Amscope 4x IQ?

Images taken in a controlled environment or with a posed subject. All subject types.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

tevans9129
Posts: 129
Joined: Thu Nov 30, 2017 8:03 am
Location: TN

Amscope 4x IQ?

Post by tevans9129 »

Searching for lighting and image quality so composition is not a priority at this time. This is using the Amscope 4x and admittedly, my technique is very much lacking but this just does not seem sharp to me. Is there any suggestions as to how I can improve the IQ?

Should I expect better from this lens or should I be looking for something different? I have a Nikon 105 f/2.8 micro AIS, a Nikon 200 f/4.0 D and a Tamron 70-200 f/2.8, will any of these work with an infinity lens? My understanding is a bellows will not, is that correct? Thanks for all comments.

Nikon D800e, PB6 closed for 3.2x, 155 slices, 30 microns, 2 second delay, 1/15, ISO 100 in Controlmynikon, Stackshot, YN300ii at 100%, SB700 @ 1:128, Styrofoam diffuser. Some sharpening performed in Sharpener Pro 3.

Image

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

Post by rjlittlefield »

We could tell better about sharpness if you could post an actual-pixels crop for us to look at.

By "actual pixels crop", I mean simply duping your image, cropping without resizing to an area that is 1024 pixels square or smaller, saving that small cropped version to a JPEG with file length < 300KB, and posting the result. I emphasize "without resizing" because we've had at least one new forum member accidentally use the Photoshop crop tool with options set so that it did resize, silently and implicitly. The results prompted some pretty confused discussion.

--Rik

tevans9129
Posts: 129
Joined: Thu Nov 30, 2017 8:03 am
Location: TN

Post by tevans9129 »

rjlittlefield wrote:We could tell better about sharpness if you could post an actual-pixels crop for us to look at.

By "actual pixels crop", I mean simply duping your image, cropping without resizing to an area that is 1024 pixels square or smaller, saving that small cropped version to a JPEG with file length < 300KB, and posting the result. I emphasize "without resizing" because we've had at least one new forum member accidentally use the Photoshop crop tool with options set so that it did resize, silently and implicitly. The results prompted some pretty confused discussion.

--Rik
Thanks Rik, here is an image as you requested. To my eye, the entire image is mushy, not crisp at all. Perhaps I should try a different subject to see if I get the same results.

Image

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

Post by rjlittlefield »

Ted, thanks for the crop.

To my eye, this looks like a typical image shot with a microscope objective on a high resolution sensor. It is heavily diffracted, due to the small effective aperture provided by the objective. This results in contrast loss that is progressively worse for finer details.

To put some numbers on this... Your D800E sensor has a pixel pitch of 0.004878 mm (35.9 mm divided by 7360 pixels). Its resolution limit, according to the Nyquist sampling theorem, 2 pixels per cycle, is 102.5 cycles/mm. Meanwhile, your 4X NA 0.1 lens is running at effective f/20 (=magnification/(2*NA) = 4/(2*0.1)). For green light, lambda=0.00055 mm, an f/20 lens has a cutoff frequency nu0 of 90.91 cycles/mm. "Cutoff frequency" is the level of detail where the lens MTF drop to zero.

So in some sense that lens is almost perfectly matched to the sensor: the optical image contrast drops to zero at almost exactly the same level of detail where the sensor becomes unable to resolve it anyway. That's a tongue-in-cheek pessimistic view, but it should give you the idea that looking for sharp pixel-level detail is an unrealistic expectation.

A more realistic view is to look at say one cycle every 4 pixels, but even there, the lens cannot have more than about 40% MTF simply due to diffraction. That is still quite a bit of softening.

The usual attack on this problem is to apply aggressive sharpening. See the discussion at http://www.photomacrography.net/forum/v ... 198#203198 and in the surrounding thread.

Because the lmage has been heavily softened by diffusion, it can tolerate levels of sharpening that would reduce an initially sharp image to junk. I'm not familiar with Sharpener Pro 3, but using Photoshop, I would treat this image to unsharp mask at about 180%, pixel width 1, based on visual appearance.

In addition to MTF falloff, this image is also suffering from a bit of veiling glare that causes blacks to turn gray. There's a small gap at the bottom of the histogram. The image can also be improved by doing a levels adjustment to essentially subtract off the veiling glare.

If you like, I can post my version of the crop to show what I mean.

One other issue is lighting. At high magnifications, there is a counterintuitive effect that hard lighting can make an image go soft. The underlying reason is some physics that we've come to call the "utilized aperture" effect. Very briefly, the use of hard lighting -- illumination coming from only a narrow range of angles -- often causes each detail of the subject to be seen by only a small part of the lens. The result is that even though the lens appears to be f/20 based on its diameter and focal length, for each detail of the subject the utilized aperture may be only f/40 or smaller. This makes a lens that was already in diffraction territory under ideal conditions act even worse.

To avoid the utilized aperture effect, you need to have quite heavily diffused illumination, with light striking the subject from a wide range of angles. Instead of the small bright reflection of your flash that we see in the eyes of this spider, well diffused illumination would produce a much wider, less intense reflection, such as you'll see for example in the photos by Thomas Shahan.

Let me know if you'd like to see that reworked image.

--Rik

Pau
Site Admin
Posts: 6052
Joined: Wed Jan 20, 2010 8:57 am
Location: Valencia, Spain

Post by Pau »

To me now it looks better, I did more or less what Rik says, although with a more modest sharpening, what do you think?
Image
Of course I you don't want to see here your image reworked I'll delete it immediately.
Pau

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

Post by rjlittlefield »

Pau, your colors are much more intense. Did you increase the saturation?

Edit: Oh, I see. Pau's version of the image appears to have accidentally gotten Adobe RGB assigned to it, where the original image did not have a color profile and therefore was implicitly sRGB.

--Rik

tevans9129
Posts: 129
Joined: Thu Nov 30, 2017 8:03 am
Location: TN

Post by tevans9129 »

rjlittlefield wrote: Ted, thanks for the crop.

To my eye, this looks like a typical image shot with a microscope objective on a high resolution sensor. It is heavily diffracted, due to the small effective aperture provided by the objective. This results in contrast loss that is progressively worse for finer details.

To put some numbers on this... Your D800E sensor has a pixel pitch of 0.004878 mm (35.9 mm divided by 7360 pixels). Its resolution limit, according to the Nyquist sampling theorem, 2 pixels per cycle, is 102.5 cycles/mm. Meanwhile, your 4X NA 0.1 lens is running at effective f/20 (=magnification/(2*NA) = 4/(2*0.1)). For green light, lambda=0.00055 mm, an f/20 lens has a cutoff frequency nu0 of 90.91 cycles/mm. "Cutoff frequency" is the level of detail where the lens MTF drop to zero.

So in some sense that lens is almost perfectly matched to the sensor: the optical image contrast drops to zero at almost exactly the same level of detail where the sensor becomes unable to resolve it anyway. That's a tongue-in-cheek pessimistic view, but it should give you the idea that looking for sharp pixel-level detail is an unrealistic expectation.

A more realistic view is to look at say one cycle every 4 pixels, but even there, the lens cannot have more than about 40% MTF simply due to diffraction. That is still quite a bit of softening.


Thanks for the explanations…hope I live long enough to digest all of it. I am a slow learner but I make up for it by forgetting quickly.
rjlittlefield wrote: The usual attack on this problem is to apply aggressive sharpening. See the discussion at http://www.photomacrography.net/forum/v ... 198#203198 and in the surrounding thread.

Because the lmage has been heavily softened by diffusion, it can tolerate levels of sharpening that would reduce an initially sharp image to junk. I'm not familiar with Sharpener Pro 3, but using Photoshop, I would treat this image to unsharp mask at about 180%, pixel width 1, based on visual appearance.


I have used the unsharp mask in PS but have come to prefer SP3. Actually, I was using Piccure+ until I had to reload my system and Piccure will not allow me to reinstall it. Seems like the company may no longer be in business as I cannot get anyone to answer emails.

My approach to sharpening has been to use the minimal amount that I can get away with. Perhaps I need to reevaluate that.
rjlittlefield wrote: In addition to MTF falloff, this image is also suffering from a bit of veiling glare that causes blacks to turn gray. There's a small gap at the bottom of the histogram. The image can also be improved by doing a levels adjustment to essentially subtract off the veiling glare.

Thanks, I will do some experimenting with that.
rjlittlefield wrote: If you like, I can post my version of the crop to show what I mean.


Absolutely, it helps me to see the approach of others and hopefully learn from it, by all means, feel free.
rjlittlefield wrote: One other issue is lighting. At high magnifications, there is a counterintuitive effect that hard lighting can make an image go soft. The underlying reason is some physics that we've come to call the "utilized aperture" effect. Very briefly, the use of hard lighting -- illumination coming from only a narrow range of angles -- often causes each detail of the subject to be seen by only a small part of the lens. The result is that even though the lens appears to be f/20 based on its diameter and focal length, for each detail of the subject the utilized aperture may be only f/40 or smaller. This makes a lens that was already in diffraction territory under ideal conditions act even worse.

To avoid the utilized aperture effect, you need to have quite heavily diffused illumination, with light striking the subject from a wide range of angles. Instead of the small bright reflection of your flash that we see in the eyes of this spider, well diffused illumination would produce a much wider, less intense reflection, such as you'll see for example in the photos by Thomas Shahan.


Lighting is one of the primary issues that I have been trying to experiment with and in my experience, there are at least 1723 variables associated with the subject. I have three speedlights and a YN300iii at present but there may be better options. I could be satisfied with images like thomasshahan, I suppose. :D
rjlittlefield wrote: Let me know if you'd like to see that reworked image.

--Rik
[/quote]

Yes, I would appreciate that and thanks for all the help and the links.

tevans9129
Posts: 129
Joined: Thu Nov 30, 2017 8:03 am
Location: TN

Post by tevans9129 »

Pau wrote:To me now it looks better, I did more or less what Rik says, although with a more modest sharpening, what do you think?
Image
Of course I you don't want to see here your image reworked I'll delete it immediately.
It looks better but still looks "mushy" to me which could be attributed to a number of things such as the condition of the subject and lighting. No, I appreciate you redoing the image and offering to help. Feel free to do so with any of my images.

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

I've found the inexpensive small light tents on eBay work well with a Strackshot rail macro system. These generally diffuse enough for most things and if you need more diffusion use a couple tents, one small one inside a larger one.

The camera/lens pokes thru the tent slit where the subject sits inside. I use strobes rather than speedights for illumination as the tents tend to eat up a lot light, and I don't have to fiddle with batteries.

Best & Happy Holidays,

Mike

tevans9129
Posts: 129
Joined: Thu Nov 30, 2017 8:03 am
Location: TN

Post by tevans9129 »

mawyatt wrote:I've found the inexpensive small light tents on eBay work well with a Strackshot rail macro system. These generally diffuse enough for most things and if you need more diffusion use a couple tents, one small one inside a larger one.

The camera/lens pokes thru the tent slit where the subject sits inside. I use strobes rather than speedights for illumination as the tents tend to eat up a lot light, and I don't have to fiddle with batteries.

Best & Happy Holidays,

Mike
Many thanks for the comments Mike. I have been thinking about strobes but since this is purely a hobby for me, I cannot see spending a lot on strobes. I have been looking at something like this one,


https://www.amazon.com/Neewer-Monolight ... hotography

What are you using and which tents if you care to share?[/url]

mawyatt
Posts: 2497
Joined: Thu Aug 22, 2013 6:54 pm
Location: Clearwater, Florida

Post by mawyatt »

I use the inexpensive Adorama Studio 300 AC (rebranded Godox SK300II) strobes. They were on sale at Adorama for $100 and free shipping before Christmas. These are manual, 300WS (about 6 times the power of a typical speed-light), proportional control modeling light, with Bowen mount and have a built in radio system that works with the R2 (Godox X1).

So far, very reliable for my use, or abuse since I usually run 300~500 stacks with only a few seconds between exposures.

Anything like these light tents. Place a few strobes above at 45 degrees and around the sides with some Dollar Store white foam panels as reflectors.

https://www.ebay.com/itm/12-32-Photogra ... kccIVk6g8g

I have a couple, 12" and 24"

Best & Happy Holidays,

Mike

Deanimator
Posts: 870
Joined: Tue Oct 23, 2012 7:01 pm
Location: North Olmsted, Ohio, U.S.A.

Post by Deanimator »

I've been running into the same issue using my Amscope 4x.

I was trying not to get too excessive with the sharpening, but now I see I'm going to have to be a quite a bit more aggressive.

I'll re-process my last stack and give it a try.
Last edited by Deanimator on Mon Dec 25, 2017 5:51 pm, edited 1 time in total.

tevans9129
Posts: 129
Joined: Thu Nov 30, 2017 8:03 am
Location: TN

Post by tevans9129 »

Tried a different subject, housefly eye, and lighting, I think this is better so feel better about the Amscope 4x. I believe work on the lighting and technique will result in more satisfactory images. Feel free to modify the image if you wish.

PB6 140, Amscope 4x, 1/8 ISO 200, 100 slices, 8 microns, YN300ii above, SB700 one side 1/32, 3/16" styrofoam diffuser, YN685 one side 1/32

Image

Deanimator
Posts: 870
Joined: Tue Oct 23, 2012 7:01 pm
Location: North Olmsted, Ohio, U.S.A.

Post by Deanimator »

tevans9129 wrote:Tried a different subject, housefly eye, and lighting, I think this is better so feel better about the Amscope 4x. I believe work on the lighting and technique will result in more satisfactory images. Feel free to modify the image if you wish.

PB6 140, Amscope 4x, 1/8 ISO 200, 100 slices, 8 microns, YN300ii above, SB700 one side 1/32, 3/16" styrofoam diffuser, YN685 one side 1/32

Image
That's a very nice image.

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

Post by rjlittlefield »

Here is my version of the rework.

Image

Yep, the fly looks good.

--Rik

Post Reply Previous topicNext topic