Moon with pixel shift resolution


RobL

Link Posted 13/03/2017 - 10:11
Last night I took a few photos of the moon with a K-1 and 150-450mm lens at 450mm, and experimented with PSR. Here are a couple of examples at the same exposure of F16, 1/320sec ISO 100 cropped and boosted in Lighroom, first with PSR and the second without, both on a tripod with remote release and autofocus:



Whilst the PSR shot has a bit more detail there is a strange artefact on the periphery in the bottom left and top right quadrants, caused either by the camera shifting slightly on the mount or the motion of the moon in that brief moment of exposure. The same thing happened at different exposures, Any ideas?

johnriley

Link Posted 13/03/2017 - 10:27
I don't know that this subject is the best to try with PSR as no matter how much the camera is kept still the moon is moving quite rapidly across the sky. Perhaps the artefact is just that movement not being compensated for by the software.
Best regards, John

Mag07

Link Posted 13/03/2017 - 11:00
Interesting. Bit hard to judge as the non PS one appears to be lighter.

Given how the moon constantly 'moves', have you thought of actually using f5.6 - f8 and a higher shutter speed to compensate a bit more? From my experience there is absolutely no visible benefit of using anything higher then f8 unless a particular lens is known not to perform too well at certain apertures but with your beauty, that is not an issue I'd think the quicker the pixel shift magic happens the better? I have been contemplating renting that lens for couple of days just to shoot the moon in PSR with the purpose of further stacking a dozen or two of exposures. Simply curious how much detail one can get that way.
'Photography...it remembers little things, long after you have forgotten....' (Aaron Siskind)

RobL

Link Posted 13/03/2017 - 11:32
Thanks, I didn't see the problem until it was on the computer screen next day. I sort of guessed that the time would be fast enough for no movement to be registered but was wrong; I started off with a much lower exposure and shifted speed and aperture until I could see details on the rear screen. F8 should have been optimum but was not so sharp so I stopped down. Mag07, the difference is probably down to me not exactly matching settings in Lightroom.

JAK

Link Posted 13/03/2017 - 11:42
Does Lightroom handle PSR raw images now (or were they jpegs?) I thought PSR raw development only worked correctly with PCU5.
John K

Mag07

Link Posted 13/03/2017 - 12:05
JAK wrote:
Does Lightroom handle PSR raw images now (or were they jpegs?) I thought PSR raw development only worked correctly with PCU5.

That's what I thought. Also, I read a comparison somewhere between using DNG and Pentax RAW format for PixelShift and the conclusion was that it's optimal to use Pentax RAW; granted, it may just be lucky coincidence or opinion based on a small sample but may be worth a shot?
'Photography...it remembers little things, long after you have forgotten....' (Aaron Siskind)

RobL

Link Posted 13/03/2017 - 13:25
Ah, that's a point. These were RAW .dng out the camera and Lightroom treated the PSR image just as normal, but maybe better results with the Pentax software which I haven't used yet.

pschlute

Link Posted 13/03/2017 - 13:32
I dont think it is in focus. For the moon i would suggest you try again but use manual focus, in live view with 100% view. You may need to turn down the lcd brightness setting (fx2 button). Also enable electroninic shutter to reduce any shake.

Here is the moon with my k1 and DA* 300mm (ignore the Clangers at the top)


Peter



My Flickr page

richandfleur

Link Posted 13/03/2017 - 13:37
Early on Lightroom couldn't deal with the pixel shift files, and a lot of artifacts were seen. Not sure if this was ever addressed.

Early pixel shift processing couldn't handle any movement, but this was addressed later on. Basically where they detect motion between shots, they just turn off pixel shift in that area, and use only one frame to provide the info.

richandfleur

Link Posted 13/03/2017 - 13:40
pschlute wrote:
I dont think it is in focus. For the moon i would suggest you try again but use manual focus, in live view with 100% view. You may need to turn down the lcd brightness setting (fx2 button).

Setting the metering mode to spot works on my camera, even if I'm using full manual mode for my actual shot. The auto metering mode selected impacts on what the rear LCD screen displays, on my K-30 at least.

RobL

Link Posted 13/03/2017 - 15:50
Just tried the image in Utility 5, I don't know if my computer was having a moment but it was incredibly slow to show the edit changes; the image in Utility 5 kept clicking from a clean edge to one that looked like a double exposure. Saving to a jpeg took about ten minutes but that reintroduced the colour abherration so I couldn't see any advantage over Lightroom. Whilst I had the PSR on motion control setting I guess the moon's motion is the cause of the problem. Thanks for all your comments.

McGregNi

Link Posted 13/03/2017 - 19:24
I agree that the moon is probably not a candidate for pixel shift, due to it being a moving subject.

I was surprised when I tried moon shots last summer .... With 300mm plus 2x TC at F9.5, I found it needed at least 1/125th sec for sharpness. That was on a tripod with remote control and mirror up release also, so as stable as I could achieve.

I wonder, can we shoot at any exposure time with PS? Or is there an optimum range of time values to bear in mind?
My Guides to the Pentax Digital Camera Flash Lighting System : Download here from the PentaxForums Homepage Article .... link
Pentax K7 with BG-4 Grip / Samyang 14mm f2.8 ED AS IF UMC / DA18-55mm f3.5-5.6 AL WR / SMC A28mm f2.8 / D FA 28-105mm / SMC F35-70 f3.5-4.5 / SMC A50mm f1.7 / Tamron AF70-300mm f4-5.6 Di LD macro / SMC M75-150mm f4.0 / Tamron Adaptall (CT-135) 135mm f2.8 / Asahi Takumar-A 2X tele-converter / Pentax AF-540FGZ (I & II) Flashes / Cactus RF60/X Flashes & V6/V6II Transceiver

RobL

Link Posted 13/03/2017 - 19:42
McGregNi wrote:
I agree that the moon is probably not a candidate for pixel shift, due to it being a moving subject.

I was surprised when I tried moon shots last summer .... With 300mm plus 2x TC at F9.5, I found it needed at least 1/125th sec for sharpness. That was on a tripod with remote control and mirror up release also, so as stable as I could achieve.

I wonder, can we shoot at any exposure time with PS? Or is there an optimum range of time values to bear in mind?

Yes as long as there is no movement over the total time of 4x the exposure speed; if the setting is 1/100 sec then the total exposure will be 1/25 sec. I don't know how effective the movement compensation is, or whether that has some adverse effect.

davidwozhere

Link Posted 14/03/2017 - 01:51
I used pixel shift to shoot the underside of a leafy canopy last autumn. The results were tack sharp except that one or two of the branches were duplicated, side by side. The wind had moved them ever so very slightly during the exposure.
Both the *istDS and the K5 are incurably addicted to old glass

My page on Photocrowd - link

RobL

Link Posted 14/03/2017 - 07:39
So, to freeze movement in PSR I assume that whatever shutter speed would normally suffice then divide by four: say 1/500 sec would require a setting of 1/2000 sec. If the ISO is increased by four stops to suit I wonder if the noise will cancel any benefit? Further experimentation needed.
Add a Comment
You must be registered or logged-in to comment.