Panasonic 7-14 - Usage tips for UWA neophyte

Started Apr 19, 2014 | Discussions thread
Anders W Forum Pro • Posts: 21,466
Re: Panasonic 7-14 - Usage tips for UWA neophyte

Michael J Davis wrote:

Anders W wrote:

Michael J Davis wrote:

Anders W wrote:

Thanks! I have tested though regrettably not in advance. Had do to some trial and error out there in the desert.

For the shot I posted (7 mm on MFT), the formulas would give us 600 / (7 x 2) = 43 s or 500 / (7 x 2) = 36 s, where the latter value comes pretty close to the 30-second rule of thumb I gave Florida.

Now let's look at the evidence produced by some of my trials, the first at 15 s, the second at 30 s, and the third at one minute. These images are downsampled to about 2.6 MP but still reasonably large.

One problem that you can readily see is that the UWA distortion makes things worse toward the edges than in the center. That distortion makes the stars a bit oblong even without any motion. At 15 s, I would say things still look pretty much OK unless you look to close and there are no problems at all in the center. At 30 s, we are at the border line, and at one minute we have clear signs of trails even in the center.

So by and large, the formula you provided seems to work pretty well. However, UWA stretching towards the edges complicates things a bit and may give us reason to be a bit more conservative.

For the record, (and understanding that the posted shots are < half size) I 'corrected' the star trails in the last shot using Focus Magic's remove motion blur. It worked well, and better if I divided the scene into two areas corresponding to different angles of 'movement'. FM will consolidate up to 20 pixels of blur in one application. In this case, the fact that the clouds are already blurred disguised any other artifacts that might be introduced.

Just a contribution to the minimising noise at lower ISOs thought.

Hi Mike,

Very interesting! Although I am well aware of Focus Magic (without yet having tried it), I must admit that I hadn't come to think of using it for this particular purpose (getting rid of unwanted star trails). If you don't mind, please post the result of your little retouching effort so that we can compare it with the original. Not that I don't trust your verbal description but if you have a convincing case to make, as you might well have, seeing the evidence will just make it even more convincing.

Star area corrected with Focus Magic in PSE.

My pleasure Anders! I used PSE because this was already a jpeg and I could apply FM within a selection. So I chose two areas the main one centering on the bright central star which required about 7 pixels at 45deg, (sorry, I didn't record it) and the lefthand side where it was about 13 pix at 30deg. As I say, by feathering the selection some 20pix above the horizon it wasn't obvious where the clouds were affected.

That looks good enough for me. Even if you can't take this too far, getting a couple of stops extra in a situation like this is more than welcome.

You said:

On top of that, I'd be very interested in getting your mini-review of what Focus Magic is good for (and not). I know it's good at dealing with subject motion and camera shake (perhaps the best app currently available for that). But how about ordinary sharpening? I am not perfectly happy with how that works in LR (without being sure there's something significantly better out there) and Focus Magic is one of the options I have been considering in that area. But I haven't got around to trying it yet, in part because the information I have seen hasn't made it all that clear what, if anything, it has to offer over LR in that department.

I've been using FM for some 4 years now, and I find that it is often the only sharpening that I apply. So many sharpening algorithms are 'tacky' and dependent upon edge-effects,they are also useless in correcting camera shake, shutter shock, and subject movement; i.e. all directional blurring.

Before FM I used, at first USM to remove anti-alaising filter effects with 3-6 Mp cameras, then I found Hi-pass sharpening, for immediate sharpening prior to printing. But now, I find that a touch of FM before printing (or web posting), improves the clarity significantly. It is especially effective at the top end of my (already good) 45-200mm panny, revealing details that aren't fully obvious at the pixel level.

Here's an example of a straight uninteresting shot at 200mm with simple adjustments in LR5 cropped to jpg in PSE; and the same shot with a simple 2 pixel application of FM. Notice the way the details on the sheep's wool jump out, plus the clarify of the DoF on the grass. View at 100%.

That looks good too. Do I understand you right that these are full-res (100%) crops?

The problem I have with LR sharpening is that whenever I try something other than the default I end up being dissatisfied with something (noise, artifacts, halos). I may work reasonably well at certain viewing distances but ... And I think I have played around long and carefully enough by now with various combinations of the four sliders (amount, radius, detail, masking) without finding anything beyond the default that I am really happy with (although some things work better than others, and I still deviate from the default at times, although I usually leave it there).

Then you have the clarity slider, which is a sort of high-pass sharpening if I understand things correctly. I sometimes use that to good effect but it's not the same thing.

200mm f5.6 ex LR

same shot with 2pixel 'blur' correction in FM on LR TIFF file.

AIUI, in focus mode it works by using a deconvoluting algorithm - inverting the equivalent of say diffraction, but you can read that on their web site.

Yes, I saw that. And it sounds promising of course.

I find the movement blur correction v useful when someone just moves their head in a portrait; in PS one can select the area affected and apply 2-4 pixel directional correction. Only if movement is linear in both senses (velocity and direction) can movement be disguised above this - hence your star trails work up to around 12 pix. Common camera shake is not directionally linear - often more of an L, J, or S shake, and this would require far more complex algorithms to correct! Shake is also of varying velocity and therefore even more complex. (That's why the oscilloscope displays of SS were interesting...!)

You are probably right that camera shake is more difficult to come to terms with in many cases.

While we're off the subject of this thread - note the DoF in the sheep shots. I think that the DoF in the first shot looks greater than the second - an illustration of another point we have discussed!

What I can agree is that the sharpening reinforces the subject isolation you get through the blurred background. And as long as you put it that way, I have no trouble agreeing with you.

But I don't feel the (close to non-existent) DoF extends further in the first crop than in the second, I don't think increasing the sharpening has the same impact on the OOF areas as using a sharper lens (which is what we were discussing then), and I think certain things might work a bit differently if you have some DoF to speak of to begin with, as in a WA landscape shot.

Hope you find that interesting. Suggest we start a new thread if you've any questions!

Yes, I'll do that or pester you with a few questions via PM if I come up with any and provided you don't mind.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
Post (hide subjects) Posted by
yyr
yyr
yyr
yyr
yyr
yyr
yyr
yyr
yyr
yyr
(unknown member)
(unknown member)
(unknown member)
(unknown member)
yyr
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow