Forcing a stack order in Zerene

A forum to ask questions, post setups, and generally discuss anything having to do with photomacrography and photomicroscopy.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

specious_reasons
Posts: 107
Joined: Tue Feb 24, 2009 12:46 pm
Location: Woodridge, IL

Forcing a stack order in Zerene

Post by specious_reasons »

I have a stack of Radiolarian, and I'm not entirely happy with the PMAX version of it. I'm playing with the DMAP settings and I think that Zerene is choosing the wrong order of the images.

Yet, I didn't find a way to force it to keep the order I designated. Is there a way to force it to keep my ordering?

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Go to Options > Preferences > Alignment and remove the checkmark on "Automatic order". This will prevent ZS from reversing the stack based on its analysis of field-of-view and image scale.

If you need to set the order very precisely, then you may also have to add the images one at a time or in small groups, using drag-and-drop or Add File(s) multiple times. Each time you drop into the Input Files panel or click the Add button inside the Add File(s) dialog, ZS will sort the newly added files by name.

By removing the checkmark on "Automatic order" and adding files one at a time, you can set exactly any order that you want.

--Rik

elf
Posts: 1416
Joined: Sun Nov 18, 2007 12:10 pm

Post by elf »

rjlittlefield wrote:Go to Options > Preferences > Alignment and remove the checkmark on "Automatic order". This will prevent ZS from reversing the stack based on its analysis of field-of-view and image scale.

If you need to set the order very precisely, then you may also have to add the images one at a time or in small groups, using drag-and-drop or Add File(s) multiple times. Each time you drop into the Input Files panel or click the Add button inside the Add File(s) dialog, ZS will sort the newly added files by name.

By removing the checkmark on "Automatic order" and adding files one at a time, you can set exactly any order that you want.

--Rik
Is there any requirement for the images to be in the order of focus depth (either in or out)?

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

For DMap, it's definitely best that the frames be in order, either front to back or back to front. If they're out of order, then the result may still look OK at first glance but it's almost certainly degraded and often there will be noticeable artifacts.

PMax itself does not care about order, but going from one end straight to the other is still the safe way to get good alignment. Sometimes when source frames are presented out of order, enough error accumulates in the alignment process that two frames focused at nearly the same depth but far apart in the stack order end up aligned slightly differently. When that happens, you get echoes.

--Rik

specious_reasons
Posts: 107
Joined: Tue Feb 24, 2009 12:46 pm
Location: Woodridge, IL

Post by specious_reasons »

Thanks, I don't know why I didn't find it.

We'll see... I have a (demo) copy of Helicon Focus, and these radiolarian pictures are the one thing that makes a better looking stacked image...

I'm trying to tweak Zerene to get similar results.

Here's HF:
Image

same stack in Zerene (Pmax)
Image

Oddly, it's only the 40X objective which gives me trouble, this is my 20X objective Zerene (Pmax):
Image

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

You'd probably like better what you get from DMap.

First, go to Options > Preferences > and set a fairly wide Estimation Radius, say around 40 if you're processing full size 10-15 megapixel images. Set Smoothing Radius at half the value of Estimation Radius (so the two sliders are lined up, one above the other).

Then run Stack > Align & Stack All (DMap).

When the method pauses midway for you to set the Contrast Threshold, run the slider so that the background turns black over most of its area, but most of the subject retains its original appearance, something like this:

Image

This should give you a final output without those nasty broad halos around the subject.

Most likely the reason the 40X objective is giving you grief is that something about its bokeh is producing harder-edged blurs. PMax turns those into inverted halos, and with that lovely featureless background, the halos are what propagate into the output.

By the way, I think the outputs you're showing don't come from quite the same set of input files. There's no way that PMax would have missed that sharply defined black dot just off the end of that right-hand spine. But it doesn't matter -- the principles are the same.

I'm guessing that the stack fed to PMax might have included more of the back of the specimen. I'm seeing a lot more detail "in the holes" with ZS PMax than with Helicon, and while it's common for HF to miss detail that PMax picks up, this stuff looks clear enough that I would not have expected HF to miss it.

Many people prefer to see just the front sides of their radiolaria, easily accomplished in either tool by feeding in a subset of the frames, or in ZS with Stack Selected.

--Rik

Edit: To correct comments about the 40X objective. What's important about the OOF blur is more the hardness of its edge, not its size.
Edit2: fix more silly confusing typos.
Last edited by rjlittlefield on Mon May 02, 2011 8:59 pm, edited 1 time in total.

specious_reasons
Posts: 107
Joined: Tue Feb 24, 2009 12:46 pm
Location: Woodridge, IL

Post by specious_reasons »

rjlittlefield wrote: By the way, I think the outputs you're showing don't come from quite the same set of input files. There's no way that PMax would have missed that sharply defined black dot just off the end of that right-hand spine. But it doesn't matter -- the principles are the same.
Oh, yes, well, I edited out all the background noise on the Pmax version of that image... :oops: Same sources, one's much more heavily edited. Good attention to detail.

Thanks for the advice, I'll be trying it out.

specious_reasons
Posts: 107
Joined: Tue Feb 24, 2009 12:46 pm
Location: Woodridge, IL

Post by specious_reasons »

I'm not sure what Helicon Focus is doing, but is does seem to be ignoring the lower layers a lot more that Zerene.

FYI:

Radius of 40, smoothing 20, full 35 image stack:

Image

The radiolarian is at a bit of an odd angle, so it's hard to cull out all the bottom, but I dropped the lowest 5. Radius of 40, smoothing 20, full 35 image stack:

Image

BTW, Zerene is performing quite nicely on my i7 950 with 4G of memory.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

specious_reasons wrote:I'm not sure what Helicon Focus is doing, but is does seem to be ignoring the lower layers a lot more that Zerene.
I don't know either. I don't normally think of ignoring detail as being a good feature in stacking, but with radiolaria I guess it can have advantages. One of the interesting features of PMax is that it's relentless about preserving detail, so even if nothing from the PMax image is useful directly, it can still serve as good guidance about where to looking for stuff that might be retouched from source or "dialed in" by adjusting DMap parameters.
The radiolarian is at a bit of an odd angle, so it's hard to cull out all the bottom, but I dropped the lowest 5.
This would be a good place to exploit Stack Selected and retouching. Run the full stack, then run a partial stack like you have here. Retouch between the two outputs to restore that backmost spine while still keeping just the front of the body. What you end up with is sort of an "artist's conception", but that's OK in this case because I think what you want is an image that conveys an idea well and doesn't require a lot of painting skills to put together.
BTW, Zerene is performing quite nicely on my i7 950 with 4G of memory.
Glad to hear. It's always possible I could do something that makes it get worse with more resources.

--Rik

Post Reply Previous topicNext topic