3D Image Enhancement made EASY – Part I [Updated April 27]

Click on image above for larger version.

Recent posts covering the Goerz '3D' Condenser have sparked interest in creating an alternative approach suitable for use with ordinary (Abbe-type) condensers.

The goals of this new effort were: (1) Low cost and Ease of construction, (2) Improvement in image contrast and apparent object depth, and (3) Ease of construction and use..

It is important to remember that the goals here are not maximum image contrast nor maximum resolution, but rather to simply achieve meaningful improvements in both (relative to normal Brightfield imaging) by use of the simplest and most generally applicable method possible. 

The method presented here appears to satisfy these goals and results in a simple addition which may be easily applied to an ordinary microscope condenser and, unlike other methods, (e.g: oblique illumination or phase contrast) may be used without any additional/special optics or concern for adjustments during use.

Also, since the light loss is minimal, the device is suitable for use on instruments having limited illumination and/or at higher magnifications than these common alternatives.

The device is constructed as a simple disk having an array of three diffusion segments, which modify the illumination in a predetermined manner and is essentially the same for all objectives, from 10x to 40x, or more. The "radial" approach both eliminates the need for adjustments when changing objectives and also allows the Condenser Iris to remain fully usable as an additional means of image control.

The construction of this Condenser "Radial Diffusion mask" is depicted in the following diagrams:

Note: Click on any above image for a larger view. 

Installation of the mask, its preferred use and example specimen photos will be presented ASAP… 

* * * * *  
Basic Construction Notes:  

Most of the construction details for this mask are non-critical. If the tape intersection point is reasonably close to the optical center of the Condenser when installed, then the device should function as intended. The precise angles are also not critical – anything approaching 90-degrees should work just fine. 

The specific type of "matte tape" to be used is also not critical. Ordinary, generic "dollar store" types seem to work about as well as anything else. As long as the tape width is about 3/4" then disks up to about 37.5mm diameter would seem feasible. Scotch (brand) Matte finish tape, however, apears to be superior to their "Satin finish" variety, which seems to exhibit less desirable diffusion characteristics, at least for this use. 

The acetate disk material is also non-critical. The basic requirements are that it be basically clear, self-supporting and thin enough to be trimmed with scissors. However, if your Condenser accepts a clear Daylight filter, then this might  serve as an alternative substrate. 

The positioning of the Opaque segment was chosen so as to not create uneven illumination with objectives of less than about 40x. However, it may be positioned closer to the disk center if contrast enhancement is desired for objectives in the 16x to 25x range also. In this case, about a third, instead of halfway, from the center to the disk edge may be tried. However, if detailed examinations with 40x and greater objectives are not the primary use, then the Opaque segment may simply be omitted with minimal loss in overall performance. (It can always be added later, if desired.) 

The Opaque segment may be created with a single strip of black PVC tape (plastic electrical tape), located on the reverse side of the disk, so it does not interfere with the matte segments. This allows the opaque segment to be re-positioned, if desired, without interfering with the other segments. 


* * * * * 
An External Mask: 

One common problem with many modern microscopes is that they often use Condensers which lack any sort of proper filter holder. Naturally, this sort of shortcoming might seem to be an issue when attempting to implement even something as simple as a Radial '3D' Mask… 

Fortunately, the design of this Mask is remarkably forgiving when to comes to placement in the Condenser, even to the extent that it can be perfectly acceptable to place the mask outside the Condenser! This can be especially true for use with objectives of >10x, where the loss in Mask performance is basically trivial. (For objectives of ~10x, the loss is most typically limited to a slightly uneven background light level.) 

For Condensers which have a significant bottom flange (e.g: most base-mount types), external mounting of the mask should be possible, as long as the width of the flange's bottom surface is sufficient to allow retention of the tape strips. 

The tape strips are simply applied across the width of the flange, such that the "inside corner" formed where the strips overlap is located approximately in the center of the objective aperture, as viewed from the eyepiece position. (For the most accurate alignment, with the least difficulty, use of the 10x objective is suggested for this process.)  The two diagrams below depict both proper tape placement and proper overall mask alignment (centering).  

(Note: Click on either image above for a larger view.)

Be aware that the tape placement should be performed only after the Condenser has been properly centered in its mount, as for Brightfield use. Once the Condenser itself is centered, any centering mechanism present should not be used to correct any misalignment of the tape strips. If the overall positioning is unacceptable, then the strips should simply be removed and re-positioned properly. Just remember that exact alignment is not essential for successful use!  

The photo below shows typical results when using the External '3D' Mask method:
(Image cropped and downsized from 16Mpix JPEG original. No sharpening used.) 

(Note: Click on above image for a larger view.) 

Additional photos using the Radial '3D Mask' will appear in the next post – so, stay tuned…


* * * * * 

The Goerz 3D Condenser – many Surprises?

The primary mystery surrounding the Goerz 3D condenser is, "What is this thing supposed to do, and, how is it supposed to do it?"

Fortunately, careful testing of this 'Mysterious' condenser has led to a better understanding of its intended workings. However, not only that but this testing has also uncovered a number of surprises surrounding its design and use.

All of this hinges on understanding the functions of strange plate located in built into the bottom of the condenser, a few mm above the Condenser Iris. This plate has a clear central strip (or, 'slit') about 3.5mm wide separating two partially-silvered areas, which exhibit about 25% light transmission, and together fill the remaining aperture area.

The key to understanding this Condenser seemingly lies in understanding the intended function of this central slit.

To better appreciate this, consider a normal condenser where the Iris is closed down such that the objective aperture (as seen in the rear of the objective) is only about 1/4 to 1/3 open.

Under these conditions the effective NA of the objective is reduced considerably (to just a bit more than 1/2 of its rated value, actually), the image contrast is increased, and also the 'depth-of-focus' in the Object plane is increased somewhat. However, all this is accomplished at the cost of object resolution, which is reduced to about 1/2 of its maximum value. Even worse, this form of masking results in an image that is objectionably harsh and prone to various undesirable artifacts.

Now, consider what happens if the normal, circular Iris opening is replaced with a slit opening of the same width, but whose length spans the full aperture of the Condenser…

In this case we might expect full resolution in the direction of the length of the slit but also an increased depth-of-focus in the direction of the width of the slit. In this way we might hope to achieve improved image contrast and 'depth-of-focus' while sustaining only minimal loss of resolution. However, although image brightness is somewhat better than with the minimally-open  iris method (above) the image is still a bit darker than normally desired and some image artifacts may still be observed.

Now, making the "sides" of this slit semi-transparent, instead of fully opaque, would seem to offer the prospect of moderating the least-desired effects – increasing image brightness and minimizing possible image artifacts, while also allowing the device to be used "normally", if desired, with little loss of imaging performance. This seems to be the Goerz approach, based on examination of the device and initial testing.

So, this is the current thinking on how the Goerz might have been intended to work – no magic, no Mystery; just some basic physical optics!

In other words, the "3D" effect seems to be mostly marketing hype for a slight, directionally-biased increase in 'depth-of-focus' resulting from the use of a somewhat unique, built-in slit aperture.

Since it was not practical to remove the critical 'slit plate' from the Goerz condenser, testing of this concept was performed on a similar (but totally normal) common NA 1.25 Abbe-type condenser.

Different Slit-type masks were compared with normal Bright field and with common COL-type masks in a preliminary effort to determine whether Slit-type masking provides any observable benefits, and, indeed, any detectable "3D" effect.

The initial results of these tests are summarized in the photos below…

(Click anywhere on the above image panel for a larger version.)

The there are two big surprises revealed in these photos: (1) despite its novel design and construction the Goerz design actually performs rather poorly as a "3D" device, barely equaling the depth-of-focus performance of a number of more common alternatives (as shown), and, (2) the real surprise lies in the number of alternative methods which easily equal, if not surpass, the performance of the Goerz device, both in terms of resolution and image contrast!

In other words, despite its novel design, elaborate construction, assortment of optics (optional tops) and unique optics-offset adjustment, the device seems to do nothing which cannot be accomplished by simple alternatives.

It only genuine attractions are its low-power (NA ~0.5) Brightfield mode (using no top), and its Darkfield modes, all of which are mildly compromised by the presence of the non-removable "3D" slit gizmo embedded in the basic design.

As far as possible "alternatives" are concerned, it may be noted in the photos that these all seem to offer results which are basically similar, differing mainly in the presentation of image details, as well as ease-of-implementation and ease-of use. All of this, of course, being potential food for a future post or two…

See the post of March 28, 2017 (below) for more on the Goerz '3D' Condenser. 

* * * * *