIS.

If something is thrashing, the autofocus seems like a much more likely explanation.

Anyway, it's simple to test - turn the OIS off and try the IBIS.