Bolex H16 RX Light metering

[carllooper]

Here's another test that can be done that doesn't involve dismantling a lens. Will require some finesse in implementation.

An f-number is defined as f / D where f is the focal length, and D is the diameter of the aperture. The use of a lowercase "f" here follows a convention used in optics, and indeed the origin of writing f-numbers with a forward slash following the f (eg. f/5.6) is to distinguish it from the f which would otherwise mean focal length, while simultaneously echoing the ratio which defines the f-number. An uppercase F is an alternative nomination for focal length, but is not an alternative nomination for aperture. Of course if someone (like myself, a naught boy) writes something like "F-stop" the intention is obviously not focal length.

So:

f-number = f / D

Now we can rewrite this equation to find the diameter of an aperture, given it's f-number and the focal length of the lens. The equation becomes:

D = f / f-number

The lens I'm looking at is one where f = 10mm (focal length) and the maximum f-number is (according to the markings) f/1.6, so we can plug these values into the above equation and obtain the aperture diameter:

D = 10mm / 1.6
D = 6.25mm

Now from this can then be created an ideal aperture from scratch (eg. a hole in some cardboard), that is independant of the lens, and has a diameter of 6.25mm. From a science fiction perspective (because I like sci-fi) we can treat this as cloning the aperture inside the lens, and teleporting it out of the lens, onto the workbench.

We are now in a position to measure light projected through this ideal aperture, using a light meter, and then compare that to the same light as measured through the lens set at f/1.6. From this we can, in principle, determine what the transmission of the lens is. Given the transmission we can then determine what the appropriate t-number should be for the lens, ie. for when it is set at f/1.6

And as an additional benefit determine if there is anything odd about the f-numbers on the lens.

As mentioned some finesse will be required in the execution and I haven't yet thought through that.

Carl

[doug]

Fascinating stuff
I've always metered my RX lenses for 1/80 sec. and they always seem to be correct when comparing with other non-RX lenses on the same RX camera. If they are indeed a wider aperture than marked, if I understand this discussion ... that means a f1.1 26mm RX lens would be something like f0.9 in reality. So in theory one could get this aperture if you fitted this lens to a non-reflex camera :!: But I've found when comparing the f1.1 lens with a non-RX f0.95 25mm Xenon lens fitted to a H16RX, I always get a brighter result with the Xenon. Maybe not as sharp though.

[doug]

The f1.1 lens I forgot to say is of course the 26mm Switar.
I should perhaps add that many non-RX lenses of do seem to have widely differing light transmission. Especially long zooms. Actual aperture can be almost a stop out.

[Tscan]

It's been a while since I used my H16 reflex... I think i was metering as a 1/60th shutter, or whatever the manual said. But i was rating my stock 2/3rds slower for sure. I would rate 100D at 64ASA and my exposures came out perfect.

[carllooper]

What would be the f-number for a given T-number that was taking into account 20% loss?

If the lens loses 20% it means it is transmitting 80%, ie. 8/10ths of the light.

To recover that 20% light, (ie. that which an ideal aperture would transmit, and which an f-number represents), we would need to increase the area of the aperture implied by a T-number, by the reciprical of 80%. But note that the result of this is not what we would use in a film shoot. If we have T-numbers then we use the T-number. It's function is to correct an f-number. Reconstructing an f-number from a T-number reconstructs the original problem for which T-numbers are designed to be the correction. So this is purely about reconstructing a problem - not a solution. With that in mind we proceed:

The reciprical of 8/10 is: 10/8

A quick sanity check on use of the reciprical can be done by rewriting the proposition and seeing if the result agrees with the intended outcome:

8/10 * 10/8
= (8 * 10) / (10 * 8)
= 80 / 80
= 100%

Yep.

So an f-number specified aperture area (fA) would be the T-number implied aperture area (TA), multiplied by 10/8:

fA = TA * 10/8

To obtain a value for TA we can use the same definition as an f-number, but with a transmission implied diameter (TD) instead:

T-number = f / TD

Rewriting this for TD:

TD = f / T-number

TD = 10 / 1.6
TD = 6.25 mm

For computing the area of the transmission implied aperture we'll need the implied radius (Tr) rather than the implied diameter (TD):

Tr = 0.5 * TD

The area is then:

TA = PI * Tr^2

We can plug this into our equation for fA:

fA = TA * 10/8
fA = PI * Tr^2 * 10/8

Now that we have fA, we can proceed to work our way back to the f-number. Given the area of a circle (in this case fA) what would be the diameter (fD) ?

Area = PI * radius^2

We can rewrite this as:

radius^2 = Area / PI
radius = sqrt( Area / PI )
diameter = 2 * sqrt( Area / PI )

So that means:

fD = 2 * sqrt( fA / PI)

Plugging this back into the definition for an f-number and expanding out the terms gives us:

f-number = f / D
f-number = f / fD
f-number = f / ( 2 * sqrt( fA / PI )
f-number = f / ( 2 * sqrt( TA * 10/8 / PI )
f-number = f / ( 2 * sqrt( PI * Tr^2 * 10/8 / PI )

Here we see the PIs cancel each other out so this can be simplified to:

f-number = f / ( 2 * sqrt( Tr^2 * 10/8 )
f-number = f / ( 2 * sqrt( (0.5 * TD)^2 * 10/8 )
f-number = f / ( 2 * sqrt( (0.5 * f / T-number )^2 * 10/8 )

Plugging in the following:

f = 10mm
T-number = 1.6

f-number = f / ( 2 * sqrt( (0.5 * f / T-number )^2 * 10/8 )
f-number = 10 / ( 2 * sqrt( (0.5 * 10 / 1.6 )^2 * 10/8 )
f-number = 10 / ( 2 * sqrt( ( 5 / 1.6)^2 * 1.25 )
f-number = 10 / ( 2 * sqrt( 3.125^2 * 1.25 )
f-number = 10 / ( 2 * sqrt( 9.765625 * 1.25 )
f-number = 10 / ( 2 * sqrt( 12.20703125 )
f-number = 10 / ( 2 * 3.493856215 )
f-number = 10 / ( 6.98771243)

Answer:

f-number = 1.493856215

Carl

ps. the bigger picture here is not accuracy per se, but consistency. Photographing a scene that is out by some fraction of a stop need not make any difference to what one is aiming at. Film has a certain level of forgiveness. But if two shots of the same scene, using two different lenses, differ in terms of their transmission, then while each shot might look just fine on their own, they won't cut together in a way that suggests the same scene (the same lighting). So it's really consistency between shots, rather than the absolute 'accuracy' of any shot that is under consideration. If consistency is not an issue (eg. a different scene, or some other reason), the above can be regarded as just an otherwise interesting technical exercise, to satisfy technical curiosity.

[Mana]

My head just exploded! 8O

[carllooper]

Correction. The last calculation suffered a typo. It should properly read:

f-number = 10 / 6.98771243

Answer:

f-number = 1.431083506

It should be noted that the implied aperture of a T-number is not a real aperture, but a virtual aperture. A pretend aperture. One that would be smaller in size than the real aperture and if used in your real lens would admit less light. In this sense the virtual aperture of a T-number could be understood as "slower" than it's corresponding real aperture. But one has to be careful here. The virtual aperture doesn't exist. And if it did, it would be one you would use in a virtual lens, one that was a perfect 100% transmitting lens (rather than a real lens) so that it admitted the same amount of light as your real aperture does in your real lens. So properly speaking a T-numbered aperture is no "slower" than the corresponding f-numbered aperture, because both numbers refer to the same thing: the real aperture in your real lens, or the virtual aperture in your virtual lens, both of which admit the same amount of light.

In other words, for the calculation done on the specified real lens, it calculates that the following numbers on the barrel of that real lens, would line up together, ie. for the same real aperture and (of course) the same amount of light being admitted.

T/1.6
f/1.43

From a usage point of view the basic idea here is that when your light meter recommends using an f-number you would use the T-scale instead of the f-scale, to set your aperture. So, for example, if the light meter recommends using an aperture of f/1.6, you would then set the aperture according to the T-scale, ie. turning the ring to T/1.6 (rather than f/1.6) which corresponds to setting an actual aperture of f/1.43 (ie. using an aperture larger in size than that recommended by the light meter, and larger by the required amount.

Carl

ps. the calculation I did is a rather long version, from first principles, but as such easier to check for errors. There are no doubt more convenient ones using logs which would make the calculation somewhat faster to do on a calculator. But of course, when marking up a lens (or testing a lens for marking consistency), you wouldn't be doing this at the last minute, on location, with the director about to call action.

[erkanumut]

Hi guys,

I never understand that Bolex didn't switched to the mirror reflex system after some time although this system is well adopted to others. Bolex perhaps wanted to stay as an amateur camera supplier except the unsuccessful PRO design made by eumig. That's why they don't need the T/stops.

You can see some Angénieux zooms marked in both of f/ and T/stops.

What a pain is this RX lens requirement! If they would switch somehow into the mirror shutter reflex, the nobody will be confused so much. Even the Soviet made handheld clockwork reflex cameras use the mirror with interchangeable lenses. On the net many lenses sold for Bolex 16 REX cameras are not RX lenses shorter than 50mm. Even there are good optics you cannot use on a Bolex REX camera like Nikon Cine-Rokkors.

That's why Canon kept away the SCOOPIC line from interchangeable lens use, because they use the beam-splitter reflex as well.

I never owned a Bolex until yet. Now, I can understand the people...

But there is an advantage of that Bolex system, you can easily reach to the prism! In a SCOOPIC or other Super8 camera, etc. it is a real headache. When somehow occurs, any dust or prism problem will cause a pain. However, the Nikon R8/10 system allows easy cleaning only for the aperture end. That's all.

With the digital camera shit, the movie cameras lost their lenses as being sold separately. And the used lenses are bloody expensive now.

Also, Bolex R8 is another thing with its so similar design and smaller C-mount lenses to H16!

I wish that SCOOPIC cameras would have the mirror and TTL metering. Anyway...

Thanks for reading!

[JeremyC]

Hi guys,

I never understand that Bolex didn't switched to the mirror reflex system after some time although this system is well adopted to others. Bolex perhaps wanted to stay as an amateur camera supplier except the unsuccessful PRO design made by eumig.!

As well as being easy to get at the prism system I think the other reason is that a mirror shutter on a Bolex would have meant extending out the lens assembly to to accommodate a mirror assembly and that would also make for a complicated prism system extruding out further from the camera body. I imagine the whole shape and balance of the camera would've changed and that would've required a complete redesign.

[erkanumut]

As well as being easy to get at the prism system I think the other reason is that a mirror shutter on a Bolex would have meant extending out the lens assembly to to accommodate a mirror assembly and that would also make for a complicated prism system extruding out further from the camera body. I imagine the whole shape and balance of the camera would've changed and that would've required a complete redesign.

I agree with you.

Sad that they didn't manage to redesign it. It would be an amazing thing to see and have a Bolex mirror reflex camera with such a build quality!

P.S. The FF distance is shorter in the Bolex H8 C-mount lenses I think.

[nikonr10]

As well as being easy to get at the prism system I think the other reason is that a mirror shutter on a Bolex would have meant extending out the lens assembly to to accommodate a mirror assembly and that would also make for a complicated prism system extruding out further from the camera body. I imagine the whole shape and balance of the camera would've changed and that would've required a complete redesign.

I agree with you.

Sad that they didn't manage to redesign it. It would be an amazing thing to see and have a Bolex mirror reflex camera with such a build quality!

P.S. The FF distance is shorter in the Bolex H8 C-mount lenses I think.

I am sure that if the Swiss was able to do redesign it there would have done , But if you think that there cameras are still working today ! now thats design for you , why change a good thing ?

[erkanumut]

I am sure that if the Swiss was able to do redesign it there would have done , But if you think that there cameras are still working today ! now thats design for you , why change a good thing ?

Of course, the Swiss could do that!!! No dispute.

Aren't the mirror reflex cameras working still today? Its a matter that no electronics are involved in for this success besides the build quality and simplicity.

You mean a good thing is to have a good design and quality only. You need an extra motor for magazine take up, a bulky motor to run it (except the EL & EBM), an extra accessory to sync it crystal controlled, looking for the RX lenses to eliminate the aberrations... Not to mention the light loss!

The Japanese, despite they copied first, showed how to design service friendly and also reliable cameras to the world, as well as the Germans and French of course.

[nikonr10]

I am sure that if the Swiss was able to do redesign it there would have done , But if you think that there cameras are still working today ! now thats design for you , why change a good thing ?

Of course, the Swiss could do that!!! No dispute.

Aren't the mirror reflex cameras working still today? Its a matter that no electronics are involved in for this success besides the build quality and simplicity.

You mean a good thing is to have a good design and quality only. You need an extra motor for magazine take up, a bulky motor to run it (except the EL & EBM), an extra accessory to sync it crystal controlled, looking for the RX lenses to eliminate the aberrations... Not to mention the light loss!

The Japanese, despite they copied first, showed how to design service friendly and also reliable cameras to the world, as well as the Germans and French of course.

AS we all know' there really is no such thing as a perfect film camera, Each has somekind of design flaw ? that we learn to work around , if not it's the human behind it , When you think what has gone in too just making a super 8 film camera , in someways/ way head of other film formats for it time in camera design ! being a limted film format as being in a cart ? apart from DS 8 , To think that super 8 film is 50 year old soon .

[erkanumut]

AS we all know' there really is no such thing as a perfect film camera, Each has somekind of design flaw ? that we learn to work around , if not it's the human behind it , When you think what has gone in too just making a super 8 film camera , in someways/ way head of other film formats for it time in camera design ! being a limted film format as being in a cart ? apart from DS 8 , To think that super 8 film is 50 year old soon .

Agreed!

I had a Nikon R10 and an R8 in mint condition. Excellent cams. Shot beautiful pictures with the R8 only. Of course I was happy with my previous Nizo. But I switched to Beaulieu.

I have an ELMO C-300, full set bought in Japan with all the mags available, afocals, etc. Canon 8-3 D8 with excellent results, and a Fujica Z-800, as well as a Scoopic 16M and an ACLII full set. That's enough for cameras. By the way, I am not a collector, I use them. Most of my stuff bought in Japan except projectors due to the voltage difference.

I am eager to have one pack of ferrania S8 in April...