Panasonic Lumix cameras & Post Focus

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

Moderators: rjlittlefield, ChrisR, Chris S., Pau

gardenersassistant
Posts: 190
Joined: Sun May 31, 2009 5:21 am
Location: North Somerset, England

Post by gardenersassistant »

Cactusdave wrote:Thanks for the tip about extracting frames Rik. Shame about ffmpeg being command line and not having a GUI. I know its 'enthusiastware', but I do regard the invention of the GUI, along with deodorant and the credit card as one of the great boons of modern life. :lol:
Pot Player will extract JPEGs from 6K post focus videos. It is free and has a GUI.

https://potplayer.daum.net/
Nick

Flickr
Blog
Journey since 2007

Rework and reposts of my images posted in this forum are always welcome, especially if they come with an explanation of what you did and how you did it.

gardenersassistant
Posts: 190
Joined: Sun May 31, 2009 5:21 am
Location: North Somerset, England

Post by gardenersassistant »

gardenersassistant wrote:
TheDocAUS wrote:Proof you can get outstanding images hand held using Post Focus on the G9: https://www.dpreview.com/forums/post/62090097

Looks like I need to work on my hand-held technique - a lot.
No you don't. :) These were "tripod assisted" - tripod used somewhat like a monopod, i.e. with hands on the camera. For rather obscure reasons I decided to keep that post very short (I tend to write rather too much for many people's taste.) I've only used this tripod assisted approach in the last couple of sessions - prior to that it was almost all hand-held (4K, with G80 and FZ330). I haven't done any like for like (hand-held vs tripod assisted) comparisons so I can't say for sure that those are any better than they would have been hand-held.
I did a like for like test this afternoon of eight subjects using 6K post focus with an Olympus 60mm macro on a G9, using available light. I first captured the subject three times using a tripod, hands-on. I then detached the camera from the tripod trying to keep the camera in the same position, and then captured the subject as many times as I felt like trying hand-held (six for most of the subjects).

For each scene I chose one of the three tripod-assisted captures. After the first subject I just chose the third "take" each time because all three looked so similar when I played them (in PhotoFun Studio, which is very convenient for having a quick look at the videos, and you can also get Exif data).

I looked at all of the hand-held captures for the scene and wanted to choose the least wobbly one, but that was tricky - for the most part they wobbled "differently" rather than obviously better/worse. I just tried to avoid the ones with really nasty looking wobbles).

For the chosen tripod-assisted one I loaded it into Helicon Focus and chose which frames to use and stacked it. I then used the same method and parameters for the chosen hand-held one. I didn't do any retouching (I generally do), so the stacks show some imperfections that I would normally try to hide with retouching. I produced TIFF files and then imported them into Lightroom where I pulled the highlights down for some of them (nothing else). I used the same highlight adjustment on both versions. I then exported JPEGs.

There are 1400 pixel high versions of the eight pairs in this album at Flickr. https://www.flickr.com/photos/gardeners ... 7324907448

There are full size versions of the eight pairs in this album at Flickr. https://www.flickr.com/photos/gardeners ... 3852820271

You can tell which stacking method and parameters I used from the end of the filename.

In one case (Example 5) there was a halo that really annoyed me in the hand-held version so I tried it with different stacking parameters and the halo went away. There are JPEGs of both hand-held stacks for that subject.

In one case (Example 7) I got a fairly horrible result for the tripod-assisted version and tried one of the other two "takes" for it, and that was significantly better (still not very good, but that was the case for the hand-held one too. I find that subjects can be quite variable as to whether they will work or not.)

As well as hand-shake, there was a slight breeze, which added some extra randomness to the exercise, much more for some subjects than for others.

My conclusions from this little exercise:

Working hand-held is much quicker and more flexible.

Framing/composition can be more precise (you can "work nearer the edges") with the tripod, and holding the same position for multiple "takes" is physically less taxing (and when you have to wait a while between "takes", waiting for a lull in the breeze, it can become quite taxing).

Working hand-held means you lose the edges of the frame because of the effects of hand-shake, with the size of the loss increasing as the magnification increases.

Working hand-held can produce some artefacts that tripod-assisted doesn't.
[This is a tentative conclusion, I noticed only one example of it (or I assume it is an example of it), the halo in Example 5, but I haven't examined the stacks in great detail. There may be more examples that I haven't noticed. Anyway, the existence of artefacts resulting from hand-shake would make sense of some of my recent experience with easier to process stacking when working tripod-assisted.]

Apart from any extra artefacts that might come with working hand-held, the image quality was really rather similar as between hand-held and tripod-assisted. The sharpness/detail might be slightly better with the tripod-assisted versions if you look very closely, but there doesn't seem to be much in it. (But it you look more carefully than I have then perhaps you will see more of a difference.)

Overall, I'm going to go on using the tripod, hands-on, but I won't be afraid to take the camera off the tripod to reach an awkwardly positioned subject more easily (or at all), or if I feel the tripod is cramping my style and I just feel like working more flexibly for a while.
Nick

Flickr
Blog
Journey since 2007

Rework and reposts of my images posted in this forum are always welcome, especially if they come with an explanation of what you did and how you did it.

zzffnn
Posts: 1896
Joined: Thu May 22, 2014 1:25 pm
Location: Houston, Texas, USA
Contact:

Post by zzffnn »

Very nice work, thank you for sharing!
I need to try more outdoor focus bracketing with my Olympus E-M10 II (which is likely a bit slower than G85's stacking mode). Maybe with an umbrella to shield against wind and diffuse light.

gardenersassistant
Posts: 190
Joined: Sun May 31, 2009 5:21 am
Location: North Somerset, England

CORRECTION

Post by gardenersassistant »

gardenersassistant wrote:
In one case (Example 5) there was a halo that really annoyed me in the hand-held version so I tried it with different stacking parameters and the halo went away. There are JPEGs of both hand-held stacks for that subject.

........

Working hand-held can produce some artefacts that tripod-assisted doesn't.
[This is a tentative conclusion, I noticed only one example of it (or I assume it is an example of it), the halo in Example 5, but I haven't examined the stacks in great detail. There may be more examples that I haven't noticed. Anyway, the existence of artefacts resulting from hand-shake would make sense of some of my recent experience with easier to process stacking when working tripod-assisted.]
I got this wrong. I processed the other two tripod-assisted "takes" of Example 5 and found that they had a similar halo problem, a problem that, as with the hand-held version, disappeared on changing one of the stacking parameters. (For the tripod-assisted stacks the problem was on the opposite side of the branch, and rather subtle and difficult to spot. It was only clearly visible when I alternated between stacks that used the different parameters). I then reprocessed the tripod-assisted "take" that I used for the comparison; it had the same issue as the other two, but even less so than them.

So from looking (somewhat superficially) at the results of this exercise, I don't have any evidence of artefacts stemming from working hand-held rather than tripod-assisted.
Nick

Flickr
Blog
Journey since 2007

Rework and reposts of my images posted in this forum are always welcome, especially if they come with an explanation of what you did and how you did it.

TheDocAUS
Posts: 110
Joined: Thu Jul 19, 2018 10:44 pm
Location: Sydney

Post by TheDocAUS »

Thanks Nick, I am reading these posts with great interest.

descall
Posts: 178
Joined: Sat Oct 01, 2011 3:12 pm
Location: United Kingdom

Post by descall »

I am having problems with the Post Focus feature on the G9 (with Oly 60mm macro). It misses the closest focus planes at the edge of the scene and can miss others in the middle of the scene, leading to focus banding in stacked images (see below, where some of the out-of-focus planes are arrowed). Note, the centre arrows below point to areas inside the inner edge of the nearest cork, not to the expected halo beyond. Has anyone had similar problems? Thanks for any thoughts. Cheers, Des

Image stacked in Helicon Focus using 'open video' option and 'Method C'.
Image

palea
Posts: 40
Joined: Mon Dec 17, 2018 8:09 pm

Post by palea »

Hi Des, post focus steps through the focus distances found at the focus points it analyzes. If the subject rapidly changes in depth within a small angle of view no point will occupy an intermediate depth, so those depths are skipped (though often I get a frame or two as the lens is moving to the next depth). It sounds like you might be expecting the sequence of fixed steps from a near focus point generated by focus bracketing. If so, autofocus bracketing a static subject like this would likely be a more suitable choice.

The matrix of focus points is fairly dense but my G7 doesn't pick up the very edges of the frame either. I find it difficult to tell from the small image posted but it looks like the band you're seeing is wider than that. In general, focus breathing means deeper frames in the stack are acquired at longer focal lengths and therefore appear smaller. Image alignment therefore squishes shallow frames, which can leave a border of artifacts as alignment isn't necessarily smart enough to realize the output and input image sizes are different. I haven't used Helicon in years---and when I did it was mainly below 0.5x---but this might be what you're seeing. I just crop the smeared border in post.

One reason I stopped using Helicon is, at the time, its user interface made reviewing and managing the frames going into the stack very tedious. Shredding the post-focus video file to jpegs and reviewing in any more performant tool is faster plus I've found it helpful in understanding what information the stacker has available to work with.
Last edited by palea on Sat Jan 19, 2019 1:07 pm, edited 1 time in total.

TheDocAUS
Posts: 110
Joined: Thu Jul 19, 2018 10:44 pm
Location: Sydney

Post by TheDocAUS »

descall

The edge is missed, because there are no AF points there.

The solution is to compose the image about 5-10% bigger, so AF points touch all areas you need in focus.

If you watch the rear screen during Post Focus you can see where there are no AF points.

The other issue is not so easy, as the image you posted is so small I cannot properly see the issue. You need at least 800x600 for us 4 eyes.

I am currently writing up some issues I have found and will post in due course, once finished.

ChrisR
Site Admin
Posts: 8662
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

I tried to find out how many focus points the G9 has. Couldn't find it, only
"Focus Bracket 1 to 999 images, focus steps can be set in 10 levels".

Maybe if you tried finer/more steps?
My compact Lumix doesn't focus in depth order, it hops about according to each point's position. There are few enough points that something could definitely be missed.
Chris R

Lou Jost
Posts: 5933
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Chris, note that "focus bracket" is different from "post focus" and does not use any focus points, it just goes through all distances from near to far in calculated steps, with user-selected overlap between steps (1-10). Post focus jumps around to catch different autofocus points.

palea
Posts: 40
Joined: Mon Dec 17, 2018 8:09 pm

Post by palea »

ChrisR wrote:I tried to find out how many focus points the G9 has.
Hi Chris, the notion of a focus point is more of a PDAF DSLR thing. It breaks down somewhat with OSPDAF (increasingly so in recent generation mirrorless). The current PDAF culmination of this trend is Canon dual pixel, where every pixel/pixel pair provides vertical focus information, and presumably it'll be extended to the other axis eventually. CDAF implementations have always worked with individual pixel to pixel differences. As such, the G9 notionally has about 20M focus points since a contrast calculation patch can be centered on any pixel (though the partial patches at the edges of the sensor might not be so great).

If Panasonic has disclosed how many patches are evaluated in the post focus process I've not come across the information but it appears to be comparable to the number of focus box positions. I don't know this number for the G9 but on my G7 it's about 6600. The on screen display during the focus process is more coarse, which I presume is partly due to slowing down the progress updates so the user can see what's going on rather than having it all be a blur because it's happening at 30+fps. It's also probably due to merging of focus point distances into DoF slabs.

From what I can tell, the evaluation of contrast patches in post focus seems to be based around a medium focus box size and its limitations at the frame edges are due to exclusions of partial boxes and the somewhat large box size. With Panasonic lenses my experience suggests there's some interaction with DFD information and the focus abilities of the lens. I'm not sure anyone's particularly looked for similar interactions on the Olympus side.
ChrisR wrote:My compact Lumix doesn't focus in depth order, it hops about according to each point's position.
The G series implementation does a pre-scan and then sweeps the points in depth order. It's my understanding GX bodies do the same.

ChrisR
Site Admin
Posts: 8662
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

In that case it's like the compact. So if there aren't enough points, you just have to hope your DOF is adequate to cover between them. How many frames in the stack then, Des?


Edit:

This post was poster before I'd seen the one above
Last edited by ChrisR on Sun Jan 20, 2019 2:41 am, edited 1 time in total.
Chris R

TheDocAUS
Posts: 110
Joined: Thu Jul 19, 2018 10:44 pm
Location: Sydney

Post by TheDocAUS »

ChrisR wrote:In that case it's like the compact. So if there aren't enough points, you just have to hope your DOF is adequate to cover between them. How many frames in the stack then, Des?
With Post Focus the lowest I have seen is 30, the highest 330 frames (so far).

ChrisR
Site Admin
Posts: 8662
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

Thanks for the clarification Palea. I'll review how the compac Lumix "seems to" work. I do know it messes things!
Chris R

Jim Haseloff
Posts: 2
Joined: Sat Jan 19, 2019 11:41 pm
Location: Cambridge, UK
Contact:

Post by Jim Haseloff »

Really enjoying the discussion - and interested in the shift to 6K mp4 collection of extended focus stacks. I've been using extended focus to image small liverwort plants, an offshoot of our research in Cambridge, where microscopy plays an important role in integrating genetic, cellular and tissue-scale behaviour in plants.

So far, this has meant 4K video capture on the Panasonic GX80, with Olympus 60mm Macro lens, extension tubes and Raynox macro adapter lenses. (More details at: https://www.hackster.io/jim-haseloff/vi ... ope-6b4f76 and https://data.plantsci.cam.ac.uk/Haselof ... index.html. I'm always suprised that there is not more chatter (like this discussion) about the usefulness of the Panasonic extended focus/video modes among macrophotographers.

Given the large array and combination of different lens systems, lighting, and camera bodies, it would be good to identify some kind of standard, readily accessible target object(s) that we could all use for comparison work. Any ideas? - perhaps this has been well-canvassed already somewhere in the forum... I saw that someone here had used a burnt match head as a target - very accessible. Perhaps placing a match head next to the edge of a millimetre ruler might work as a standard? - or something similar, commonplace, with a bit of texture and detail.

I'm still playing around with different lenses, spacing and adapters - so a better reference would be useful. Also, I'd like to make the image capture as portable as possible for field work. I've been using a Vanguard Alta Pro 2+ 263AT tripod with GH-100 pistol grip ball & socket head - which is great, but it should be possible to shrink this - I'm looking at adding short boom arms to mini-tripods. At the other end of the spectrum, there seem to be some opportunities for better "studio" use with cameras that are supported by the Panasonic tethering software. It would be great to better automate the image collection process - so that we can start building extended focus time-lapse movies - with the prospect of HDR and other imaging modalities...

Thanks again for the interesting discussions!

Jim Haseloff
Twitter: @jimhaseloff


Image

Image

Post Reply Previous topicNext topic