Cinevate Flip Module - Retail Version

Beat Takeshi said:
boner-ranch.jpg



BOING!!!!
Click to expand...

LOL! Hey Dennis my flip never did arrive :)
 
K, now that we have the retail enclosure done, we'll be doing a new series of tests regarding light loss. At this point, a decent estimate is 0 stops from the adapter/flip itself, and .9 to 1.2 stops lost with a 50mm f1.4 attached, and CF1L installed. The bare Brevis adapter shows about a .5 stop gain with no lens attached. Although no tests have been done on the Letus with a lens attached, we're still projecting a .5 to 1 stop advantage in light efficiency.

Setup is still possible in minutes. Once lens mount micro-collimation and rear XY adjust is done, expect to be shooting in seconds...just like now. We've got a quick release system in the works which will make attachment even easier that it is now. You'll be able to go from fully rigged on rails, to bare cam in about 10 seconds.
 
Last edited:
Wow ! Why the camera sits so high ?
I can't wait to see the first sgpro owner that tried this.
I think sgpro with this flip unit is too much adaptors on 1 camera.
Why can't we make some digital converter, that flips the image? In the jvc hd200/250 and xl h1 they have this option in the menu.
 
Dennis,

Are you saying the flip adapter won't reduce the light at all !?!

I don't think there is a full stop loss with the Brevis and even a 2.8 prime, but I should check my test numbers again. Its actually kind of tricky to measure, because if you just go by say the f-stops on the video lens, those can be quite inaccurate in my experience.
 
Dennis, we keep going back and forth about this.. there is a variable in your test your not accounting for. ITS PHYSICALLY impossibly for the bare brevis to amplify light by ANY stops..... Newtons law of the conservation of energy says so.

ITs actually impossible for it to lose 0 stops... i accept your .5 stop adapter loss claim. If the brevis did amplify light by half a stop, you my friend should recieve a nobel prize. :)
 
It is absolutely impossible for there to be more light exiting an adapter than there is entering it, that's completely true and Timur's spot on with that observation.

But it is possible to measure the amount of light exiting the adapter and come to the conclusion that it's 100% of the light entering it, or even more. And while that's technically possible, it'd just be extremely undesirable, and mistaken. You can magnify the amount of light hitting a certain point, making that point appear brighter, but it'd happen by concentrating the available light into a small hotspot. Think of a magnifying glass -- it can amplify the amount of the sun's energy to the point where it can start a fire, but the result is that you get a small pinpoint of hot light.

What would be technically impossible is to magnify past (or even preserve) 100% light transmission while maintaining an even field of light distribution. All lens surfaces are going to cause light transmission loss, that's just the way the universe works. You cannot gain energy by shining light through an adapter -- if you did, as Timur said, you'd win a Nobel prize and create a perpetual energy machine -- heck, we could solve the world's energy needs just by stringing adapters together in a six-mile-high chain and shining the results onto a solar panel... :)

If an optical path appears to be magnifying the amount of light, it'd have no choice but to be producing a hot spot in the center and vignetting in the corners. If you measured the hot spot through a light meter you could maybe show a gain, but you'd really have to throw the unit up on a waveform monitor, shooting a white card, and look at the light falloff at the edges. If you're seeing a flat line that's good, if you're seeing a "hill" (higher in the center, lower at the edges) then that's bad and an indicator that there's vignetting going on.

Having a vignetting hotspot wouldn't be gaining light, you'd just be moving it around. The only way to actually gain light would be to add a light source inside the adapter.
 
Last edited:
regardless in my eyes it appears the brevis 35 with flip will cost about $1300 but allow for different focus screens and more lens adapters than the letus.

The letus is also a limited time offer at $1200 too AND includes a rod support.

I think if these systems stick around $1200 then theyll be a great offer at only a stop or two of light loss
 
I've always described the light loss issue in terms of what shooters would expect to see on their camera with and without the adapter given the same lighting, distance and framing. How/where that light loss get's attributed is a matter of mathematics, but the fact remains that with a 50mm f1.4 lens attached, users can expect to see a relative difference of .5 to .7 stops with the non-flipped Brevis. I agree with the solar panel analogy 100%, however what we're comparing here is a bare camera vs a camera with a 35mm lens projecting onto a small screen (a break in the convential chain of optics), which is then relayed to the video camera lens. If you project an LCD projector onto a 20 ft screen, then compare it to the same LCD scaled to a screen 10 ft across, the smaller screen will look much brighter. It's the same amount of energy, but over a smaller area...so it appears brighter when viewed from the same distance. From my research here, it seems very clear that the light loss equations work as expected after the imaging element, but to impose them before is to suggest that the ground glass/spinning disk/imaging element has an aperture and an f stop value...and it doesn't.

What we're seeing here with the retail flip module + CF1L is 10 to 14 IRE using a 50mm f1.4 lens mounted, (HV20) and with the XHA1, about 20 IRE drop with an 85mm f1.8. In other words, framing the same scene with and without the adapter attached on those two lenses would show an average drop of 13 and 20 IRE respectively. That equates (and correct me if I"m wrong) to .7 of a stop with the HV20/50mm f1.4, and a full stop with the XH-A1/f1.8 85mm attached. This is better than we had modelled optically.
 
I find the ways in which people are measuring the amount of light loss quite a bit tricky.
Among the variables
1. the gamma of the video camera is not linear from 0 to 100 ( so if you are measuring say a white between 70 - 100 a 15 IRE change may be quite different than at dark grey below 40.
Best to measure movement in the middle of the Grey scale but even that's inaccurate because:

2. the various gammas have radically different slopes. Probably Cine D is the flatest but to tell the truth all my tests have ben done with HD Norm because that's where I keep my camera. But either way they will throw the measurements off.

3. If you look at the IRE response to opening up the 35mm f stop, I have seen (odd as hell) that wider openings get less response and as I close the iris it gets more and more pronounced. 1 stop between f2 and f2.8 is much less than between 4 and 5.6. Often opening up wider than f2 or even f2.8 has little affect.

My guess is that this varies from diffusing screen to screen and perhaps may depend also on the video camera, the adapters optics and the distance to the diffusing screen. But i can't say i've tested the variables - I just know they are weird.

I say this just to caution everyone not to get too caught up on light loss numbers . They are general aproximations and very valuable at that but only aproximations - your mileage may vary.

I trust the shutter speed more than anything else as an objective measure. If I can get the same IRE reading say from a naked HVX @ 1/60 and then an adapter (at a given fstop & lens) at say 1/30. I know that they are 1 stop different all other variables are silenced for that moment at least.

But it might be possible for one adapter to be faster with a 1.4 primes and another to be faster or equal at 2.8.

Take it all with a grain of salt because being really precise can drive you nuttty.

The best test is the same lens on the same HVX using different adapters side by side. But those tests have to be extremely careful. Actually I would do them all at f2.8, f2 and f1.4


Having said all that mouthful. I do find Dennis's numbers for the CF1L in the ball park.
 
yea but you can simply measure it. get a white wall. light it evenly, and shoot with the waveform. The white wall will read as a solid stripe across. Then put a brevis/sg/LEx etc.. on and measure the difference with a lens that has a known light loss factor.... its really very simple..... the new stop rating, minus the Lens lightloss, is the adapters loss in transmisison of light.
 
Timur, I don't find anything exact about that. You can do a decent comparison between 2 adapters that way, but you can't come up with absolute numbers.

For example how are you measuring the light loss?

Are you comparing the IRE difference - that is not neccessarily linear as I explained above.

What do you mean by a lens with a known light loss factor?

The "new stop rating" on what: the video camera or the 35mm lens? - I don't find either HVX fstops or the 35mm adapters f stops neccessarily linear in these rigs. A lot of ordinary exposure rules just don't make sense in this world.

I've done that kind of test - thought I saw say a 1 stop difference then double checked by holding an ND 3 over the faster adapter. Yikes didn't get what I expected. Compared by changing the shutter speed -different again. Tried comparing IRE changes that i thought were one stop with a different way of changing the exposure or measuring it and the results were not what I expected. Tried measuring the f stop change in HVX and it was different wide open than at more closed stops.

I did it with a friend, we are both old guys with lots of years behind us and we sat there scratching our heads.

I've compared 3 Brevis screens with Redrock and SGPro, yet my results have differed from some of Dennis's on another video camera with different primes. I would never post my results as absolute.

I've been saying this again and again on this forum. There are many more variables here than are immediately apparent. Once again I emphasize - these are general numbers and your mileage may vary.

It doesn't mean you can't compare these guys and the generalities are very useful. But you can't simply say this adapter loses 1/2 stop and this one loses 1 stop when they were tested under different conditions.

Lenny Levy
 
Last edited:
Testing light loss between two adapters is quite easy, repeatable, and totally predictable. It just requires using a dimmer, a light meter, and a waveform monitor (or, with the HVX/DVX, you can get very close by using the marker instead of a waveform monitor.)

Point at a flatly-lit, evenly-lit white card. Adjust the camera iris so that it reads out at 50 IRE on the waveform monitor. Take a meter reading on the white card and write that down.

Then put on an adapter with no lens. Fill the screen with the white card (fuzzy, yes, but the light entering the adapter is still the light, whether focused or not) and look at the waveform monitor. It'll be reading less than 50 IRE. Adjust the dimmer to bring the light level back up to the point where the waveform monitor is showing 50 IRE again. Now use the lightmeter to meter the light hitting the white card. Compare this new reading against the last reading -- the amount of stop difference is the amount of light loss of that particular adapter.

If you want to measure the effect all-in, with the 1.4 lens attached, the same process applies. Just meter before, put on the adapter, raise the light level using the dimmer until you hit 50 IRE, and then meter again. The difference between the two lightmeter readings is your adapter/lens chain loss.

If, while doing this, you put the waveform monitor in single-line mode, you'll also be able to tell if the adapter is causing hotspot/vignetting issues which would cloud the true light transmission test results. If it's a big hill with a hotspot in the center and darkened edges, how do you tell which part is the true light transmission? You'd have to average somehow.
 
There is a definite difference between a focussed image and one out of focus with respect to light loss. It's not huge, but it's there. In our case it has to do with how much light is directed back to the video lens. The angle of light striking the diffusion medium has a lot to do with how it's directed behind the unit. It's also why I believe there is some degree of latitude compression observed by adapter users. It's more like a selective filter (scatter) than compression but that's how it manifests itself. What I can say with complete confidence is that existing Brevis users, and those renting it where light loss is a concern, will not have to worry about appreciable light loss with the flip. I can also say that what I'm seeing with the scope is about what we see on the camera with regard to relative light loss.

Barry, you do indeed need to average the IRE as it will drop on the edges regardless of adapter. If you zoom in to academy format using a 35mm SLR lens, this can be avoided on likely every adapter...but that's cheating (my humble opinion). Our goal here is to maximize field of view with every lens, good bad or ugly. Most 35mm SLR lenses show considerable edge falloff when used at wide open apertures on film...and this cannot be practically eliminated by the adapter. These 50mm lenses are showing considerable falloff: http://www.kenrockwell.com/nikon/50-comparison/falloff.htm It's not something to get too concerned about as we don't shoot white walls that often...but we do need to be aware of it when testing the adapter systems.

We've started doing all of our testing with a laptop and Adobe Onlocation CS3 so I'll be doing extensive testing over the next few weeks to put numbers behind the observations. The split screen function makes it fairly easy to test a scene with and without the adpater, and have identical framing on both. Scopes are nice too :-) I don't agree with the lens-off method of testing, but I am going to run a series of tests that way. Quoting the relative difference between a scene with and without the adapter/lens is the easiest (again, my opinion) for most folks to deal with. It's also simple to plan lighting if you're assuming a given light loss expectation. We've also made the effort to install a CAT5 component HD distribution system to get output raw HDV (via HDTV output) to a 42" 1080P panel for evaluation.
 
Last edited:
I agree with Barry that testing without a lens is innacurate . Way too many variables.

However the dimmer idea is great. Why didn't I think of that? Thanks Barry. Its an absolute measure of light that bypasses all the non linear exposure controls in the camera and its easy.

I would just do it with a lens on. Also I caution you if you are doing it with fast lenses test it at 1.4, 2 and 2.8 as they will not be linear and will respond differently on different adapters and screens. If you haven't compared these guys side by side you will not have experienced it but it is a fact.

I haven't tried different cameras but i suspect that's part of the equation on the open fstops as well. The lightloss at a given 35mm fstop with say an HVX may actually be different with another camera and a different 35mm fstop. That's why I don't like absolute numbers. But side by side comparisons will be accurate and the dimmer idea is great.

Regarding the white card flat screen. I suspect you'll never get it completely flat as most lenses have some fall off which you've seen when trying to shoot green screen with a 2/3" camera.
 
Last edited:
What we've been doing is similar to Barry, but without a dimmer.

1. Attach adapter to camera using 50mm f1.4 lens and set up scene (in our case, a product photo booth with set lighting). Grab a few seconds of footage using Onlocation including scope and waveform. The scene is white backdrop with several dark objects scattered across.

2. Remove adapter, leaving camera powered, exposure locked, and on tripod. Use split screen function of Adobe Onlocation to frame the identical scene with the bare camera. Capture footage and compare waveform monitor readings.

20 IRE is 1 stop. Make sense?
 
Is this leading to a test of many different adapters, or just a test of the Brevis with and without the Flip Module and findings released sometime thereafter? Thanks.
 
OK I just checked the IRE on my HVX starting with a white wall at 50IRE with the shutter at 1/250.

If I close the shutter 1 stop (1/500) it goes down to about 33 or 34, if I lower it 2 stops (1/1000) it was closer to 20.
If I open it 1 stop to 1/60 it goes up to about 74. If I raise another stop I'm into my knees so its all meaningless. This varied a bit between HDNorm , SD Norm and Cine D, Didn't look at Low maybe that would have been better.

I conclude that it's not linear even 1 stop below and 1 stop above 50 IRE. Maybe its linear 1/2 stop above or below.

My suggestion - use a 600W light, breakdown and get a $6.99 dimmmer from a hardware store and wire it up in the middle of a cheap zip cord with a few wire nuts and test it with a lightmeter - i.e. Barry's method with lens installed.

- Lenny Levy
 
Dennis Wood said:
20 IRE is 1 stop. Make sense?
Click to expand...

No no no no no... there's no direct correlation between a set # of IRE and an f-stop across any cameras. It's all up to the individual camera's latitude and also its gamma curve.

There are 109 IRE points distributable across a camera's entire dynamic range. If you put a RED One up there, it'll have 11+ stops of dynamic range, yet the entire signal will still be calibrated to 109 IRE on the monitored output. So does that mean it's about 10 IRE points per stop? Maybe, depends on how linearly those IRE points get distributed (i.e., how much of a toe or knee is there in the gamma?)

Take another example -- plug a Sony FX1 in and point it at the same scene. It has maybe six stops of dynamic range if you're generous. I seem to recall 5.5. But its signal will still be mapped to 109 IRE, so does that mean that 1 stop on an FX1 = 20 IRE? Maybe. Again, depends on the gamma.

So no, you can't measure that way with any expectation of accuracy, at least without standardizing that you will only test on one specific camera, and that you go through and map the IRE-to-F-stop correlation beforehand. Without doing that, you can have no assurance that you're getting reliable information.

The light meter way will give you a direct and highly accurate indication of the exact amount of light involved. And compensating to a specific IRE level should mean you're hitting the exact same point in the gamma curve on that particular camera. So the method I described above will be, I believe, a lot more accurate and repeatable.
 
I concur with Barry 100%.
Use the dimmer and a lightmeter though I would do it with a lens on the adapter.

I think this should be an accurate test that you could use to compare different adapters even without having them side by side as long as the caveat was with "X" camera and "y" 35mm lens @ a given f stop, there is "z" light loss.

And - always test at both 2.8 and more open f stops.

BTW - the test without a lens sounds interesting, I just don't know enough about these things to trust it would be the same with a lens.

Does that make sense?

(Man I am spending way too much time on this forum- my apologies. I need a life or at least some work!)

- Lenny Levy
 
Last edited:
You must log in or register to reply here.
Back
Top