Please Teach Me Sharpening

Here you can discuss topics such as focus stacking, stitching and other techniques that relate to the processing of micrographs.
Message
Author
User avatar
rnabholz
Posts: 3066
Joined: Thu Jan 08, 2015 10:11 pm
Location: Iowa USA
Contact:

Please Teach Me Sharpening

#1 Post by rnabholz » Sat Jun 04, 2016 3:05 pm

Ok everyone, I need your help please.

It is in regards to what level of sharpening is reasonable, desireable, enhances instead of detracts, etc.

I have generally had no issue in more conventional photography with the process, but I am finding that as I push the capabilities of the camera and microscope optical systems further towards their limits, that I am not sure where the boundary is as far as using these techniques to get the most from these images.

Toward that end, below you will find a recent shot captured with a 100x Oil in darkfield (DF which poses even more of a challenge IMHO). It is straight from the camera, a Canon 70D, just resized to conform with the forum size limits - no other changes. The subject was about 54 microns in length.

EDIT: At the mention below that it may be helpful to start with a full size image, I have made it available for download. It is a 3.5mb JPEG. You may get it here:

http://www.homebuiltastronomy.com/Sharp ... alFull.JPG

I am requesting and inviting you to download the image, apply your skills to sharpening the image and repost it in this thread. Please include the software used, steps taken, and specific parameters where appropriate (like Unsharp Mask Radius, Amount and Threshold).

I hope all of you will consider joining in and taking a shot. The worst that can happen is that we all learn something.

Thanks

Rod
Attachments
SharpeningTrial.JPG
SharpeningTrial.JPG (265.14 KiB) Viewed 9078 times
Last edited by rnabholz on Sat Jun 04, 2016 6:39 pm, edited 2 times in total.

User avatar
zzffnn
Posts: 2710
Joined: Sat Jun 20, 2015 3:57 am
Location: Houston, Texas
Contact:

Re: Please Teach Me Sharpening

#2 Post by zzffnn » Sat Jun 04, 2016 3:17 pm

Rod,

I am not qualified to teach you on that and don't think I can make that image much better. That dense diatom is likely a challenge for any darkfield rig and any operator. Some dense/thick subjects just don't work well in darkfield, as you already knew, it is not unlike phase in that aspect.

IMHO, the only possible way to improve on that diatom is mounting it better and trying different iris setting, which you already did your best there.
Selling my Canon FD 200mm F/2.8 lens

User avatar
75RR
Posts: 6243
Joined: Sun Oct 12, 2014 2:34 am
Location: Estepona

Re: Please Teach Me Sharpening

#3 Post by 75RR » Sat Jun 04, 2016 4:06 pm

I think you will continue to have problems with sharpening while you continue to over magnify your images.
On my monitor your diatom is 26cm, if actual size is 50µm then 26cm/0.005mm = 5200
My scale images: viewtopic.php?f=6&p=27920#p27809 are also 50µm wide but only 10cm on the screen i.e. 10cm/0.005mm = 2000
Zeiss Standard WL (somewhat fashion challenged) & Wild M8
Olympus E-P2 (Micro Four Thirds Camera)

User avatar
rnabholz
Posts: 3066
Joined: Thu Jan 08, 2015 10:11 pm
Location: Iowa USA
Contact:

Re: Please Teach Me Sharpening

#4 Post by rnabholz » Sat Jun 04, 2016 4:19 pm

zzffnn wrote:Rod,

I am not qualified to teach you on that and don't think I can make that image much better. That dense diatom is likely a challenge for any darkfield rig and any operator. Some dense/thick subjects just don't work well in darkfield, as you already knew, it is not unlike phase in that aspect.

IMHO, the only possible way to improve on that diatom is mounting it better and trying different iris setting, which you already did your best there.
Thanks zz.

User avatar
rnabholz
Posts: 3066
Joined: Thu Jan 08, 2015 10:11 pm
Location: Iowa USA
Contact:

Re: Please Teach Me Sharpening

#5 Post by rnabholz » Sat Jun 04, 2016 4:21 pm

75RR wrote:I think you will continue to have problems with sharpening while you continue to over magnify your images.
On my monitor your diatom is 26cm, if actual size is 50µm then 26cm/0.005mm = 5200
My scale images: viewtopic.php?f=6&p=27920#p27809 are also 50µm wide but only 10cm on the screen i.e. 10cm/0.005mm = 2000
Thanks 75, I see your point.

I did not mean to rule out resizing as an option in the sharpening process, so with your good point in mind you or anyone would like to take a run at it, I welcome consideration of the parameter as part of the process.

Thanks

Rod

billbillt
Posts: 2885
Joined: Mon Nov 10, 2014 10:01 pm

Re: Please Teach Me Sharpening

#6 Post by billbillt » Sat Jun 04, 2016 4:30 pm

Hi Rod,

Here is my attempt at sharpening with my very limited ability with this... I can't tell a lot of difference..
Attachments
SharpeningTrial 1.jpg
SharpeningTrial 1.jpg (165.96 KiB) Viewed 9056 times

User avatar
KurtM
Posts: 1514
Joined: Tue Jun 09, 2015 12:08 am
Location: League City, Texas
Contact:

Re: Please Teach Me Sharpening

#7 Post by KurtM » Sat Jun 04, 2016 4:50 pm

Here's my entry, using the Light Zone Relight tool to remove fuzz, and and then the Sharpen tool. I got pretty agressive with it because I want to see what's in that blue field, is it striation or punctua? Not sure the question got answered, but a little more information was coaxed out anyway. Results would be much better starting with full size photo file.
Attachments
SharpeningTrial_lzn.jpg
SharpeningTrial_lzn.jpg (166.56 KiB) Viewed 9051 times
Cheers,
Kurt Maurer
League City, Texas
email: ngc704(at)aol(dot)com
http://sawdustfactory.nfshost.com/microscopes/
https://www.flickr.com/photos/67904872@ ... 912223623/

User avatar
zzffnn
Posts: 2710
Joined: Sat Jun 20, 2015 3:57 am
Location: Houston, Texas
Contact:

Re: Please Teach Me Sharpening

#8 Post by zzffnn » Sat Jun 04, 2016 5:11 pm

I prefer Kurt's version. Kurt, did you adjust contrast? Please try it, if you have not.

Does Kurt's get closer to the visual view, than other photos in this thread, Rod?

Also I forgot to mention:
Did you try different and individual exposure levels, without HDR? I think I use my camera at -2 EV with darkfield and point compensation. Overexposure (white color) can drown out tiny dots and lines.
Selling my Canon FD 200mm F/2.8 lens

User avatar
rnabholz
Posts: 3066
Joined: Thu Jan 08, 2015 10:11 pm
Location: Iowa USA
Contact:

Re: Please Teach Me Sharpening

#9 Post by rnabholz » Sat Jun 04, 2016 5:18 pm

billbillt wrote:Hi Rod,

Here is my attempt at sharpening with my very limited ability with this... I can't tell a lot of difference..

Thanks for the submission Bill. Can you tell me what you used to do it and what steps you used. Folks may be interested. Thanks again.

Rod

User avatar
rnabholz
Posts: 3066
Joined: Thu Jan 08, 2015 10:11 pm
Location: Iowa USA
Contact:

Re: Please Teach Me Sharpening

#10 Post by rnabholz » Sat Jun 04, 2016 5:23 pm

KurtM wrote:Here's my entry, using the Light Zone Relight tool to remove fuzz, and and then the Sharpen tool. I got pretty agressive with it because I want to see what's in that blue field, is it striation or punctua? Not sure the question got answered, but a little more information was coaxed out anyway. Results would be much better starting with full size photo file.
Thanks Kurt, it looks good.

What does the "relight tool" actually do, you mentioned removing fuzz? Is a levels tool that reduces halos? And you did just straight shapening, not unsharp mask?

Your point about the possibilities being different starting with a full size file is a good one. I have edited the opening post to give the option for downloading the original file for those who would like to start from there. It is available here:

http://www.homebuiltastronomy.com/Sharp ... alFull.JPG

Thanks for participating, and feel free to take a run at the big file

User avatar
rnabholz
Posts: 3066
Joined: Thu Jan 08, 2015 10:11 pm
Location: Iowa USA
Contact:

Re: Please Teach Me Sharpening

#11 Post by rnabholz » Sat Jun 04, 2016 5:27 pm

zzffnn wrote:I prefer Kurt's version. Kurt, did you adjust contrast? Please try it, if you have not.

Does Kurt's get closer to the visual view, than other photos in this thread, Rod?

Also I forgot to mention:
Did you try different and individual exposure levels, without HDR? I think I use my camera at -2 EV with darkfield and point compensation. Overexposure (white color) can drown out tiny dots and lines.
Yes. I think Kurt's does show more of what I was seeing at the eyepiece.

Regarding exposure, yes, shooting darkfield I find an exposure that does not blow out the highlights, and usually bracket at least 2 stops on both sides. Generally the underexposed side is gives the best result.

HDR is disabled when I am shooting with the 70D.

Thanks zz

Rod

User avatar
gekko
Posts: 4701
Joined: Sun Oct 12, 2014 7:38 am
Location: Durham, NC, USA.

Re: Please Teach Me Sharpening

#12 Post by gekko » Sat Jun 04, 2016 5:46 pm

Rod, having seen your bird pictures, I would say that you know more about this than most people, including me of course. However, I'll bite and see what I can do, probably no better than any of the people who tried it.
Having said that, I will make a few observations that in my experience help (or at least I think theyhelp) and some of them have nothing to do with this particular issue. What I say may be heresy to some, including experts who know much more about digital imaging than I, but this has been what I learned, including from my own experience. Much of this is likely what every one does anyway, and some of it has been described by others above, but I wanted to put all my thoughts together.
I would like to stress strongly that none of this is "expert" advice, and clearly the images I produce are nowhere near as good as most of the members either occasionally or routinely produce, including, for example, seb Vasselle, 75RR, JimT, John B, and others that don't come to mind right now.

1. When taking a picture, I sometimes note any colors that are rather different from what I'm used to (e.g., alga that are more yellow than green or more blue than green, and the like) so in case the colors come out obviously different I can try to adjust them.

2. If you are using the RAW file, this point doesn't apply to you. I nowadays save both the RAW and the jpeg files. I use the jpeg unless it is way overexposed or underexposed. I set the camera noise filter to OFF. My camera has three levels of noise filtering: Standard, Low, and Off. I found that setting it to Off results in sharper pictures to begin with. In case I have to use very high ISO, I try to reduce the noise as necessary during post processing, and usually only in certain areas (background, mostly) because when there is much detail in the image, the noise is usually not too bothersome. In other words, some noise and sharp is I think better than no noise and not sharp. Sharpening after too aggressive noise filtering does not result anywhere in as good an image as not filtering noise and a little sharpening, I think. Also, and this too is applicable only to the jpeg file, I set the camera options for sharpness, and contrast to -1 (range -2 to +2) and saturation to 0.

3. In computer science, they used to say "GIGO" (garbage in, garbage out). If an image is not sharp to begin with (bad focus or blur due to shake), it goes into the trash can.

4. Avoid overexposure at all costs, and try to avoid underexposure. If it is not possible to avoid both, underexpose. Pixels that are overexposed are pure white and nothing can be done about them after the fact (well, almost nothing: if not too drastic, reducing the exposure in the RAW file may bring back the lost pixels). Underexposed pixels can be brought up, but are likelly to be very noisy.

5. I agree with what 75RR said about not overdoing the image size. In my hands, with my monitor, and with my bad vision, I find that I can get away with a total magnification of about 2500 to 3000 times or even slightly more using 40x or lower objectives. With the 100x, I prefer to stay under 2000x. These magnifications are calculated as size of the object on the screen divided by actual size (of course using the same measuring units, say µm.

Below, when I mention menu items and the like, they refer to the program I use (Photoshop Elements). Other programs may have different names for the functions, etc.
6. Editing the image (you may do it differently and better, but this is what I do). For each operation I use a new layer by duplicating the current layer, then edit the new one. I've never used adjustment layers (I don't even know if Elements has them or not) because I don't know much about them.

First, if I had forgotten to set white balance in the camera before taking the pictures, I set it as the first step in post processing: I use the eyedropper (either in the Levels menu, or, more usually, in Enhance/Adjust Color/Remove color cast) to click on something I know is neutral gray or white (usually the background where there is nothing to look at). If I'm using the RAW file, I use the eyedropper in the RAW converter.
Second, if necessary, as when the image is too dark or too light, I adjust levels. Frequently, if necessary, I bring up shadows very slightly so things that look completely black look more like they did through the eyepiece.
Third, usually the image looks much duller than it did through the eyepiece, in which case I increase contrast (some increase may have already happened when I adjusted levels, above).
Now the image should look good.
Fourth, I crop the image if desired, then save the cropped image as a .PSD file so all layers are saved.
Fifth, I downsize it, then sharpen it. I explained my understanding why sharpening is necessary for digital images (and this could be carried out automatically by the camera software, or in post processing). I sharpen using either unsharp mask or the high pass filter. In either case, I duplicate the layer and sharpen the new layer using a radius of 0.3 to 0.5 in unsharp mask (no more than 0.7 in high pass filter). These values are what I use. This give only a slight amount of sharpening, and is not usually adequate. So I duplicate the layer again, filter slightly the increase in noise produced by the previous sharpening, and repeat the process. I work with the image zoomed to 2x so I can see what I'm doing, but people with good vision can look at it full size. I stop when the sharpening results in a harsh rather than smooth image, or one with "shiny" details, or where halos, not unlike those we see when the condenser aperture is closed too far) begin to appear, and I back off to the previous layer, and if necessary to the one before that. I want sharp but just enough so it doesn't look blurry. My "standard" of sharpness has changed over time as I get a better feel for what looks sharp enough. Less is better than more.
Sixth, I save the image (as .PSD), flatten it, change the color space from Adobe RGB to sRGB, and save it again as a jpeg file, which is what I upload to the forum.
This whole procedure takes much longer to describe than to do. It takes only a few minutes to actually edit an image.

Any criticism or suggestions for improvement will of course be much appreciated (I can sometimes learn...)

EDIT: I said earlier " images I produce are nowhere near as good as most of the members either occasionally or routinely produce, including, for example, seb Vasselle, 75RR, JimT, John B, and others that don't come to mind right now." Well, clearly I somehow managed to omit KurtM, Rod, Seb28, and I'm sure I still omitted others that don't come to mind right now.
Last edited by gekko on Sun Jun 05, 2016 2:42 pm, edited 2 times in total.

User avatar
KurtM
Posts: 1514
Joined: Tue Jun 09, 2015 12:08 am
Location: League City, Texas
Contact:

Re: Please Teach Me Sharpening

#13 Post by KurtM » Sat Jun 04, 2016 6:10 pm

Okay, here's my results with the full size image. I adjust contrast, sharpened, then played with saturation and luminosity. Didn't take a lot of time, maybe a full minute. But I always breeze through the processing of images, because otherwise I end up overthinking it and chasing my tail round and round.

In Light Zone, the tools I use for the above are:

1. Relight tool. This adjusts contrast. The Zone Mapper tool is another and even more powerful contrast adjustment tool, and oftentimes I use them both. You really need to watch the Light Zone tutorial videos (Youtube) to understand what the tools do and how to use them. I also use the Relight tool to increase detail and reduce fuzz. This is the nomenclature they use; any sharpening is really just very localized contrast adjustment. Light Zone gives a couple different tools for sharpening, just as it does for contrast adjustment, and both may be used at once to fine tune the effect.

2. Sharpen tool. I don't understand this on a real technical level, have never understood the term 'unsharp mask', so can't explain the what/how/why of it. All I can say is experience has taught me that using the Sharpen tool alone results in a relatively coarse effect, and I like the final look much better when I start the sharpening process in the Relight tool first. Again, best to watch the Light Zone tutorial vids.

3. Hue/Saturation tool. This gives sliders (as all the LZ tools do) to adjust Hue, Saturation, Vibrance, and Luminosity, and I play with them to see what happens. Sometimes I like what I get, and other times I just close out of the tool without keeping any of the changes it gave. The same is true of all the tools, of course, but I gotta say I use Sharpen at least 80% of the time generally, and more like 100% of the time when dealing with close-up photos like birds, macro shots, photomicrographs, and that sort of thing.

Finally, I resized to 1024 pixels in Picasa 3.
Attachments
SharpeningTrialFull_lzn 1024px.jpg
SharpeningTrialFull_lzn 1024px.jpg (80.48 KiB) Viewed 9025 times
Cheers,
Kurt Maurer
League City, Texas
email: ngc704(at)aol(dot)com
http://sawdustfactory.nfshost.com/microscopes/
https://www.flickr.com/photos/67904872@ ... 912223623/

User avatar
KurtM
Posts: 1514
Joined: Tue Jun 09, 2015 12:08 am
Location: League City, Texas
Contact:

Re: Please Teach Me Sharpening

#14 Post by KurtM » Sat Jun 04, 2016 6:15 pm

gekko wrote:If an image is not sharp to begin with (bad focus or blur due to shake), it goes to the trash can.

Avoid overexposure at all costs, and try to avoid underexposure.
^^^ These are my own personal cardinal rules.
Cheers,
Kurt Maurer
League City, Texas
email: ngc704(at)aol(dot)com
http://sawdustfactory.nfshost.com/microscopes/
https://www.flickr.com/photos/67904872@ ... 912223623/

billbillt
Posts: 2885
Joined: Mon Nov 10, 2014 10:01 pm

Re: Please Teach Me Sharpening

#15 Post by billbillt » Sat Jun 04, 2016 6:53 pm

I see Kurt's is better than the one I did... I used Irfanview and did the "sharpen" button several times.. I know very little about it.. I use it mainly to resize photos..

BillT

User avatar
mrsonchus
Posts: 3574
Joined: Tue Feb 03, 2015 9:42 pm
Location: Cumbria, UK

Re: Please Teach Me Sharpening

#16 Post by mrsonchus » Sat Jun 04, 2016 7:14 pm

In interesting thread Gents, I've downloaded the two sizes and may have a little meddle with them tonight if I get a chance - it's a really good image to practice with as it's a difficult one to enhance for sure. :)
John B

User avatar
zzffnn
Posts: 2710
Joined: Sat Jun 20, 2015 3:57 am
Location: Houston, Texas
Contact:

Re: Please Teach Me Sharpening

#17 Post by zzffnn » Sat Jun 04, 2016 7:16 pm

1.JPG
1.JPG (265.14 KiB) Viewed 9006 times
2a.jpg
2a.jpg (130.57 KiB) Viewed 8979 times
3kurt.jpg
3kurt.jpg (166.56 KiB) Viewed 9004 times
1st is original, 2nd is mine, 3rd is Kurt's.

I like a compromise between mine and Kurt's. My screens are not adjusted - on my computer screen, I lean towards mine own from the middle ground; on my cell phone screen, I prefer Kurt's.

I did all my adjustments on my computer screen. It seems that Kurt used more sharpening than I did, and my computer screen reacts more negatively (than my cell phone screen) with sharpening artifacts.

This shows that personal preference and screen both play a role in image editing.

I used:

FastStone IMage Viewer
Colors - edit lighting -
contrast +94 (maximum possible is +100)
Saturation -60 (minimum possible is -100)
Highlights -22 (minimum possible is -100)

Then sharpen/Blur
Amount +28
Radius +1.5
Last edited by zzffnn on Sat Jun 04, 2016 9:13 pm, edited 11 times in total.
Selling my Canon FD 200mm F/2.8 lens

User avatar
gekko
Posts: 4701
Joined: Sun Oct 12, 2014 7:38 am
Location: Durham, NC, USA.

Re: Please Teach Me Sharpening

#18 Post by gekko » Sat Jun 04, 2016 7:27 pm

Here is my version. Not very different from the other versions already posted. This is not an easy one, expecially that I'm not sure what it looked like under the microscope. I did make it smaller, though, and this always helps make it look sharper, although one may lose some resolution.
Image
Last edited by gekko on Thu Jun 09, 2016 8:20 pm, edited 2 times in total.

billbillt
Posts: 2885
Joined: Mon Nov 10, 2014 10:01 pm

Re: Please Teach Me Sharpening

#19 Post by billbillt » Sat Jun 04, 2016 7:34 pm

Hi Gekko,

Looks to me yours is the best rendition.. I didn't think about the smaller size.. That was a good idea!..

BillT

User avatar
KurtM
Posts: 1514
Joined: Tue Jun 09, 2015 12:08 am
Location: League City, Texas
Contact:

Re: Please Teach Me Sharpening

#20 Post by KurtM » Sat Jun 04, 2016 7:47 pm

Agree, Gekko's shows max rez. But at the cost of looking overcooked...?

Which begs the question: what are we looking for? Are we trying to get the maximum information out of the image, or overall most pleasing picture? Both are perfectly legitimate goals as far as I'm concerned.

Sometimes images show things the eye was unable to discern through the oculars (especially true in astrophotgraphy). This makes the camera a useful observational tool, which post processing software renders even more powerful.

NICE work, gekko! 8-)
Cheers,
Kurt Maurer
League City, Texas
email: ngc704(at)aol(dot)com
http://sawdustfactory.nfshost.com/microscopes/
https://www.flickr.com/photos/67904872@ ... 912223623/

User avatar
rnabholz
Posts: 3066
Joined: Thu Jan 08, 2015 10:11 pm
Location: Iowa USA
Contact:

Re: Please Teach Me Sharpening

#21 Post by rnabholz » Sat Jun 04, 2016 8:08 pm

Thanks All.

Gekko - Excellent points all in your first post. Completely agree, overexposure is deadly (as zz mentioned as well)

Your other point about Garbage in Garbage out is one of the issues I struggle with here. In my bird photography I can look at an image from 50 paces and tell you if it is sharp or not. ( small exaggeration) As you say, it will never be sharp if it doesn't start with a sharp capture.

I photomicroscopy, I don't think that I have those chops yet. I don't have the feel if what I have is technically well captured or not, and therefore struggle knowing how hard to push the post processing without looking like I am trying to make a silk purse out of a sow's ear.

The discussion of size is helpful as well. Again something I have a feel for in my other photo endeavors, still developing here.

Kurt - the question of intended use is absolutely another vegetable to add to the stew. A scientist may be looking for a specific aspect of structure and would be willing to overlook the "negative" effects of over-doing some aspect of processing to bring clarity to his driving interest.

As pointed out by 75 and gekko, the reduction in size does help regardless of the other aspects of the processing.

Keep em coming everyone and thank you.

Rod

billbillt
Posts: 2885
Joined: Mon Nov 10, 2014 10:01 pm

Re: Please Teach Me Sharpening

#22 Post by billbillt » Sat Jun 04, 2016 8:16 pm

I took the liberty of resizing the photo gekko had worked on.. It does not look TOO bad being enlarged..
Attachments
test.jpg
test.jpg (69.3 KiB) Viewed 8980 times

User avatar
gekko
Posts: 4701
Joined: Sun Oct 12, 2014 7:38 am
Location: Durham, NC, USA.

Re: Please Teach Me Sharpening

#23 Post by gekko » Sat Jun 04, 2016 8:52 pm

Kurt, you were aright, I did sharpen it a lot more than I normally do (5 iterations instead of 1 or 2). That was the only way to get the details reasonably sharp. But then normally the original image would be much sharper, and if it weren't, I would not bother trying to fix it (it cannot be really fixed). If the view through the microscope was clear, sharp, and crisp, and the picture came out like the original image, then I think it is a waste of time trying to fix it. If, on the other hand the view through the microscope was not sharp and clear, I wouldn't try to take a picture in the first place. Both the image through the eyepiece has to be crisp and clear, and the image taken by the camera has to be nice and sharp. Then, when the image is downsized, it can be sharpened to give good results. Otherwise it is a waste of time trying to rescue it.

User avatar
gekko
Posts: 4701
Joined: Sun Oct 12, 2014 7:38 am
Location: Durham, NC, USA.

Re: Please Teach Me Sharpening

#24 Post by gekko » Sat Jun 04, 2016 8:56 pm

BillT, thank you for your attempt, but as you know, downsizing an image throws away information, but enlarging it does not add back the information that was lost in downsizing it. So, in effect, we get the equivalent of empty magnification. At least that is what I think.

User avatar
KurtM
Posts: 1514
Joined: Tue Jun 09, 2015 12:08 am
Location: League City, Texas
Contact:

Re: Please Teach Me Sharpening

#25 Post by KurtM » Sat Jun 04, 2016 9:37 pm

Good stuff gekko ... but in defense of the original image, I know from experience that this particular diatom is a frustratingly difficult one both to image, and to observe visually. Like so many diatoms, there's tantalizing detail seeming right on the verge of resolution, and we pull out all the stops trying to coax it out. Heck, that's a big part of what makes diatoms such a fascinating pursuit, and why they're highly regarded as test object for optics.

Rod picked a seriously tough one for us...
Cheers,
Kurt Maurer
League City, Texas
email: ngc704(at)aol(dot)com
http://sawdustfactory.nfshost.com/microscopes/
https://www.flickr.com/photos/67904872@ ... 912223623/

User avatar
rnabholz
Posts: 3066
Joined: Thu Jan 08, 2015 10:11 pm
Location: Iowa USA
Contact:

Re: Please Teach Me Sharpening

#26 Post by rnabholz » Sat Jun 04, 2016 9:49 pm

Sorry guys, didn't mean to vex you, but I think that is another contributing factor to my troubles.

These darn diatoms do in fact push my equipment's capabilities. I am not used to the idea that my camera just might not be able to deliver what I am expecting from it and that is a reasonable outcome.

Toss in the slight spreading of point sources that comes with darkfield, and a fellow starts to doubt himself.

That is not to say I have everything mastered, and that it is all the equipment's fault, not by a long shot. And that is why I am interested in everybody's approach.

Thanks for the efforts so far and the great discussion.

Hope others will still chime in.

Rod

User avatar
gekko
Posts: 4701
Joined: Sun Oct 12, 2014 7:38 am
Location: Durham, NC, USA.

Re: Please Teach Me Sharpening

#27 Post by gekko » Sat Jun 04, 2016 9:54 pm

Kurt, good point. I can see the problemwhen dealing with super difficult diatoms. And Rod is using his 100x so that's that as far as resolving more detail under the light microscope (except for going to shorter wavelength).

User avatar
mrsonchus
Posts: 3574
Joined: Tue Feb 03, 2015 9:42 pm
Location: Cumbria, UK

Re: Please Teach Me Sharpening

#28 Post by mrsonchus » Sat Jun 04, 2016 10:13 pm

KurtM wrote:Agree, Gekko's shows max rez. But at the cost of looking overcooked...?

Which begs the question: what are we looking for? Are we trying to get the maximum information out of the image, or overall most pleasing picture? Both are perfectly legitimate goals as far as I'm concerned.

Sometimes images show things the eye was unable to discern through the oculars (especially true in astrophotgraphy). This makes the camera a useful observational tool, which post processing software renders even more powerful.

NICE work, gekko! 8-)
Agree 100% - that's exactly how I feel about pp - for me it's about 80% for information and about 20% for aesthetic attractiveness - I prefer to produce images that are about 40% information (and of course 60% aesthetics) as a std 'on-screen' version for forum upload etc, and to 'go all out' for information at about 80% for say plant structural fine-details e.g. to show the existence of several discrete layers of a secondary cell-wall where without this level of pp they would hardly or only ambiguously be discernable - the pp can 'say' "yes, they are there as you suspected" as it were.

To show the details rather than simply the existence of such a feature as say the 'hook' on a stomatal-pore's entrance, I would lean towards sharpening using a high number-of-levels-difference, just to bring out the already potentially clear edges.

Then there's what I would call my 'impressionistic' technique, such as I have posted of B & W 'psuedo SEM' images to show the 3D nature of chromosomes in what I refer to as their 'basket' within an about-to-divide cell...

Many, many options are available to us lucky folk, the 'straight from the camera's jpeg routines' option is beginning to look more and more limited to me, only tonight I've been practicing with Canon .CR2 image RAW pre-jpg processing and have learned a huge amount already - and it's all good!

A great discussion, interesting to see what others think - I think my first step with diatom frustules (empty that is) would be to convert to B & W, then perhaps clip the high end of the histogram to remove lost pixels that have blown-out. The I think I would go for the edges with the 'high number of levels' sharpening to avoid attempting to sharpen mid-tones that haven't much of a role it would seem to me with diatom images when searching for detail and resolution....

All supposition from me, but I'm going to give it a go anyway - it would be very useful to have RAW images available........... :)
John B

User avatar
mrsonchus
Posts: 3574
Joined: Tue Feb 03, 2015 9:42 pm
Location: Cumbria, UK

Re: Please Teach Me Sharpening

#29 Post by mrsonchus » Sat Jun 04, 2016 10:18 pm

rnabholz wrote:Sorry guys, didn't mean to vex you, but I think that is another contributing factor to my troubles.

These darn diatoms do in fact push my equipment's capabilities. I am not used to the idea that my camera just might not be able to deliver what I am expecting from it and that is a reasonable outcome.

Toss in the slight spreading of point sources that comes with darkfield, and a fellow starts to doubt himself.

That is not to say I have everything mastered, and that it is all the equipment's fault, not by a long shot. And that is why I am interested in everybody's approach.

Thanks for the efforts so far and the great discussion.

Hope others will still chime in.

Rod
Yes, it's a frustration that we probably all feel rather keenly... :cry:
Those point-distortions (point-spread-functions) can, if described correctly mathematically, be reversed with a deconvolution function I think, but I suspect that may require resources beyond our means... :cry:
An analysis of your objective + optical-train's artifacts would presumably need to be made to construct the functions, although this isn't my area at all, just something I read a while ago now.
John B

User avatar
gekko
Posts: 4701
Joined: Sun Oct 12, 2014 7:38 am
Location: Durham, NC, USA.

Re: Please Teach Me Sharpening

#30 Post by gekko » Sat Jun 04, 2016 10:20 pm

rnabholz wrote:Sorry guys, didn't mean to vex you, but I think that is another contributing factor to my troubles.

These darn diatoms do in fact push my equipment's capabilities. I am not used to the idea that my camera just might not be able to deliver what I am expecting from it and that is a reasonable outcome.

Toss in the slight spreading of point sources that comes with darkfield, and a fellow starts to doubt himself.

That is not to say I have everything mastered, and that it is all the equipment's fault, not by a long shot. And that is why I am interested in everybody's approach.

Thanks for the efforts so far and the great discussion.

Hope others will still chime in.

Rod
As far as I know, if your camera-microscope optical "connection" is proper, and the sensor image is parfocal with the eyepiece image, the camera should be able to give you almost as good a result as you see through the eyepieces. You have an excellent sensor with much more than enough resolution. So if the image is significantly worse than the view through the eyepiece, maybe you should look into that (starting with brightfield in order to remove one factor; also DF is harder to judge, and harder to expose properly because I think the contrast can be very large.)
Having said all that, other than this image and one of a circular diatom, I thought your images are superb and for darfield images, I think they are more than superb. So I wouldn't let one or two pesky diatoms aggravate me. Maybe try brightfield on those. Sorry. I'm rambling...

Post Reply