Huge focusing problem on Panasonic S1R

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

Lou Jost
Posts: 5945
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Huge focusing problem on Panasonic S1R

Post by Lou Jost »

I am using a Sigma 70mm Art lens on this camera, attached via a Sigma MC-21 smart adapter. Sigma certifies this combo is compatible with the S1R, and Sigma is part of the L mount consortium, so this is not a third-party lens or adapter.

I had hoped to use the S1R and 70mm lens for automatic focus bracketing like I do with my Olympus PEN-F. The literature seemed to indicate it could do that, but no, that turned out to be not true, because Panasonic and Olympus mean different things by "focus bracketing."

OK, I figured I can get around that by tethering the camera and making a software robot that clicks the forward focus arrow at a set time interval. Nope, Panasonic's tether program doesn't let me focus manually via the focusing arrows (what are they for, then?)

Well, as a last resort it should be possible (though tedious) to focus manually, moving the lens' focus ring in tiny steps while the camera takes pictures. That surely can't fail, right?

Wrong. The Sigma lens is a focus-by-wire lens, and it is sensitive to the speed of rotation of the focusing ring. The camera has two options for the focus ring. You can set it to be "non-linear" ("focus responds according to rotation speed") or "linear" ("focus responds to rotation angle"). However, this lens behaves the same for both settings. It does not change focus at all unless you turn it fast. This is crazy. Even in the llinear mode, I can turn the focusing ring through 90 degrees in five seconds and the focus still doesn't move!!!

Not only are small steps impossible, but ANY precise focusing is impossible. Once you are close to correct focus, you can't make slow careful adjustments, because nothing happens. You have to jerk it around to make it move!

So I am afraid this camera has been a bitter disappointment. Maybe this can be fixed by software updates. I hope so. Or maybe native Panasonic lenses would behave differently. I do like the image quality and pixel-shift mode. But whoever designed this should be fired.

Anyone else have this problem? Seems like a pretty big problem to have gone unnoticed.

Edit: Here is someone else with a similar problem using a native Panasonic L lens:
https://www.l-camera-forum.com/topic/29 ... -90-280sl/

Macro_Cosmos
Posts: 1511
Joined: Mon Jan 15, 2018 9:23 pm
Contact:

Post by Macro_Cosmos »

Sounds like a pretty massive oversight, the 70mm Sig is one weird cat for sure.

A friend of mine acquired an S1R and it died within a week, showing some weird error. Luckily the responsive reps replaced his camera on the same day, plus to them.

Maybe this isn't an issue with their lenses that aren't focus by wire? This is definitely an odd issue and a pretty big oversight.

clarnibass
Posts: 130
Joined: Fri Jun 10, 2016 11:33 pm

Post by clarnibass »

Just a quick question about this camera, can you use flash in the high resolution mode? I mean off camera flash with the trigger on camera. Couldn't find that exact spec.

Lou Jost
Posts: 5945
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

No. But forum member Mike Wyatt's custom-made stacking controller can get around that!

chris_ma
Posts: 570
Joined: Fri Mar 22, 2019 2:23 pm
Location: Germany

Post by chris_ma »

hmm, seems like assisted manual focus doesn't work very well with most lenses.

It seems more like a problem with the compability of the l-mount alliance though, from what I've read the (some?) panasonic l-mount lenses can be switched over to full manual by pulling down the focus ring, so it would be interesting to know if the manual focus works better there.
clarnibass wrote:Just a quick question about this camera, can you use flash in the high resolution mode? I mean off camera flash with the trigger on camera.
As Lou writes, you can't without external software control.
It should be noted that one main advantage of flash, freezing camera shake, will not work in high-resolution mode anyway since the 8 frames are a fraction of a second apart (I'd say about 1/2 sec for 8 images total by ear) , so even if you had a flash which could fire that fast your setup has to be static during that time. so using LEDs seems not like a big disadvantage in that case.

I'm extremely happy with my first tests with the S1R, but I'm focusing by rail with manual lenses, so I didn't run ito the same trouble as Lou.
The image quality of very small pattern is way above any normal camera and highest of any FF sensor I've ever seen.
the new Fujifilm GFX100 is about the same resolution, slightly sharper but with a bit more aliasing and around 3x the price (plus harder to find lenses due to magnification and coverage).
chris

Lou Jost
Posts: 5945
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

I agree with Chris that the image quality seems comparable to some medium format cameras.

"..Panasonic l-mount lenses can be switched over to full manual by pulling down the focus ring"

Yes but I think they still focus by wire, so they could have the same problem.

clarnibass
Posts: 130
Joined: Fri Jun 10, 2016 11:33 pm

Post by clarnibass »

chris_ma wrote:As Lou writes, you can't without external software control.
It should be noted that one main advantage of flash, freezing camera shake, will not work in high-resolution mode anyway since the 8 frames are a fraction of a second apart (I'd say about 1/2 sec for 8 images total by ear) , so even if you had a flash which could fire that fast your setup has to be static during that time.
Would something like that controller mentioned be able to allow that? Let's assume the flashes definitely can't take eight photos that fast. Having the flash for each exposure is the difference between it working or not, no?
Half a second for eight frames suggests a pretty fast shutter speed. Do you use it to hopefully have less blur from any vibration (is there any)? Because with continuous lights how fast can you get it at (assuming) the lowest ISO?
It has to be static during the high res exposures regardless of light used, no?
chris_ma wrote:The image quality of very small pattern is way above any normal camera and highest of any FF sensor I've ever seen.
the new Fujifilm GFX100 is about the same resolution, slightly sharper but with a bit more aliasing and around 3x the price (plus harder to find lenses due to magnification and coverage).
I imagine you mean resolution for the same frame? So the smaller sensor can even be an advantage. How would you even get the same frame on the medium sensor?
Edit: Just checked that Fuji, the sensor is larger to a degree that I guess could be a disadvantage for some macro, it's actually much smaller than any film medium camera I've used. Seems sort of like a tweener.

chris_ma
Posts: 570
Joined: Fri Mar 22, 2019 2:23 pm
Location: Germany

Post by chris_ma »

clarnibass wrote:Would something like that controller mentioned be able to allow that?
I haven't tried any of this, so this is all from my theoretical understanding, so anybody who has hands-on experience and spots any thinking errors please jump in:

the problem is you can't control the sequence of the 8 shifted exposures, so you'd have to have a dark room and use long time exposure (still giving you dark pictures) and time the flash with an external controller to hit one flash per exposure.
Let's assume the flashes definitely can't take eight photos that fast. Having the flash for each exposure is the difference between it working or not, no?
jup. you need the same exposure for each frame of course.
Half a second for eight frames suggests a pretty fast shutter speed. Do you use it to hopefully have less blur from any vibration (is there any)?
with a bright enough light source you could have each exposure at say 1/1000sec which would eliminate blur on each frame, but if the camera moves between the 8 exposures the camera software will very likely introduce errors from combining them.
so in reality you'll need to have a stable enough setup not to shake during the whole process. my tests were around 0.5x to 1.4x magnification with shutter speeds from 1/100sec up to 1/10sec on a pretty solid setup but wobbly table and I got excellent results as long as I kept my breath during the picture collecting process :)
Because with continuous lights how fast can you get it at (assuming) the lowest ISO?
with very bright lights you can get up to the max shutter speed per single capture (1/8000 or probably even 1/16000 since it's in electronic shutter mode).
if you try that on 100 ISO you better bring some sun glasses though :)
It has to be static during the high res exposures regardless of light used, no?
exactly, that's why I think flash will only bring a minor improvement (might test that in the next days if I have time)
chris_ma wrote:the new Fujifilm GFX100 is about the same resolution, slightly sharper but with a bit more aliasing and around 3x the price (plus harder to find lenses due to magnification and coverage).
I imagine you mean resolution for the same frame? So the smaller sensor can even be an advantage. How would you even get the same frame on the medium sensor?
Edit: Just checked that Fuji, the sensor is larger to a degree that I guess could be a disadvantage for some macro, it's actually much smaller than any film medium camera I've used. Seems sort of like a tweener.
not sure I understand your questions right, but generally speaking small sensors can be an advantage in macro work because it's easier to find good lenses at low cost (smaller magnification and smaller image circle needed).
the advantage on a larger sensor is that you get better dynamic range (contrast from brightest to darkest) with the same resolution (larger pixel pitch, which also means the lenses can be a bit lower resolution), or higher resolution with the same dynamic range.

in the end, it all kinda balances out a bit, but on the high end large sensors will be better.

for example:
let's say we'd want to capture an object which is 18mm wide.
with a M4/3 camera with 6000pixel wise sensor resolution, you'd need a lens that magnifies 1:1 and only covers a 23mm diagonal.
that means you can probably find one with large aperture and the moderate magnification means less problems with diffraction due to effective aperture. on the other hand, it needs to resolve 3micron on the sensor level, which is tricky.

if you would use the same pixel count on a full frame camera, you'd need a lens for 2x magnification that covers a 44mm diagonal. that means a bigger lens with larger aperture (more diffraction problems), but on the other hand it only needs to resolve 6microns on the sensor level, so that puts less demands on the resolving power. in the end what you mainly get is more dynamic range (and with a top end lens probably slightly higher detail)

if we do the same thing with a full frame camera with 100MP, we'd get the same dynamic range and possibly much higher resolution, but finding a lens that covers the sensor and resolves 3microns at 2x magnification will be difficult (ie expensive).
this is about what the S1R does (although they claim 187MP), with the added benefit of strongly reduced aliasing/moiree problems on finest pattern detail.

with the Fujifilm GFX100 you'd need a lens with mag of around 2.5x ,that covers 55mm, and resolves a pixel size of about 3.8micron. not easy to find, but if you go the full way (let's say a printing nikkor) it would definitely outclass any M4/3 camera, even one with pixel shift.

just some thoughts
chris
Last edited by chris_ma on Sun Jul 14, 2019 6:40 am, edited 1 time in total.

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

I believe you can do this with a specialized controller and maybe strobe. Some strobes (not the inexpensive ones) have the ability to recycle in under 100ms with a very fast output (<100us), the key is maintaining exposure between the 8 bursts in 4 seconds. This should be possible with a prefire of a couple flashes before the actual 8 exposures, the idea is to maintain an average capacitor voltage for each exposure. May need to tighten up the control sensors level in the strobe though.

The controller could be designed to fire the strobes a couple times before the camera trigger, then fire 8 strobe triggers synced up with the camera after the initial delay for the camera.

Don't have a S1R, but if someone wants to send me one I'm sure I could get this to work when I have some time. :D

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

Lou Jost
Posts: 5945
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

You have as long as you need to charge the flash; you just need to set the shutter speed to a number longer than the flash recycle time. But you will need nearly exact flash intensities at each shot in the group of eight shots. I think this could be hard to accomplish. And you don't have access to those shots to adjust them individually.

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

Lou,

The high speed strobes utilize an IGBT which gates off the ionization current to the Xeon bulb, so the energy storage capacitors don't discharge almost completely like a normal strobe when the ionization extinguishes, the ionization is terminated by the IGBT not a low capacitor voltage. The capacitor voltage drops a small amount, then the recycle begins, this is why I mentioned to fire off a couple flashes before the use of the 8 exposures, so the discharge and recycle will average out somewhat in the prefixes and the 8 exposures will be more uniform.

The "Ready Indicator" on most strobes/speedlights is just a simple capacitor voltage minimum level comparator, it just indicates the capacitor voltage is greater than a preset amount, not that it's regulated at a specific voltage. Why I mentioned you might need to tighten up the comparison limit.

For example: If you have a 600WS strobe at 1/64 power (~ 9WS). The capacitor voltage is usually ~ 400 volts DC, so this equates to 7500uF of capacitance (E=1/2 CV^2). Then dE/dV = CV, so for a 1/64 power flash the capacitor voltage will drop only ~ 3 volts to deliver the flash power. This should recycle quickly. All this tells us that the larger the strobe available output power rating and the lower the required flash power the better.

The Adorama Rapid 600 claims a recycle time as fast as 50ms and optical 0.1T as fast as 51us. This is what I would try and use for this test.

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

Lou Jost
Posts: 5945
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Even if it is not too difficult to arrange for flash in hi-res mode, as people mentioned above it doesn't help stop movement in this application. If there is any movement or vibration, you'll get eight perfectly sharp images that don't match up exactly, and the final computed result will be garbage. IAnd if you have a perfectly stable set-up, then you might as well use continuous light. So I don't recommend using your valuable time for this, Mike.

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

So the camera doesn't do any image alignment!! Had thought that there was some rudimentary sort of quick alignment process going on to create the final rendered image.

Without alignment then the only way this would work is if all the images are perfectly stable wrt vibration and movement. So shutter speed, flash, and all the things we do to prevent vibration from blurring our images is toast!!

Agree, this flash method won't help and the only apparent solution is to have an absolutely stable subject and setup without any sources of vibration getting thru.

Best,
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike

Lou Jost
Posts: 5945
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

That's right, Mike. Indeed it would be hard to do the kind of alignment required and still be able to use an exact computational algorithm to calculate the true values of the quarter-pixel "virtual pixels". Any alignment process is an approximation, with some error, and the error is sure to screw up an algorithm that assumes the subject does not change at all during the exposure. I don't know what the algorithm is, but it probably requires the use of small differences in intensity between shots, and those small differences have to have a potentially large effect on the image, else fine sub-pixel detail would not show up in the hi-res shot. If the small differences are caused by alignment errors, the whole thing will blow up.

I've got tons of hi-res images full of checkerboard artifacts because of vibrations.

chris_ma
Posts: 570
Joined: Fri Mar 22, 2019 2:23 pm
Location: Germany

Post by chris_ma »

mawyatt wrote:So the camera doesn't do any image alignment!! Had thought that there was some rudimentary sort of quick alignment process going on to create the final rendered image.
yes the camera expects the camera to be stable.
there is a mode that if the subject moves the problematic area will be rendered by a single image instead of combining the 8 images (which would lead to artefacts otherwise)

I guess the high res mode was mainly intended for studio and landscape photography, where the stable camera paradigm makes sense (actually there are cameras now which also can do this with handheld shots, but that's because the sensor shift takes into account the camera movement and not the stitching algorithm).

I'd say that's a fair assumption, because if the camera is shaky you'd get distortions due to the rolling shutter which would make things very very hard to align in software.

now very short flash exposures might be a way to get rid of rolling shutter distortion, but who in his right mind would get the idea to shoot 8 exposures pixel shifted half a pixel with flash and then stitch them together in post ;)
chris

Post Reply Previous topicNext topic