Frustulia Filter Faceoff Fun

[rnabholz]

In the ongoing quest for the quintessential green filter for improving achro and phase images, today I ran a quick test of the following:

1. A Lee #124 Gel Filter as graciously provided by Fan
2. A American Optical Green filter as found in the filter wheel of the Ortho Illuminator
3. A Meade Green Interference Filter - as seen in Ebay item #231233865897. Intended for CCD Imaging and Visual use.

The target is a Frustulia, from a strew slide, sample taken this summer from the Maquoketa River in Iowa.

The platform was my AO 10, Ortho Illuminator, 100x Plan Achro Iris Oil, 1096b Toric Darkfield condenser, Objective and Condenser Oiled, Canon 70D, Single frames. Exposure adjusted as necessary for filtered shots, Green images converted to grayscale, resized to 1024 on the long side, and unsharp mask applied identically to all images.

First the unfiltered shot

Frustulia 91µm Unfiltered

Maq River Frustulia 3968 Color.JPG (85.07 KiB) Viewed 5128 times

The AO Ortho Illuminator Filter

Frustulia 91µm AO OEM Filter

Maq River Frustulia 3976 AO OEM.JPG (88.09 KiB) Viewed 5128 times

The Lee 124 Filter

Frustulia 91µm Lee 124 Filter

Maq River Frustulia 3973 Lee 124.JPG (82.91 KiB) Viewed 5128 times

The Meade Green Interference Filter

Frustulia 91µm Meade Green Interference Filter

Maq River Frustulia 3970 Meade Interference.JPG (82.09 KiB) Viewed 5128 times

[75RR]

In the ongoing quest for the quintessential green filter for improving achro and phase images ...

Think the first one (The AO Ortho Illuminator Filter) is best.
Have you tried them with Brightfield and Oblique yet? They should work even better then I think.

[zzffnn]

Thank you for doing this, Rod.

Just by looking at the images, AO > Lee > Meade > unfiltered, to my eyes.

Were your camera exposures (brightness compensation, ISO and aperture) comparable, between those three images? Did you set brightness compens /ISO/F the same and let camera auto select shutter? I am guessing that may be the best, as your 700D has minimal shutter shake.

I could be totally wrong, though I am guessing:
If you set camera to auto everything, or you manually set brightness compens/ISO/F/shutter to be exactly the same, then the more efficient filter may have an unfair advantage (different green filters' green pass efficiency may be different). What do you think?

At least, my camera's auto exposure does not work well in darkfield.

I am not picking on you - it is just my nerdy curiosity.

[rnabholz]

Thanks 75. We agree about the AO being the winner. I have to admit I was surprised at the poor showing of the Meade. I may have to revisit that one, maybe I missed the focus a bit.

I have not used them on brightfield or oblique yet. Will do soon.


Thanks zz.

Agreed on the order.

I always run the camera in Manual. I adjusted only the shutter speed to balance the exposures across the varying brightness of the filters.

The shutter speeds ran from .6 to 1.3 seconds with the AO being brightest and the Lee the darkest.

They may not be perfectly matched exposure wise, but I do think the are representative of the filters reduction of CA.

In any case, a place to start a conversation.

[zzffnn]

Rod,

The filters may not differ that much. All those images are great, by the way.

I was just wondering if the main difference we are seeing here was caused by slightly different exposure level / subject brightness.

With my limited experience, I found that:

1) changing the exposure compensation changes darkfield images A LOT, so that parameter should he kept the same, if one wants to compare things:

2) my camera always wants to use higher iso than I want to, especially for darkfield. I always keeps it at 200. Noise may be an issue for dark images, if I let camera go auto.

[rnabholz]

The ISO was at 200 for all of the shots.

I guess that I am judging the sharpness of the dots, which hopefully have been stripped of color fringes caused by the achro objectives.

In order to get a feel for that I do think that the exposures need to be nearly even or reasonably close. In addition to a marked color difference for each, there was a full stop of exposure difference across the range of filters.

I agree, underexposure will block up the detail quickly in darkfield, in this case that would defeat our ability to judge the sharpness. That is the reasoning I applied in making exposure adjustments.

At the end of the process my purpose is to always end up with a properly exposed image, not in this case to document the density of, or energy passed by the filters, so that is the approach I took for the test. I offer it for what it is worth understanding my goals.

I hope it provides some insight.

Thanks.

Rod

[zzffnn]

Thanks, Rod.

Your results are very helpful to me. I don't have the AO filters, but I am glad to see that the easily available and cheap Lee filter can get close.

[Rudix]

Interesting experiment Rod!

Looks like color filters can be a useful tool to bring out the most detail in a subject. I agree the AO filter seems best with the Lee a close second.

I am not surprised there was a difference in illumination levels, not all filers are created equal. Depending on the metering mode used on the camera the auto exposure could be quite accurate. If the camera has a spot metering option (I do not know the 70D), and you put the spot on the subject, it should be close. The normal "average" metering does not work that well with a smallish bright subject on a dark background, it is easy to burn highlights.

I wonder if a camera with no low-pass filter would reveal more detail? Can you see more detail with the eye than you can see with the camera? I know from experience in normal photography that my camera can capture far more detail than I can see, looking at images at 100% size often show up unexpected details missed by the eye but here the resolution of the optics might well be the limiting factor.

Would post processing be able to get more detail from the images? I suppose not if the optics are the ultimate limitation but advanced image processing as used in other fields can produce amazing results. I have spent a lot of time developing image processing software for other fields of photography, will be interesting to see what, if anything, can be done with images like these. Ideally the image should be captured in RAW mode on a high resolution sensor without a low-pass filter. Best results would most likely be obtained using the "Expose to the right" technique.

Thanks for sharing with us, informative and thought provoking!
Rudi

[rnabholz]

Interesting experiment Rod!

Looks like color filters can be a useful tool to bring out the most detail in a subject. I agree the AO filter seems best with the Lee a close second.

I am not surprised there was a difference in illumination levels, not all filers are created equal. Depending on the metering mode used on the camera the auto exposure could be quite accurate. If the camera has a spot metering option (I do not know the 70D), and you put the spot on the subject, it should be close. The normal "average" metering does not work that well with a smallish bright subject on a dark background, it is easy to burn highlights.

I wonder if a camera with no low-pass filter would reveal more detail? Can you see more detail with the eye than you can see with the camera? I know from experience in normal photography that my camera can capture far more detail than I can see, looking at images at 100% size often show up unexpected details missed by the eye but here the resolution of the optics might well be the limiting factor.

Would post processing be able to get more detail from the images? I suppose not if the optics are the ultimate limitation but advanced image processing as used in other fields can produce amazing results. I have spent a lot of time developing image processing software for other fields of photography, will be interesting to see what, if anything, can be done with images like these. Ideally the image should be captured in RAW mode on a high resolution sensor without a low-pass filter. Best results would most likely be obtained using the "Expose to the right" technique.

Thanks for sharing with us, informative and thought provoking!
Rudi

Hi Rudi

Glad you found it interesting.

The extremes in darkfield can be a challenge for a camera meter, after uneven results using other modes, I have settled on Manual. A test shot or two usually dials things in pretty well.

I think the camera is doing a pretty good job of capturing what I am seeing visually, but I have no doubt that there is a Photoshop hired gun that could bring out more than my rudimentary skill permit.

Thanks for the interest and your comments.

Rod