Mitakon Creator 85mm f/2.8 1-5X Super Macro
Only a few lenses designed for system cameras are capable of a continuous range of magnifications above 2x without requiring bellows, extension rings or additional optics. Among these lenses, the Canon MP-E 65 mm f/2.8 1-5x Macro Photo, as its name suggests, provides a continuous magnification range from 1x and 5x. It is the most expensive among such lenses, but its image quality is generally regarded as excellent only up to 3x, and just acceptable at 5x. It needs electronic signaling between camera and lens, and therefore it is mainly usable on Canon DSLR cameras, as well as a number of mirrorless cameras via a "smart" lens adapter that translates the necessary signals between camera and lens.
This lens, hereby referred to as the 85 mm 1-5X for brevity, provides a broad and continuously variable range of magnifications matched by only another current lens, and an exceptionally long working distance. All this in a relatively compact package and at a price substantially lower than the only other lens for system cameras that offers the same magnification range. It would be unrealistic to expect that these characteristics are not accompanied by some compromises in image quality. Even the more expensive Canon MPE 65 mm 1x-5x cannot compete with the best fixed-magnification copy lenses like the Printing Nikkor 105 mm at 1x, or infinity corrected microscope objectives like the Mitutoyo M Plan Apo series.
Therefore, the question I want to answer with these tests is not whether the 85 mm 1-5x is as good as these super performers (I already know it is not), but rather whether it provides usable images in a more limited context, e.g., the routine macro work typically performed at museums and research institutions, where digital images only need to be good enough to print on good-quality paper at sizes of at most 15-20 cm (most often, significantly smaller than this), or to fill an HD screen. In this context, a compact and relatively inexpensive kit containing an ordinary macro lens that reaches 1x, plus the 85 mm 1-5x, covers the whole magnification range from infinity focus to 5x with just a single lens swap. Field macrophotography is also an application where this lens can be advantageous, since lens-swapping is best avoided in this case.
The 85 mm 1-5x in practice
It appears from the main specifications that this lens was designed to compete with the Canon MP-E 65 mm, and perhaps also to show that Zhongyi could do better than the embarrassing failure of the Mitakon Creator 20 mm f/2 to provide its initially intended magnification range. However, the first production batch of the 85 mm 1-5x turned out to provide a low contrast and decidedly poor image resolution. A quick redesign by Zhongyi largely corrected these problems, and the Mark II model of this lens visibly differs from the initial release in the lower diameter of its front element (slightly more than half the original diameter of the Mark I).
Zhongyi's explanation of the reasons of the faults of the Mark I model is not quite clear to me. For one thing, the explanation only talks about the diameter of the front element being wrong. Since this element is followed at a short distance by the diaphragm, and subsequently by more optics, it seems unlikely that changing only the diameter of the front element would suffice. The diaphragm must also have been changed by reducing its maximum opening, and the following optics probably as well. Unless the front element of the Mark I model was twice as wide as necessary (in which case, the equally wide diaphragm would not begin to reduce the effective aperture of the lens until stopped down by more than two stops, which should have been noticed by early testers), reducing its diameter is also likely to reduce the lens speed.
All specimens from the defective first batch of lenses should have been returned to the maker and replaced with Mark II specimens at the cost of the maker, but some sellers might still be stocking a few Mark I specimens and selling them (knowingly or otherwise) to unsuspecting customers. Mark I specimens might also lurk among second-hand offerings of this lens.
Interestingly, several months after introducing the Mark II redesign, the Zhongyi web site still shows pictures of the Mark I version. The text on the same web page has been updated for the Mark II (with some exceptions mentioned below). Perhaps for this reason, the majority of online shops still display the same pictures of the Mark I version.
The user instructions that accompany this lens are extremely unhelpful. They are just generic guidelines like "If you find mechanical breakdown in lens or dust inside the lens, please reach to us with the contact information in the instruction, rather than disassemble the product by yourself.Wrong operation may leads to permanent damage." [errors and mistypings are original]. The most technical information includes generic compatibility advice on third-party lenses with a handful of the most popular camera brands (which likely you are already aware of, as far as your camera is concerned). The instructions contain no reference whatsoever to using this specific lens.
Useful reviews and tests of this lens are not numerous. Several reviews are YouTube videos, which I generally find slow-going and a big waste of time. I can visually scan a written review and find the interesting points (if there are any) in one-fifth to one-tenth of the time I must spend wading though a YouTube video. Many of these YouTube lens reviews are so insufferably poor that I never make it to their end. It appears that the creators of these video reviews firmly believe that filming a reasonably good video review is far quicker than writing one, while in reality the exact opposite is true. For these reasons, I am not recommending any video reviews of this lens.
This lens is available in a variety of DSLR and mirrorless mounts. Whenever a legacy or manually operated lens is available in Nikon F mount, I tend to choose this mount for two reasons:
Since this is a fully manual lens without a built-in "chip", the camera must be configured to "Shoot without a lens" or equivalent setting.
Mechanical and optical build
The optics of the 85mm 1-5X are mounted in an almost totally metal barrel that feels solid and well-built. Total length (mounting flange to front of filter mount) of the Nikon F version is 122 mm with the lens at 1x, and extends to 166 mm at 5x. The front optical group is fixed at the front of the extending barrel. The aperture ring is located at the front of the magnification ring, which replaces the focus ring in this type of lens. In general, when using this type of lens, you should first select a magnification on the magnification ring, then adjust the distance to the subject in order to bring the latter into focus. Using the magnification ring to focus changes the magnification, which is generally undesirable in scientific and technical photography. It also changes the subject framing, which is undesirable in all types of photography.
At 789 g including the lens collar, this is a heavy lens. It has no visible plastic parts except for the long matte black sleeve surrounding the rear element.
All lens markings are engraved or laser-etched, and unlikely to be erased by handling the lens over the years.
This lens contains 12 elements in 8 groups. Zhongyi does not publish the optical schemes of its lenses, and I found none online. The rear optical element is mounted approximately 53 mm in front of the lens mounting flange, which means that almost half of the lens barrel is empty of optics. Looking through the rear of the lens, it is possible to see the perimeter of multiple lens elements and/or metal rings, a strong indication that the contrast of this lens is reduced by inadequate internal baffles and light traps. No lens shade was mounted at the front of the lens while taking the above picture, which is therefore indicative of how the lens behaves without one.
The front element diameter is 18 mm and the rear element roughly 15 mm, with intermediate elements even smaller. It would have been possible to reduce the diameter of the lens barrel, or at least of the portion at the front of the magnification ring, to one-half of the present.
The magnification ring is rather stiff to turn. Considering that it is only used to choose the magnification and not to focus, and that its stiffness prevents the lens from extending when mounted vertically, this is not a bad thing. The machined surface of this ring is ample enough to provide a good grip with three or four fingers.
The filter mount diameter is 58 mm, exactly like the Canon 1-5x lens. Zhongyi repeated the same mistake as Canon, missing an opportunity to design a front of the lens barrel as narrow as possible to make it easier to illuminate the subject. However, the much higher working distance of the Zhongyi lens and the lower overall diameter of its lens barrel reduce the negative effects of this fault. Incidentally, no one I know of ever uses filters on this type of lens, except possibly a "protector" to prevent soiling the front element. The Mark II lens could have been redesigned to use a 20 or 25 mm filter and an accordingly narrow front of the barrel.
The LED ring light provided with the lens is the same shiny aluminum contraption that accompanied the Mark I version, and reduces the working distance by almost 16 mm (plenty of working distance remains, so this is not a problem). The throat of the central opening in the ring light is matte black with concentric light traps, but is a bit too wide to work as a lens shade. It should have been redesigned to match the smaller front element of the Mark II model. If a UV filter/protector is used on the lens in the field to protect its front element against water drops and accidental contact, it may be a good idea to mount this filter between lens and ring light.
The shape of the machined slot for the USB connector suggests a USB-C socket, but the socket in the slot is instead Micro-B USB. A 1 m white USB cable is provided. A black cable may be less conspicuous, especially with reflecting subjects. The ring light can be powered by a USB power bank (not supplied) in the field.
The diffuser mounted in front of the LEDs seems to work well, and the individual LEDs are not visible through its surface. The light is relatively bright, but given the long working distance, the main usefulness of this ring light is to facilitate framing and focusing. Without extra illumination, it is hard to see the subject even on the LCD live view. Especially in the field, electronic flash (with a suitable diffuser/reflector mounted on the camera or lens) remains a necessity.
It has been reported that this ring light is fragile. Perhaps the USB socket is its weakest point. A well-planned ring light for this lens could use the full difference in diameter between the 58 mm filter mount and the 18 mm front element to incorporate multiple concentric rings of LEDs, which would make illumination quality more even and stronger than the current narrow doughnut.
A narrow lens shade, attached to the filter mount of the lens (or in front of a filter) cuts out off-axis illumination and slightly improves contrast. The above figure shows a combination of lens shade and step-down rings that does not vignette even on full frame. The front opening of the lens shade is narrowed down to 28 mm by a step-down ring. A single step-down ring, alone, can be used as a lens shade in cases where keeping the maximum working distance is necessary.
My tests with even narrower lens shades show that a 25 mm female filter mount at the front of the lens shade does not visibly vignette on full-frame. The 25 mm thread is the smallest I found in the form of a step-down ring.
Using a 25 mm long stack of extension tubes for RMS-mount microscope objectives as a substitute for the lens shade also produces no visible darkening in the corners at any magnification and aperture on full-frame, in spite of the narrow opening (around 18 mm). From a geometry point of view, RMS extension tubes do slightly restrict the cone of light entering the lens, especially at 1x, but the edges of this lens shade are apparently too much out of focus to become visible. However, RMS extension tubes from different makers can differ in their inner diameter, so results with this type of lens shade may vary. It may also be necessary to paint the interior of these extension tubes flat black, especially if not threaded. Flocking the interior of the extension tubes is likely not practical, given the thickness of the flocking material. RMS extension tubes are potentially better at improving contrast than wider lens shades, but if you wish to use these extension tubes you should be prepared to do some experimenting.
A conical adapter with RMS thread at its narrow opening (or a wider thread for different microscope objectives, e.g. 24 mm) may also be satisfactory as a lens shade, but most of these adapters require a further adapter ring to screw on the filter mount of this lens. It is usually necessary to flock the interior of this type of adapter, since most of them have shiny interior surfaces. Also in this case some experimentation is necessary. RMS conical adapters, in particular, may be too long and vignette a little.
The shoe of the lens collar is compatible with Arca clamps. It also has a 1/4"-20 threaded socked at the bottom. Since the threaded socket is meant for mounting the lens on a tripod head, the lens shoe does not have the small retaining bolt at either end of the Arca-compatible plate, and loosening the Arca clamp while the lens is mounted on it allows the lens to completely slide out of the clamp. This has disastrous consequences if the lens is mounted vertically or sharply inclined and the clamp is loosened without holding the lens with the other hand. It would be a nice touch for Zhongyi to provide a threaded M3 hole at either end of the shoe bottom, and two M3 bolts and an Allen key in the lens package to allow users to mount these safety bolts if desired.
Camera-based image stabilization is unlikely to help at the magnification range offered by this lens. In particular, to really help, with these lenses IS should be able to detect where along the lens axis the center of rotation is located. Without this essential piece of information, the camera cannot shift the sensor by the amount required to counteract this movement. Therefore, you should switch off IS entirely.
Whenever possible, you should also attach the lens shoe to a focusing rack mounted on a solid tripod. In the field, you may be able to get away with hand-held shooting with electronic flash by shooting numerous images of the same subjects, if you can accept a low percentage of keepers.
Actual magnifications and working distances
Since the magnification ring rotates continuously without intermediate clicks or detents, the only repeatable magnifications are at the 1x and 5x ends of rotation of the magnification ring.
If precise magnification of a given subject must be documented, the most reliable way with this lens is to take pictures of the subject and of a ruler (e.g. a microscope resolution test slide) without changing anything else in the setup. Alternatively, with some subjects it may be possible to measure e.g. the length of the subject with calipers. A less repeatable method is to set the magnification in correspondence of one of the magnification settings displayed on the magnification ring. These positions are displayed on the magnification ring as integers from 1:1 to 5:1, and the colons between digits can be used to align the corresponding magnification to the index mark on the lens barrel fairly accurately. The magnification markings are linearly spaced along the magnification scale, but some guesswork is unavoidable for non-integer magnifications like 1.5x.
Magnification is given by
m = L' / L
where m is magnification, L the actual length of the subject, and L' the imaged length.
Based on images of a ruler with half-mm ticks shot with a Sony A7R II camera (with active sensor width 35.8 mm), the actual magnifications provided at each nominal marking on the magnification ring are:
I am aware that, in at least one case, an actual magnification different from the above has been reported for the 1x setting on the Mark II lens. Nonetheless, the above are my measurements on my specimen of the lens.
The working distances specified on the Zhongyi web site are also somewhat different from my measurements. This is perhaps a side effect of the redesign of the lens. It is possible that the working distance specifications on the Zhongyi web site refer to the Mark I version, and have not been updated for the Mark II. Either way, these working distances are remarkably high, especially at 5x. The working distance of the Canon MP-E 65 decreases from 100 mm at 1x to 42 mm at 5x, which is less than half of the 85mm 1-5X at the same magnifications.
One notable behavior is that switching between 4x and 5x requires no adjustment in focus, or just a micrometric adjustment. While changing magnification between 4x and 5x, the image first becomes unfocused, and then returns into focus. This is caused by the front of the lens extending toward the subject when focusing from 4x to 5x, and is probably just a lucky coincidence, rather than a design feature.
Lens aperture and speed
Pupils and pupil ratio
According to the Zhongyi web site, this lens has "near telecentric performance" (I assume, on the subject side). A telecentric lens on the subject side has the entrance pupil located at infinity behind the lens. In addition, the diameter of the front element of a telecentric lens at 1x must at least equal the diameter of the image circle. Neither condition is true in this lens.
The entrance pupil as seen through the lens front is located roughly 15 mm behind the front optical element. The distance between the front element and aperture does not change with magnification. At best, the lens behaves "almost" like a telecentric lens in the sense that its perspective rendering is equivalent to that of a lens of long focal length, but it cannot be called telecentric any more than a telephoto lens can. The lens is not telecentric on the image side, either.
The rear pupil is deep within the lens barrel and moves toward the subject with increasing magnification. The pupil ratio, or pupil magnification, is given by
p = pout / pin
where p is the pupil ratio, pout is the diameter of the exit pupil and pin the diameter of the entrance pupil. In the Mark II version of this lens, the pupil ratio is far from unity (as well as difficult to measure), and increases from roughly p = 2.7 at 1x to p = 4.3 at 5x.
Effective apertureThe pupil ratio is exceptionally high for a macro lens, and cannot be ignored in calculations of effective aperture, which therefore must be computed as:
A' = A ((m / p) + 1)
where A' = effective aperture, A = nominal aperture, and m = magnification.The aperture scale on the lens barrel displays f/values between 2.8 and 22. These nominal aperture values only apply at infinity, where this lens cannot focus, so they are only theoretical figures. Thus, with the aperture fully open, and for the moment taking the f/2.8 lens speed at face value, computing the effective aperture with the above formula yields A' = f/3.8 at 1x, and A' = f/6.1 at m = 5 (instead of f/6 and f/17, respectively, had p been unity). There is no doubt that the high pupil ratio was intentionally designed as a way to decrease the effective aperture at a given magnification.
At this point, the question that needs answering is whether the nominal f/2.8 speed is a realistic specification. This question is especially relevant since the redesign leading to the Mark II model substantially changed the diameter of the front element (and, quite likely, of the diaphragm and of at least some of the internal elements), but the official lens specifications were not updated.
Empirical measurements of effective lens aperture
To answer this question from an empirical point of view, I took out-of-focus pictures of a standard white card from the ColorChecker Passport with the 85 mm 1-5x fully open at 1x and 5x settings. The blurring of the out-of-focus images effectively homogenizes the images and eliminates any surface texture of the target. I illuminated the test target with electronic flash at a manually set power chosen to produce a medium-gray image far from both white and black saturation. I then shot a series of pictures of the same card with the same flash power with a Nikon AF Micro Nikkor 105 mm f/2.8 focused at infinity, stopped down at half-stop intervals from f/2.8 to f/22. From this image series, I selected the image quantitatively closest in brightness (see next paragraph) to each of the two test exposures from the 85 mm. The two selected test images from the Micro Nikkor correspond to the actual effective aperture of the 85 mm fully open at the two given magnifications.
While not an exact method, results should be off by no more than half a stop. The Godox AD200 electronic flash used for this test, when configured at manual power, produces repeatable results with discharge-to-discharge differences of less than 1/10 of a stop (my measurements), and therefore contributes negligibly to errors. Quantitatively comparing the different images involves averaging the values of the R, G and B channels among pixels in an area of the image, then averaging the three channels together. The observed pixel-to-pixel differences affect, at most, only the two least-significant bits of the 8-bit values of each channel. Finally, the test images from the Micro Nikkor that most closely matched in averaged brightness the two test images from the 85 mm 1-5x were selected, and the corresponding apertures of the Micro Nikkor were taken as the effective apertures of the 85 mm.
Results of this comparison indicate that, fully open at 1x, the 85mm is equivalent in effective aperture to the Micro Nikkor at f/8, while the 85 mm fully open at 5x is equivalent to f/19. The measured values at these two magnifications differ from each other by approximately two and a half stops. The two effective apertures computed above (f/3.8 and f/6.1), on the other hand, differ from each other by roughly two stops. The inconsistency of half a stop may be due to measurement errors. The most striking result, however, is that the pair of empirically measured effective apertures is much slower than the respective pair of computed ones (by two and a half stops and three stops, respectively). This indicates that the nominal aperture at infinity of the Mark II lens is not f/2.8, but somewhere between f/4.8 and f/5.6.
Theoretical calculation of nominal lens speed
From a theoretical point of view, taking at face value the 85 mm focal length, a front element diameter of 18 mm gives a nominal speed of f/4.8. This is easily computed as a = fl / df, where fl is the focal length of the lens and df the diameter of the front element. This formula actually provides a lower limit for the lens nominal aperture, and the actual aperture can be higher than this value, e.g. in retrofocus lenses (but never lower, because the outer rim of the front element limits the optical surface through which the lens can collect light). The f/4.8-f/5.6 speed computed from the measured effective aperture in the preceding section agrees well with the theoretically computed speed in the present section, and is roughly one and a half to two stops slower than the f/2.8 lens specification by Zhongyi.
The original front element diameter of the Mark I model is close to 33 mm (based on measurements of the diameters of front element and filter mount on pictures of this model). Based on the definition of lens speed, this corresponds to f/2.6 for the Mark I - not too different from the specified f/2.8. Thus, it seems likely that correcting the problems with the Mark I model has resulted in the loss of roughly 11/2 to 2 stops of lens speed, and a corresponding increase in diffraction blur of the lens fully open. This does not mean that image quality of the Mark II is worse than in the Mark I, because the problems of the Mark I were not caused by diffraction. It only means that, had both the Mark I and Mark II been diffraction-limited, the Mark I, fully open, would outperform the Mark II.
The presence of a fixed correcting optical group at the rear of the helicoid-mounted optical assembly of the lens that controls magnification makes it likely that the focal length of the lens actually changes at different magnification settings. This introduces a further uncertainty factor. Knowing that the position of the entrance pupil is fixed with respect to the front of the lens (15 mm behind the filter mount flange, see above) simplifies things. Since the working distance (W) relates to focal length (f) and magnification (m) as:
W = f ((p / m) + 1)
plugging in the actual magnifications, working distances and pupil ratios worked out above, and solving for f, we obtain:
The computed focal length at nominal 1x is not too different from the specification of 85 mm. The value at 1x may differ from 85 mm because of the Mark II redesign. It may also be that 85 mm is a "nominal" focal length for the lens (theoretically) focused at infinity, which reduces to 71 mm at 1x. The computed focal length at nominal 5x becomes even shorter, as typical of macro lenses that have fixed correcting groups or internal focus.
Since we arrived to these focal lengths through a chain of measurements and calculations that likely introduced new numerical errors at each step, these values are only indicative, and we must resist the temptation to use these focal lengths to attempt and refine the calculations of nominal and effective apertures etc.
Incidentally, there must exist, somewhere, a few beta-test prototypes of the 85 mm 1-5x distributed to professional photographers (unless they have all been returned to Zhongyi). Initial tests of these specimens were reported to give very good results in spite of their (presumably) wide front element, so it cannot be completely excluded that Zhongyi might be able to find another solution to the problems of the Mark I model, that does not involve a loss of lens speed. If this solution involves a much higher production cost, however, we might never see this hypothetical Mark III model unless we can convince Zhongyi that they would be able to sell this version in spite of a substantially increased price.
This test uses a glass ruler with ticks at 10 μm intervals or 100 μm intervals, as specified in the figure captions. The images are 1:1 pixel crops of the center portion. The test camera is an Olympus E-M1 Mark II set at 400 ISO and fully electronic shutter. Testing was performed on a Zeiss 57420 measuring microscope stand with added Newport UMR 8.25A linear stage and 500 μm per turn plunger micrometer for fine focus, which gives a sub-μm focusing precision. Focus was established on the LCD preview at maximum magnification (14x as indicated by the camera).
This test is indicative of the lens resolution when used on a Micro 4/3 camera. A separate test was carried out on a full-frame camera.
At 1x, the image at f/2.8 slightly resolves the 10 μm line pairs (with a resolution around 2 pixels per line pair), but at f/4 this detail is already lost. On the other hand, the 100 μm-spaced ticks in these images look slightly sharper at f/4 than at f/2.8. In the rest of the images, f/2.8 and f/4 give roughly the same visible sharpness.
Since I do not have a finer ruler, I performed additional tests at 1x and 5x with silicon chips (not shown). Also these images are relatively blurry, and agree with the results shown above. In general, on Micro 4/3, f/2.8 to f/4 (as indicated on the aperture ring) should be used for best resolution. f/5.6 can be used as a compromise between resolution and DOF. f/8 and higher should be avoided, unless the image is meant to be published at low resolution on the web. As a whole, however, image resolution is relatively poor on Micro 4/3, and many better lenses are available at one or another of these magnifications.
This test was carried out on a Sony A7R II with 42 Mpixel full-frame sensor. The subject is a steel ruler with etched and blackened markings, illuminated by electronic flash with a large beauty dish as diffuser placed as close as possible. The moderate color shift from warmer to cooler at increasing apertures and at 1x-2x is caused by the electronic flash gradually washing out the ambient illumination as the flash power increases. Switching off the ambient illumination entirely would eliminate this shift. Focus was set on the flat portion of the ruler's surface, and the deeply etched furrow at the center of the images is not fully in focus except at its edges.
As expected, the physically larger pixels of the A7R II produce a better impression of sharpness, at least up to 3x. At 1x and f/2.8 to f/4, however, the anti-aliasing algorithm of the camera struggles with very fine detail of the subject and produces numerous small artefacts. This probably indicates that the lens outresolves the Bayer matrix of the sensor, a fact known to promote color moiré and other artefacts. The problem disappears at higher apertures and at 2x and higher magnification, where the lens no longer outresolves the sensor.
It is useful to remember that a 20 Mpixel Micro 4/3 sensor has approximately the same pixel size as a 70 Mpixel full-frame sensor. f/5.6 is still good at 1x and may be acceptable at 2x. f/4 should not be exceeded at 3x to 5x if maximum resolution is needed. However, f/11 and even f/16 may still be good enough for printing at relatively small sizes, e.g. in scientific journals.
The two above images are not pixel crops. The image shot at 5x was cropped on both sides to eliminate about one-third of the image area, then cropped vertically and reduced to a width of 900 pixels for web presentation. The image of the same subject shot at 3x in the same conditions was cropped to show the same area as the preceding image, then reduced to the same width. The image shot at 5x shows more detail, in spite of both images being substantially reduced in size. Contrast at 5x is also better than at 3x. This shows that the higher magnification adds real detail of the subject instead of just empty magnification. Therefore, it is a good idea to fully use the magnification range of this lens, even though the images at higher magnifications are perceived as blurrier than at lower magnifications.
This lens does not perform well on Micro 4/3, because of the small physical pixel size of this format. Presumably, it would work a little better on APS-C cameras, but not by much. Since the 85 mm 1-5x is designed to cover full-frame sensors, this is the only use for which I recommend this lens.
The above uncropped image was shot with the 85 mm at 5x on full frame, reduced in size for web publication, and color saturation was moderately increased to conform to my visual perception of the subject. It looks acceptable on the screen (except for a little sensor dust), although a professional photographer probably should not tell a customer that this image was shot at effective f/76 (i.e. f/11 as shown on the aperture ring). Opening up the diaphragm of course gives sharper images, but DOF drastically reduces (which is why focus stacking is practically obligatory with these subjects).
The above images are crops of the center of the preceding image (albeit not 1:1 pixel crops, but reduced to roughly 1/3 of the original resolution). The difference in sharpness between the f/2.8 and f/11 settings is quite visible, even after reducing the original images.
Even on full frame, image quality (both in terms of resolution and color contrast) cannot compete with the best fixed-magnification lenses. However, image quality on full frame is acceptable for most real-world uses (other than pixel-peeping and printing at poster size). The broad magnification range and exceptionally long working distance at magnifications above 1x make therefore this lens interesting, in spite of its shortcomings. Paired with a conventional macro lens that reaches 1x (for example a 200 mm for even longer working distance at 1x, or a 100 mm for portability and ease of use), this kit gives access to all magnifications up to 5x.
The Laowa 25mm f/2.8 2.5-5X Ultra Macro, coupled with a macro lens that continuously focuses from infinity to 2x (e.g., the Laowa 100mm f/2.8 2x Ultra Macro APO) could be an alternative. The working distance of the Laowa 25 mm (40 mm at 5x) is much shorter than in the 85 mm, but the image quality of the Laowa 25 mm is said to be better, and the loss of the magnification range between 2x and 2.5 x with this kit is not a major problem. However, most current native macro lenses for mirrorless cameras that reach 1x unaided (and a couple that slightly exceed this magnification) offer electronic focusing and associated functionality like in-camera focus stacking and focus bracketing, while the Laowa 100 mm 2x macro is fully manual.
Although not quite relevant to this review, the lens was shipped from Germany and delivered to me in a mangled DHL package (above) that looks like someone stepped on it. Thanks to the large size of the package and its padding, as well as the strong box used by Zhongyi to package this lens, nether lens nor box bear a trace of the mishap. Still, this mangled package is better than the Amazon package shown below, which arrived only days earlier in the state shown in the picture (Amazon refunded the damaged item, which had been cut together with the package).
The black lens box sports the Zhongyi Optics logo, also repeated on the butterfly-type front lens cap (which does not sit well on the lens and easily falls off). The last two kanji of the logo, joined together for aesthetic reasons, are 光学, which unsurprisingly translate to "optics" (I know because in Japanese this word happens to be written exactly the same as in Chinese). The lens, however, carries only a "Zhongyi" logo in Latin (sometimes called English) script.
This box seems very solid and is in fact thick and quite heavy, far better than the disposable cardboard and egg-crate packaging often used for far more expensive lenses. Only the bottom half of the box is padded with form-fitted synthetic rubber foam. The ring light is packaged in a plastic bag and a fragile cardboard box without any brand or text, and may be made for Zhongyi by another company. The lens box is not sealed in any way (or at least, mine wasn't). The plastic bag containing the lens is also not sealed, and does not contain the customary silica-gel package that is supposed to protect the lens from humidity that may penetrate the box during storage and shipment.
The Zhongyi Mitakon Creator 85mm f/2.8 1-5X Super Macro, Mark II, is a fully manual lens that offers a continuous 1x to 5x magnification range and an exceptionally long working distance. This lens was quickly redesigned by Zhongyi after the first distributed batch turned out to give poor image quality. The redesign caused the nominal aperture to increase from f/2.8 to roughly f/4.8-f/5.6, and the effective aperture fully open to roughly f/8 at 1x and f/19 at 5x. Because of the relatively low image resolution and contrast, this lens cannot compete with the best fixed-magnification lenses and microscope objectives, but can find a specialized niche, for example in the field, where a high range of magnifications and a high working distance are desirable, and swapping lenses is better avoided.
A narrow lens shade slightly improves contrast. The lens is sharp up to 3x and still usable at 5x on full frame, but on Micro 4/3 can be called sharp only at 1x and f/2.8.