Cinevate Brevis Flip 35mm Lens Adapter Review (Part2 of 2)
By Matt Jeppsen

Cinevate
www.cinevate.com
(647) 723-2664 (Canada)

The following is Part 2 of FreshDV’s two-part test and review of the Cinevate Brevis flip adapter. Today we cover Image Performance and Workflow, Accessories and Miscellaneous and finally Summary and Conclusions. You can read Part 1 here. Download the supplementary archive of all referenced charts below.

 Brevis Flip Review Supplementary Chart Pack: Download

Image Performance and Workflow
Having a flip option solves one of the biggest issues in the 35mm adapter shooting workflow. Without an adapter that inverts the image correctly, users are forced to find inventive ways to monitor and frame their images. Many arrive upon the solution of a production monitor like the Ikan or Marshall, mounted upside-down on an articulated arm. On some cameras you can use a strong neodymium magnet to engage an LCD’s forward-facing flip, while others even do minor surgery on a camcorder to engage that switch. I should note that with any 35mm adapter it is highly recommended that you use a high-quality external monitor to check critical focus…when you are shooting wide open on a fast telephoto lens DOF can be as shallow as mere inches and minor focus errors will ruin a shot. That being said, it is not impossible to use only your camcorder’s LCD, particularly when it is as sharp as the EX1 or even the Z1U. And using an adapter that flips the image really opens up this option for users.
The other area of workflow that the flip improves on over the Pre-flip is post-production. When shooting with a non-flip adapter, footage has to be manually inverted in your editing software. This is relatively simple to accomplish, but can introduce other challenges. For instance, when you place inverted footage on a Final Cut Pro timeline, the thumbnails on the clips remain upside down. This is a minor annoyance, but it really can impact the speed of editing if you have learned to lean on that particular feature. Shooting with the Brevis Flip is elegant in the sense that you don’t have to mess with these workarounds and patches. It just works.

35mm adapter design in general has seen a lot of research and development in the past few years. As such, the image quality has seen numerous improvements over early designs. The Brevis Flip is no exception. There are generally three major areas that shooters care about with lens adapter designs; light loss, sharpness, vignetting, and chromatic aberration. Note that sharpness is never perfectly consistent across the frame of ANY 35mm adapter. Some designs show significant resolution drops on the edges and in corners of the frame. The same is true of chromatic aberration and fringing artifacts, these issues tend to be more obvious on the edges. Vignetting and portholing is also inherent in all 35mm adapters, I’ve never seen one that didn’t exhibit these issues to some extent. It is important to note that the 35mm lens itself can have a dramatic effect on these behaviors, particularly with the issue of vignetting. For this review I thoroughly tested the Brevis Pre-Flip and Flip in these important areas of image performance. Towards the end I’ll also take a look at image strobing and grain.

Light Loss
All 35mm adapters have a certain amount of light that is “lost� in the design, this is a fact demanded by physics. Consider that the light loss of a 35mm adapter must be added to the native sensitivity (or insensitivity) of the camcorder it is attached to. You are also limited by how fast your 35mm lenses are. These three factors can really add up to quite a bit of light lost, which means that shooting becomes more difficult without good lighting gear. So the native light loss of an adapter is a very important factor. In testing with 50mm f/1.4, I found the Brevis Flip to exhibit about a half-stop of light loss at f/1.4. What’s interesting is that the Brevis Flip adapter loses the same amount of light as the Brevis Pre-flip…there is no change. That is very impressive…the addition of an optical array hasn’t adversely impacted light performance.

As some users have noted, the Brevis Pre-flip has a unique non-linear response to lenses and light. I’ll not delve deeply into that here, it has already been thoroughly discussed at the DVXuser forums and also in a video released by Cinevate. The one part I DO wish to touch on is that the non-linear response that users have noted seems to also have a slight effect on light loss. At f/1.4, f/2, and f/2.8 both the Brevis Flip and Pre-flip exhibit between .5 to .6 stops of light loss. At f/4 it measures about 3/4 of a stop (.75). This is consistent with user reports that the Brevis doesn’t seem to “get any brighter� when they open up their lens from f/4 to f/2.8 and further. So there is no apparent light performance advantage for users shooting with fast lenses (in this case anything brighter than a f/2.8). But it cuts both ways…being that there are fewer benefits in buying fast glass (with the exception of shallower DOF), users can shop the slower, cheaper glass for their kit. A 50mm f/2 is significantly cheaper than a 50mm f/1.2. On the other hand, it would be preferable if the Brevis could take full advantage of faster lenses. To answer some of the questions raised in regards to this behavior FreshDV is going to be conducting some further testing on this unique light response curve. When we’ve had time to reach some conclusions we will release more information on exactly what we think is going on. Suffice to say that most of the Brevis user reports on non-linearity in the upper range of a lens match perfectly what we have measured in the lab.

Resolution and Sharpness
One of the side effects of shooting with a 35mm adapter is the noticeable loss in image sharpness. Depending on what you are trying to accomplish, this can be either highly desirable or a major concern. Many suggest that the softening effect of 35mm adapters is one of the major reasons they look so filmic in contrast to straight video. Others want to eke every little bit of sharpness and resolution out of their image workflow. I tend to fall somewhere in the middle of that argument.

For this test I used a DSC Labs Multiburst Resolution chart. This amazing test pattern is capable of measuring up to 1200 lines of resolution in Line Pairs per Picture Height (LPPH). The two test cameras used were setup at a common baseline profile. For instance, on the PMW-EX1 test camera, the Detail and Crispening settings were at “0� which is essentially neutral. The HDR-FX1 test camera used a sharpness setting of “7� which is in my experience as close to neutral detail as you can get. In all tests we first baseline tested the camcorder, and then attached the Brevis adapter and measured it’s performance. It’s interesting to note that the camcorder and recording format used with the adapter has a direct effect on measurable resolution. Another major factor noted was the amount of zoom-in on the camcorder; specs and capabilities of fixed lenses on camcorders vary widely between manufacturers, and each must be setup a little differently when mounting a 35mm adapter. For instance, on the Sony FX1 you have to zoom in to at least Z82 to clear the edges of the Brevis Flip imaging element. With the Flip unit you can in fact zoom in further, all the way to Z99 (the Pre-flip limits you to Z71). However, I noted that sharpness tends to suffer on the Flip unit the farther in you zoom. This is not so much the fault of the fixed camcorder lens, as they tend to remain fairly consistently sharp through the majority of the lens. This must have to do with a combination of the lens adapter optics and the 35mm lens itself.

When counting lines of resolution, we at FreshDV always try to err on the side of caution. Our numbers take into account the banding effect when viewed on a monitor and at what point the line pairs tend to blend together. I want to note something interesting about the line counts of the PMW-EX1. I rated the stock EX1 Fujinon lens at a relatively low 750 Vertical and 675 700 Horizontal lines of resolution (in 1080/24p HQ mode). However, the “apparent� or “discernable� resolution of the EX1’s image is actually much higher, closer to 1000 x 950. This effect is most pronounced on the EX1 test camera, the FX1 charts don’t show nearly as much “apparent� resolution over the actual line counts.

I found that the EX1 delivers optimal sharpness with the Brevis Flip when the camcorder lens is zoomed to Z75. This is nearly as wide as you can safely set it without getting the image element edges in frame. Configured as such with the CF1Le element and a 50mm lens at f/4 it delivers 700 Vertical and 700 Horizontal lines of resolution. “Apparent� sharpness is somewhere in the low 900’s. As is always the case with adapters, chromatic aberration and fringing is slightly more pronounced vs the stock camera lens. Image corner and edge sharpness drops a bit, but remains well within reasonable limits and discernibly crisp past 725 lines. I determined that the FX1 delivers optimal results with the Brevis Flip when zoomed in to about Z82. This is very close to the threshold where the imaging element edges are in frame (you may find you need to zoom in a little further). Configured with the CF1Le element and a 50mm lens at f/4 it measures 550 Vertical and 475 Horizontal lines of resolution. “Apparent� sharpness is in the area of 625 x 575. Chromatic aberration on the edges of the frame is noticeably increased and edge sharpness suffers a bit but remains crisp a bit past 525 lines. Changing from the CF1Le imaging element to the CF1L has little effect on center frame resolution, but edge sharpness suffers slightly.

For a detailed breakdown of the Brevis Flip, Pre-flip, CF1L and CF1Le imaging element performance, see the attached Figure 1 resolution matrix chart. Convenient comparison crops from the Brevis Flip + EX1 resolution charts are compiled into Figure 2. And if you wish to analyze resolution performance further, we’ve archived all our FX1 and EX1 raw uncompressed TIF charts in one big supplementary ZIP download. Those charts show the sharpness differences at various camcorder zoom levels.

Vignetting and Edge Light Falloff
In my experience, all 35mm adapters exhibit some vignetting on the image edges. This is sometimes called portholing. There are a number of factors that contribute to this; the adapter design and optics, the performance and quality of the attached 35mm lens, the focal length of the 35mm lens, and the zoom-in factor of your camcorder, to name a few. I tested the Brevis Flip equipped with the CF1Le imaging element to determine how evenly light was distributed across the frame. To do this I set up a plain white background and lit it evenly with soft sources, ensuring light levels were consistent across the width of the frame. Using a waveform monitor to show IRE percentage levels, I baselined the FX1’s stock lens at three different zoom settings. By examining the waveform’s IRE curve and noting the percentage numbers on the far left, middle, and far right, I was able to determine a measurable representation of vignetting (or edge light falloff) in the system. I then mounted up the Brevis Flip, attached various focal length 35mm lenses, and noted the IRE curve at each of the three FX1 zoom settings. The lenses used were a 35-105mm Tokina zoom lens set to 35mm, a 105mm Nikon, and a 50mm Nikon. For a visual representation of the 50mm lens waveform at Z82, see Figure 3.

These test result numbers clearly illustrate that longer focal length lenses on the Brevis Flip tend to show a much flatter curve across the frame. The 105mm at f/4 and the camera set to Z82 exhibited just 2% IRE falloff from the center to the outer edges of the frame. That is about as even as you can expect out of a 35mm adapter…it’s just 1% different from the stock lens at that zoom level. And it’s important to note that the consistency of the lens itself is a major factor; at Z82 the 50mm frame left dips 4% IRE, while frame right shows a 10% falloff. But at Z94 the 50mm frame edges showed a much more respectable 3% IRE drop at both edges. As I said before, there are many factors that contribute to vignetting. Zoom-in distance on wide lenses is even more of a factor than on long glass; the 35mm lens showed relatively even light falloff on the edges that progressively increases as the camera is zoomed out wider. At Z94 the 35mm only shows 5% IRE edge loss, but back the camera out to Z82 and watch it drop a whopping 23% IRE.

Now it seems that the easy answer to vignetting is to just zoom in as far as you can, right? That would be true if all you care about is vignetting. When you study the Figure 1 matrix and realize that image resolution tends to drop as you zoom in the camera, it becomes more of a balancing act between sharpness and an evenly illuminated frame. Also, the farther you zoom in, the less field-of-view you have to work with. Suddenly that 50mm lens doesn’t frame like a true “normal� focal length 50mm lens, so standard composition practices and aesthetic choices are impacted. There is no perfect solution to this issue, only personal decisions that must be made. You have to decide what you can and can’t live with and modify your adapter configuration accordingly. Thankfully the Brevis Flip enables many more zoom distance options over the Pre-flip unit, which limits FX1 users to Z71 only. To dig a bit deeper into these test results, the details for all the lenses are available along with the resolution matrix in Figure 1.

Chromatic Aberration and Edge Sharpness
Chromatic Aberration (CA) and softening at the image edges are both common issues with 35mm adapters. As I said in the vignetting section, both issues are present in every adapter I’ve seen footage from. It’s just a matter of how much, and under what circumstances. As with the other image areas discussed above, these two concerns are directly affected by the camcorder’s zoom-in setting and the lenses in use. But in my opinion, the biggest contributor to CA issues is the achromat lens in the adapter optics. I was pleased with the Brevis Flip’s new achromat, as it created significantly less image CA at all camera zoom levels versus the Pre-flip unit’s achromat. This improvement is particularly noticeable in the FX1’s footage.

I found that on both test cameras sharpness was noticeably higher at wider zoom settings. On the EX1 at Z75 the corners are discernably “crisp� at around 725 lines of resolution. But increasing the zoom-in factor to Z80 drops that “apparent� resolution to around 675 lines. CA also appears to increase slightly. The same is true for the FX1 camera, resolution at the edges drops and CA artifacts significantly worsen as you zoom the camera in further. There is a direct correlation between camera zoom-in distance and edge sharpness.

Comparing the CF1Le and CF1L imaging elements on the EX1 showed no real differences in edge sharpness, perhaps because the EX1 is so sharp that subtle softening tends to be hidden in all that detail. FX1 edge sharpness performance was slightly better with the CF1Le element. As stated earlier, CA is less apparent with the Flip unit mounted on the FX1 vs the Pre-flip. As a side note, the CF1Le option installed into a Brevis Pre-flip also improves edge sharpness over the CF1L element. For a closer look at the EX1 edge performance with the Brevis Flip, study the corner comparison crops in Figure 2. You can peruse through FX1 charts in our full downloadable archive of supplementary charts.

Strobing and Grain
Brevis adapters use an oscillating imaging element design that vibrates the image in short, effectively random movements. It’s not completely random in my opinion, there is in fact a basic pattern of movement. Now if the adapter’s imaging plane didn’t move fast enough or with enough intensity you would probably see what looks like grain in the image. Oscillating adapter users will also sometimes point out fixed pattern noise, when the grain structure is vibrating in what appears to be a consistent pattern. Grain and noise is most obvious when shooting at higher f-stops like f/8 and f/16 (and particularly when the image frame is darker). It can also manifest at times when shooting against the sun. That is the most likely way you would spot the oscillation pattern, by stopping down and shooting against the sun or a clear blue sky. Fortunately the Brevis Flip (and Pre-flip) comes with the option to tune the image plane oscillation. Using a little care (and in a clean environment) you can open up the adapter tube and adjust a screw that increases or decreases oscillation amplitude. If you are seeing any of the above-mentioned image issues you’ll likely have to tune your adapter and retest until the problem clears up. Thankfully we are given this option, though I wish there was a way to do it without opening up the adapter and fiddling with a jeweler’s screwdriver. Tuning the adapter in this manner will also allow you to shoot at higher electronic shutter speeds without strobing. In my experience, you can shoot up to shutter speeds of around 1/125 without noticeable image flicker or strobing. Some will no doubt find that up to 1/250 is acceptable, particularly in well-lit environments when the lens is wide open. What this means is that you are slightly limited in the ways that you can control light in very brightly lit environments, like a sunny exterior shot. If you wish to shoot down in the lower f-stop ranges to enhance shallow DOF, you’ll have to use ND filters on the 35mm lens or in a mattebox. It’s not recommended, but you can also use your built-in camcorder ND filters. They work fine in a pinch and in my experience have little or no effect on the quality of the image.

Accessories and Miscellaneous
The Brevis Flip is sold without any rod support system, adapter support bracket or baseplate. So when buying the Brevis Flip (or any adapter for that matter) I suggest you also budget in a support system. Cinevate does in fact offer their own accessories in the form of solid carbon 15mm rods in various lengths, an adjustable riser camera baseplate, and of course the adapter support bracket. All their current offerings are standard 15mm compatible, so you are free to mix and match brands as you please. Having tested the new 15mm solid carbon fiber rods, I came away very impressed. They are amazingly stiff and quite lightweight. I prefer the 45cm (18�) variety. Unless you already own an adjustable riser plate, you’ll probably need one for the Brevis Flip. If you use the Flip unit in it’s lower vertical configuration the camera will have to be lifted about 3� over the rods. The Brevis adjustable riser plate can handle this. Whether or not you choose the aforementioned Cinevate gear, you will definitely need the Brevis mount to support and lock the imaging tube to your rods. The new tool-less revision of the mount offers a little over 2.25� inches in vertical adjustability, and the top Brevis tube clamp can be removed in seconds without breaking out an allen wrench. It works equally well on Cinevate rods and other 15mm brands as well.

Battery life of the Brevis Flip is the same as the Pre-flip; upwards of 20 hours. The internal rechargeable battery can charged in about an hour with the plug-in AC adapter. As the battery cannot be serviced or swapped in the field, Cinevate sells an optional adapter plug that allows the use of standard commodity batteries. With the power on, the Brevis does create a minor vibration and a small amount of noise. This vibration has never affected the image in my use. However, in very quiet shooting environments I can hear the motor running in my built-in camera mic. This can be addressed with a good shotgun or off-camera mic, but is certainly worth noting. The Brevis Flip comes from Cinevate with a 1-year warranty. In my personal experience as a Brevis Pre-flip owner, the company has been quite responsive to the community and keeps a good eye on their user forum.

Summary and Conclusions
The Brevis Flip is a very solid 35mm adapter choice. It is well suited to a wide variety of small-format HD camcorders and really excels on the extremely sharp Sony PMW-EX1. Setup and configuration is reasonably simple as 35mm adapters go, and the modularity and adjustability of the entire system is a very strong point in my experience. It is also very solidly built despite the relatively light weight. I like the fact that Cinefuse elements can be swapped in and out to change image characteristics, and the super-long battery life and short recharge times are impressive. I wish that the battery was not of the permanent internal variety, but give props to Cinevate for providing a solution (albeit optional) for emergency power in the field.

While I greatly appreciate the ability to fine tune the adapter oscillation, this feature would be far more useful if it could be tuned on the exterior of the adapter tube on the fly. Image quality is very good as 35mm adapters go, particularly considering the optical array required to flip the unit. One would expect a hit in quality when adding more glass vs the Pre-flip unit, but the Flip actually improves general image quality and without any further expense of light. That is an impressive accomplishment. The non-linear light response behavior is a bit puzzling, but once understood is sort of a non-issue unless you have a major investment in fast lenses. Edge to edge image sharpness is well within the range of acceptability when the camera is properly configured, though certainly not perfect. As noted earlier, I have my doubts at how perfect this can ever be in 35mm adapter designs. The CF1Le element seems to have improved the area of edge sharpness versus the CF1L option (and is a worthy upgrade to the Pre-flip unit alone).

Setup and configuration on the Flip unit is much easier than the Pre-flip; for one, the image is upright. You’d be surprised how much simpler it is to adjust the unit with a properly inverted image. The second reason is that the optics of the Flip offer much more zoom-in distance, you can set the camera from Z82 all the way in to Z99 (you are limited to Z71 on the Pre-flip unit with the FX1). So it’s much simpler to get critical focus on the grain of the imaging plane, and then readjust zoom levels as desired. As discussed in the vignetting and resolution sections, zoom-in factor has a major effect on image performance. I have found that I am always adjusting to attain the perfect balance of resolution, edge sharpness, vignetting, and field-of-view. Thankfully the Brevis Flip allows this much adjustability, and users can determine what settings are acceptable for their own needs. In fact, that is really one of the best things the Brevis has going for it…a completely modular design and configurability throughout the system. The Brevis Flip is available as an upgrade to the Pre-flip for 569.25 or as a complete bundle for 1405.80. More info at www.cinevate.com.

After much testing and use in the field, I’ve decided to rate the Brevis Flip unit at 4 out of 5 on our Fresh scale.

Additional Notes
Zoom Numbers: Sony FX1 users will probably notice that I used “Z� numbers to note the zoom level of the camera lens, yet the FX1 only offers seemingly-unrelated numbers on it’s external zoom ring (and no measurable percentage scale on the display). However, the “pro� Z1U is the fraternal twin brother of the FX1 and one of the features it adds is a scale of Z numbers that relate to zoom percentage. I was able to match the two camera’s identical external zoom rings and note the corresponding Z scale on the Z1U’s LCD in relation to the FX1 zoom ring markings. For reference, that complete correlating scale is shown in Figure 4. For a visual reference of what Z82, Z94, and Z98 should look like on the FX1 zoom ring, see Figure 5.

EX1 charts corner shadow: The Brevis Flip test unit arrived with a small chip missing from the lower-right corner of the upper prism array. See Figure 6. Cinevate was aware of this defect and informed us prior to shipping the unit. As this was one of the first flip units off the production line, we chose to go ahead with the test and work around the issue. The defect had no noticeable effect on our FX1 test results, but showed up in the EX1 charts as a slight softening and shadow in the extreme lower-right corner of the frame (and only in that corner). As such, we disregarded that particular corner of the EX1 frame when compiling test results.