Does Letus Extreme lose more light than Brevis

The fact that the Brevis does not lose light in a linear fashion that corresponds to the taking lens is a well known, commonly observed fact amongst our users. It is very, very hard to believe as you watch an aperture go from f1.4 to f2.0 to f2.8 (50% reduction from f1.4, then 50% area reduction again from f2.0 to f2.8). We are not defying physics here at all...the imaging elemnent is a break in the optical chain. The physics at work at the imaging plane simply do not correspond to conventional thinking regarding aperture and light loss.

Lenny measured the Brevis at just over a stop faster then the Letus EX with the taking lens at f2.8. I can't speak for the Letus, but his findings are absolutely consistent with our own findings here. With the Cinevate flip, the numbers would be closer to .7 stops at f1.4 and virtually the same, .7 stops as the 50mm lens is stopped to f2.8.

Don't take my word for it. Attach a Brevis to any camera and connect it to a waveform monitor. Stop the lens done from f1.4 to f2.8 and you will see exactly what I, and every one of our customers is talking about. Douglas, there is nothing I've said in this thread that any Brevis user could not confirm for themselves in about 10 seconds of testing :-) This is why we really think adapters should be tested for light loss with a standard f1.4 50mm mounted at f1.4, f2.8 and f4.0

Cheers.
 
Last edited:
Look guys,
I've been through these discussions before. Awhile ago I noticed that different adapters had different DOF's. Actually Dennis first mentioned it to me and I thought he was nuts. "That's impossible" guys like you argued... "DOF has to due only with f stop , etc, etc, etc." I was even arguing with Barry about it. Now everybody here seems to recognise that its true, because they've looked for it and can see it themselves. Its an observable fact. Theory comes after abservation.

Well its the same deal with linearity, and all you have to do is open your eyes and compare adapters and exposure. Its simply an observable fact.

Yes in a normal situation if you close an iris one stop you reduce light 1 stop. However the adapters are different because you are looking at the image through a diffusing screen that has its own properties and you are looking at it from a distance away so how that screen scatters light affects everything from DOF , to exposure, to gamma, and nature of the Bokeh.

Measure it yourself. It may be different on different video cameras, but on an HVX I promise you that a pre-flip Brevis CF1L will only very barely react to the change in light from 1.4 to 2.8 on the 35mm lens. Its simply a fact. Other units haver their own linearity "footprints".

My own theory is that because the Brevis is a less dense screen it simply doesn't scatter the light from the outside of the wider iris enough to make a difference when observed from a distance. (similar to vignetting perhaps.) The flip optics could change that entirely. Its both good and bad. You get no advantage from a fast lens, but you also have a very fast adapter at f2.8.

The other units each have their own characteristics but these need to be taken into account when comparing speed among other factors.

BTW I've been in this business for many years and used to run a film rental department so i'm used to lenses and collimation issues.

A number of these issues I think deserve a sticky so we don't keep going around in the same circles. Back focus is another. I would love to see one for DOF, Linearity, and back focus. I've suggested it before but no one seems to respond.


Lenny Levy
 
The amount of light that is transmitted to the ground glass is governed by the aperture of the 35mm lens. It determines how much light gets through. If you stop it down by one stop, 1/2 as much light gets through. The adapter has nothing to do with this, this is all happening at the lens level, before the adapter even factors into it.

If the claim is as I understand it to be, it's simply not possible and not rational. Here's a simple example:

Let's say there's a light source that's outputting 8,000 footcandles. We have a lens that's set to f/4.0 (for sake of clarity let's pretend there's no light lost by the lens, and f/4.0 = t4.0 in this case.) The amount of light exiting the lens will be, for sake of argument, 500 footcandles.

That's a mathematical fact, a given. The amount of light getting through the lens is restricted by the lens' aperture. If we stop that lens down to f/5.6, then 250 footcandles will exit the lens. If we stop it down to f/8, 125 footcandles will exit the lens. Make sense?

Okay, so now we introduce an adapter into the equation. The adapter is primarily a ground glass screen, with perhaps a diopter and maybe a condensor lens involved. This optical path is going to result in some light loss, especially the ground glass portion. So let's say the ground glass and other components eat up a full stop of light, right? So, if our lens is transmitting 500 footcandles, the adapter's optics will eat up 250 footcandles and transmit 250 footcandles.

Simple math, simple light transmission, right?

So, if we close down the aperture on the 35mm lens by one stop (to f/5.6), we'll get 250 footcandles (instead of 500) exiting the lens. The adapter will see 250 incoming footcandles and transmit 125. So the total relationship of incoming light to exiting light stays constant.

If we close down another stop, to f/8, we'll now have 125 footcandles exiting the lens, and 62.5 exiting the adapter.

If we close down another stop, to f/11, we'll have 62.5 footcandles exiting the lens, and 31.25 footcandles exiting the adapter.

What has been claimed is that the Brevis will, when the 35mm lens is stopped down, actually be faster than other adapters, and the more you stop down the lens, the more the gap grows (i.e, the Brevis will grow to be 2 stops faster than other adapters.) How? How is this possible? The amount of light entering the adapter has been consistently, mathematically, chopped in half. Yet, somehow, the Brevis must be "manufacturing" or "creating" light in order for this "performance gap" to increase. It's receiving a limited quantity of light that's directly cut in half for every stop that the 35mm lens has been stopped down -- how can it possibly "widen the gap" to appear to be two stops faster than another adapter? It can't.

The only possible way I can think of for there to be any sort of observable effect for this would be that there's some sort of condensor lens that's amplifying and magnifying a hot spot in the center of the frame (like a magnifying glass takes a wide spread of light and concentrates it into a tiny pinpoint which is super-hot). If that's what's going on, then I could see results like these claims showing up, but I would also expect it to be accompanied by a significant amount of portholing/vignetting.

It is a very curious claim that's being made, and one that does not seem scientifically supportable. I also would like to test this. Anybody in the San Diego area with a Brevis who'd be willing to donate it for a day so I can try some experimentation to get to the bottom of this?
 
Added: I have also noticed some nonlinearity in the very-most-open stops, being f/1.6 to f/2.0 on an HVX (without any adapter involved). At those most-open stops the lens doesn't appear to respond completely linearly; stopping down from 1.6 to 2.0 doesn't result in 2/3 of a stop of light reduction. As to why, the only explanation I can come up with is that the HVX CCD probably uses microlens technology over each pixel, and the extra light entering the lens through the wider aperture is hitting at such an angle that it's not being "seen" as efficiently by the microlens.

However, none of that has any bearing on how an adapter works. The adapter looks at a constant display (the ground glass screen) and light is transmitted through that at a constant rate of scatter, due to the pattern of the ground glass. As to why we see nonlinear response with all the adapters from 1.4 to 2.0, I don't know that answer yet, but I haven't seen an adapter that performs differently, they all seem to ignore anything more open than about 2.0.
 
Barry,
I love you but your totally out on a weak limb here and basically misusing the scientific method. We don't use science to disprove observed facts - that's medieval thinking really. We use science to make sense of observed phenomena. This is an observed fact. Just because you haven't seen it doesn't mean that others haven't. They have.

My partner gave me the same routine for a month unitl I showed it to him. Its a fact - you can come up with a theory of your own to explain it but I think mine isn't half bad.

If you measured the light at the film plane I think it would be linear as a film or chip would react, but you are seeing only what's visible from a distance away as transmitted by a dirty piece of glass. If the glass was smooth enough in fact you wouldn't see an image at all. Adapter optics are a wrinkle all their own.

Do you still disagree that different adapters display differring DOF? That's also an observed fact.

By the way I don't think that the Brevis is manufacturing light, that may be why its hard to understand - I think its the opposite - it only sees the light at 2.8 and doesn't see anything from the wider iris. Seems quite reasonable to me - plus of course - it is a fact!

I would expect it to be accompanied by more vignetting but it doesn't actually seem to be, but that may be due to all the work dennis has done on picking screens - I don't claim to understand it and haven't tested nearly all the parameters.

Hold it though I just saw your last post - are you agreeing that it happens on the others from 2 to 1.4. Why are arguing about it then. Same deal on Brevis, just more pronounced.
By the way in my experience the Redrock was more linear. My theory- denser screen- more linear - more scattering from the wider iris. Just a cheap hypothesis though.

Lenny Levy
 
I said the same thing: From 1.2 to 2.0 there's a hair of difference. 2.0 to 2.8 it shows up on the LetusEX.

And, as far as light loss in general is concerned, I suspect the reason is different zoom settings between the adapters. Might have to gather them up and see how much light's loss from the HVX itself between the different zoom settings.

The good thing is with the flips out, every single test should be able to zoom to the exact same number for each adapter, ruling the camera out where light is concerned.
 
Last edited:
How is a camera connected to a waveform monitor not scientifically supportable? It measures both edge falloff and differences in light levels so I can say with certainty that the scope starts as a straight line in our test (lit white board) and stays that way as you stop down the lens. Not to quibble, but what I'm stating has been well documented by our users. From f1.4 to f2.8, there is little to no change, and from f2.8 to f4.0, you'd see about half a stop. (Lenny has indicated that the Letus EX is losing just over one stop from f1.4 to f2.8...which suggests the Brevis at that same aperture would require half as much light.) The camera zoom settings are irrelevant in this testing because they are not touched as the taking lens is stopped down from f1.4 to f2.8, to f4.

On the other hand, when testing a bare XH-A1, and a bare HV20, a change from f2.8 to f4.0 did show a 20 IRE change consistent across both cams, when the camera was set up at 50 IRE to start. That's exactly what our optical engineer had predicted, assuming we tested at the middle of the gamma curve.

I should stress that I have never, (and will likely never) use or test any other adapter out there...so I am definitively speaking based on extensive testing with the Brevis system only. We have way too much respect for our customers to make any of this up :-) I'm mystified as to why the discussion is consistently so hotly debated given that there are now a large number of Brevis owners making the exact same observation as we've been maintaing for 18 months or so. What I will commit to is a clip that will settle the discussion for good...it's actually done already and incorporated into Cinevate's "Video Univerity" series on image tuning. I"ll have Paul pull the clip tomorrow which shows the scope, camera LCD, and me stopping the taking lens down.
 
Last edited:
Wow Barry, this sounds like something that you've been thinking about in the back of that big ol' head of yours for some time now. I would love to see it go to the front burners of your mind for a day or two.

If I had one I'd drive down with it for you to but the best I can do is ask if anyone I know with one will ship it to you. Thanks for your posts.

And thank you too Dennis and Lenilenapi.

EDIT: Thanks Kholi, what you point out about the zoom setting sounds promising to my uneducated ears... Barry?
 
Dennis, just to clarify... does this phenomenon remain with the flip module mounted?
I think you've said a few times that it does just not as pronounced... is that right?

Sorry... I am trying to keep up with you guys
 
I think the difference in speed is generally due to the density of the screen though zoom plays a part as well. You can measure how much zoom matters by testing it without an adapter. I think you'll find if you start at say 2.8 on the video camera you'll see very little difference as you zoom at least until the very long end. At wide open there is greater affect.

Lenny Levy
 
Lenilenapi said:
I think the difference in speed is generally due to the density of the screen though zoom plays a part as well. You can measure how much zoom matters by testing it without an adapter. I think you'll find if you start at say 2.8 on the video camera you'll see very little difference as you zoom at least until the very long end. At wide open there is greater affect.

Lenny Levy
Click to expand...

When you get a chance, actually put it to the test with the adapters on and tell us what you get. Camera @ 2.4 or2.8 is fine, a lot of the time that's as far as you can get the iris on the HVX down when doing indoor shooting, anyway.

Open would be just as good.

Better YET: Why hasn't anyone RECORDED this in action? We can see what YOU see when you hit the record button, so please show us how this works. Record the stopping down from 1.2 to 2.8, then 2.8 to 4 so we can see how little it makes a difference first hand.
 
US, the flip module changes nothing with regard to the linearity of the response to aperture changes (or lack of!). You just need to add .1 to .2 stops to the non-flip numbers :-) The hidden benefit in all of this to Brevis users is that you can confidently stop your 35mm lens down to the f2.8 to f4 range (where most are noticeably sharper edge to edge) without losing much in the way of light.
 
Last edited:
Excellent, thanks for saying it like that for my "simple mind" :)

Is there any chance you could mail a used Brevis that you might not be using to Barry for testing? I don't think I have anyone willing to let go of their adapter :)
 
US, that's nothing that has been rattling around, that's just top-of-the-head stuff.

There are several claims that are made that just don't make sense. I like things to make sense. There was once a claim that a Brevis was actually *brighter* than not using the adapter. This is, of course, impossible. Then there was a claim that in the Philip Bloom test, the more you stopped down the Brevis, the more the difference between it and an M2 (i.e., maybe the Brevis lost 1 stop at f/4, and the M2 lost 2 stops, but by the time you got to f/8 the Brevis was losing maybe 1/4 stop and the M2 was still losing 2 stops -- I'm making up these numbers but the concept jibes with how I remember the info being presented). That's what I'm calling into question here -- it does indeed seem completely impossible. There has to be some other factor at work.

As for taking users' reports, I don't find that's the most reliable indicator of factual accuracy; just like the pollsters found out in the New Hampshire primary. People are out there saying that the HV20 is "better in low light" than a DVX -- they'll swear by it. Even though it's absolutely completely entirely wrong. There were people "testing" and declaring that the Red had no rolling shutter artifacts, and that the EX1 had no rolling shutter artifacts, even though they most definitely both do have them.

If people are happy with whichever product, great for them, that's the state we'd like to see all users achieve. I'd much rather see people walk away from the endless discussions and debating and just get out and start putting their products to work. But I also am a stickler for accuracy; I'd like to know what is actually happening and why.
 
Oh, and as for adapters for testing -- we tried to put together a call to all manufacturers for adapters so we could get to the bottom of all this. Not surprisingly, nobody offered an adapter -- except Cinevate! They were the only ones who said they would participate, but the timing didn't work out, and that was a while ago. But Dennis was willing to put his up for comparison, which speaks well for his confidence in his product.
 
Barry, You're right that claims are made all the time on here that are complete nonsense. Personally i have never made any claims for any adapter like those you mentioned.

I was personally curious enough about all the BS about these adapters that I bought three of them - actually 4 including the Redrock.

The only thing I would appreciate is that when responsible people ( I consider myself responsible) make observations, especially when they that are supported by others, you don't just dismiss them by calling them unscientific or impossible. Wait till you're able to make your own observations.

Lenny Levy
 
So I guess Dennis might send you an adapter now even if it is not alongside all other popular ones. We trust your word, at least I do. I have not made up my mind yet, but it in fact the Brevis is at all faster at 2.8f than all others then I think I will be sold as soon as I see some of that flipped HD that people seem to be extremely happy with... even if I have to get the flip for it to be 100% sharp, which is something I wasn't planning on. If all goes well with my plan I will need the best 2.8 performance I can get so I can get the 28-70mm and 80-200mm zooms as I keep thinking about. Of course there are several friends I've made here that are trying to steer me towards primes instead, but I am stubborn and will only go prime if I absolutely positively have to. It is because of this that I am so interested in this folk myth being tested... I am cautiously optimistic about it too. Not knowing a 48th about lenses as you do I am solely basing my limited optimism in the fact that I have personally heard two completely separate people say something of the sort... so I will wait for some of that Barry posting that I enjoy only as much as I used to enjoy reading about Autobahn escapades on M3's on the pages of Road & Track.

Not that you write like those middle aged crisis ridden folks... hehehe... you are much more objective.
 
FWIW, I agree with Lenny (and others) about the Brevis' non linear response.

Barry, if you don't know, please don't say you know. "Statements of fact" here and on the EX1 board are leading to much confusion and angst. Your opinions carry much weight.
 
I'm going to shoot a little demo today that will make this visually clear. I accept that it's very hard to believe, incredibly counter-intuitive, and I would not have accepted the observation had I not quantified it myself. If nothing else, it will spark a great discussion, and I hope illustrate a different way to measure light loss on adapters in general :-) Our R&D budget has been extensive, and with six different imaging elements to shoot with, we have extensive data on the issue of light loss and linearity. CF1Le and CF1 are purpose designed to achieve maximum light efficiency, particularly as you stop down the lens. They are also a very good film match with regard to bokeh in our tests beyond f2.8

The difference Phil Bloom demonstrated is here: http://web.mac.com/philip.bloom/Shootout/Page_4.html Note that the taking lens was at f8 ...and check out the HVX settings. The Brevis grab is still overexposed by a stop or two in that shot!

For f1.4 to f2.8, CF2Le and CF3Le are a better match to film bokeh...but lose more light. The important thing is that our users can choose what works best. With our next CINEFUSE release, anyone doubting the importance of drop-in imaging elements will likely change their mind :-)
 
Last edited:
Great! Dennis, thank you for doing this... I hope you can host it in 720p! Let me know if you need bandwidth I should have a few thousand GB left over hehe...

Secondly, when you said "CF1Le... particularly when you stop down the lens... very good film match with regard to bokeh in our tests beyond 2.8" and "for f1.4 to f2.8, CF2Le and CF3Le are a better match to film bokeh... but loose more light".
If I were to use only the 2.8f zoom lenses, would I be in that bokeh sweet spot with the CF1Le wide open or should I still go with the CF2Le which you recommended once when you didn't know about my uncommonly slow lens choice?

Dennis Wood said:
With our next CINEFUSE release, anyone doubting the importance of drop-in imaging elements will likely change their mind :-)
Click to expand...

When is this happening? And what should we expect?

StMad, I think Barry will get to the bottom of this. His posts were only questioning the possibility of this unexplained and unprecedented behavior. He is writing in the tone of someone searching for an elusive explanation for something which down right weird as heck. Since he expressed his willingness to test this phenomenon I trust he will continue to weight in on discussions regarding it and in doing so he will again be carrying the free-marketing factor in his posts. Also, as you no doubt know, he will be making use of his unbiased approach to render his opinions useful to the community as he has been doing since way before you and I joined back when the HVX was a rumor.
I really hope he can get his hands on one, because if he records this, not only will we get a great guide on how to take advantage of it and an in depth explanation as to it's pros and cons if there are any, but better yet, he might become so curious as to why it does happen that he won't stop till he has a scientific explanation for it..?
 
You must log in or register to reply here.
Back
Top